Detrusor Muscle in TUR-Derived Bladder Tumor Specimens: Can We Actually Improve the Surgical Quality?

Introduction

Transurethral resection of bladder tumor (TURBT) is one of the most common surgical procedures in urologic practice. ^1,2 In this context, ∼70% of newly detected cases of bladder cancer (BCa) are shown to be nonmuscle invasive (NMIBC) and thus are routinely treated with TURBT. Accurate BCa pathologic staging allows the application of an adequate follow-up protocol and potential adjuvant treatment according to tumor grade and invasiveness. Moreover, a complete resection of all tumor tissues significantly reduces the risk of BCa recurrence and/or progression. ¹ For all the aforementioned reasons, it seems clear that TURBT must be carried out carefully following precise proceedings criteria. ³ Conversely, published data show a range of residual tumor detection at repeated TUR between 33% and 76%, ^4,5 thus suggesting that TURBT is far from being considered a simple procedure to be taught. To this regard, Pan and Soloway ³ suggested a TURBT checklist aimed at providing a point of reference for surgeons while enhancing the quality of the procedure itself. This intraoperative flowchart outlines the importance of mandatorily including the muscularis propria of the bladder in every resection. Similarly, current clinical guidelines stress the need for a pathologic evaluation of the muscle layer, especially for high-grade and pT1 tumors. ^1,2 Indeed, rates of tumor understaging at first resection have been shown to be higher for pT1 tumors, and the absence of detrusor muscle (DM) in the resected specimen was found to be an important source of staging error ^4,6,7 with muscle-negative resections being understaged at rates up to 49% compared with 14% for muscle-positive resections. ⁷ Therefore, DM sampling has been considered a surrogate marker of resection quality, even if others have suggested that including DM in low-grade papillary tumors may not be critical. ⁸

Published series assessing TURBT quality showed rates of DM absence (DM−) at final pathology ranging between 18.8% and 52.1%. ^{9 –12} These high rates of surgical inadequacy ³ raise the question of what could reasonably be considered an acceptable surgical error in oncologic terms. These observations prompted us to conduct an intradepartmental analysis to assess rates of and predictors of DM− at final pathology, in a series of consecutive TURBTs performed in the real-life setting at a tertiary–referral academic center. Moreover, we tested the impact of DM− on postoperative tumor recurrence.

Materials and Methods

We retrospectively collected data from 437 consecutive TURBTs performed at a single tertiary–referral academic center between November 2013 and December 2014. Complete perioperative data regarding surgeon experience (staff physicians vs residents), sampling modality (resection in one piece vs separate base resection), and tumor characteristics (number, size, appearance, and location) were retrieved for all patients. Similarly, pathologic characteristics at final report were recorded, including tumor grade, tumor stage, and DM status (present vs absent). Tumor size was stratified according to the dimension of the largest diameter (>3 cm vs ≤3 cm); flat appearance was defined as a lesion with minimal thickening of the epithelium or completely plain; multifocality was defined as more than one lesion. According to the policy of our institution, residents were strictly supervised by an attendant senior surgeon while performing any procedure; in this context, the senior surgeon oversaw the resident during the entire surgical procedure from the positioning of the patient on the operating bed and the setting of the instrument, thus guiding the young surgeon in training with active comments and remarks over the properly called surgical time. For the specific purpose of this study, only TURBTs with malignant diagnoses were considered. To reduce potential clinical confounds, pT2 diseases were not included in the predictive analysis; similarly, surgeons who performed fewer than three procedures throughout the considered time span were also excluded from the final analyses. A sample of 302 procedures was thus considered.

Patient follow-up was based on a first cystoscopy scheduled 3 months after the resection while subsequent surveillance was tailored to the individual risk of tumor recurrence and progression. Only histologically confirmed tumor recurrences were considered. Patients without an event were censored at the last date of follow-up. Likewise, patients lacking follow-up were excluded from the survival analysis.

According to clinical guidelines, ¹ all surgical resection procedures were performed with the intent to obtain DM, with separate base resection being performed for larger tumors where one-piece resections were not feasible. All procedures were performed with white light cystoscopy and a standard monopolar cautery resectoscope system. Patients with high-grade and/or pT1 disease were submitted to early re-TURBT within 6 weeks of the first resection.

Specimens were analyzed by dedicated urologic pathologists; pathologic staging was determined according to the 2009 TNM staging system, ¹³ and grading was determined according to the 2004 WHO classification. ¹⁴

The primary endpoint of the present study was to assess rates and predictors of DM− at final pathology, mainly focusing on the potential impact of surgical volume. Moreover, as a secondary endpoint, we looked at the eventual effect of DM− on patients' postoperative recurrence-free survival throughout the follow-up.

Descriptive statistics was used to analyze rates of DM− in the overall sample; differences in proportions were tested using the chi-square test (χ ²). Univariable analysis (UVA) and multivariable analysis (MVA) logistic regression models tested the association between predictors (e.g., tumor grade, type of resection, TUR status, tumor size, multifocality, morphology, tumor stage, and surgical experience) and DM− at final pathology. Kaplan–Meier analyses were used to depict the probability of disease recurrence over the follow-up; a log-rank test was used to compare recurrence-free survival rates of patients with DM presence or absence at final pathology. Statistical analyses were performed using SPSS statistical software, v 20.0 (IBM Corp., Armonk, NY). All tests were two sided, with a significance level set at 0.05.

Results

BCa was found in 302 cases. Table 1 lists pre- and intraoperative data regarding the overall cohort of procedures according to DM+ vs DM− status. Overall, DM was present in 273 (90.4%) of the specimens. No differences were found in terms of DM− rates between staff physicians and residents (p = 0.6; χ ²: 0.3). Similarly, sampling modality did not significantly differ between the two groups (p = 0.2; χ ²: 1.9) (Table 1).

Tumor stage was pTa, pT1, and pT2 in 179 (60.6%), 63 (21.9%), and 53 (17.5%) procedures, respectively; DM− was found in 63% and 37% of pTa and pT1 specimens, respectively. According to tumor characteristics, the rate of DM− specimens was higher for flat (namely, carcinoma in situ) compared with papillary lesions (p = 0.00; χ ²: 16.2). Conversely, no difference was found between groups in terms of tumor size and number of lesions. Dome lesions were more frequently resected without the inclusion of DM compared with lesions resected from all other bladder walls (Table 1). Table 2 depicts rates of DM+ and DM− surgical resections for each surgeon; overall, DM− procedures ranged from 0% to 20% per surgeon.

At UVA and MVA, only flat tumor morphology achieved independent predictor status for DM−, after accounting for clinical, pathologic, and anatomic variables (Table 3).

Overall, follow-up data were available for 210 (69.5%) patients, with a mean (SD) length of 9.89 (3.18) months. At Kaplan–Meier analysis, recurrence-free survival rates did not statistically differ between patients with DM+ or DM− (Fig. 1). Similarly, no difference was found between the two groups when patients were stratified considering only either pTa or pT1 tumors at final pathology (Fig. 1).

OPEN IN VIEWER

FIG. 1.

Kaplan–Meier estimates of disease recurrence according to detrusor muscle presence or absence at the final specimen, in the overall population (A) in pT1 only (B) and in pTa only (C). DM = detrusor muscle; DM + = detrusor muscle present; DM− = detrusor muscle absent. Color images available online at www.liebertpub.com/end

Discussion

In essence, TURBT is widely considered a simple surgical procedure; in contrast, TURBT could harbor well-known pitfalls, mainly involving the ability to achieve a complete tumor eradication with adequate tissue sampling to allow for an accurate staging of the disease. ^3,15 Indeed, NMIBC is characterized by a well-known risk of disease recurrence and progression to muscle-invasive disease, ranging from 15% to 61% and 0.2% to 17%, respectively, according to different risk categories based on pathologic and morphologic disease characteristics. ^1,2 As a whole, poor resection quality with incomplete tumor removal may lead to a less optimal treatment, thus contributing to the wide heterogeneity of patients' prognoses. ¹⁵ For instance, in a survey including a total of 2410 patients from seven European Organization for Research and Treatment of Cancer (EORTC) trials, the rate of BCa recurrence at first postoperative cystoscopy greatly varied among institutions, ranging from 3.5% to 20.6%, irrespective of disease characteristics. ¹⁶ This led the authors to speculate that the adequacy of the procedure, representative of surgeon skill and surgical experience, may greatly impact recurrence rates.

The importance of high-quality TURBT has been stressed by several authors in recent years. ^3,15,17,18 Pan and Soloway ³ suggested a detailed checklist with a meticulous analysis of all surgical steps; the authors warned that the typical operative team for a TUR in an academic setting is usually composed by junior residents who often perform the procedure without sufficient preparation and adequate senior supervision. Similarly, Richards et al. ¹⁵ recently analyzed the important steps of TURBT along with any recent advances of the procedure, thus outlining the importance of the technical skills needed for this procedure, an aspect often considered by junior surgeons to be trivial. Of relevance, especially for high-grade tumors, both surveys considered a deep resection of the bladder to be a crucial step in allowing for a detailed DM layer assessment, especially for high-grade tumors. ^3,15 Moreover, a minor thermal damage, which was able to reduce potential difficulties in the pathologic evaluation of the bladder specimens, has been shown using bipolar TUR ¹⁹ ; however, the choice of different energy sources, such as monopolar cautery or bipolar thermovaporization, should be regarded according to the surgeon's preference. ³ In this context, Herr ^6,7 considered the ability to obtain DM in the resection specimen as a proxy of surgical quality, along with the completeness of the resection and recurrence at the resection site. In a range between 9% and 49% of the cases, NMIBCs stage was underestimated at first resection; Herr ⁷ reported that 19.8% of initially staged Tis, Ta, and T1 diseases were proven to be muscle invasive and the absence of muscle in the resected specimen was considered to be responsible of 49% of staging errors. ⁷ Similarly, Dutta et al. ²⁰ showed understaging in 64% and 30% of patients whose tumors either lacked or presented DM in the specimen, respectively.

To this aim, we assessed rates and predictors of DM− in a series of specimens from TURBTs performed at a tertiary–referral academic center. Our interest was fueled by the evidence of recently published series showing rates of DM− after TUR ranging between 18.8% and 52.1%. ^{9 –12} These proportions may be considered too high, with the realization that better results may be achieved through the strict application of teaching programs and the utilization of modern equipment. ¹⁷ We found a rate of DM− specimens of roughly 10%, which is considerably lower compared with currently published series ^{9 –12} (Table 4). In a cohort of 225 patients treated with complete first TURBT, Mariappan et al. ¹⁰ found an overall DM− rate as high as 52.1%, although up to 83% of their resected tumors were staged pTa compared with 60.6% of tumors in our series, thus suggesting that a potentially more conservative surgical approach during the resection itself could have influenced the overall rates of DM− specimens. Our current findings of DM− TURs are actually even lower compared with the 18% rate of specimens without muscle layer recently shown by Shindo et al. ¹² in their series, including only TUR positive for pT1 BCa (Table 4), which appears to be quite relevant given the importance of including DM sampling in pT1 disease.

As a second major finding, we found that only flat tumor morphology achieved independent predictor status for DM− at resection. Multifocality, T-stage, tumor size, and surgeon experience but not morphology, have been variably associated with DM− at TURBT in previous series ^{9 –12} (Table 4). These findings are hardly comparable, however, given the different population of patients included in the studies, together with the differing definitions applied for the morphology categorization and the sizing of the lesions. Moreover, surgeon experience did not impact resection quality in our series; indeed, we did not find any significant difference in terms of DM−/+ rates between senior and junior surgeons. Conversely, Mariappan et al. ⁹ demonstrated a twofold higher risk of DM− resections for junior surgeons compared with senior surgeons. It is possible, however, that our policy of strict resident supervision by senior surgeons during surgical procedures could explain these findings, as was equally demonstrated in a previous study by Shoshany et al. ¹¹ In this context, considering the number of procedures performed by residents, we found a substantial difference between our study and other series, which reported a higher rate of TURBT performed by junior surgeons compared with ours, which could possibly explain our lower rate of DM− cases. However, we did observe poor intersurgeon heterogeneity in terms of DM+/− rates (Table 2), with a maximum of 20% of DM− specimens for a single surgeon, which is anyhow almost comparable to the lower currently published rate of DM− at TURBT. ^{9 –12}

Finally, despite the importance of DM sampling in the management of superficial BCa, we failed to demonstrate any potential impact of DM presence or absence on histologically proven disease recurrence. On the contrary, a significant association between higher rate of early recurrence and the absence of DM at final specimen was demonstrated in older published series. ^{9 –12} We consider that the lack of data regarding possible early adjuvant treatments received by patients after TURBT, together with the relatively lower number of DM– TURs included in the current sample, could have influenced our survival analysis.

These findings are of remarkable importance from the perspective of both surgical quality and training. Indeed, while previous data showed that a common surgical procedure such as TURBT could be affected by an undeniable risk of understaging and incompleteness of the resection, we conversely demonstrated that a higher quality resection could be achieved, lowering the rate of procedures without DM sampling below 10%. As a whole, the current data also provide evidence that the impact of surgeon experience could be largely attenuated with the proper supervision and training of residents.

Our study is not devoid of limitations. First, the retrospective design, with the possibility of a misclassification of tumor morphology and/or sizing, together with a potentially different approach of surgeons performing the resections of likely low-grade lesions, could undermine the significance of these findings. However, disease characteristics are routinely collected and included in our electronic database using the same criteria for all patients. Second, we did not observe a significant difference in terms of DM− specimens between low- and high-grade diseases. Moreover, the lack of a pathologic revision for each specimen could represent a pitfall, given the potential variability in terms of DM identification between different pathologists; however, all specimens were analyzed by dedicated experienced uropathologists, thus ensuring adequate reliability to our results. Finally, given the small number of patients, the rate of DM− procedures could have been underestimated, although previously published series reported similar results. ^{9 –12} Overall, these data certainly deserve a validation with a larger sample.

Conclusions

Given the higher risk of recurrence and progression associated with an incomplete and understaged TURBT, all efforts must be made to increase the overall quality of surgical procedures. In this context, the proportion of DM− procedures should be invariably lower than previously shown; we observed that up to 9 out of 10 procedures were comprehensive of DM sampling at final pathology, thus demonstrating that a 10% could be realistically pursued in the everyday clinical setting. Moreover, the lack of significant differences between surgeries performed by residents or staff physicians suggests that TURBT quality may be independent of surgical expertise in settings where young residents have been adequately trained.