Materials and Methods

Criteria for case selection—The medical records database of the Purdue University Veterinary Teaching Hospital was searched to identify dogs that had a diagnosis of TCC of the urinary bladder or urethra between January 1, 2003, and December 31, 2008. Dogs were included in the study if a bladder or urethral mass was identified, a biopsy of the mass was attempted at the Veterinary Teaching Hospital by use of transurethral cystoscopy or surgical cystotomy, and the diagnosis of TCC was confirmed histologically.

Medical records review—Data obtained from the medical records included sex, neuter status, breed, body weight, location of the tumor, biopsy method, number of biopsy procedures attempted, and quality of biopsy (ie, diagnostic vs nondiagnostic quality). A diagnostic-quality biopsy was defined as a biopsy sample that allowed a board-certified veterinary pathologist to make a definitive diagnosis of TCC. Nondiagnostic quality biopsies included samples with insufficient tissue, blood contamination, or artifactual damage to tissue caused by the biopsy procedure that precluded accurate histologic evaluation. Slides from all biopsy samples were reviewed by a single pathologist (JAR). The experience level (ie, board-certified specialist vs house officer) of the clinician obtaining the biopsy was also recorded for dogs undergoing transurethral cystoscopy.

Transurethral cystoscopy—Male dogs were placed in left lateral recumbency with the right hind leg abducted. The penis was extruded by an assistant and cleansed with dilute chlorhexidine gluconate prior to introducing the cystoscope. In male dogs, cystoscopy was performed by use of a 65- to 70-cm-long flexible ureteroscope with a diameter of 2.5 mm (7.5F)^a or 2.8 mm (8.5F)^b and a 1.2-mm (3F) working channel. Female dogs were placed in dorsal recumbency. The perivulvar hair was clipped, and the area was aseptically prepared with dilute chlorhexidine. The vestibule was flushed with dilute betadine solution prior to introducing the cystoscope. A 30-cm-long, 4-mm (12F)-diameter rigid cystoscope^c with a 19F sheath and a 1.7-mm (6F) working channel was used in female dogs weighing > 20 kg (44 lb). An 18-cm-long, 2.7-mm (8F)-diameter rigid cystoscope^d with a 15F sheath and a 1.7-mm (5F) working channel was used in female dogs weighing 5 to 20 kg (11 to 44 lb). A 13-cm-long, 1.9-cm-diameter rigid cystoscope^e with a 1.2-mm (3F) working channel was used in female dogs weighing < 5 kg.

In all dogs, biopsy samples were obtained by use of 3 or 5F flexible wire biopsy forceps^f or a wire stone retrieval basket.^g The diameter of the biopsy instrument used for each dog was dictated by the diameter of the working channel of the cystoscope. All transurethral cystoscopic biopsies conducted before October 1, 2004, were performed by use of pinch biopsy forceps,^f whereas pinch biopsy forceps and a wire stone retrieval basket^g were used for biopsy after October 1, 2004. Cystoscopic biopsy success rates before and after October 1, 2004, were therefore compared. Video recordings of transurethral cystoscopic procedures were reviewed, when available, to document the instruments used for biopsy. When biopsy samples were obtained by use of wire stone retrieval baskets, an attempt was made to loop the basket around the base of papillary projections of the bladder or urethral mass. The stone basket was then closed to ensnare and tear off a portion of the mass. The cystoscope or ureteroscope, with the basket still extended beyond the end of the scope, was then withdrawn through the urethra rather than attempting to pull the basket and biopsy sample back through the working channel of the scope.

Surgery—All surgical biopsies were performed with dogs in dorsal recumbency under general anesthesia. For simple cystotomy and partial cystectomy, the abdomen was incised along the ventral midline from the xiphoid cartilage to the pubis, and a full exploratory celiotomy was performed in most patients. Bladder tissue was biopsied via blade excision or excised with Metzenbaum scissors, and the bladder was typically closed in a simple continuous or simple interrupted pattern by use of 2–0 or 3–0 synthetic absorbable suture.^h For laparoscope-assisted cystotomy, a 1-cm incision was made at the umbilicus and another 1-cm incision was made 4 to 6 cm to the right of midline and approximately 6 cm cranial to the pubis. An open technique was used to place a 5- to 10-mm cannula at or near the umbilicus, and the abdomen was insufflated with carbon dioxide to a pressure of 12 mm Hg. A laparoscope was introduced through the cannula in the umbilical incision, and a visual abdominal exploration was performed. A second 10-mm cannula was inserted in the right caudal abdominal quadrant, and Babcock forceps were used to grasp the apex of the bladder. The bladder was then elevated to the body wall, and the cannula was removed. Two stay sutures were placed through the bladder wall and used for retraction or to secure the bladder to the skin. A 1-cm incision was made between the stay sutures, and excess urine was suctioned from the bladder lumen. A 5-mm screw-tipped trocarⁱ was inserted, and a 2.7-mm rigid cystoscope with a 15F sheath^d was inserted through the trocar to perform a full cystoscopic exploration. Masses identified cystoscopically were typically biopsied by use of biopsy forceps, or the bladder wall incision was extended and the mass was biopsied with Metzenbaum scissors. The bladder wall was closed by use of 2–0 or 3–0 synthetic absorbable sutures in a simple interrupted pattern. The bladder was replaced, and the abdominal cavity was insufflated and reexamined prior to closure. Carbon dioxide was then allowed to escape from the abdomen and the abdominal wall, and skin incisions were closed routinely.

Statistical analysis—Data were expressed as mean ± SD, and values of P < 0.05 were considered significant. Only data from the first biopsy attempt were used for analysis. Data on adequacy of biopsy for histologic diagnosis, biopsy method, TCC located in bladder (trigone), TCC located in urethra, clinician experience level (house officer or board-certified specialist), sex, neuter status, and date (before or after October 1, 2004) were placed in a contingency table and compared for the cystoscopic versus open surgical biopsy techniques by use of a Fisher exact test.^j Body weight was compared for the biopsy techniques by use of ANOVA after log transformation.^k Logistic regression was used with adequacy of biopsy for histologic diagnosis (1 = diagnostic quality; 0 = nondiagnostic quality) as the binary outcome variable of interest; biopsy method, TCC located in trigone, TCC located in urethra, clinician (house officer or board-certified specialist), sex, neuter status, and date as dichotomous dummy variables; and body weight as a continuous variable.^l

Results

Signalment—Ninety-two dogs were included in the study. Thirty-five were male (32 castrated and 3 sexually intact), and 57 were female (55 ovariohysterectomized and 2 sexually intact). Mean ± SD body weight was 19.9 ± 11.8 kg (43.78 ± 25.96 lb; range, 4.7 to 54.5 kg [10.34 to 119.9 lb]; median, 16.1 kg [33.8 lb]). Breeds represented included mixed (n = 25), Scottish Terrier (12), Shetland Sheepdog (9), Beagle (9), West Highland White Terrier (5), German Shorthaired Pointer (3), German Shepherd (3), Labrador Retriever (3), Dachshund (2), Miniature Schnauzer (2), and 1 each of Alaskan Malamute, American Eskimo, Basenji, Border Collie, Chesapeake Bay Retriever, Chihuahua, Chow Chow, Collie, English Springer Spaniel, Fox Terrier, Golden Retriever, Maltese, Miniature Pinscher, Pembroke Welsh Corgi, Chinese Shar-Pei, Shih Tzu, Siberian Husky, Staffordshire Bull Terrier, and Weimaraner.

Tumor location—Most dogs in this study had advanced disease, and tumors were frequently large, multifocal, or diffusely located throughout the lower urinary tract. Seventeen male dogs and 26 female dogs were noted to have tumors at multiple sites in the lower urinary tract. A tumor was present in the trigone of the urinary bladder in 13 male dogs and 23 female dogs. A urethral tumor was present in 4 male dogs and 26 female dogs. Prostatic involvement was presumed on the basis of ultrasonographic findings, including asymmetric prostatic enlargement, heterogeneous prostatic echotexture, or prostatic mineralization, in 4 male dogs. Despite these findings suggestive of prostatic TCC, prostatic biopsies were not attempted in any of the dogs in this study because all dogs with suspected prostatic involvement also had concurrent urethral or bladder masses.

Biopsy method—Twenty-nine of the 35 male dogs had biopsies performed once, 5 dogs had biopsies performed twice, and 1 dog had biopsies performed 3 times. Cystoscopy was performed as the initial biopsy method 20 times, and a diagnostic-quality biopsy sample was obtained 13 times for an overall success rate of 65% in male dogs. Surgical biopsy was performed as the initial biopsy method 15 times and resulted in a diagnostic-quality biopsy sample in all 15 cases. Surgical biopsy procedures included 6 partial cystectomies, 6 laparoscope-assisted cystotomies, and 3 cystotomies performed via celiotomy.

Fifty-five of the 57 female dogs had biopsies performed once, and 2 female dogs had biopsies performed twice. Cystoscopy was performed as the initial biopsy method 52 times, and a diagnostic-quality biopsy sample was obtained 50 times for an overall success rate of 96% in female dogs. Surgical biopsy was performed as the initial biopsy method 5 times, and a diagnostic-quality biopsy sample was obtained in all 5 patients in female dogs. Surgical biopsy procedures included 4 cystotomies and 1 laparoscope-assisted cystotomy.

Video recording of the cystoscopy procedure was available for review in 13 male dogs. However, the biopsy instrument could be seen in only 4 of these videos. Two biopsy samples were obtained with a wire stone retrieval basket, and 1 was of diagnostic quality. Two biopsy samples were taken with pinch biopsy forceps, and 1 was of diagnostic quality. Video recording of the cystoscopy procedure was available for review in 20 female dogs. However, the biopsy instrument (pinch biopsy forceps) could be seen in only 1 of these videos, and in this instance, a diagnostic-quality biopsy was obtained. This small data set precluded statistical analysis of the effect of biopsy instrument.

Predictors for obtaining a diagnostic biopsy sample—Application of a Fisher exact test to selected categorical variables indicated that only sex was significantly associated with obtaining a diagnostic biopsy sample (Table 1); specifically, diagnostic samples were obtained more frequently in females (96%) than in males (65%; P = 0.024). Contingency table analyses were then conducted by use of Cochran-Mantel-Haenszel P values adjusted for sex. Adjusted analyses indicated that breed (P = 0.0021) and biopsy method (P = 0.011) were predictive of obtaining a diagnostic biopsy sample, with mixed-breed dogs and transurethral cystoscopy (vs surgical cystotomy) being associated with an increased chance of obtaining a nondiagnostic biopsy sample. Location of the TCC in the trigone region of the bladder or urethra, experience of the clinician obtaining the biopsy, date of biopsy (before or after October 1, 2004), and neuter status were not predictive of biopsy success.

Logistic regression was applied to explore the effect of sex, breed, biopsy method, body weight, presence of TCC in trigone or urethra, clinician experience, date of biopsy, and neuter status. Preliminary examination of the data indicated quasicomplete separation of the data across sex. Exact univariate logistic regression was therefore used to estimate median unbiased estimates for the logistic regression procedure (Table 1). Exact logistic regression confirmed the findings from contingency table analysis that sex, breed, and biopsy method were predictors for obtaining a diagnostic biopsy sample.

Discussion

The major finding of the study reported here was that transurethral cystoscopy is an effective initial method for obtaining tissue biopsy specimens in dogs with TCC of the bladder or urethra. This was especially true in female dogs, in which the diagnostic success rate for cystoscopic biopsy was 96% in this study.

The lower diagnostic success rate of transurethral cystoscopic biopsy in male dogs (65%) highlights some of the unique challenges of cystoscopy in the male dog. Early attempts at cystoscopy in male dogs were limited by the necessity to perform a perineal urethrostomy incision to allow introduction of a rigid cystoscope.¹⁶ Increased availability of flexible urethrocystoscopes has permitted routine endoscopic examination of the entire male urethra and bladder; however, the narrow diameter of the male urethra limits the size of the cystoscope and biopsy instruments that can be used.^13–16 Smaller biopsy instruments limit the amount of tissue that can be procured and may therefore decrease the likelihood of obtaining a diagnostic-quality sample. Cystoscopy is also more challenging in male dogs because of greater difficulty in precisely directing and positioning a flexible cystoscope relative to a rigid one.¹⁴ These factors may have negatively impacted the success of cystoscopic biopsy of tumors in male dogs, compared with that in female dogs, in the present report.

Prior to analyzing the data in the present report, we had anecdotally noted an apparent disparity in the number of diagnostic-quality samples obtained in male and female dogs with TCC. It was assumed that this disparity was related to inadequate tissue samples obtained by use of smaller diameter pinch biopsy forceps in male dogs. In an attempt to address this problem, our group began to routinely use a wire stone retrieval basket to obtain biopsy samples in most dogs with TCC. A wire stone retrieval basket can be used to ensnare and tear off large papillary tumor growths and may therefore obtain better quality tissue samples than does a pinch biopsy forceps. Unfortunately, for many dogs in the present study, it was not possible to retrospectively determine whether a pinch biopsy forceps or a wire stone retrieval basket had been used for tumor biopsy. The wire stone retrieval basket was first used for tumor biopsy at the Purdue University Veterinary Teaching Hospital on approximately October 1, 2004. When the data were stratified on the basis of this date, there was no significant difference in the likelihood of achieving a diagnostic-quality biopsy. Therefore, we cannot make any conclusions regarding the utility of individual cystoscopic biopsy instruments in obtaining tissue samples from dogs with TCC. Although it is our impression that the wire stone retrieval basket allows for larger biopsies to be obtained and increases the likelihood of success, the results of the present study do not allow for these conclusions. A prospective study in which samples obtained with each biopsy instrument are evaluated concurrently is suggested.

In the present study, there was no influence of body weight on the likelihood of obtaining a diagnostic-quality cystoscopic biopsy sample. This is notable, given that larger biopsy instruments are used in heavier dogs and larger instruments should theoretically increase the diagnostic yield of biopsy. This may be explained by the fact that the variation in patient weight in the present study was not large enough to influence the cystoscopic instruments used. The working channel and biopsy instrument used with the flexible ureteroscope were the same size (3F) for all male dogs in this study, regardless of body weight. For the female dogs, the biopsy instrument used with the rigid cystoscope was the same diameter (5F) for all dogs weighing > 5 kg. This weight range encompassed 94% (49/52) of all female dogs in which cystoscopy was performed. Thus, there was little variability in the size of the instruments selected for use for cystoscopic biopsy on the basis of patient body weight and this may have accounted for the lack of a significant influence of body weight on cystoscopic biopsy success rate.

There was no detected effect of clinician experience (ie, board-certified specialist vs house officer) on the likelihood of obtaining a diagnostic-quality biopsy in the patient population in the present study. Cystoscopy requires a degree of technical proficiency to perform, and there is reason to assume that less experienced house officers would have more difficulty obtaining diagnostic-quality biopsy samples than would more experienced specialists. However, at our hospital, it is routine for a board-certified specialist or more experienced resident to be present for every cystoscopy performed. Therefore, it is unlikely that any of the cystoscopies credited to house officers in the medical record were done without the supervision or assistance of a more experienced clinician.

Surgical cystotomy was a highly effective biopsy method in the present study, with a 100% diagnostic success rate. This result was anticipated because surgical techniques can easily enable large representative tissue specimens from bladder masses to be obtained. Surgical biopsy has the drawback of greater invasiveness and also carries the risk of complications such as dehiscence of the bladder incision with subsequent uroabdomen and tumor seeding of the abdominal cavity or surgical incision.^5,10–12 Although these are all documented complications of surgical biopsy of TCC, few major surgical complications were reported in the patients in the present report. One male dog experienced minimal leakage from the urinary bladder following cystotomy, as confirmed by positive contrast cystography. No specific treatment was undertaken, and follow-up positive contrast cystography showed the leak to have spontaneously resolved 24 hours later. No other major complications, including tumor seeding of the abdominal cavity or surgical incision, were noted for the patients in the present study. Minor complications, including transient increase in hematuria, stranguria, and pollakiuria, were common following surgical and cystoscopic biopsy and typically improved within 24 to 48 hours.

The first report of cystoscopic examination of the bladder of dogs was by Vermooten¹⁷ in 1930. Barringer¹⁸ first reported the use of cystoscopic examination and biopsy of chemically induced TCC in dogs in 1947. Since then, cystoscopic evaluation of dogs with spontaneous TCC has been described by several authors.^13,14,19 Previous reports have not commented extensively on the diagnostic challenges associated with cystoscopic biopsy. Specifically, cystoscopic biopsy samples are usually small and superficial in nature, particularly compared with biopsy specimens that are obtained surgically. They may not contain sufficient viable tumor cells with which to make an accurate histologic diagnosis. This may be especially true in tumors characterized by extensive necrosis or inflammation. Cystoscopic biopsy specimens may also be subject to substantial crush artifact during the process of obtainment. However, despite these limitations, it was still possible to obtain diagnostic-quality biopsy samples in 65% of male dogs and in 96% of female dogs in the present study. Transurethral cystoscopy should therefore be considered an effective means of procuring tissue biopsy specimens in dogs with TCC.

We are unaware of similar studies evaluating the utility of transurethral cystoscopic biopsy in human patients with TCC. This is not surprising as the typical clinical signs of TCC and the goals of cystoscopy in humans and dogs are quite different. Nearly all dogs with TCC have advanced disease (ie, stage T₂ or greater; Appendix) at the time of evaluation and are not candidates for surgical treatment.¹ The purpose of cystoscopy in these dogs is to obtain tissue for an accurate histologic diagnosis, thereby permitting optimal medical treatment. In contrast to the situation in dogs, at least 70% of humans with TCC have superficial lesions (ie, stage T₁ or lower).²⁰ Gross, painless hematuria is the most common symptom in humans with TCC, and cystoscopy is routinely performed in these patients to ascertain the cause for the hematuria.²⁰ Superficial TCC in humans may be strongly suspected on the basis of cystoscopic visual inspection alone; however, biopsy is still necessary to diagnose, stage, and grade the cancer.²¹ Biopsy in these patients is usually accomplished through TURBT. This method of cystoscopic biopsy is used because TURBT specimens are intended to include the underlying muscularis propria layer to best assess the depth of tumor invasion and therefore the tumor stage. Patients with a diagnosis of superficial TCC by TURBT may be considered candidates for adjuvant intravesical chemotherapy. Patients with a diagnosis of muscle-invasive TCC made by means of TURBT typically undergo radical cystectomy. Although TURBT is the typical method for locally staging TCC in humans, it does have limitations. Cheng et al²² reported that 52% of tumors resected by means of TURBT were understaged by use of this method, compared with cystectomy samples taken from the same patients. Thus, although cystoscopic biopsy is used primarily to confirm a diagnosis of invasive TCC in dogs, it serves to diagnose, stage, and direct further treatment for TCC in humans.

Another notable difference between cystoscopic procedures in male dogs versus humans is the fact that in men, 19 to 24F rigid cystoscopes may be used,²³ which are similar in diameter to the largest rigid cystoscope used in female dogs in the present study. However, in male dogs, cystoscopy must be performed by use of flexible ureteroscopes. This difference in instrumentation and approach allows for larger biopsy samples and use of the TURBT approach in humans, whereas this is impossible by use of a transurethral approach in male dogs.

The retrospective nature of the present study precluded the meaningful comparison of other variables related to cystoscopic and surgical biopsy in dogs with TCC. In particular, the cost and anesthetic time associated with each procedure could not be determined in an unbiased fashion. Many dogs in this study underwent additional imaging studies, biopsy of additional organ sites, or placement of urinary tract stents during anesthesia for transurethral cystoscopy or surgical cystotomy, and these would all have affected anesthetic time and procedural costs. It is our impression that cystoscopic biopsy may require less anesthetic time and is less costly than surgical cystotomy. However, a prospective study in which dogs with similar stages and severity of disease are anesthetized only for cystoscopic or surgical biopsy would be necessary to better define the typical cost and required anesthetic time for either of these procedures.

Similarly, a prospective study would also provide a better means for making direct comparisons between the diagnostic utility of cystoscopic and surgical biopsy. Although the data in the present report indicate the likelihood of achieving a diagnostic biopsy with either method, neither surgical nor cystoscopic biopsy can be inferred to be a superior method on the basis of these data. The decision to perform cystoscopy or surgery was likely nonrandom in these patients, and surgery was likely pursued in most cases because it was necessary (eg, urinary tract rupture was present at the time of initial patient evaluation) or because previous attempts at cystoscopic biopsy had failed. Patient body weight might also have biased the decision to pursue surgical cystotomy or cystoscopy in these patients. Cystoscopic biopsy of TCC is typically more challenging in small patients and is often impossible in patients weighing < 5 kg. Although a diagnostic-quality biopsy was successfully obtained cystoscopically in patients weighing as little as 4.7 kg in the present study, we cannot discount the possibility that patient body weight may have biased the decision to pursue cystoscopic or surgical biopsy in these patients. A prospective study in which dogs are randomly assigned to biopsy via surgical cystotomy or transurethral cystoscopy or a study in which both biopsy methods are used in all dogs is necessary to ascertain whether surgical cystotomy or transurethral cystoscopy is a superior biopsy method in dogs with TCC.