15 Donald J. Meuten1 and Travis L.K. Meuten2 1 North Carolina State University, USA 2 Colorado State University, USA Primary renal neoplasms are uncommon in domestic animals; they are usually malignant in dogs, cats, and horses and benign in cattle. In dogs approximately 70% are epithelial, 25% mesenchymal, and 5% nephroblastoma. Nephroblastoma also has a unique origin in the thoracolumbar segments of the spinal cord in young dogs. Primary renal tumors are usually unilateral but may be multiple or bilateral and can also have a multicentric origin in cattle and dogs. By a wide margin the most common neoplasm in the kidneys of all domestic animals is lymphoma, and these are not primary. Nodular dermatofibrosis is a rare inherited disease in German shepherd dogs and is associated with renal cell adenocarcinomas, adenomas, renal cysts, and concurrent fibrous nodules in the dermis and subcutis. The disease is caused by a mutation in the tumor suppressor gene, folliculin (FLCN), which was discovered in dogs. The majority of neoplasms that originate in the kidney or especially the urinary bladder are easy to diagnose via gross patterns and histology. Immunohistochemical (IHC) markers that further help confirm tissue of origin are PAX8, napsin A, CD10 (neprilysin), and uromodulin for renal origin and uroplakin, a specific marker for proteins in urothelium. Renal carcinomas are some of the few tumors that dual mark with cytokeratin and vimentin. Depending on the report, between 60 and 85% of primary renal neoplasms in dogs are of epithelial origin and 60–70% in cats (Tables 15.1–15.3). Although 85–90% of the epithelial neoplasms in dogs and 75–100% in cats are classified as malignant this designation is subjective and may have led to over classification of carcinomas. A 2015 study reports the value of mitotic counts (MC) to predict biologic behavior in dogs with renal cell carcinoma (RCC) and provided survival times based on the MC.1 Reported metastatic rates are approximately 60–70% in dogs,1,2 50% in cats, and metastases are expected in horses but are uncommon in cattle. Most renal cell tumors in cattle are greater than 2 cm in diameter, and less than 10% metastasized but they were all classified as carcinoma.3 Renal cell tumors are malignant in horses, but relatively few cases have been studied.4,5 The cells of origin are proximal or distal convoluted tubular epithelium or the collecting duct.6,7 The combined data from five sources totaled 894 primary renal neoplasms in dogs (Table 15.2), and indicated 70% were epithelial, 25% mesenchymal, and 5% mixed; 86% were classified malignant and 7% benign. We removed nephroblastomas from this classification, as it was not always clear from reports if they were benign or malignant.8–11 It is recommended that readers refer to the original reports to retrieve specific information as it is difficult to both summarize and provide enough details to be useful. Furthermore all studies vary in numbers and percentages, and the smaller the overall number of patients in a report the more selection biases are introduced. Hopefully the trends, patterns, and percentages are close enough between studies to make generalizations. Given these inherent features, approximately 65% of primary renal tumors in dogs are carcinoma, 25% are sarcoma, and metastases are expected in the majority of these cases. Table 15.1 Renal neoplasia a 125 lymphoma, 120 adenocarcinoma, 91 vascular. b 196 lymphoma. Table 15.2 Primary renal neoplasia in dogs Studies: 1 Veterinary Medical Data Base. 2 Klausner, J.S. and Caywood, D.D. (1995) Neoplasms of the urinary tract. In Canine and Feline Nephrology and Urology (eds. C.A. Osborne and D.R. Finco). Williams & Wilkins, Philadelphia, PA, pp. 903–916. 3 Klein, M.K., Cockerell, G.L., Harris, C.K., et al. (1988) Canine primary renal neoplasms: A retrospective review of 54 cases. J Am Anim Hosp Assoc 24:443–452. 4 Baskin, G.B. and Paoli, A.D. (1977) Primary renal neoplasms of the dog. Vet Pathol 14:591–605. 5 Bryan, J.N., Henry, C.J., et al. (2006) Primary renal neoplasia of dogs. J Vet Intern Med 20:1155–1160. Nephroblastoma (6%) was not included in the calculation of benign versus malignant. a One fibroleiomyosarcoma. b Chondroma, osteoma, myxoma. c Malignant fibrous histiocytoma. d Four teratoma; 1 hamartoma. e Undifferentiated carcinoma. f Renal sarcoma. Table 15.3 Primary renal neoplasia in cats Studies: 1 Klausner, J.S. and Caywood, D.D. (1995) Neoplasms of the urinary tract. In Canine and Feline Nephrology and Urology (eds. C.A. Osborne and D.R. Finco). Williams & Wilkins, Philadelphia, PA, pp. 903–916. 2 Osborne, C.A., Quast, J.F., et al. (1971) Renal pelvic carcinoma in a cat. J Am Vet Med Assoc 159:1238–1241. 3 Henry, C.J., Turnquist, S.E., et al. (1999) Primary renal tumors in cats. J Feline Med Surg 1:165–170. Nephroblastoma (18%) was not included in the calculation of benign versus malignant; most are malignant. From the archival material in the Veterinary Medical Data Base (VMDB) 579 canine and 252 feline renal tumor diagnoses were retrieved. For dogs, 187 (32%) were primary renal neoplasms; 140 were epithelial and almost all were designated malignant (97%), 40 were mesenchymal and 70% of these were designated malignant (Tables 15.1 and 15.2). There were twice as many secondary tumors in dogs (392 tumors, 68%); 121 were hemangiosarcoma, 120 were carcinoma, and 91 were lymphoma. Some of the hemangiosarcomas could be primary. In cats, 30 of 252 (12%) were primary tumors and 222 (88%) were secondary, of which 196 were lymphomas (Table 15.1). The canine population for this study was approximately 467,000, and the feline population was 144,900, for incidences of 0.16% and 0.2215% in dogs and cats, respectively, which is similar to published rates. Two reports summarizing 73 cases in cats reported renal carcinomas were the most common tumor (28 tumors, 38%), followed by nephroblastomas (18 tumors, 25%) and sarcomas (15 tumors; 21%).10,12 This is similar to the data in Table 15.3, in which 46% of 120 tumors were renal carcinoma, 18% nephroblastoma, 15% sarcomas, and 9% transitional cell carcinoma (TCC). Archival material from the VMDB (Table 15.1) indicated a higher prevalence of epithelial tumors (77%); 23 of 30 primary renal tumors in cats were epithelial, and all of these were malignant; 7/30 were mesenchymal (23%). There were 252 diagnoses of feline renal tumors, and more than 95% of all tumors were classified as malignant; most of these were lymphoma. A report of relatively few cases in cats indicates renal cell tumors are more frequent than in dogs.13 Renal cell tumors are rare in other species. Most primary tumors are epithelial, and they are largely reported as benign in cattle3 and malignant in horses.4 In 586 tumors from cattle in Brazil there were nine renal tumors, six carcinomas and 3 adenocarcinomas.14 The VMDB population was 73,195 cattle, with 16 tumor diagnoses, 3 primary renal and 13 secondary; 7265 sheep with two tumors, both secondary; and 7764 goats with two tumors, one each primary and secondary. Twenty‐nine tumor diagnoses were retrieved through the VMDB from a population of 134,268 horses; 11 were primary and 18 secondary. From 1953 to 1976 four cases were found in 3633 equine autopsies (0.11%) at Cornell: two RCCs, one renal adenoma, and one mesenchymal tumor.15 The reported incidence of 0.11% for equine renal tumors and 0.055% for RCC is similar to citations in the older literature of 62 in 40,000 necropsies.15 A 2009 report summarizes findings in 27 horses with primary renal tumors; 23 cases were from reports in the literature and 4 were new cases. Metastases were reported in 19 of these horses.4 Most articles about renal tumors list the different histologic diagnoses but generally do not separate clinical information by histological diagnosis. Those that provide this information by tumor diagnosis indicate the differences are slight or none and the number of animals in each group are small. The information on clinical characteristics and incidence in this chapter is reflective of renal tumors in general, rather than for specific histological types of tumors. There are no specific clinical signs or laboratory data to suggest a primary renal tumor is present in an animal. Hematuria is the most common laboratory abnormality and obviously there are many more likely causes for this than a tumor in the kidney(s). A few cases of polycythemia have been reported in the dog, cat, and horse. Azotemia is not common and is not expected, as rarely would neoplasia (other than lymphoma) involve greater than 75% of the total renal mass. This is a rare tumor in domestic animals, and when found it is usually an incidental lesion at autopsy or slaughter because these tumors are clinically silent.3,15,16 Of 894 primary renal cancers in dogs, there were 26 (3%) classified as adenoma and in cats 4/120 (3%) (Tables 15.2 and 15.3). Extrapolating from the incidence of renal carcinomas the actual incidence of renal adenomas in dogs and cats is probably less than 0.1%. Adenomas are reported in all species and they are less common than carcinomas. Progression of hyperplasia to adenoma to carcinoma is not known in domestic animals. However in dogs with hereditary dermatofibrosis there appears to be a progression of renal lesions from cysts to hyperplasia, adenoma, and carcinoma. It is logical this would occur in other renal epithelial tumors. There are no studies on actual incidence, age, breed, or sex predilection for renal adenomas, although there is some data for renal tumors in food animals. Over an 11‐year period, 20 renal cell tumors were identified in approximately 13,500 cattle processed for slaughter.3 All 20 cows were adults, 2–20 years of age. Nineteen of them had multiple tumors; 11 of these were visualized grossly, 8 others were microscopic, and 19 were classified histologically as carcinomas. Only one tumor was classified as an adenoma (8 × 10 mm single lesion in one kidney), however, only 1 of 19 carcinomas metastasized suggesting, despite the classification as carcinoma, they behave in a benign manner. The authors discussed the difficulties of distinguishing adenoma from carcinoma. They suggested the renal tumors were carcinomas, based on the multiplicity of tumors, and indicated they may develop in multiple sites within the kidneys. The occurrence of metastases in only one cow was clearly a different pattern from the high rate of metastases seen in renal carcinomas in humans, dogs, cats, and horses. A survey of 6706 cattle over a 45‐year period in Brazil yielded 586 tumors, of which 9 were in the kidney and none were adenomas (6 carcinomas, 3 adenocarcinomas).14 Adenomas are discrete, solitary, tan to white tumors located in the renal cortex (Figure 15.1A,B). In dogs and cats they are small (usually <2.5 cm) but size of the tumor is dependent on the biologic behavior of the neoplasm and size of the dog or cat. A 2.5 × 2.5 cm renal tubular tumor in a cat did not metastasize or invade adjacent tissues and had a benign histologic appearance with a low MC. This cat also had hypertrophic osteopathy.17 In horses and cattle, adenomas can be large (>6 cm) and have central areas of hemorrhage and necrosis. Tumors may only be found if the kidneys are sliced in multiple areas. On cut surface they are well demarcated and bulge, and only larger neoplasms have discolored necrotic centers. Bilateral and/or multiple adenomas occur, especially in dogs18 and cows.3 An inherited disease in German shepherds is associated with dermatofibrosis and renal adenomas. The adenomas in these dogs are almost always multiple, and there can be concurrent cysts or renal adenocarcinoma, partly dependent on the age of the dog and how long the lesions have had the opportunity to progress.18 Renal adenomas are non‐encapsulated but sharply demarcated from the adjacent cortex. They are composed of well‐differentiated tubules and acini that may be subclassified as tubular, papillary, or solid, based on the major histological pattern: central or elongated lumina (tubular type), papillary growths of varying sizes that project into lumens (papillary type) (Figure 15.2A), or solid sheets. Mixtures of all three types can occur, and this is also seen with renal carcinomas. Cytological and nuclear features are uniform and benign. A single layer of cuboidal epithelial cells with ample eosinophilic cytoplasm will line tubules or papillary projections. Nuclei are single, placed centrally or basally, and have a single nucleolus; mitotic figures are not observed, or found rarely. There is a paucity of supporting stroma, usually consisting of thin strands at the base of tubules. Corpora amylacea are common in renal cell tumors of cattle. Renal mucous gland cystadenomas were described in a 35‐year‐old horse. Grossly, the lesions were cystic, 1–4 cm in diameter, and found in the medulla and cortex of one kidney.19 Histologically, the tumors were well differentiated, compressed adjacent parenchyma, and were lined by tall epithelial cells that contained mucinous material, which was also present in the lumina of the cysts. The origin of the tumor was designated as the mucous glands common in the pelvis and proximal ureter of horses. The histological distinction of renal adenoma from carcinoma is subjective for tumors in the 2 cm range, and because carcinomas may be well differentiated, they each have similar histologic types and neither has a capsule. As pathologists, it seems likely we favor a diagnosis of well‐differentiated carcinoma over adenoma for tumors in the “gray zone” and especially if multiple tumors are present. The easiest solution is to count mitotic figures.1 Regardless of the designation, if the MC is <10/10 HPF the biologic behavior of the tumor is likely to be benign and survival time long (years), at least in dogs. Adenomas are small, sharply demarcated, noninvasive, and composed of well‐differentiated epithelium with no or few mitoses. Carcinomas that are large and infiltrative and have cellular and nuclear pleomorphism are easy to classify. Gradations of the two will make distinction difficult in some cases but MC has been shown to correlate with survival in dogs.1 An arbitrary cut‐off of 2 cm as a criterion of malignancy has been used: less than 2 cm adenoma, greater than 2 cm carcinoma. The use of size is based on older information from renal tumors in humans and is not a sound generalization for tumors in animals.20 Certainly small tumors, less than 1 cm, are likely to be benign in any species, but neoplasms larger than 2 cm may also have a benign course.17 In dogs and cats this generality is a guide and in cattle most renal tumors are greater than 2 cm in diameter but they rarely metastasize. The best criteria to distinguish benign versus malignant are metastases, increased MC, invasion, anaplastic cellular and nuclear features, as well as historical data for each species. Neither adenomas nor carcinomas have a capsule. Multiplicity of gross tumors suggests a malignant classification or metastases from a nonrenal organ (lung, prostate, or mammary). However, multiplicity of renal tumors is also associated with benign biologic behavior in several species. These are malignant epithelial tumors without features of embryonal differentiation. They have also been classified as RCC, malignant nephroma, clear cell carcinoma, hypernephroma, and Grawit’s tumor.6 Many tumors will have well‐differentiated tubules or papillary structures that make their epithelial identification easy. Solid variants and sarcomatous regions also occur and one of the most common histologic characteristics is a mixture of various cell types. Renal carcinoma is an uncommon tumor in domestic animals; however, it is the most common primary renal tumor in dogs, cats, and horses. The reported incidence for dogs is 1.5 in 100,000 and for cats 0.7 in 100,000.21 Other reported incidences include 0.3–1.5% for dogs8 and 0.2–0.5% for cats.12 A retrospective study found 4 (0.05%) in 8149 canine tumors and 3 (0.23%) in 1299 feline neoplasms, suggesting that in this limited number of cases renal carcinoma was 4.5 times more frequent in cats than in dogs.13 Table 15.2 summarizes five studies, and of 894 canine neoplasms, 510 were classified as RCC (57%), 48 (5%) as transitional (urothelial) cell carcinoma, 11 (1.2%) as squamous cell carcinoma (SCC) and only 2 as undifferentiated carcinoma. There were 601 epithelial tumors and 569 (94%) were classified as malignant. Overall 86% of the epithelial and mesenchymal tumors were classified malignant. One study reported pulmonary metastases were detected in 10/61 (16%) at the time of diagnosis and at the time of death 37/49 (76%) had metastases.2 Metastases were found in 70% of the dogs with carcinoma, 88% with sarcomas, and 75% with nephroblastomas.2 Table 15.3 summarizes three studies in cats, and of 120 neoplasms 55, 46% were classified as RCC, 11 (9%) as urothelial cell carcinoma (UC), and there were no undifferentiated carcinomas. A search from databases of four veterinary colleges and one private referral practice during a 6‐year interval identified 19 primary renal neoplasms in cats.5 There were 13 renal carcinomas (11 tubular; 2 tubulopapillary), 3 UC, and 1 each nephroblastoma, hemangiosarcoma, and adenoma.5 Abattoir surveys report rates of 8.5, 4.3, and 0.9 renal tumors per million animals in cattle, pigs, and sheep, respectively.22 During an 11‐year period, 20 renal cell tumors were identified in 13,500 cattle processed for slaughter. Nineteen were classified as carcinoma even though only 1 tumor metastasized; 1 tumor was classified as an adenoma, a single 8 × 10 mm mass.3 The diagnosis of carcinoma was based on multiplicity of tumors. A survey of 6706 cattle in Brazil over a 45‐year period reported 586 tumors, 9 of which (6 carcinomas and 3 adenocarcinomas) were in the kidney; all 3 adenocarcinomas metastasized to lung and renal lymph node and no information was provided on the other 6.14 There were 35 tumors in the urinary bladder.14 Tumors in the kidneys of horses are rare; if primary or secondary tumors occurred frequently they would be found on routine autopsies. Only 4 cases were found in 3633 equine autopsies (0.11%) at Cornell over a 23‐year period: 2 RCCs, 1 renal adenoma, and 1 mesenchymal tumor.15 The reported incidence of 0.11% for renal tumors and 0.055% for RCC is similar to citations in the older literature of 62 in 40,000 necropsies.15 A 2009 report summarized clinical characteristics in horses with primary renal tumors. Twenty‐three cases were from reports in the literature and 4 were new cases.4 The diagnoses were renal carcinoma 18, renal adenocarcinoma 6, and 1 each tubular adenocarcinoma, papillary adenocarcinoma, and transitional cell carcinoma (TCC) in the renal pelvis. Metastases were identified in 19 of the horses. Renal carcinomas are reported to occur in middle‐aged male dogs (mean age 8–9 years),2,8,9 but can be seen in dogs less than 6 years of age and tumors in younger dogs may behave more aggressively.1 The only breed predilection is German shepherds, with hereditary dermatofibrosis and renal tumors. Most reports suggest a male predominance in dogs, approximately 2:1. In humans, the incidence of renal tumors is 2–5 times greater in males.6 In cats, renal tumors (all types) tend to occur in the 8‐ to 11‐year‐old age group (range 2–13 years). Some studies report tumors in cats are more common in males10 and others report equal numbers in male and female cats.5 Horses with renal tumors range in age from 4 to 25 years old (mean 13 years); no breed or sex predilections are reported.4,23,49 Twenty cows with renal cell tumors ranged in age from 2 to 20 years.3 Clinical signs are vague and by the time dogs present, between 20 and 50% will have a palpable abdominal mass and about the same will have evidence of pulmonary metastases via imaging.1,2 Some dogs will have cachexia, which is associated with a negative outcome. In all species the chief complaints and clinical signs associated with primary or secondary renal tumors are nonspecific. In dogs and cats, reported clinical problems are an abdominal mass (20–50%), weight loss, pollakiuria, and nonspecific problems such as lethargy, vomiting, and anorexia.2,5,8–11 By the time these tumors produce clinically detectable problems in dogs, cats, and horses the tumor is advanced. Metastatic disease can be detected with thoracic radiographs in approximately 15–50% of dogs with renal carcinoma at initial presentation.2,24 Abdominal radiographs will detect a mass in 80% of the cases, and in 50% of these the mass can be identified in a kidney.9,10 Ultrasound or intravenous pyelograms are reported to correctly detect a mass in 80–100% of the cases, depending on the host and location of the tumor.2,9 There are no specific clinical pathology abnormalities that suggest a renal tumor is present. Several paraneoplastic syndromes have been reported with renal tumors: polycythemia vera, paraneoplastic leukocytosis, hypertrophic osteopathy, hypoglycemia, and hypercalcemia. The most common abnormalities are hematuria (50–100%) and pyuria (50%).2 Proteinuria is reported in a high percentage of cases, but the majority or all of these patients have concurrent hematuria and/or pyuria. Hemorrhage with or without inflammation is the most likely explanation for proteinuria in these cases. Distinction between glomerular and hemorrhage‐induced proteinuria for patients with renal tumors is not reported. Azotemia is present in approximately 25% but probably involves superimposed prerenal (dehydration) or concurrent renal disease, as the majority of the total renal mass is unaffected if the only lesion is renal neoplasia.2,9 Lymphoma would be one of the few tumors that could destroy enough functioning renal mass to cause renal azotemia. One report provides information on clinical characteristics, laboratory data, and diagnostic evaluation by histologic groups.9 Most results are similar across histologic groups; notable differences were that hematuria was more common with TCC than with renal tubular cell tumors, and none of the eight dogs with TCC had detectable metastases at the time of initial diagnosis. Searching for tumor cells in urine is generally futile but occasional successes are possible. Anemia is reported in approximately one‐third to half of the animals with renal tumors. Mechanisms are not identified but several factors are probably superimposed, including anemia of chronic inflammatory disease, hemorrhage, and hematuria; however, decreased erythropoietin seems very unlikely as there is so much remaining renal mass. Secondary absolute polycythemia has been reported infrequently with renal tumors in dogs and cats.5,25,26 The increased red blood cell mass is likely due to the production and secretion of erythropoietin or erythropoietin‐like peptide from the tumor. Packed cell volume will be increased in the range of 60–70% in dogs and will return to reference range after tumor removal. This syndrome has been seen with renal adenoma, carcinoma, nephroblastoma, TCC, fibrosarcoma, and lymphoma as well as non‐neoplastic lesions such as cysts and hydronephrosis.5,25,26 Paraneoplastic leukocytosis of >100,000 WBC/μL has been reported in a few dogs with renal tumors.2,27 In one case production of granulocyte–macrophage colony‐stimulating factor by the tumor cells was documented and the marked leukocytosis resolved after nephrectomy.27 This dog also had concurrent hypertrophic osteopathy, which has been reported in a few other dogs with primary renal tumors. This unusual bone lesion is seen more frequently with tumors in the urinary bladder. Hypercalcemia was attributed to the production of parathyroid‐related protein (PTHrP) by a renal angiomyxoma28 and hypercalcemia has been associated with a chromophobe carcinoma in a dog.29 Increased serum gamma glutamyltransferase (GGT) was present in a dog with renal adenocarcinoma and it decreased gradually following nephrectomy and then increased associated with growth of pulmonary metastases.30 Bilirubin and other hepatic enzymes were not increased and albumin and urea nitrogen were not decreased. This observation is interesting and the patterns of a substance increasing, decreasing, and increasing with presence, removal, and recurrence of a tumor type is how hypercalcemia of malignancy was first discovered. GGT is located in the brush border of proximal convoluted tubules and has been measured in the urine of dogs to indicate nephrosis. The authors cited two similar cases in dogs that had increased serum GGT and renal tumors. Increased serum GGT and isoenzymes of GGT are seen in some cases of renal adenocarcinomas in people. In a fairly large series of renal tumors in dogs GGT was increased in 6 of 64, but alkaline phosphatase (ALP) was increased in 21 of 65.2 Hepatic enzymes in different combinations (ALP, alanine transaminase (ALT), aspartate transaminase (AST), GGT) may be increased in 10–25% of dogs due to secondary liver diseases or stress. It will be interesting to see if others can identify cases of a unique increase in GGT with primary canine renal tumors. A report of 27 horses indicated that ages of affected horses was 4–25 years old (mean 13 years); weight loss, anorexia, hematuria, and colic were the most common clinical problems. Clinical histories were varied and not specific. Routine laboratory data was not specific. The tumors were epithelial, malignant, and metastases were present in the majority of horses.4 Hematuria was present in 94% of horses with a urinalysis and hypoglycemia was present in 19%. Paraneoplastic hypoglycemia is reported in a horse with RCC.31 Rectal palpation and/or imaging studies revealed a mass or an enlarged kidney in the majority of these horses. Four of five horses had pulmonary nodules seen in radiographs of the thorax. Nineteen had metastases at autopsy; the most common sites were lungs (16) and liver (14) and several cases had metastases to bone and/or muscle.4 In ruminants, the tumors are usually asymptomatic or at least they are undetected ante mortem. Most carcinomas are unilateral, but they can be bilateral and multiple; neither kidney is more predisposed.3,4,18,32,33 They are well‐demarcated masses, located in the cortex, yellow, tan‐brown to cream colored, often located at one pole, and they vary considerably in size, from 2 cm in diameter to occupying greater than 80% of one kidney (Figure 15.1C). Large neoplasms that invade the renal capsule and enter the retroperitoneal space and encompass and/or invade the adjacent adrenal gland are easy to classify as carcinoma. The smaller the primary tumor the more difficult it is to differentiate carcinoma and adenoma. Large masses have expected areas of necrosis and hemorrhage and are friable due to little supporting stroma. They may also invade the renal pelvis or enter blood vessels (Figure 15.1C,D). Rarely, they infiltrate through the retroperitoneal space and spread by implantation metastasis through the abdomen but can occur in all species. Multiplicity of renal tumors occurs dogs, cattle, and humans and smaller tumors may only be found during microscopic examination of random sections. Nineteen of 20 cows with renal cell neoplasia had multiple tumors; 11 of these were visualized grossly, and 8 others were microscopic. Canine tumors can be cystic, and the cysts contain variable amounts of clear or red‐brown fluid. The report of 27 horses indicated that the tumors were unilateral: 15 left kidney, 8 right kidney (4 not specified). Four of these horses had tumor in the contralateral kidney, but these were interpreted as metastases. Bilateral renal tumors occur in other species and there is no clear means to distinguish multicentric origin from metastases. None of the retrospective reports indicated the primary tumor in horses invaded other parts of the urinary tract. Renal carcinomas can be subdivided into histological and cytological types. However, to predict biologic behavior and estimate survival times for dogs with RCC use MC.1 Histologic types of RCC seen in animals are solid, tubular, papillary, or cystic (multilocular cystic). A study of 70 dogs indicated 34% were solid, 21% papillary, 24% tubular, 4% multilocular cystic, 9% clear cell, and 9% chromophobe. Forty percent of the total had mixed patterns and 7% had sarcomatoid features. Tubular is the most common type in other animals, and clear cell type accounts for 75% of RCC in humans. Each histological type can be classified cytologically as chromophobic, eosinophilic, or clear cell, and mixtures of all three are typical. The predominant histologic and/or cell type should be used to classify the tumor and areas of sarcomatoid change and cysts can be described. The classification used in human tumors is based on numerous cases that have long‐term follow‐up and molecular profiles correlated with morphology. We can characterize the morphology of tumors in animals but in veterinary pathology it is very difficult to gather large case series that have enough overall tumors and examples of each variant, to avoid sample bias. The largest study of RCC in dogs used case materials gathered from the United States and Canada to find approximately 200 cases of renal tumors over a 16‐year period, from which 70 cases had adequate tissues, clinical data and follow‐up information that was archived or could be determined. The number of cases per group was small, but the types of RCC that appeared to have prognostic value in dogs were: clear cell carcinoma (6 cases), which had decreased median survival time (MST) of 87 days, increased risk of death and 3 of 6 had metastases; multilocular cystic tumors (3 cases) which appeared to have longer survival time and a more benign course.1 Clear cell type RCC is seen more frequently in laboratory animals and human beings, whereas in dogs only 6/70 (9%) were clear cell carcinoma in the same study,1 and they are observed rarely in cattle. They are easy to recognize with H&E by their “clear” cytoplasm, which is due to high glycogen and lipid content (Figures 15.3 and 15.4). They are usually in solid rather than tubular RCCs. Cell borders are distinct, cytoplasm is abundant, and nuclei are round and dense, providing the appearance of a well‐differentiated tumor. The clear cytoplasm and solid growth that resembles adrenal cortical tissue led to the name hypernephroma. Foci of clear cells can be found in many renal cell tumors if they are searched for and will help suggest renal origin if the location of the primary tumor is uncertain. In dogs they are reported to have cytoplasmic staining for vimentin and are negative for CD117 (KIT).1 They are derived from the proximal convoluted tubules, hereditary, and associated with mutations in the VHL gene.34 Approximately 75% of renal carcinomas in humans are clear cell type, making them the most common renal tumor in humans by a wide margin, and as in dogs they shorten lives and metastasize. Chromophobic cell variants have cuboidal cells with granular, lightly to intensely eosinophilic, abundant cytoplasm and distinct cell membranes. They stain positively for colloidal iron1,29 and are immunoreactive to CD117 (KIT), encoded by the c‐KIT proto‐oncogene.1 They may be in found in any histologic type of RCC. They often form trabeculae of varying sizes, with or without lumina. They should be distinguished from oncocytomas, which are solid tumors filled with cells of similar appearance. Chromophobe RCC was documented in a 12‐year‐old female husky that had hypercalcemia.29 The tumor was 8 cm in diameter and it contained a mixture of epithelial and mesenchymal cells. The epithelial cells had abundant, cloudy eosinophilic cytoplasm that resembled cells seen in chromophobe RCCs of humans and in oncocytomas of humans and dogs. Differentiation between oncocytomas and chromophobe RCCs may require histochemistry, IHC, and electron microscopy. Oncocytoma and chromophobe RCC may be the same tumor. Chromophobe RCCs stain positively for colloidal iron and KIT, negatively for periodic acid Schiff (PAS) and should not be packed with mitochondria. Oncocytomas should be negative for colloidal iron but positive for PAS and should contain numerous mitochondria. In the aforementioned chromophobe RCC in a husky, the epithelial component was negative for vimentin and PAS and positive for colloidal iron and KIT. A subpopulation was positive for CD10 and results for cytokeratin were not reported.29 The only cells positive for vimentin were the sarcomatoid elements. Because chromophobe RCC resembles oncocytoma the authors encouraged separation of these differentials. Chromophobe RCCs and oncocytomas are negative for vimentin. Chromophobe RCCs, collecting duct carcinomas, and oncocytomas are derived from collecting ducts and may look similar histologically. The eosinophilic cell type is a variant of chromophobe tumors in which the cytoplasm is intensely eosinophilic. Cells are cuboidal shaped, cell borders are distinct, and they form trabeculae of various widths, usually without lumina. This is the most common variant in cattle (19/20 cattle tumors are eosinophilic, 1/20 are clear cell). The eosinophilic variant and oncocytomas look similar with H&E. Oncocytomas have more numerous mitochondria, an absence of cytoplasmic vesicles, and they are positive to cytokeratin but negative to vimentin.32 The majority of RCCs in humans and dogs are cytokeratin and vimentin positive, which is a unique characteristic.6,32 Chromophobic and eosinophilic tumors stain light blue with histochemical stains for colloidal iron.6 Tubular RCCs have numerous tubules and acini of various sizes that are the predominant histologic features. Epithelial cells lining tubules range from well differentiated to anaplastic (Figure 15.2E). Lining cells are usually singular. More anaplastic tumors have cells piling together irrespective of adjacent cells and may invade adjacent tissues. Some lumina will contain homogeneous secretory material and/or sloughed cells. There may be regions with papillary formation or solid proliferations devoid of lumina. Stroma is minimal and desmoplasia absent. These are common in cattle and although designated carcinoma they do not metastasize. Papillary tumors have distinct branching papillae projecting into clear spaces of varying sizes. They look well differentiated at first observations with low and medium magnifications. Papillae may be small and singular within a space or have multiple branches (Figure 15.2A–D). In cross‐section they appear as a circle with cells lining the outside and a central core of stroma and a blood vessel. Solid tumors contain epithelial cells apposed closely together. They do not form tubules, acini, or papillae, although foci with these structures may be seen in regions of the tumor. Cell borders are usually distinct and cytoplasm stains lightly. There is considerable variation in cell sizes between different tumors but within one tumor, cells and nuclei tend to be fairly uniform. This is a common pattern in dogs (Figures 15.3 and 15.4). Cystic and multilocular cystic are likely variants in which cysts are the predominant feature. Cysts are of various sizes and may be empty or contain a lightly stained homogeneous product (Figure 15.2G). They are lined by benign‐appearing epithelial cells. Counting mitotic figures, or at least producing reliable counts, is problematic in these tumors, as a majority of the tumor may be acellular or paucicellular, clear or fluid‐filled spaces. MC should be started in regions that have mitotic figures or if none are found then start in a solid region. Crossing regions with cystic spaces is inevitable. Multilocular cystic is a variant used in human RCC and has been used in dogs. The canine tumors were described as “multiple, variably sized cystic spaces separated by fibrous septa lined by cuboidal epithelium” and 3 of 70 cases were so designated.1 Although results are not statistically significant, multilocular cystic type may be associated with longer survival times.1 Sarcomatoid features can be found in some renal tumors (10%) and when seen there is an obvious transition from one of the epithelial patterns into a spindle cell tumor with high cellular and nuclear density. A collecting duct carcinoma in a dog has been described histologically, histochemically, and via IHC and compared to the same diagnosis in humans.35 The tumor contained papillary and tubular structures and others may have considered this a tubular RCC. Without histochemistry or IHC, the derivation from collecting ducts would not be appreciated. The neoplasm had cellular and nuclear pleomorphism, a high mitotic count, and it invaded regionally and metastasized widely.35 Collecting duct carcinomas are highly malignant in humans. Interstitial stroma ranges from mild, with just enough stroma for tubules to rest on, to marked desmoplastic reactions. Occasional RCCs will have marked desmoplasia encasing tumor cells. If there is retraction of the connective tissue from the tumor cells it gives the false impression on intralymphatic emboli. Capsules are absent or incomplete. Tumors tend to grow by pushing at the edges more than by infiltrating. When tumors infiltrate and extend beyond the kidney, the diagnosis of carcinoma is straightforward. Hemosiderin, proteinaceous secretions, and corpora amylacea are features of renal cell tumors in cattle. Histologic diagnoses in 27 horses with primary renal tumors were 18 carcinomas, 6 adenocarcinomas, and 1 each of tubular adenocarcinoma, papillary adenocarcinoma, and TCC in the renal pelvis.4 In dogs, 50–70% metastasize,2,8,9 whereas 70% metastasize in horses,4 50% in cats,5,10 and only 5% in cattle.3 The most likely sites are lung, liver, and regional lymph node. Other sites reported are serosal surfaces and ipsilateral adrenal gland.8,9 Metastases to the skin is an unusual site for any tumor. However, this site has been reported with renal tumors and urothelial carcinomas in dogs. Metastases are common in horses and are widely disseminated.4,23,36 Four of five horses had pulmonary nodules seen in radiographs of the thorax. Nineteen had metastases at autopsy; the most common sites were lung (16) and liver (14) and several cases had metastases to bone and/or muscle. The high rates of metastases in most species is attributable to the biology of the tumor and the finding that RCCs are clinically silent, so by the time they are detected the tumor has months or years to develop its full malignant potential. In humans, bone is a common site for metastases of RCC and RCC is one of the most common tumors in humans to metastasize to bone.37 One of the earliest reports indicated that 37 of 54 dogs with renal tumors survived less than 21 days; however, it is often not clear how decisions from owners influence survival data.9 Some dogs in that study survived 8–24 months,9 while a median survival of 16 months with a range of 0–59 months is reported in another.2 Another study in dogs indicated 20 of 29 carcinomas, 14 of 16 sarcomas, and 3 of 4 nephroblastomas had metastases at death.2 All groups of renal neoplasms in this study exhibited malignant behavior and no difference in survival could be detected between groups of tumors or within histologic subtypes.2 The survival ranges for carcinoma was 0–59 months (median 16 months) and for sarcoma 0–70 months (median 9 months). Metastases may be as fully differentiated as the primary tumor. Less‐differentiated varieties have the expected features of reduced cytoplasmic area, indistinct cell borders, nuclear crowding, multiple cells piled together, various sizes and shapes of nuclei, mitoses, and vesicular chromatin (Figure 15.3D). Some tumors will have very high MC (up to 60 mitoses/10 400× fields, 2.37 mm2). The more differentiated varieties resemble adenoma and have one or two layers of well‐differentiated eosinophilic cells lining tubules (Figure 15.2E). Long survival times of 4–5 years in some dogs with a diagnosis of RCC suggest a propensity to diagnose primary renal epithelial tumors as malignant, but their clinical behavior may be benign. Using MC will help identify RCCs with different biologic behaviors.1 In a study of 70 dogs the authors used MC cut‐offs of <10, 10–30, and >30 in 10 400× fields using oculars with FN22 and therefore a field‐of‐view area of 2.37 mm2. Dogs with an MC of <10 (n = 36) had a significantly increased median survival time (MST) of 1184 days, dogs with intermediate MC (n = 14) had an MST of 452 days, and dogs (n = 20) with tumors >30 had an MST of 147 days and an increased hazard ratio for experiencing death or euthanasia, regardless of histologic or cytologic classification or nuclear grade. MC was the only independent prognostic variable found using multivariate analysis. MC should be used when evaluating a biopsy or nephrectomy specimen to help predict biologic behavior, at least survival times in dogs. Some of the low‐grade RCCs with 0–3 MC may have been considered adenoma, Regardless of adenoma versus low‐grade carcinoma, MC provides a useful prognostic parameter that is easy to determine. The nuclear grading system is one designed for RCC in humans that uses defined, subjective nuclear and nucleolar characteristics that are not as easy to determine as MC.38 When applied to the dogs of this study it revealed different patterns of outcome: grade 1 – 6 dogs, MST not reached; grade 2 – 28 dogs, MST 1065 days; grade 3 – 31 dogs, MST 379 days; and grade 4 – 5 dogs, MST 87 days.1 The trend is clear that as nuclear and nucleoli variabilities increase survival times decrease. The 6 dogs with grade 1 RCC had survival times of 136, 362, 451, 718, and 955 days, and 1 dog was alive at 3703 days. All material was selected from biopsy or nephrectomy specimens. Metastasis and recurrence data could not be determined for all cases. Metastasis detected by imaging at the time of initial diagnosis was noted in 5/44 (11%), and an additional 10 developed radiographically detectable pulmonary nodules in 2–23 months (15/44 (34%)). Fuhrman et al. defined four nuclear grades (1–4) in order of increasing nuclear size, irregularity, and nucleolar prominence. The study found that nuclear grade was more effective than other parameters examined in predicting development of distant metastasis following nephrectomy of humans with various RCC.38 Future studies should employ this grading scheme on animal tumors to confirm its utility and determine if it and MC are needed to predict behavior of RCC in animals. Presently, determining the MC and type of RCC should suffice for biopsy or nephrectomy specimens. Short survival times are reported in horses (0–1 year),4 but for all species an obvious reason for short MSTs is euthanasia at the time of diagnosis, which can be for a variety of reasons. Overall, the prognosis for renal carcinomas in dogs, cats, and horses is poor, as the majority of cases has or will develop metastases. However, long‐term survival is possible with nephrectomy when tumors are unilateral, not large, and have not metastasized.1,2 It can be difficult to decide if a small tumor is benign or malignant. Large tumors are often easy to classify because the tumor has already metastasized or invaded the renal pelvis, capsule, or vessels. Although tumor size has been stated to be a criterion (<2 cm adenoma, >2 cm favors carcinoma), this is an oversimplification from the older literature.20 A single tumor less than 2 cm in diameter is likely to be benign in all species, but renal tumors in cattle3 behave in a benign fashion even though tumors can be as large as 60 cm. Only 1 of 20 renal cell tumors studied in cattle metastasized; most had minimal cellular atypia yet were classified as carcinoma based on their multiplicity. Two of the 20 had marked cellular pleomorphism; one of these metastasized and one did not. The one tumor classified as an adenoma was 8 × 10 mm and was confined to one kidney. Characteristics that favor carcinoma are large size, large areas of necrosis, infiltration, invasion of parenchyma or vessels, increased MC, and nuclear atypia. Only MC and nuclear grade are correlated with long‐term follow‐up studies.1 Classification by histological or cytological subtype is not predictive of biological behavior in domestic animals. Multiplicity of primary renal cell tumors is not a reliable criterion as this occurs in 33% of canine and in 95% of bovine cases. If a renal tumor metastasizes to the opposite kidney, there are invariably metastases in other organs. Most tumors located in the kidney are easy to diagnose via H&E sections and gross patterns. Diagnostic challenges can be to determine the biologic behavior of a primary renal tumor or to find the primary if there are multiple epithelial tumors. Adenoma and tubular RCCs resemble similar tumors that originate in lung, mammary, and prostate glands. If lesions are confined to the kidney, especially the cortex, and/or regional lymph nodes, a diagnosis of primary RCC is straightforward. Renal tumors that are multiple and bilateral yet do not metastasize to other locations are primary renal cell tumors, probably of multicentric origin. If a renal tumor metastasizes to the opposite kidney, there are metastases in other organs. When metastases are widespread, the distinction is less easy and the criterion more subjective: location of the largest tumor suggests that organ as the primary site, and metastases to the kidney will be multiple, not singular. Renal carcinomas invariably arise in the cortex, and lesions in the medulla favor a metastatic tumor. If the primary source is unknown and IHC is desired, it is worth noting that, while not specific, dual positivity for cytokeratin and vimentin is suggestive of renal origin. Finding an IHC pattern of positivity for uromodulin, PAX8, napsin A, and neprilysin confirms renal origin. Foci of clear cells are also consistent with renal origin. Corpora amylacea are features of renal cell tumors in cattle. An unusual feature of most renal cell tumors is that they dual mark with cytokeratin and vimentin. This is also seen in tumors that originate in prostate, ovary, or mesothelium.39 In RCCs this may be due to their origin from primitive renal blastema or an epithelial–mesenchymal transition. An epithelial–mesenchymal transition may assist tumor cells to dissociate enabling invasion and/or metastasis and may have prognostic significance.40 Kidneys originate from mesenchymal blastema, which reacts positively to vimentin but not cytokeratins. The blastema differentiates into epithelium and loses some vimentin reactivity as tubules and parietal epithelium develop positive staining to cytokeratins. The coexpression of vimentin and cytokeratins may represent the embryonic mesenchymal blastema, or an epithelial–mesenchymal transition in neoplastic cells, or an origin from renal stem cells. Only the parietal epithelium of Bowman’s capsule and distal tubules of adult mammalian kidney express CK19. Uromodulin, a unique protein synthesized by the kidney41 has been a useful IHC marker for identifying and studying renal cell tumors in humans, cattle, and dogs. Two studies found positive uromodulin immunoreactivity in all renal cell tumors in cattle (20) and dogs (13).3,42 The canine tumors were unilateral and single masses, 9 were clearly epithelial and 4 were solid tumors, 1 of which was classified as sarcomatoid. All 13 were positive for uromodulin, 11 were also positive for vimentin, 12 positive for KIT, 9 positive for wide‐spectrum screening cytokeratins, and 7 were positive for carcinoembryonic antigen.42 Only distal convoluted tubules in the non‐neoplastic tissue adjacent to tumors were positive for uromodulin. Normal stromal cells were positive to vimentin and the cytoplasm of tubules, ductules, and urothelium were positive for cytokeratins. One tumor, a tubulopapillary carcinoma did not stain positively with any of the four epithelial antibodies used, but it was positive for vimentin and uromodulin. All three solid tumors were negative for cytokeratins but the sarcomatoid tumor was positive. The two papillary–cystic carcinomas were the only two tumors that were negative for vimentin. Two tumors, both classified as papillary–cystic had low MC of 5/10 HPF, whereas the average MC of the other 11 was 36/10 HPF. Clinical assessment at the time of nephrectomy did not reveal metastases but there were no follow‐up data, clinical data, or autopsies to correlate IHC or histologic results with clinical outcomes. The presence of KIT in 12 of 13 tumors suggests that mutations of the c‐KIT proto‐oncogene may have a role in renal oncogenesis. KIT expression is found in chromophobe renal tumors in humans and was present in the canine renal tumors regardless of histologic type. Only 3 of 13 tumors had CD10 (neprilysin) reactivity, which is in the brush border of renal tubules and can be found in types of renal carcinomas in humans, in 90% of clear cell types, and approximately two‐thirds of papillary tumors.42 The IHC pattern for a collection duct carcinoma in a dog has been reported and compared to the human counterpart.35 Although the numbers are small, uromodulin is a useful “IHC marker” to identify a renal cell tumor in dogs and cattle. It will be necessary to know if any other epithelial tumors stain positively or if uromodulin is specific for renal epithelium. Napsin A is proteinase expressed in the lung and renal proximal convoluted tubules and has been used in human RCCs to help identify renal origin.43 Neprilysin (CD10) is expressed in lymphoblastic tumors, germinal cells, and proximal convoluted tubular epithelium. In humans, CD10 is present in 90% of clear cell renal carcinomas and approximately two‐thirds of papillary tumors, but in the canine study it was only present in approximately 20% of the tumors.42 Sensitivity and specificity will need to be determined for the different antibodies in different species to help determine their diagnostic utility. Renal tumors are so uncommon in animals it will be difficult to accumulate enough cases with follow‐up data to determine if IHC helps determine prognosis or select treatments as it does in humans. Differentiation of RCC from UC can be done by anatomical location and histology. RCCs are in the cortex versus pelvis for UCs. Melamed–Wolinska bodies are characteristic of UC, widespread metastases favor UC, tubules and acini favor RCC, but there will always be some gray zone cases in which the distinction is difficult. IHC with uroplakin III (UP III) and p63 should identify UC, whereas uromodulin, neprilysin, napsin A, and PAX844 are all useful to identify renal origin (Figure 15.4). These antibodies and their application to tumors in domestic animals need to be thoroughly evaluated. In human tumors wide panels of antibodies are often needed to differentiate tumors that look similar.44,45 This is a rare tumor in animals, usually benign, composed of “oncocytes”. The histogenesis is not clear, but these tumors may arise from the intercalated cells of collecting ducts, at least in humans.6,32 The case reported in the dog was bilateral and invaded adjacent muscles; there were no clinically detectable metastases, but an autopsy was not performed.32 There is also one other report of an oncocyte‐like RCC in a German shepherd with dermatofibrosis.46 In humans, oncocytomas are considered benign.6 Microscopically, these neoplasms consist of solid areas, nests, or cords and tubules of closely packed, round to polygonal, monomorphic cells with intensely eosinophilic, granular cytoplasm. Nuclei are round to oval with coarsely stippled chromatin and a prominent nucleolus. There may be anisocytosis and anisokaryosis and bi‐ or multinucleate tumor cells. The diagnosis of renal oncocytoma is based on distinctive granular, eosinophilic cytoplasm; a positive PAS reaction; immunoexpression of cytoplasmic cytokeratin, but not vimentin; abundant mitochondria; and an absence of ultrastructural vesicles. The main differential consideration is the eosinophilic variant of a chromophobe RCC, which has ultrastructural cytoplasmic vesicles, few mitochondria, and is positive for colloidal iron.29 A canine chromophobe RCC was PAS negative, positive for colloidal iron, and had a sarcomatoid element that was positive for vimentin; cytokeratin and electron microscopy were not reported. It is our opinion that an oncocytoma is a default diagnosis for a tumor that has large, uniform eosinophilic cells packed with mitochondria and which can be found in different organs (e.g. endocrine, muscle). They are senescent cells but have a cell of origin. For the larynx, we know previously diagnosed oncocytomas are rhabdomyomas. Oncocytomas and chromophobe RCCs are both derived from intercalated cells of the collecting ducts and look very similar with H&E, therefore future studies should attempt to determine if these are, in fact, the same tumors. This unique and rare syndrome is hereditary in German shepherd dogs and produces multiple subcutaneous fibrous nodules that may precede or be concurrent with uterine leiomyomas, and/or multiple renal cysts, adenomas, or adenocarcinomas, unilateral or bilaterally.1–4 The entity is caused by mutations in the folliculin gene (FLCN, previously BHD gene) located on chromosome 5 and it has a dominant mode of inheritance.4–6 The genetic defect was first identified in dogs and then later in humans with Birt–Hogg–Dubé syndrome, an inherited renal cancer. Inactivation of this tumor suppressor gene is one of the critical steps in this disease and a similar disease in rats.5 One study linked the heritage of 43 dogs with this syndrome to one male.3 It has also been reported in the golden retriever, boxer, and mongrels and is seen more frequently in females. The average age of affected dogs is 8.5 years, with a range of 5–11 years. There is a report of the subcutaneous lesions occurring without renal tumors in a dog.7 The subcutaneous nodules vary from a few to hundreds and range from millimeters in diameter to large masses greater than 5 cm in diameter. They are well‐delineated, non‐encapsulated nodules of benign fibroblasts and associated collagen. The dermal portions tend to blend in with adjacent collagen, and the subcutaneous portions are well circumscribed. They can be found anywhere on the body but are most frequently seen along the limbs, back, and head. They produce a distinct bulging, palpable, mass in the subcutis, but larger masses may have an ulcerated surface. If lesions comparable to these are found, then additional tests should be performed to search for renal and/or uterine masses. Ten of 11 German shepherd bitches with this disease had multiple uterine leiomyomas.1 In the kidneys the epithelial tumors are bilateral, multiple, and/or cystic (43 of 45 dogs). They are sharply delineated, bulge on cross‐section, and vary from tan‐white to gray. Sizes range from a few millimeters to greater than 10 cm for the solid tumors and greater than 25 cm for the cystic portions. Non‐neoplastic cysts contain gelatinous, clear to red‐brown fluid and may rupture into the peritoneal space. Cysts may be the earliest lesion of this disease in dogs and they appear to progress through steps of hyperplasia, adenoma, and adenocarcinoma. Similar observations have been made in a rat model of this disease. Neoplastic transformation and growth of the tumors appears to be slow until the dogs are several years old. Some lesions may only be microscopic, and the smaller gross lesions consist of well‐differentiated epithelial cells that line tubules or cysts with occasional papillary projections into luminal spaces. Cysts are such a characteristic component of these tumors that they are often referred to as cystadenocarcinomas. Like other primary renal adenocarcinomas, these tumors have areas of tubular, papillary, or solid growth patterns, often admixed in one tumor. An oncocyte‐like variety has also been described.2 The solid areas tend to be more anaplastic; the cells are pleomorphic, ranging from cuboidal to spindle shaped, with bizarre nuclei and numerous mitotic figures. Metastases resemble the primary tumor, and they were detected in 10 of 23 dogs autopsied. The most common locations of metastases were sternal and renal lymph nodes, but metastases were also found in the peritoneum, liver, spleen, lung, pleura, and bone.1,2 Even though the skin lesions are benign they may be too numerous to treat. Renal lesions may also be benign (cysts and adenoma) but adenocarcinomas are expected. All these tumors occur rarely in the kidneys and when present usually arise from urothelium in the pelvis.8 Urothelium retains the embryonic potential to differentiate into glandular epithelium (secreting mucus) and squamous and transitional epithelium. In the urinary bladder UC may be subclassified and graded but this has not been applied to renal UC. Undifferentiated carcinoma is a term used when the tumor cells are so poorly differentiated that the cell of origin cannot be determined. It is not applied to renal cell cancers that can be classified but are anaplastic. Paraneoplastic leukocytosis of >100,000 WBC/μL was reported in a dog with UC of the renal pelvis. In this dog production of granulocyte–macrophage stimulating factor by the tumor cells was documented and the marked leukocytosis resolved after nephrectomy.9 This dog also had concurrent hypertrophic osteopathy. Papillomas are rare (1%), and they consist of papillae lined by one or a few layers of well‐differentiated cuboidal or columnar epithelial cells that form a thin covering of transitional epithelium on small fibrous septa. These structures are of various lengths and project into the lumen of the pelvis. Papillary growths may occasionally be seen macroscopically in the pelvis. UCs originate in the pelvis or ureter and have identical histological features to TCC in the urinary bladder. Approximately 5% of primary renal tumors in dogs are UC and 9% in cats (Table 15.2). SCCs (1–2%) occur in the pelvis, and they have intercellular bridges and keratinization sufficiently developed to justify this diagnosis. Undifferentiated carcinomas are epithelial neoplasms that do not differentiate into recognizable tubular, transitional, or squamous epithelium and are rarely diagnosed. This term should not be used for anaplastic neoplasms that can be classified. Confirmation of their epithelial origin may require cytokeratin. Of the 812 renal cell tumors in dogs only 2 were classified as undifferentiated carcinoma, there were none in cats and none in reports on horses and cattle. This probably reflects the histology of renal tumors, in that nearly every renal carcinoma has a focus of clear cells, tubules, or other identifying characteristics that help define renal cell origin or provide a histologic classification. Poorly differentiated tumors that lack clear epithelial characteristics may have been considered sarcoma. Some carcinomas can be so poorly differentiated and produce such a desmoplastic response that they can only be differentiated from a sarcoma with the use of cytokeratin and vimentin. RCCs can stain positively with both or only with cytokeratin. True sarcomas are positive with vimentin but not cytokeratins. Names for this neoplasm include Wilms’ tumor, embryonal adenosarcoma, embryonal nephroma, and nephroblastoma.10–12 The tissue of origin is the metanephric blastema; stromal cells and blastema develop from a common stem cell.10 The blastema normally differentiates into nephron units and cells that do not differentiate undergo apoptosis. If these events fail then remaining cells are potential sources for neoplastic transformation. The tumor is a triphasic mixture of embryonic epithelium (glomerular buds and tubules), undifferentiated blastema, and myxomatous mesenchyme (stroma) in various amounts. Glomerular structures never predominate but their presence, along with the age of the animal, and renal location are keys to the diagnosis. The tumors arise from neoplastic transformation during nephrogenesis or from nephrogenic rests that persist postnatally. The latter has been reported in a dog with nephroblastoma and polycythemia.10 In people, the Wilms’ tumor gene (WT1 on chromosome 11) is a causative factor, as are two other genes, and there are strains of rats that develop nephroblastoma.13 Nephroblastoma can also form in the thoracolumbar junction and is seen most frequently in German shepherd dogs, but has also been observed in other breeds and crossbreds.13–18 The spinal cord tumors form embryonic glomeruli, tubules and rosettes (Figure 19.27); they react positively to polysialic acid (present in embryonic renal cells and nephroblastomas in humans) and stain for Wilms’ tumor gene product (WT1).13 These tumors likely develop from remnants of renal rests trapped between the dura and developing spinal cord (see Chapter 19). In chickens, nephroblastomas are associated with an oncovirus. This is a congenital neoplasm, and many develop during fetal life but are not detected until later in life when a clinical problem becomes obvious. They have been reported in the bovine fetus. Approximately 92% of nephroblastomas in swine are present in animals less than 2 years of age, and 77% are present at 1 year of age, suggesting that some do develop later in life.11 Most of these animals are asymptomatic and the tumor is discovered at autopsy/slaughter. The tumors appear to be more common in males by a 2:1 ratio of males to females (although no difference was reported in another study of swine). Case reports in dogs also indicate a male predominance.10–12,19 The tumor occurs in cats and looks grossly like lymphoma, which is a more likely differential, even if the cat is young. Nephroblastoma is the most common primary renal neoplasm in children, swine, chickens, and fish.20 It is the second most common primary renal tumor in cats (21/120; 18%) (Table 15.3) and the third most common in dogs (55/894; 6%) (Table 15.2). The reports cited in Tables 15.2 and 15.3 are survey studies and therefore this data likely reflects a true prevalence in dogs and cats rather than a propensity to report an interesting tumor. Estimates of these tumors are 4.4–20 per 100,000 pigs in the United States and 0.35 per 100,000 in the United Kingdom.11 Authors speculated that the incidence would be even higher in abattoir surveys of swine if all suspected lesions had been submitted for microscopic confirmation. They have been reported in all the other domestic animals.10,19,21 Characteristically, the neoplasms are unilateral, single, and at one pole and are located in the cortex. They may be confined to the kidney or extend through the capsule, where they adhere to the body wall or mesentery. Exceptions to this appear as bilateral tumors, multiple tumors, and invasion into the pelvis. They usually occupy a large proportion of the affected kidney (Figure 15.5A) and may be large enough to compress abdominal viscera. Tumors greater than 50 cm in diameter and weighing over 34 kg are reported in breeding age sows.11 The natural and cut surfaces are lobulated, meaty to firm, white to tan with cystic areas and other areas discolored yellow, gray, or red. They resemble lymphoma, which is a more likely diagnosis than nephroblastoma, even in young animals. Rarely, fat, muscle, cartilage, and bone may be present. The presence of these tissues has led to the terms adenosarcoma, mixed tumor, and sarcocarcinoma. The critical feature is a disorganized mixture of embryonal epithelial (tubules and glomeruli), blastema, and mesenchymal tissues (Figure 15.5B–D), the most diagnostic of which are embryonic glomeruli. Tufts of epithelium invaginate into a lumen to form glomeruloid structures and searching sufficient sections will eventually identify a few that are worthy of the diagnosis or a photograph. When the tufts are cut in cross‐section there is a solid ball of cells in the center of the lumen, and cells lining the space have little to no visible cytoplasm and create a rim of “naked nuclei” (Figure 15.5C). Glomeruli and tubules will be in various stages of differentiation. Embryonic glomeruli are surrounded by irregular tubules that have lumens of various sizes; some form small acini or tubules, and others are elongated and dilated into a collecting drainage‐like system (Figure 15.5D,E). The undifferentiated blastema cells often are the predominant cells but they are not diagnostic in themselves. A classic pattern is a proliferation of blastema cells, in the center of which are partially or fairly well‐developed tubules and glomeruli. All of these structures are encompassed by variable amounts of immature, lightly basophilic, loose mesenchymal stroma (Figure 15.5B,C). Foci or large regions of undifferentiated blast‐like cells, with no visible cytoplasm will be present throughout the tumor. These foci can contain a few open spaces and appear to form lumens. Cystic structures are present in lesser numbers and are lined by cuboidal epithelium or squamous epithelium, with or without the presence of mucus, sloughed epithelial cells, or keratin. The mesenchyme is loose, areolar, and myxomatous, has a light basophilic hue, and is rarely dense enough to be birefringent. In ruminants the mesenchymal elements tend to be equally as developed as the epithelial components. There can be regions with herringbone patterns and dense fibrous proliferation similar to a fibrosarcoma. Differentiation and/or metaplasia into muscle may be present, and less frequently there is formation of cartilage and/or bone. A study in swine reported the IHC characteristics of five nephroblastomas. This article22 should be read for details, as they used numerous antibodies, but to summarize: Tumors in people are reported to be positive for desmin but negative for other markers of muscle. The blastemal and stromal elements in humans are positive for vimentin, and the epithelial components are cytokeratin positive. Cytologic evaluation reveals a mononuclear cell tumor17 that resembles lymphoma (Figure 15.5 F). Presumably the round cells are the undifferentiated blastema cells. The exfoliated cells may be monomorphic or bimorphic cells. If bimorphic, then this is a distinguishing feature from lymphoma as lymphomas are consistently monomorphic. The larger cells are lymphoma‐like but some are in clusters and appear adhered to one another. Lymphoma cells should be individualized and not adhered. Nuclei are large, somewhat irregular in shape, fill the cytoplasm and have immature vesicular chromatin. There may be a second population of round cells that have more cytoplasm, eccentric nuclei, and in which the cytoplasm appears to contain a secretory product that imparts an eosinophilic or amphophilic color to the cytoplasm. In swine and poultry, metastases are rare, but in dogs and cats, metastasis is expected in >50% of cases. Epithelial and mesenchymal components may be present in metastases. Histological and cytological criteria for estimating the biological behavior are not well established but high mitotic activity, invasion, anaplasia, and sarcomatous differentiation indicate malignant behavior. Well‐differentiated tubules and glomeruli may indicate a less aggressive growth. If the host is a species other than swine, metastases are anticipated. Likely sites are regional sublumbar and mesenteric lymph nodes, lungs, liver, and the contralateral kidney. These are rare tumors in the kidneys of domestic animals that contain cellular components from all three germ layers. Most mixed tumors are classified as nephroblastomas. The presence of gut, lymphoid, sweat glands, and hair favor a germ cell neoplasm over nephroblastoma. Teratomas are described and illustrated in Chapter 16. Primary neoplasms may arise from mesenchymal tissues in the kidney. The most common tumors are undifferentiated sarcoma, fibroma/fibrosarcoma and hemangioma/hemangiosarcoma (Table 15.2), however cases of leiomyoma/leiomyosarcoma will be found and rare examples of lipomas, osteoma, chondroma, or their malignant counterparts are reported.1–3 Approximately 5% of primary renal tumors in dogs are undifferentiated sarcomas and 15% in cats (Tables 15.2 and 15.3). These diagnoses were primarily made on H&E sections, and whether they would remain sarcomas or undifferentiated after the application of immunohistochemistry is speculative. Mesenchymal neoplasms that cannot be classified otherwise are placed in this group, and confirmation can be obtained by negative cytokeratin and positive vimentin staining characteristics. These tumors have variable amounts of stroma, which can be abundant and can isolate tumor cells. Hemangiosarcoma can be secondary or primary. When hemangiomas or hemangiosarcomas are solitary, then stating that their origin is the kidney is easy. When hemangiosarcoma is widely disseminated, it becomes difficult to impossible to distinguish multicentric origin from metastatic lesions, but the latter is more likely. Hemangiomas are confined to the kidney and hemangiosarcomas usually grow through the renal capsule and cause considerable hemorrhage, and a large proportion of the total mass is non‐neoplastic. Histologically, they appear as blood vessel tumors and stain with antibodies against factor VIII. There is a wide discrepancy in percentage of hemangiosarcoma (HSA) in dogs between studies cited in Table 15.2. This may be due to the difficulty in separating primary from secondary tumors in dogs. They are rare in all other species. In cats, only 1/120 tumors was an HSA (Table 15.3). HSA is a common tumor of the spleen, right atrium and subcutis in dogs. It is less common as a primary tumor in the retroperitoneum or liver and kidney. Approximately 10% of primary renal tumors in dogs, and less than 2% in cats, were classified as primary HSA (Table 15.2). There may be some biases in that two of the canine retrospective studies reported a greater percentage of HSA than the others. Perhaps the distinction of primary versus secondary was not always clear. Visceral HSA has a poorer prognosis than subcutaneous HSA. Retroperitoneal HSA in dogs is associated with survival times of less than 40 days and 8/9 had metastasis.4 Primary renal HSA has a better prognosis than other visceral HSA and therefore should be differentiated. The overall number of dogs and the number of dogs in the various subgroups reported is small but hopefully these retrospective results will be confirmed in studies with more dogs. MST of approximately 14 dogs with renal HSA that underwent nephrectomy was 278 days (range 0–1005 days), the 1‐year survival rate was 29% and only one dog was reported to have detectable pulmonary metastases at the time of diagnosis.3 Dogs that had hemoperitoneum only survived approximately 60 days. Anemia and hematuria were the most common clinical pathology abnormalities. The anemia was mild and even if hemoperitoneum was present, the anemia was not severe (22–26% PCV). All dogs had a detectable mass in the kidney via abdominal radiography (14/14) and/or ultrasound (9/14). The median tumor diameter via ultrasound was 3.3 cm (range 1.8–14 cm). Ten tumors were in the left kidney and 4 in the right. Metastases to abdominal organs and lymph nodes were not present at the time of exploratory surgery. Four dogs were euthanized because of confirmed (4/10) progression of the cancer: metastases were present in lungs 4, liver 2, and contralateral kidney 1. Six dogs were euthanized because of suspected progression of the cancer based on imaging or clinical examination, but none of these had an autopsy or histopathology performed. The numbers are small but it appears that dogs with renal HSA have lower metastatic rates and longer survival times than those with splenic and other visceral forms of HSA. Fibroma and fibrosarcoma account for approximately 5% of primary renal cell tumors in dogs and 2% in cats (Tables 15.2 and 15.3). As a primary tumor they are rare in all other species. They appear as their counterparts in other locations (Figure 15.6A).1–3 A spindle cell neoplasm with pointed nuclei, no visible cell borders, birefringent matrix, collagen that stains appropriately histochemically (Masson’s or Van Gieson), and if applied, positive immunoreactivity to vimentin are some of the salient features. A report on four cases of fibroma in dogs indicated they are well demarcated, singular or multiple, and usually located at the corticomedullary junction.1 Renal interstitial cell tumors have all the characteristics described for fibroma. The distinguishing features are the presence of cytoplasmic lipid droplets in interstitial tumors, which may require ultrastructural study to document and the matrix should be alcian blue positive with interstitial tumors and negative for fibromas.5,6 Renal interstitial tumors are located at the corticomedullary junction, tend to be multiple, and are believed to arise from renal interstitial cells that contain prostaglandin, arachidonic acid, and a neutral antihypertensive lipid that may lower arterial blood pressure. In humans, the tumors are cyclooxygenase‐2 (COX‐2) positive.7 Macroscopically and microscopically with H&E they are indistinguishable from fibroma. In human tumors there is controversy as to whether these are neoplastic or hyperplastic lesions.6 Mesoblastic nephromas, a variant of nephroblastoma, are rare congenital mesenchymal tumors derived from the metanephric blastema seen in children and reported in young dogs.8–10 The gross appearance is a white to tan, solitary mass in the cortex. Their histology resembles fibroma and renal interstitial tumors. They are benign growths of fibrous tissues that vary in their cellular density and are reported to be positive for collagen histochemical stains, alcian blue, and vimentin and should be negative for COX‐2, factor VIII, desmin, and von Willebrand factor.8 The few cases reported in dogs were found as incidental lesions. An angiomyxoma was described in a dog with paraneoplastic hypercalcemia.11 The tumor was approximately 4 cm in diameter and extended from the cortex to medulla. The tumor was composed of differentiated spindle cells, abundant myxoid stroma, and entrapped non‐neoplastic renal elements. Tumor cells labeled with vimentin and blood vessels within the tumor for von Willebrand factor. Hypercalcemia resolved post nephrectomy and increased concentrations of PTHrP returned to reference range. In humans, this is a rare tumor thought to originate from myofibroblasts. Chemical, physical, viral, and genetic etiologies are associated with renal cell tumors in animals and humans.1–17 Chemical carcinogens known to induce renal cell tumors in animals include nitrosamines,8,17 aromatic amines1 (dyes, rubber, coal, gas industries), nitrosureas,8 triphosphates,1 cadmium,3 aflatoxin,4 and lead.7 Estrogen administration is reported to cause renal carcinoma in the Syrian hamster.10 Compounds associated with renal carcinomas in humans are asbestos,1 cigarette smoke,9 coffee,1,9 phenacetin,9 diuretics,9 hydroquinone,18 and analgesic abuse.11 Viruses are a cause of renal cell tumors in chickens (avian leukosis – oncornavirus), leopard frogs (Lucke adenocarcinoma – herpes virus), and gray squirrels (pox virus).12–14 Renal tumors are common in budgerigars: up to 25% of all tumors in this host are of renal cell origin, and retrovirus sequences have been detected in these tumors.15 There is an association between nephrotoxicity and nephrocarcinogenicity in laboratory animals,18 and there is a higher incidence of renal cell tumors in humans with end‐stage renal disease and cysts.1,17 An important factor in renal cell neoplasia is gender: males have a higher incidence in humans (two‐ to five‐fold greater in males) and laboratory animal models, and nearly all reports suggest a male predominance in dogs (approximately 2:1).1,2,16 Renal cancer in humans is not considered a single disease. It has various histologic types and defined clinical courses, different responses to therapies, and different genetic mutations.19 It is reasonable to assume this will be demonstrated for renal cancers in animals; however, these assumptions do not always hold true for different species. Hereditary renal cancer in humans is associated with the VHL (von Hippel–Lindau) suppressor gene,20–22 mutations in MET gene and mutations in the BHD (Birt–Hogg–Dubé) gene (folliculin or FLCN).19 Clear cell histologic subtype is the most common renal carcinoma in humans and it is associated with inactivation of the VHL gene in the majority of affected patients. However, investigators could not identify mutations in the VHL gene in 13 dogs with renal carcinoma or 2 with renal sarcoma.23 Inactivation of the VHL gene was not prevalent in these dogs with renal carcinoma; however, other genes and mutations could be involved. For example, mutation of MET proto‐oncogene occurs in people with hereditary papillary renal carcinomas and a germline mutation of MET has been described in rottweilers.24 The prevalence of this is approximately 70%, perhaps contributing to the high incidence of cancer in this breed.13 Mutations in the tumor suppressor gene folliculin (BHD gene) are a critical step for the development of dermatofibrosis and renal tumors in dogs similar to Birt–Hogg–Dubé syndrome in people and rats.25,26 Twelve of 13 canine renal tumors were positive for KIT and it was speculated that mutations in the c‐KIT proto‐oncogene may be part of renal oncogenesis in dogs.27 KIT expression is found in chromophobe renal tumors in humans. The tumor suppressor gene WT1 on chromosome 11p13 is the “Wilms’ tumor” gene and is associated with nephroblastoma.1,28 Chromosome 3 translocations are associated with clear cell tumors in humans, and the short arm of chromosome 3 (3p) is associated with nonpapillary renal carcinoma, while trisomy of chromosome 17 is associated with papillary renal carcinoma.1 The Ras oncogene family is associated with nephroblastomas and N‐nitrosoethylurea‐induced renal carcinomas in rats but is not critical in the genesis of RCCs in humans.1,29 Mutations of the tumor protein p53 gene (TP53, also known as p53) are more important in the progression step of renal carcinogenesis than initiation but are not common alterations in renal neoplasia. In humans, there are multiple cytogenetic and molecular alterations associated with phenotypic variants of renal epithelial tumors.1,20,21 Wistar rats have familial renal adenomas,30,31 as do German shepherd dogs with renal tumors and nodular dermatofibrosis.32 The Eker rat is a model for hereditary renal carcinoma in which a single suppressor gene, the Tsc‐2 gene on chromosome 10q12, has been identified and is responsible for cancer induction.18,31, Heterozygote animals develop multicentric renal cell adenomas and carcinomas by 1 year of age. The trait is autosomal dominant, and homozygotes are lethal, with death of the fetus at approximately 13 days of gestation.31 Metastatic neoplasms are twice as common as primary neoplasms in dogs and seven times as frequent in cats (see Table 15.1). The most common metastases in dogs are hemangiosarcoma, adenocarcinoma (unspecified primary), and lymphoma; in cats the most common tumor by a wide margin is lymphoma (some of these may be myeloproliferative tumors). Most neoplasms metastatic to the kidneys also have metastases in the lungs. An exception to this is lymphoma, in which pulmonary tumors are very rare. If lesions in the kidney(s) are metastatic, then there are metastases elsewhere. Problematic tumors are metastatic epithelial tumors from prostate, mammary or pulmonary sources, which can be difficult to distinguish from primary renal cell tumors. Size of the tumors, locations of tumors, histology, regional lymph nodes involved, and the low prevalence of primary renal cell tumors are usually the deciding factors. IHC markers can help identify cancers of unknown primary site (CUPS): uromodulin has been used to confirm renal cell origin, and clear cells and epithelial tumors that dual mark for vimentin and cytokeratin suggest renal origin. Lymphoma is the most common tumor in the kidneys of cats and is one of the most common tumors in kidneys of all species.1 When present in the kidneys it is not confined to this location and lymphoma will be present in lymph nodes and other lymphoid tissues if it is searched for. The neoplasm can cause azotemia by destroying over 75% of the parenchyma or by a unique location in the excretory pathway obstructing the outflow of urine. It can be associated with nonregenerative anemia, rarely polycythemia (erythropoietin stimulation), or hypercalcemia. In all species lymphoma forms multiple, bulging, soft, white‐tan masses of varying sizes (Figure 15.7A). Histologically, they are “typical” lymphoma cells, characterized by round, blastic nuclei, little visible cytoplasm, and no supporting stroma, dissecting through the interstitium. They dissect through the renal parenchyma, surrounding and isolating tubules and glomeruli as they form variably sized masses (Figure 15.7B). Usually there is little necrosis or hemorrhage. They could be confused with other poorly differentiated neoplasms, histiocytic tumors, or myeloproliferative disease, but usually the distribution in lymph nodes, histology/cytology, and, if needed, IHC are sufficient to establish a diagnosis. We are not aware if there is a pattern that favors B‐ versus T‐cell differentiation for lymphomas in the kidneys of animals. Adrenal cortical carcinomas or pheochromocytomas occur with frequency in the kidneys from circulatory metastasis or direct extension. Hamartoma has been described in an 8‐month‐old heifer with a 0.5 cm mass at the corticomedullary junction.2 The lesion was a mixture of tubules, spindle cells, collagen, glomerular‐like structures, and blood vessels. Telangiectasia is described in Welsh corgi dogs and consists of non‐neoplastic proliferations of blood vessels in both kidneys and other organs, including duodenum, brain, vertebrae, subcutaneous tissue, and spleen.3 In the kidney they produce multiple bulging red nodules that grossly look like hemangiomas or hemangiosarcomas. The distinction from hemangioma is that the blood‐filled spaces in telangiectasia are lined by simple endothelium, one cell layer thick, and there is no proliferation of endothelial cells along the blood vessels or between thin trabeculae separating the cavernous spaces (Figure 15.8). Knowledge that the vascular proliferations are in a Welsh corgi certainly biases the diagnosis. They congregate at the corticomedullary area. They probably represent malformations and have been compared to hemangiomatous syndromes in humans. Granulomatous interstitial nephritis can produce gross lesions that resemble neoplastic nodules. Diseases that mimic this appearance are feline infectious peritonitis, white spotted kidney disease and hairy vetch in cattle, and Halicephalobus gingivalis (Micronema deletrix) in horses. Distribution of other lesions and histology usually make the identification of these diseases easy. Halicephalobus gingivalis is more common in Europe than in the United States, but occurs worldwide and may produce large renal granulomas that grossly can be indistinguishable from neoplasms (Figure 15.9A). The lesions in the horse depicted in Figure 15.9 were bilateral and resulted in renal failure. There are concurrent microscopic lesions in the nervous system.4 Commonly affected tissues are brain, spinal cord, kidney, oral, nasal and lymph nodes, but lesions may be disseminated through the body. Histologically, the lesions are typical granulomas with intralesional larvae. Feline infectious peritonitis (FIP) will produce small to large neoplastic‐like masses in kidneys that grossly mimic lymphoma (Figure 15.9C). Flat, plaque lesions are more characteristic of FIP, however some granulomas will be round and bulge from the surface. FIP lesions follow blood vessels. Pyogranulomatous lesions in peritoneum, liver, thoracic cavity as well as vasculitis will correctly identify the disease with or without knowledge of serum proteins, coronavirus titers or PCR for coronavirus. Solitary or multiple (polycystic) congenital cysts may be confused with cystic tumors grossly, but microscopic assessment clearly distinguishes them. Cysts contain transparent, amber to yellow fluid and are lined by a single layer of epithelial cells that are usually compressed and elongated. There is no proliferation of lining or papillary projections. In West Highland white terriers a polycystic renal and hepatic disease is reported to be autosomal recessive.5 The cysts arise from the collecting ducts. Acquired cysts are part of inflammatory or neoplastic renal diseases that form due to obstruction of outflow of the glomerular filtrate. Cysts are often present in nephroblastomas and primary adenocarcinomas. A fibroepithelial polyp 12 cm in diameter completely filled one kidney in an 11.6 kg dog. The gross appearance looked like a botryoid tumor but histopathology revealed the mass was not neoplastic and had a fibrous core covered by a single layer of non‐neoplastic urothelium (Figure 15.9B).6 Tumors of the renal pelvis and ureter are rare; when present, they are invariably urothelial carcinomas (UC, TCC) or the same with areas of squamous differentiation. UCs are much less common in the renal pelvis than they are in the urinary bladder, where they are the most common neoplasm by a wide margin. When in either location they can cause hydronephrosis and spread to the lower urinary tract by implantation metastases. Occasionally they can disseminate throughout a ureter as plaques or small raised nodules (see Figure 15.13A). In the pelvis they invade the medullary crest and start to ascend to the medulla while eroding ventrolaterally through the capsule and into the perirenal tissues (see Figure 15.13C). Metastases are usually present by the time a diagnosis is established or an autopsy is performed. Other primary or metastatic tumors can localize in these regions, but they are uncommon. Papillomas are rare, occur in the pelvis, and are characterized by papillae lined by one to five layers of mature transitional epithelium that covers a thin fibrous septum. They are of variable length and some may be macroscopically visible. Hemangiomas are reported to occur in the renal pelvis and ureter of cattle with enzootic hematuria. Leiomyosarcoma1 and a giant cell sarcoma2 of uncertain cell origin have been reported in the ureter of dogs. Archival material in files at North Carolina State University (NCSU) contained rare examples of leiomyosarcoma and UC in dogs and lymphoma in a cow ureter. Neoplasia of the urinary bladder is common in dogs but uncommon in all other species. An exception is cattle in endemic areas where bracken fern (Pteridium spp.) grows and is reported to produce epithelial and mesenchymal bladder tumors in as many as 25% of cattle. Essentially all tumors that may occur in the bladder can be found in cattle that live in these geographic areas. The most common tumors in these cattle are papillomas, UC, fibroma, and hemangioma. Approximately half of the cases will have multiple tumors; the metastatic rate is about 10% (see Table 15.7). Renal tumors are not more common in these cattle so clearly the carcinogens are excreted and/or concentrated in the urine where they remain in contact with the bladder mucosa. Bovine enzootic hematuria (BEH) is a clinical entity with multiple lesions, which result in bloody urine in up to 90% of cattle that graze bracken fern. Approximately 5–8% of sheep developed UCs after grazing ferns related to bracken in Australia. Outside these geographic niches, the incidence of bladder cancer is much lower, at 0.01–0.1% of cattle in abattoirs. Tumors of the bladder and urethra account for 0.5–1.0% of all canine neoplasms and 2% of all malignant canine neoplasms. UC or TCC is the most common tumor by a wide margin.2–7 Primary tumors of the bladder are much more common than secondary tumors, metastasis to the urinary bladder is a rare occurrence in all species. Metastases are located in the wall of the bladder unless the primary was prostatic, in which cases some will spread to the bladder mucosa. Table 15.4 indicates that 812 of 845 canine urinary bladder tumors in dogs were primary (96%) and 38 were secondary. In cats there were 63 tumors, of which 56 were primary (89%) and 7 were secondary. Literature reviews and Tables 15.4–15.7 summarize the different tumors seen in the urinary bladder of dogs, cats, and cattle. For dogs and cats, UC/TCC is the most common tumor and it is highly malignant. Table 15.4 Urinary bladder neoplasia a 656 TCC, 83 undifferentiated carcinoma. b 45 TCC, 7 undifferentiated carcinoma. Table 15.5 Canine primary urinary bladder tumors Studies (UC = TCC): 1 VMDB: Veterinary Medical Data Base, 845 tumors, 812 primary. 2 Osborne, C.A., Low, D.G., et al. (1968) Neoplasms of the canine and feline urinary bladder: Incidence, etiologic factors, occurrence and pathologic features. Am J Vet Res 29:2041–2053. 130 tumors, 126 primary; 3 Norris, A.M., Laing, E.J., et al. (1992) Canine bladder and urethral tumors: A retrospective study of 115 cases (1980–1985). J Vet Intern Med 6:145–153. 4 Valli, V.E., Norris, A., et al. (1995) Pathology of canine bladder and urethral cancer and correlation with tumor progression and survival. J Comp Pathol 113:113–130. 5 NCSU: North Carolina State University, 99 tumors, 94 primary. 6 Klausner, J.S. and Caywood, D.D. (1995) Neoplasms of the urinary tract. In Canine and Feline Nephrology and Urology (eds. C.A. Osborne and D.R. Finco). Williams & Wilkins, Philadelphia, PA, pp. 903–916. a 5 leiomyoma and 3 fibroleiomyoma. b Additional diagnoses were: 1 myxoma; 6 neurofibroma. Table 15.6 Feline primary urinary bladder tumors VMDB: Veterinary Medical Data Base 56 + 63 cases in the literature. Table 15.7 Bovine urinary bladder tumors with enzootic hematuria Studies: 1 Pamukcu, A.M., Price, J.M., and Bryan, G.T. (1976) Naturally occurring and bracken‐fern‐induced bovine urinary bladder tumors clinical and morphological characteristics. Vet Pathol 13:110–122. 139 cattle; 289 tumors. Each tumor is listed individually; 1 round cell sarcoma; 35% had only an epithelial tumor, 9% only a mesenchymal tumor and 55% mixed tumors; 10% metastasis of carcinomas. 2 Carvalho, C., Pinto, C., and Peleteiro, M.C. (2006) Urinary bladder lesions in bovine enzootic haematuria. J Comp Pathol 134:336–346. 433 cattle; 870 lesions. 3 Ozkul, I.A. and Aydin, Y. (1996) Tumours of the urinary bladder in cattle and water buffalo in the Black Sea region of Turkey. Br Vet J 152:473–475. 608 had only epithelial or only mesenchymal and 455 were mixed tumors; 1 carcinosarcoma; a NR, not reported. CIS was seen in many cases but only with other lesions; diagnoses of carcinosarcoma, or tumor not specified are not included. b 9 fibrosarcoma. Cats have subtypes of neoplasms in the bladder comparable to those in dogs, with similar percentages: of 119 cases 104 (87%) were epithelial (TCC 67%) and 15 mesenchymal (13%).5,7,8,9 Approximately 10% of urinary bladder tumors in dogs are mesenchymal, and they are split 50:50, benign and malignant (Table 15.4). Smooth muscle tumor is the most common primary mesenchymal bladder neoplasms in dogs and cats, and hemangioma and fibroma are in cattle, at least those that are exposed to bracken fern (Table 15.7). There are not many examples of bladder cancer reported in horses, and SCC is the most common. Paraneoplastic diseases associated with bladder or urethral tumors include hypercalcemia (rare), cachexia, hyperestrogenism, hypereosinophilia,10 hypertrophic osteopathy, and polycythemia. Hypertrophic osteopathy occurs with some regularity in animals with bladder tumors and is reported with UCs, adenocarcinomas, rhabdomyosarcomas, and embryonal nephromas. The association of hypertrophic osteopathy with rhabdomyosarcomas in the urinary bladder of young dogs is reported frequently but overall this is an uncommon tumor. Most of the clinical information available is for tumors in the urinary bladder of dogs, and data are usually not separated for specific types of tumor.3–5 Some of the clinical signs are nonspecific (weight loss, weakness, lameness, dyspnea, etc.) and some are at least referable to the urinary system: hematuria, dysuria (95/115; 84%), pollakiuria (37%), abdominal pain (10%), and incontinence (9%) were reported in dogs with bladder or urethral tumors.6 Cats have the same problems, many of which are referable to lower urinary tract disease – a common initial diagnosis.8,9,10,11 Concurrent urinary tract infections are present in approximately 30% of cats and dogs. For both dogs and cats the greater likelihood that diseases other than cancer are the cause of the lower urinary tract signs and clinical pathology abnormalities often leads to delayed diagnosis of the UC, allowing the tumor to grow, infiltrate, and/or metastasize. Urinalysis will have abnormalities in 90% of dogs with epithelial or mesenchymal tumors of the urinary bladder or urethra, but these abnormalities are nonspecific.2–7 The most common are hematuria (>75%), pyuria (50%), proteinuria (30%), and bacteriuria (30%). Hematuria is the most common clinical pathology abnormality associated with UC but there are numerous and more common differentials for hematuria such that even when UC is the cause of hematuria it is seldom diagnosed when hematuria is first detected. The delay in diagnosis provides time for this aggressive neoplasm to infiltrate and metastasize. Hematuria is due to concurrent cystitis and/or physical disruption of blood vessels, either in the tumor or from contact and/or invasion of the tumor into adjacent parenchyma. Proteinuria is somewhat misleading as it is due to concurrent hematuria and pyuria. There are no reports that specify the proteinuria in dogs or cats with renal or bladder cancer as glomerular in origin. Dogs receiving chemotherapy for UC may have a positive bacterial culture in the urine in approximately 50% of the cases.12 The percentages for hematuria, pyuria and bacteriuria are different but the same abnormalities are present in cats. Many cats are older and have concurrent chronic renal disease, which may cause azotemia. If the tumor obstructs urine outflow this may cause postrenal azotemia. Hypercalcemia has been reported with a few tumors of the lower urinary tract.6 The likely mechanism is paraneoplastic production of hormones or cytokines, but hypercalcemia due to neoplasia of the urinary bladder is rare. Increased liver enzymes (ALP 27%, ALT 17%) are reported with tumors of the bladder and urethra in dogs.6 The mechanism is not known but is probably secondary to corticosteroid‐induced stress or concurrent and unrelated hepatic problems as the dog is older. The increases are mild and the percentage of dogs with increases in these enzymes might be comparable to any malignancy or a wide variety of other diseases. Azotemia is present in approximately 15% of dogs with bladder or urethra tumors. If azotemia is not due to concurrent renal disease or dehydration then it is secondary to obstruction of the outflow of urine, resulting in postrenal azotemia. This is more common in cats. Invasion through the wall of the bladder by the tumor or rupture of the bladder and production of uroabdomen is rare. If uroabdomen is present then the expected changes are azotemia, hyponatremia, hypochloremia, hyperkalemia, hyperphosphatemia, and creatinine and urea nitrogen concentrations greater in the abdominal fluid than serum. If present, the anemia is nonregenerative, and there are multiple mechanisms superimposed. The two most important of these are the anemia of chronic inflammatory disease and blood loss in the urine. Cancer will cause the anemia of chronic inflammatory disease just as effectively as a chronic infectious disease. Obviously there are many more likely causes of anemia and hematuria than bladder carcinoma. However, if the anemia is unexplained, the hematuria does not respond to treatments for cystitis and the patient is older, then these subtle clues should suggest consideration for bladder cancer. Certainly if the patient is a Scottish terrier or a mass is identified in the bladder then the index of suspicion for UC is high. Hypereosinophilia has been reported in one 14‐year‐old cat with UC.10 Cytology and/or biopsy are presently the tests of choice to confirm UC/TCC. If a representative sample is obtained and the pathologist is experienced in cytology then cytology is diagnostic. There are histologic grading schemes for UC/TCC but these are not applicable to cytology. These classifications are important but less so than in establishing the diagnosis, and cytology will accomplish this. How to obtain the sample is a clinician’s preference but is absolutely critical to our microscopic interpretation. Like any biopsy, the ability of the clinician/surgeon to access the true lesion and obtain a sample of sufficient size and quality are as important as or more so than our microscopic assessment. If a good sample is obtained, the diagnosis of UC is easy in the majority of cases. The quality of the sample is more critical than the pathologist. Small tissue samples must be placed in cassettes or cloth designed to prevent loss during processing. In dogs, approximately 30% (29/96) of transitional cell tumors can be diagnosed from urine cytologic examination, 80% (10/13) from prostatic or urethral washes, and 90% (20/22) from percutaneous fine‐needle aspirational (FNA) cytology.6 Methods for obtaining tissue through the urethra include cystotomy, cystoscopy, and catheterization.3,4 Biopsies obtained by transurethral cystoscopy from 92 dogs with UC produced samples of diagnostic quality in 96% of female dogs but just 65% of male dogs.4 The smaller urethra in males limits the size of the instruments that can be used which in turn limits the size of the sample that is obtained. Inadequate sampling because of physical constraints is the likely reason for the lower percentage of diagnostic samples in male dogs. Multiple options are available to obtain cells and/or tissue and it is at the discretion of the clinician to decide which works best for them. The larger the specimen the easier it is to be definitive and in a larger sample there will be sufficient tissue for ancillary testing, if needed. However, a diagnosis of UC can also be made with confidence from samples with only a few hundred cells. Looking for tumor cells in a urinalysis is the least invasive and least effective technique. When tumor cells are found in a voided sample of urine a diagnosis can be established but this is effective in only 25–30% of cases. Positive findings with this technique can rule in the diagnosis of UC but negative results do not rule out this differential. The best method for examining suspected cells in a urine sample is to collect a fresh sample, prepare a concentrated preparation, make a film of the sediment, and stain with a Diff‐Quik type stain. Do not diagnose from a wet mount, Sedi‐Stain preparation, or new methylene blue as these stains enhance nuclear details. These stains can make the nuclei and nucleoli of hyperplastic urothelial cells so prominent that an anatomic pathologist may call them neoplastic. A cell block of the sediment can also be processed and stained with H&E. Cell blocks can be sectioned multiple times to search for tumor. These sections can also be used for IHC and therefore have multiple uses to help establish a diagnosis. Cytological confirmation of tumor cells in the urine seems a logical diagnostic aid but must be interpreted cautiously in suspected bladder tumors, especially if concurrent inflammation is present. Inflammation of the urinary tract stimulates hyperplasia of urothelium, making the distinction of hyperplasia from dysplasia or neoplasia difficult in some cases (Figure 15.10). Urine is harsh on cellular details and cells immersed in urine will acquire artifacts that are amplified over time. The diagnosis of UC is based on the recovery of numerous, large, anaplastic epithelial cells. However, if inflammation is present, the anaplastic cytological changes must be marked, especially for an anatomic pathologist to make the diagnosis. For effective cytologic interpretation, it is important to recall that inflammation can also induce cellular atypia (Figure 15.10 F). Likewise, UC may evoke some inflammation, and it is preferable to examine a second sample of urine when there is no inflammation or to recommend another method to retrieve cells or tissue to establish the diagnosis. Correlating cytologic interpretation with clinical data, especially imaging data, is often critical to the diagnosis. It is easier to make a definitive diagnosis of UC from FNA cytology that enters the suspected tumor. These preparations are of higher quality and have fewer artifacts than samples prepared from voided urine. In cytological preparations the tumor cells will be in clusters and/or individual, will be extremely large (>40 µm diameter), and will have marked cytologic and nuclear variability consisting of various sizes and shapes of cells, nuclei, and nucleoli (Figure 15.10A). Usually there is little or no inflammation in these FNA samples. The more numerous the cytologic abnormalities and the less evidence of inflammation, the more likely the cells are neoplastic. If only a few of these cytologic abnormalities are identified and there is inflammation, then the cellular atypia is more likely dysplasia or hyperplasia of transitional epithelium. Search for cells with large cytoplasmic vacuoles or inclusions. These vacuoles and signet ring formations are characteristic of UC/TCC, making the diagnosis easier (Figure 15.10E). Some vacuoles will appear empty and others contain homogeneous or stippled eosinophilic material. The colors may vary slightly with different stains. These cytoplasmic inclusions are referred to as Melamed–Wolinska bodies. They are characteristic for urothelial cells, and when seen in large anaplastic cells or morulae are diagnostic for UC (Figure 15.10B–D).13 These structures are so characteristic of UC that if seen in cytologic preparations from other locations, such as a lymph node, skin, or abdominal or pleural fluid then UC should be searched for. In histological preparations they appear very similar and will be clear or contain eosinophilic material of varying density. They are PAS positive and will stain positively for uroplakin. Cytoplasmic vacuoles and signet ring cells can be seen cytologically and histologically in other carcinomas such as mammary and gastrointestinal, but Melamed–Wolinska bodies are characteristic for UC/TCC. If the diagnosis from cytology is equivocal then consider repeating, consulting with a cytologist, preparing a cell block, aspirating a draining lymph node, and correlating with other data. Most importantly, determine if there is a mass in the trigone region of the bladder. Immunocytochemistry (uroplakin) can identify the cells as urothelial but it does not distinguish normal, hyperplasia, or dysplasia from neoplasia. Markers of UC are also available and new molecular markers to detect chromosomal abnormalities or gene mutations are being developed. Reports state that FNA cytology sampling of suspected tumors should not be performed if a UC is a differential because of the possibility of seeding tumor cells into the abdominal wall.3,14,15 This has happened in a few dogs, but tumor seeding by FNA is uncommon to rare (6/357 cases)15 and as of yet cystocentesis has not been proven as a cause.12,15 Implantation of UC in the abdominal wall is more commonly caused by surgical procedures to remove or debulk the primary tumor, perhaps as high as 10%.15 The most likely explanation for tumor in the abdominal wall is seeding of tumor cells by the surgical or diagnostic procedure.16 However, lymphatic spread or an unusual spread via ligament attachments are possible.15 The number of cases is small, but it is thought that UC established in the abdominal wall worsens an already poor prognosis. If a neoplasm of the urinary bladder is suspected then the likelihood of it being a UC in a dog or cat is high, which means the patient has a malignant tumor that will shorten its life. The clinician must balance the need to establish the correct diagnosis, or determine if bacterial cystitis is present,12 with the unlikely possibility of tumor seeding via FNA. Cancer‐screening tests are not common in veterinary medicine, but UC/TCC is one of the few tumors for which markers have been developed or are used as models for UC in humans.17–28 UC is one of the most malignant tumors in veterinary medicine and unfortunately at the time of initial diagnosis 20% of dogs have detectable metastases and few dogs survive past 1 year, despite a variety of treatment options. An accurate, noninvasive marker that detects UC in the urine is needed to identify the presence of this tumor early, enabling early therapeutic intervention and to monitor treatment efficacy. Cytology or biopsy is the gold standard of diagnosis, to which all innovative diagnostics are compared. Biomarkers that can detect UC, especially non‐muscle‐invasive UC, have been and are being developed to recognize human patients with this neoplasm, especially those at high risk of developing UC.18,22,27,28 Most markers try to detect abnormalities in exfoliated cells in the urine or detect proteins that are released from neoplastic cells into the urine of all UC patients (high sensitivity). Biomarkers should be unique to urothelial neoplasia and should not be associated with urothelial hyperplasia or inflammation (high specificity). However, false positives and false negatives are seen with any urine‐based diagnostic test for multiple reasons. Proteins of small molecular weight or that are positively charged could pass through the glomerular filtrate and therefore are a source of tumor antigens in urine that could originate from neoplasms located elsewhere in the body. Other potential false positives may be due to concurrent diseases in or outside the urinary tract that increase urinary concentrations of the measured biomarker. Interfering substances in the urine that limit the detectability of the assay can result in false negatives. The best tests will have high sensitivity and specificity with high negative and positive predictive values. There are over 15 biomarkers used to detect UC in humans, some of which have FDA approval.18 A few of the screening tests that can be performed on samples of urine include basic fibroblast growth factor, bladder tumor–associated antigen, fluorescence in situ hybridization, minichromosomal maintenance protein‐2, microsomal RNA, and calgranulins.17–28 Although there are other epithelial and mesenchymal tumors in the bladder, all present markers are designed to recognize tumors of the urothelium, the most common and malignant of which is UC. Basic fibroblast growth factor (bFGF) is a proangiogenic peptide used as a marker for urologic and non‐urologic tumors in human beings and has been detected in high concentrations in the urine of dogs with bladder cancer.17,18 Although the numbers of dogs were small, one study demonstrated significantly higher concentrations of bFGF in dogs with bladder cancer than in normal dogs or dogs with urinary tract infection (UTI).17 Results are expressed as ng/g creatinine, and the median concentration of bFGF was 2.23 in normal dogs, 2.45 in dogs with UTI, and 9.86 in dogs with bladder cancer. One dog with bladder cancer did not have increased concentrations of bFGF, and one dog with UTI had concentrations of bFGF comparable to those of dogs with cancer; 86% of dogs with cancer could be correctly identified by increased concentrations of bFGF, and 90% of dogs with UTI did not have increased concentrations. The commercially available ELISA test kit uses a monoclonal antibody to recognize natural and recombinant human bFGF.17,18 Another commercially available test is the bladder tumor–associated antigen (BTA), which was developed for use in humans but a veterinary version is also available (V‐TBA).18–21 The assay detects a glycoprotein antigen complex that is partly of host basement origin and partly of tumor origin. These protein complexes are released into the urine of patients with urogenital cancer and are detected with qualitative colorimetric tests. Results are not quantified: they are either positive or negative. Urine samples should be centrifuged for optimal results. Three separate studies in dogs report similar results; the test has high sensitivity but lower specificity.18–21 It is capable of recognizing approximately 90% of the dogs with urogenital cancers (high sensitivity) but it has a high false‐positive rate of 50–65% in dogs with non‐neoplastic diseases of the urinary tract. False‐positive results are seen with pyuria, hematuria, proteinuria, and glucosuria, all of which are common abnormalities in dogs and cats without bladder cancer. The test has a relatively low false‐positive rate of 10–15% in healthy dogs and in dogs that are sick but due to non‐urinary tract diseases; specificity is approximately 85% for these two populations.20,21 A negative test result indicates that approximately 90% of the dogs tested do not have cancer in the urogenital tract. Combining sensitivity and specificity values with prevalence data for UC in geriatric dogs results in a negative predictive value of 0.999 (99.9% of dogs with a negative test result do not have UC).21 Therefore this dipstick test is useful as a screening test in house to rule out urinary cancer. A negative test result in a geriatric dog would avoid further expensive diagnostic tests. However, a positive test result could be cancer or a non‐neoplastic urinary tract disease and therefore this test is not useful to rule in cancer of the urogenital tract. The positive predictive value was 0.028, therefore less than 3% of dogs with a positive test result will have UC.21 This test can rule out the likelihood of UC but it cannot rule in UC. Like any diagnostic test, the results must be interpreted along with all other clinical and laboratory data. Of the available diagnostic/screening tests for canine UC, the V‐BTA qualitative latex agglutination test probably has been evaluated the most thoroughly, with a large number of urine samples and statistical analyses that not only look at sensitivity and specificity but also at positive and negative predictor values. These latter indices are valuable to help us determine how useful tests are in various diagnostic settings. Fluorescence in situ hybridization (FISH) is used to detect chromosomal aneuploidy in cells present in the urine of humans with suspected bladder cancer (Figure 15.11).23 The FISH‐based Urovysion® Bladder Cancer kit detects gains in chromosomes 3, 7, 17, and losses of 9p21 (CDKN2A locus) which are prevalent among human UC tumors. Molecular changes may precede morphologic changes, providing the means to screen or detect bladder cancer early and monitor recurrence. Cells obtained in a urine sample are fixed on glass slides, probed with fluorescent oligonucleotides complementary to the four genomic regions of interest, and at least 25 morphologically abnormal cells are evaluated and the mean counts from the four regions reported. The test is not to replace traditional diagnostic methods but is used as an aid to establish a clinical diagnosis or to determine if further testing is warranted.18 Canine UCs have a molecular signature characterized by gains in Canis familiaris chromosomes (CFA)13, CFA36, and a loss in chromosome 19 that were not present in several other canine tumors.23 Tissue samples from 31 canine UCs revealed 97% gained CFA13, 84% gained CFA36, and 77% lost CFA19. In urine, 100% of the samples gained CFA13, 92% gained CFA36, and 63% had loss in chromosome 19; 100% showed at least 2 aberrations and 50% showed all 3 aberrations as assessed by FISH.23 Evaluation for these chromosomal abnormalities indicated over 90% specificity and sensitivity when compared to urinary bladder from dogs with no known diseases of the bladder (Figure 15.11). Most of the tumors studied were invasive UCs, as are over 90% of canine bladder tumors, or the biopsy samples were too small to determine degree of invasiveness. Furthermore, at least one of these chromosomal aberrations was detected in cells in the urine from 100% of dogs with UC confirmed by histopathology or cytology (n = 24).23 These results indicate that FISH performed on urine samples could be a noninvasive screening or diagnostic test for UC. However, the test requires expertise and is expensive. A polymerase chain reaction (ddPCR) assay has also been used to detect copy number aberrations in tissues and urine from dogs with UC.62 The test was excellent when used on tissues and it also detected 67% of the dogs with UC when used on urine.62 If tests can be used on urine samples to differentiate dogs with UC from dogs with diseases that produce similar symptoms (cystitis, urolithiasis, polypoid cystitis, hyperplasia, or dysplasia of urothelium) we will have valuable diagnostic tools and means to evaluate therapeutic efficacy. Another potential screening test for UC is the identification of BRAF mutation in urine. Activation of a BRAF mutation was recently found in canine UC tissue and was detectable in the urine of dogs with UC.3,24 The mutation was found in CFA16 in greater than 80% of 62 naturally occurring canine UC cases.24 Survivin is a protein that inhibits apoptosis and it has been found in UC tissues from humans and dogs and is another potential marker for UC.25 Telomerase prevents telomere shortening, which permits tumor cells to avoid apoptosis. Telomerase activity had been detected in human UC, a canine UC cell line, and in the urine of dogs with UC.25 Calgranulins are antimicrobial proteins that are used to detect and study inflammatory and neoplastic diseases.22 They are part of the S100 family of calcium binding proteins (S100A8, A9, A12) and in different concentrations in vitro they appear capable of promoting or inhibiting carcinogenesis. Proteins S100A8/A9 (calprotectin) are overexpressed in bladder and prostate cancers in humans and rodents and are part of a gene signature complex to recognize invasive carcinomas. An immunoassay was used to measure calgranulins in urine samples from normal dogs, dogs with urinary tract infection (UTI), and dogs with UC or prostatic carcinomas.22 The numbers of dogs were small but the results suggest the concentration of this biomarker or the ratio of S100A8/A9 to A12 may prove useful to detect and/or follow the progression of these tumors. Concentrations distinguished dogs with urinary cancer from normal dogs but there was overlap of concentrations in dogs with UTI (false positive). The ratio of S100A8/A9 to A12 helped to identify and distinguish the dogs with UTI. The anionic charge of calgranulins indicates they should not pass the glomerular barrier but if there was concurrent glomerulonephritis, common in geriatric dogs, this barrier could be compromised. It is not known if this immunoassay can detect diagnostic abnormalities in the early stages of neoplastic transformation. Concentrations should be normalized to urine‐specific gravity or concentration of creatinine; blood contamination as evaluated did not affect measurements.22 Expression of specific microRNAs were found to be significantly higher in tissue samples of canine UC as compared to urinary bladder tissue that was grossly normal or that had inflammatory disease of the urinary bladder.27 The interpretation of these results was that certain microRNAs may act as oncogenes in the pathogenesis of canine UC and that assays for these microRNAs in tissue, blood, and urine may have diagnostic utility or could be used to predict responses to treatments.27 Another potential biomarker for UC is minichromosome maintenance proteins, which are critical for replication of DNA and to control replication to once per cell cycle. Assays for these proteins and the immunocytochemical identification of these proteins in exfoliated cells from human patients with UC have been developed.28 A cyclic peptide cancer‐binding ligand, PZL4, will bind to human and dog UC cells, but does not bind to urothelial cells from normal dogs or dogs with chronic cystitis.29 This ligand can be attached to micelles and deliver imaging labels or therapeutic drugs to the urothelial cancer. This technique has the potential to enhance diagnostic imaging or treatment as PZL4‐decorated micelles will attach to the surface of canine UC cells and enter the cells.29 Biomarkers and other diagnostic tests should have high sensitivity (percentage of patients with the tumor or disease are positive), high specificity (percentage of patients without the tumor are negative), and high positive and negative predictive values.18 The biomarkers summarized above report high sensitivity rates of 90% or greater (dogs with UC test positive). However, high specificity rates are only seen when results are compared to “normal” dogs, meaning the tests can distinguish normal dogs from dogs with bladder cancer, but other diseases in the urinary bladder may cause false‐positive results. Tests such as the BTA (bladder tumor antigen) test should prove useful as a screening test to rule out cancer of the bladder, since 99.9% of dogs with a negative BTA test result do not have UC.21 However, biomarkers need to distinguish dogs with UC/TCC from dogs with diseases that look similar, such as polypoid cystitis, UTI, or urolithiasis, then they could be used as diagnostic tests or provide a means to follow patients receiving treatments. Most diagnostic tests can recognize a disease when it is fully developed and can distinguish affected patients from normal individuals. We need tests that recognize UC early in its course and that can distinguish affected patients from individuals with non‐neoplastic diseases of the urinary bladder, which share similar symptomatology. Ideally, biomarkers would be tested on models or spontaneous cases that are determined to be early in the development of UC. These biomarkers also do not need to be stand‐alone tests, they can be used to complement cytology and other diagnostic procedures. Screening and diagnostic test results need to be correlated with the gold standard of histopathology. Grading schemes can be applied to the histopathology specimens and these need to be correlated with long‐term follow‐up studies with autopsy data to know if grades are predictive. In human oncology the goal is to diagnose UC when it is noninvasive and it is logical that the treatment of animals would be more successful if UC were diagnosed before it invaded muscle layers. However, the etiologies and molecular abnormalities of UC may differ between species such that UC in dogs is directed to invasive forms from their initiation. Until UCs are recognized, early treatment regimens will be handicapped. Tumors of the bladder and urethra account for 0.5–1.0% of all canine neoplasms and 2% of all malignant canine neoplasms. Approximately 90% of urinary bladder neoplasms in dogs, cats, and horses are of epithelial origin and are malignant (Tables 15.4, 15.5, and 15.6). By a wide margin UC is the most common tumor in all species. In dogs approximately 75% are UC and some studies report over 90%; in cats they account for almost 90% of bladder neoplasms. The terms urothelial carcinomas (UC) and transitional cell carcinoma (TCC) are synonymous and the diagnosis is usually straightforward. Benign tumors of the urinary bladder are rare, less than 4% of primary bladder tumors in dogs and cats are benign. Benign epithelial tumors are rarely found as incidental lesions at autopsy or in surgical pathology submissions other than in ruminants grazing on bracken fern. By the time an epithelial neoplasm produces sufficient problems to be diagnosed clinically, the tumor is advanced and 20% of dogs, cats, and horses will have clinically detectable metastases. Autopsy data indicate that the rate of metastasis from UC/TCC increases to 50–90%, the majority of these go to lungs and regional lymph nodes but metastases can be found in every organ if they are searched for. Their propensity for metastases, combined with their resistance to treatments, makes this one of the most malignant tumors in veterinary medicine. Grading schemes for UC/TCC are described in the literature and in this chapter; however, we need follow‐up data to know if these accurately predict clinical outcomes in dogs. The present grading schemes work well for the variety of lesions seen in cattle grazing on bracken ferns. Most studies on domestic animals do not separate the clinical characteristics for each of the types of tumor, and therefore most of the reported data is for bladder neoplasms in general or bladder carcinomas. In this chapter information is provided for bladder carcinoma, primarily UC/TCC and limited information is specific for other tumors of the bladder. Terminology and classification of proliferative lesions in the urinary bladder have undergone multiple changes since the last edition of this book. The change in nomenclature to favor UC in humans was not a clear consensus. The term urothelial was favored to transitional as it reflects the specific histology of urothelium, avoiding the less specific term transitional. The report stated “because there was a slight majority opinion in favor of the term ‘urothelial,’ it was adopted as the preferred term, but ‘transitional’ can be used synonymously”.1 We should switch to urothelial carcinoma (UC) for animal tumors but make sure our consumers (clinicians, oncologists) know this diagnosis is synonymous with TCC. Older literature that provides information about clinical data, incidence, clinical pathology, treatments, survival and metastatic rates are for TCC and therefore these initials will appear in this chapter when referring to specific references. This chapter summarizes published information based on prior classifications and provides information about the newer classifications and grades. Other changes were designed to modify how UCs are graded. The original morphological classification for human urothelial tumors was reported in 1973 by the World Health Organization (WHO), modified in 1998,2 2004, 2010,1 and 2012.3 The 1998 classification was applied to 100 lesions found in the urinary bladder of dogs.4 The canine material was archival and was retrieved from collections over a 10‐year period and the results published in 2006. The same WHO classification was applied to 400 tumors and tumor‐like lesions present in the urinary bladders of cattle from Italy and the results published in 2010.5,6 All of these publications found lesions in animals that were similar to those in humans and found the classifications in the WHO report to be useful and appropriate, with some exceptions. The majority of cases in dogs were, as expected, high‐grade UC (51/100 cases) and there was only one low‐grade UC and one papilloma.4 There were 47 non‐neoplastic lesions: polypoid cystitis and fibroepithelial polyps. The majority of human urothelial tumors are low grade, noninvasive forms and the studies in humans base histologic distinctions on hundreds to thousands of cases with clinical follow‐up that helps determine which subtypes are predictive and therefore worthy of division. In 2010, modifications of the WHO report were suggested1 and in 2012 additional modifications were proposed.3 All of these publications are excellent and well illustrated, and investigators should refer to these for how classifications have been modified and why they were changed. New classifications and modifications are inevitable but they make summarizing and comparing information between studies difficult and they make prior studies outdated. Until the next modifications appear, we prefer those published in 2012 and summarized in Box 15.1. Criteria to grade these tumors need to be clearly defined and the distinguishing features easily recognizable such that cut‐offs are clear and there is consistency between pathologists. Prospective studies and new studies that use retrospective materials need to use one grading scheme or do statistical comparisons between grading schemes. In veterinary medicine our studies in dogs will be hindered by the fact that the great majority of UC/TCC are high grade. The number of cases assigned to papilloma or low grade will be so few that follow‐up data will be biased. However, cattle with BEH have a multitude of lesions, including low‐grade UC, mesenchymal tumors, and non‐neoplastic urothelial proliferations that are better suited to a more complex grading scheme.5,6
Tumors of the Urinary System
RENAL TUMORS
Epithelial tumors
Primary (%)
Total
Primary (%)
Epithelial
Mesenchymal
Both
Secondary
Canine
579
187 (32)
140 (75)
40 (21)
7 (4)
392 (68)a
Benign
4 (3)
12 (30)
Malignant
135 (97)
68 (68)
Feline
252
30 (12)
23 (77)
7 (23)
222 (88)b
Benign
0
0
Malignant
23
7
Studies
1
2
3
4
5
Total (%)
Tumor, n
187
523
54
48
82
894
Carcinoma (%)
124 (66)
280 (54)
35 (65)
31 (65)
40 (49)
510 (57)
Adenoma
4
16
1
5
0
26 (3)
UC (TCC)
12
18
5
4
9
48 (5)
Papilloma
1
0
3
2
0
6 (0.7)
Nephroblastoma
6
40
2
2
5
55 (6)
SCC
0
11
0
0
0
11 (1.2)
Fibroma
9
4
1
0
0
14 (1.7)
Fibrosarcoma
5
40
0
2
0
47 (5)
Lipoma/sarcoma
0
7/1
0
0
0
8 (1)
Hemangioma/sarcoma
0
7/67
1
1
0/12
88 (10)
Leiomyoma/sarcoma
3
1/3
0
0
0/5a
12(1.3)
Lymphoma
0
0
1
1
0
2 (0.2)
Rhabdomyosarcoma
1
1
0
0
0
2 (0.2)
Undifferentiated SA
22
17
3
0
1
43 (5)
Other mesenchymal
0
5b
0
0
3c
8 (1)
Other diagnoses
0
5d
2e
0
7f
14 (1.7)
Epithelial
141 (75)
325 (62)
44 (82)
42 (88)
49 (60)
601 (67)
Mesenchymal
40 (21)
158 (30)
6 (11)
4 (8)
28 (34)
236 (26)
Nephroblastoma
6 (4)
40 (8)
2 (4)
2 (4)
5 (6)
55 (6)
Benign
17 (9)
35 (7)
5 (9)
8 (16)
0
65 (7)
Malignant
164 (88)
438 (84)
47 (87)
38 (80)
82
769 (86)
Studies
1
2
3
Total (%)
Tumor, n
71
30
19
120
Carcinoma (%)
31 (44)
11 (37)
13 (68)
55 (46)
Adenoma
3
0
1
4 (3)
UC (TCC)
6
2
3
11 (9)
Nephroblastoma
11
9
1
21 (18)
SCC
3
0
0
3 (2.5)
Fibrosarcoma
2
0
0
2 (1.7)
Lipoma
2
0
0
2 (1.7)
Leiomyosarcoma
2
0
0
2 (1.7)
Hemangiosarcoma
1
0
1
2 (1.7)
Undifferentiated SA
10
8
0
18 (15)
Epithelial
43 (61)
13 (43)
17 (90)
73 (61)
Mesenchymal
17 (24)
8 (27)
1 (5)
26 (22)
Nephroblastoma
11 (16)
9 (30)
1 (5)
21 (18)
Benign
5 (7)
0 (0)
1 (5)
6 (5)
Malignant
55 (78)
21 (100)
17 (90)
93 (78)
Adenoma
Incidence
Gross morphology and histological features
Carcinoma
General considerations
Incidence
Age, breed, and sex
Clinical and clinical pathology characteristics
Gross morphology
Histological features
Growth, metastases, and survival
Diagnostic considerations and immunohistochemistry
Oncocytoma
References
Nodular dermatofibrosis and renal cell tumors
Gross morphology and histological features
Urothelial (transitional) cell papilloma and carcinoma, squamous cell carcinoma, and undifferentiated carcinoma
Histological features
Embryonal tumors
Nephroblastoma (embryonal nephroma)
General considerations
Age, breed, and sex
Incidence
Gross morphology
Histological and immunohistochemical features
Growth and metastasis
Teratoma
References
Mesenchymal tumors
General considerations
Undifferentiated sarcoma
Hemangioma and hemangiosarcoma
Fibroma and fibrosarcoma
Renal interstitial cell tumors
Mesoblastic nephroma
Angiomyxoma
References
Etiology
References
Metastatic tumors
General considerations
Lymphoma
Adrenal tumors
Tumor‐like lesions
Hamartoma
Telangiectasia
Granulomas
Cysts/polyps
References
TUMORS OF THE RENAL PELVIS AND URETER
TUMORS OF THE URINARY BLADDER AND URETHRA
Primary
Total
Primary (%)
Epithelial
Mesenchymal
Secondary
Canine
845
812 (96)
752 (93) *
60 (7)
33 (4)
Benign
0
35 (47)
Malignant
752
25 (53)
Polyp
28
Feline
63
56 (89)
52 (93) **
4 (7)
7 (11)
Benign
0
0
Malignant
52
4
Polyp
3
1
2
3
4
5
6
Total
Tumors, n
812
126
115
110
94
1164
2421
UC, n (%)
656
43
100
100
81
797
1777 (74%)
Undiff. carcinoma
83
15
2
1
3
42
146 (6)
Adenocarcinoma
7
6
7
6
2
78
106 (4)
SCC
6
11
2
0
3
50
72 (3)
Adenoma
0
0
0
1
0
3
4 (0.2)
Papilloma
0
22
0
0
0
36
58 (2)
Leiomyoma
29
8a
3
3
2
12
57 (2)
Leiomyosarcoma
0
4
1
0
1
50
56 (2)
Fibroma
4
5
0
0
0
12
21 (1)
Fibrosarcoma
2
4
0
0
0
21
27 (1)
Hemangioma
2
1
0
0
0
2
5 (0.2)
Hemangiosarcoma
0
0
0
0
0
18
18 (1)
Rhabdomyosarcoma
9
1
0
0
2
20
32 (1.3)
Sarcoma
14
6
0
0
0
16
36 (1.5)b
Epithelial
89%
Mesenchymal
11%
Total
%
Tumors
119
Epithelial
104
87
Mesenchymal
15
13
UC (TCC)
80
67
Undiff. carcinoma
12
10
Adenocarcinoma
4
3
SCC
5
4
Adenoma
1
1
Papilloma
2
2
Leiomyoma/sarcoma
7
6
Hemangioma/sarcoma
4
3
Rhabdomyosarcoma
1
1
Sarcoma
1
1
1
2
3
Total
%
Tumors, n
198
870
1063
2131
UC
49
359
66
474
22
Papilloma
34
84
359
477
22
Papilloma low malig.
NR
55
NRa
55
3
Undiff. carcinoma
2
NR
1
3
0.1
Adenocarcinoma
19
20
NR
39
2
SCC
4
13
NR
17
0.7
Adenoma
5
8
1
14
0.6
Epithelial total
1079
51
Fibroma/sarcoma
3
18
397b
418
19
Hemangioma
75
256
198
529
25
Hemangioendothelioma
3
14
20
37
2
Hemangiosarcoma
0
43
21
64
3
Leiomyosarcoma
4
0
0
4
0.1
Mesenchymal total
1052
46
Clinical signs/clinical pathology
Urinalysis
Chemistry
Anemia
Cytology or biopsy
Markers
References
Epithelial tumors