Role of immediate radical cystectomy in the treatment of patients with residual T1 bladder cancer on restaging transurethral resection

Authors


Correspondence: Guido Dalbagni, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, USA.

e-mail: dalbagng@mskcc.org

Abstract

What's known on the subject? and What does the study add?

  • Bladder cancer patients with lamina propria invasion (T1 disease) and residual T1 disease on restaging transurethral resection of bladder tumour (re-TURBT) are at a very high risk for recurrence and progression. Despite this risk, most patients are treated with a bladder preserving approach and not immediate radical cystectomy (RC).
  • In this study we have shown that a quarter of patients with T1 bladder cancer and residual T1 on re-TURBT who are treated with immediate RC are found to have carcinoma invading bladder muscle at RC and 5% have lymph node metastases. We have also found that >30% of patients treated with deferred RC after initial bladder-preserving therapy harbour carcinoma invading bladder muscle and almost 20% of these patients have lymph node metastases. Thus, immediate RC should be considered in all patients with T1 bladder cancer and residual T1 on re-TURBT.

Objective

  • To report the overall survival (OS) and cancer-specific survival (CSS) of patients with residual T1 bladder cancer on restaging transurethral resection of the bladder tumour (re-TURBT).

Materials and Methods

  • We performed a retrospective review of 150 evaluable patients treated for T1 bladder cancer with residual T1 disease found on re-TURBT between 1990 and 2007.
  • Patients were treated with immediate radical cystectomy (RC) or a bladder-preserving approach (deferred or no RC).
  • A univariate Cox proportional hazards regression model was used to test the association between treatment approach and survival.

Results

  • Residual T1 bladder cancer was found in 150 evaluable patients, of whom 57 received immediate RC and 93 were treated with a bladder-preserving approach.
  • Fourteen out of 57 patients receiving immediate RC and 8/26 patients receiving deferred RC had carcinoma invading bladder muscle in the RC specimen. Three out of 57 and 5/26 patients had lymph node metastases in the RC specimen.
  • Median follow-up was 3.74 years.
  • Thirty-nine patients died during follow-up, 16 from bladder cancer. There was no significant association between immediate RC and CSS (hazard ratio [HR] 1.15, 95% confidence interval [CI] 0.43–3.09, P = 0.8) or OS (HR 0.79, 95% CI 0.4–1.53, P = 0.5).

Conclusions

  • Because of the low number of events we cannot conclude whether RC offers a survival advantage in patients with residual T1 bladder cancer on re-TURBT.
  • Since a quarter of patients had carcinoma invading bladder muscle, RC should be considered in these patients.
  • A larger, preferably randomized, study with longer follow-up is needed.

Introduction

Approximately 70–80% of patients with bladder cancer present initially with non-muscle-invasive disease, and 20% of these involve the lamina propria (stage T1) [1, 2]. It is recommended that patients with stage T1 bladder tumours undergo a restaging transurethral resection of the bladder tumour (re-TURBT) before deciding on treatment, since 50–70% of these patients continue to harbour viable tumour and 20–30% will be upstaged to carcinoma invading bladder muscle [3, 4]. Patients for whom residual T1 bladder cancer is found at re-TURBT constitute an especially high-risk group and have an 88% chance of recurrence and an 82% chance of progression to carcinoma invading bladder muscle within 5 years. By comparison, patients with <T1 pathology (T0, Ta or Tis) during re-TURBT have a 58% recurrence rate and a 19% progression rate at 5 years [5].

There is an ongoing debate over the role of early radical cystectomy (RC) for patients with T1 bladder cancer, because some studies have shown a survival advantage while others have not [6-10]. Our aim was to report the cancer-specific survival (CSS) and overall survival (OS) of patients with clinical stage T1 bladder tumours who had residual stage T1 disease on re-TURBT and were initially treated with either immediate RC or a bladder-preserving approach.

Materials and Methods

After obtaining institutional review board approval, we performed a retrospective review of our institutional database to identify patients treated between 1990 and 2007 for clinical T1 bladder cancer and residual T1 disease found at re-TURBT. Diagnosis of clinical T1 disease was based on cystoscopy, bimanual examination, radiographic examinations and pathological evaluation. Patients whose pathology slides were unavailable and patients with histology other than TCC were excluded, but patients with TCC with aberrant differentiation were included. Of 523 patients with T1 bladder cancer who underwent a re-TURBT, 157 were found to have residual T1 disease. All tumours were staged according to the TNM system of the Union for International Cancer Control and graded according to the WHO/International Society of Urological Pathology 1998 classification system [11].

Multifocality was defined as a history of Ta tumours or concomitant papillary tumours. Multifocal T1 disease was defined as the presence of concomitant T1 tumours. For inclusion in the study, the re-TURBT had to have been performed within 3 months of a patient's initial T1 diagnosis. RC was defined as immediate if performed <3 months after the re-TURBT and deferred if performed >3 months after a re-TURBT in a patient initially treated with conservative management whose tumour recurred or progressed. Patients were followed with cystoscopy, urine cytology and imaging every 3–6 months.

A landmark time analysis was performed using a landmark time of 3 months after the date of re-TURBT, so that immediate RC status could be ascertained. Patients who died within 3 months of re-TURBT were excluded, leaving 150 patients in the final cohort.

Kaplan–Meier curves were calculated to determine the difference in CSS between the two treatment approaches. Although we originally planned multivariable Cox proportional hazards regression to test if the treatment approach had any effect on CSS and OS, we were restricted to a univariate approach owing to low event numbers. Statistical analyses were conducted using stata 11 software (StataCorp, College Station, TX, USA).

Results

The descriptive characteristics of the 150 included patients are shown in Table 1. The re-TURBT procedure was performed at the Memorial Sloan-Kettering Cancer Center for 76% of patients (114/150) and at an outside institution for 24% (36/150). The median (interquartile range [IQR]) time from the diagnostic TURBT to the re-TURBT was 32 (22–46) days. The treatment approach was chosen by the patient after consultation and discussion of the different treatment options with the treating surgeon.

Table 1. Characteristics of patients with residual T1 disease on re-TURBT, stratified by treatment.
 Deferred or no RC N = 93Immediate RC N = 57
  1. *Multifocality was defined as a history of Ta tumours or concomitant papillary tumour. Multifocal T1 disease was defined as having concomitant T1 tumours. Grade was unknown for one patient without immediate RC. §Percentage of evaluable patients (missing data in 26 patients).
Characteristics at initial T1 diagnosis  
Median age (IQR), years65 (55–73)67 (61–73)
Female, n (%)21 (23)12 (21)
Multifocal, n (%)  
Any multifocal*25 (27)12 (21)
Multifocal T113 (14)8 (14)
Carcinoma in situ, n (%)26 (28)12 (21)
High grade, n (%)92 (100)57 (100)
Aberrant differentiation, n (%)4 (4)9 (16)
BCG treatment, n (%)  
Before T1 diagnosis5 (5)6 (11)
After T1 diagnosis60 (65)0 (0)
Muscle in specimen from re-TURBT, n (%)  
No16 (17)13 (23)
Yes77 (83)44 (77)
Incomplete resection on re-TURBT, n (%)§8 (11)14 (29)

Of 150 patients with T1 bladder cancer and residual T1 disease on re-TURBT, 57 (38%) received an immediate RC and 93 (62%) were initially treated with a bladder-preserving approach. Sixty out of 93 (65%) of patients not treated with immediate RC were treated with intravesical BCG. The remaining 33 were treated with partial cystectomy and cystoscopic surveillance (four patients), lost to follow-up after re-TURBT (eight patients) or did not get any adjuvant intravesical therapy (21 patients). Of the latter 21 patients, 10 eventually underwent RC, 10 were followed without need for RC and one patient progressed to metastatic bladder cancer. Twenty-six out of 93 (28%) patients were treated with deferred RC after failing conservative treatment at a mean (range) of 8 (3–29) months after re-TURBT. Four patients were treated for progression to carcinoma invading bladder muscle, whereas 22 were treated for recurrent non-muscle-invasive bladder cancer. Of the latter 22 patients, five (23%) were upstaged to carcinoma invading bladder muscle at cystectomy. Overall, carcinoma invading bladder muscle was found in the RC specimen of 25% of patients (14/57) treated with immediate RC and 31% of patients (8/26) treated with deferred RC, with a single patient who had carcinoma invading bladder muscle before deferred RC found to have no residual tumour in the RC specimen. Three out of 57 (5%) patients treated with immediate RC and 5/26 (19%) patients treated with deferred RC had pathologically proven lymph node metastases.

Aberrant differentiation was found in 9/57 patients (16%) treated with immediate RC, of whom two had TCC with micropapillary features, four had TCC with squamous differentiation, two had TCC with glandular differentiation and one had a nested variant TCC. Of 93 patients treated initially with a bladder-preserving approach, four (4%) had aberrant differentiation. A single patient had TCC with micropapillary features, two had TCC with squamous differentiation and a single patient had TCC with glandular differentiation.

None of the patients treated with immediate RC received neoadjuvant chemotherapy, whereas 3/57 (5%) were treated with adjuvant chemotherapy (four cycles of gemcitabine and cisplatin in two patients and six cycles of gemcitabine and cisplatin in the third patient). Four out of 26 patients (15%) treated with deferred RC received chemotherapy. One patient received neoadjuvant chemotherapy (four cycles of gemcitabine and cisplatin), two patients received adjuvant chemotherapy (four cycles of gemcitabine and cisplatin), and one patient was treated with neoadjuvant chemotherapy (gemcitabine and cisplatin) that was discontinued after the second cycle because of thrombocytopenia and auditory acuity decline, and adjuvant chemotherapy with four cycles of carboplatin and paclitaxel.

Other than complete resection at re-TURBT, presence of aberrant differentiation of the tumour and treatment with BCG after re-TURBT with residual T1 disease, no obvious differences in clinical characteristics were observed between the two treatment groups (Table 1). The higher rate of BCG treatment in the bladder-preservation group was expected, as patients not treated with immediate RC are likely to be treated with BCG [12]. Overall, 60/93 patients (65%) not treated with immediate RC were treated with intravesical BCG after the diagnosis of T1 bladder cancer. The rate of incomplete resection at re-TURBT in the group of patients treated with immediate RC (29%) was higher than the rate of incomplete resection in the group of patients treated with bladder preservation (11%). This difference can be explained either by a decision of the surgeon not to resect aggressively a tumour that seemed likely to require a cystectomy, or by assuming that more advanced tumours are larger and less amenable to complete resection. Either explanation supports the notion that the patients treated with immediate RC had tumours with worse features. The higher rate of aberrant differentiation in patients treated with immediate RC also supports the notion that patients in this group had worse features and were thus chosen to be treated with immediate RC.

At last follow-up, 39 patients had died, 16 of them from bladder cancer. The median (IQR) follow-up for surviving patients was 3.74 (2.2–6.0) years. The median (IQR) follow-up for surviving patients treated with immediate RC was 4 (2.4–5.8) years and 3.5 (2.2–6.1) years for those treated with bladder preservation initially.

There was no significant association between treatment approach (immediate RC vs. bladder preservation) and improved CSS or OS on univariate Cox regression analysis (hazard ratio [HR] 1.15, 95% CI 0.43–3.09, P = 0.8; and HR 0.79, 95% CI 0.4–1.53, P = 0.5, respectively). Kaplan–Meier estimates of CSS are shown in Fig. 1. The 5-year CSS was 89% (95% CI 77–95%) in the bladder-preservation group and 85% (95% CI 66–93%) in the immediate RC group. The 5-year OS was 73% (95% CI 59–83%) for patients treated with bladder preservation initially and 72% (95% CI 53–84%) for patients treated with immediate RC.

Figure 1.

Kaplan–Meier curve of CSS among patients with clinical T1 disease on re-TURBT. (Immediate RC = solid line, No immediate RC = dashed line).

Discussion

Most patients with high-risk non-muscle-invasive bladder tumours are managed initially with a bladder-preserving approach, and only a minority of patients receive immediate RC [12]. These patients are treated with intravesical immunotherapy or chemotherapy after resection of the bladder tumour, followed by cystoscopic and urine cytological examinations [13]. The rationale for performing re-TURBT before making a treatment decision in patients with stage T1 bladder cancer is twofold: to reveal carcinoma invading bladder muscle that was under-staged during the initial resection, and to resect residual viable tumour [3, 4]. Patients with residual T1 bladder cancer on re-TURBT have worse outcomes when compared with patients with lower stage residual tumour. These patients are at a considerable risk of recurrence and progression to carcinoma invading bladder muscle [5]. This observation has led several authors to recommend that such patients be treated with immediate RC rather than a bladder-preserving approach [5, 6] and to highlight the low use of this treatment method [12]. Herr and Sogani [6] observed in a previous report that patients with high-risk non-muscle-invasive bladder cancer have better survival when treated with early RC than delayed RC, although the definition of early RC in that report was different from the definition of immediate RC used in the present series (within 2 years of initial BCG treatment in their report vs. within 3 months of re-TURBT in the present series). Moreover, a substantial number of patients in their series (61%) were treated after progression to carcinoma invading bladder muscle [6]. By contrast, other reports either failed to show a survival advantage in patients treated with RC for T1 bladder cancer when compared with patients treated with a bladder-preserving approach [14], showed inferior survival in patients treated with RC [10], or showed that incorporating quality of life into the analysis reduces or eliminates the survival advantage of RC [9].

Patients with residual T1 disease found at re-TURBT have a risk of progression of up to 82% at 5 years [5]. We hypothesized that this group of patients would benefit from immediate RC, but as the cohort in the present study was relatively small and the number of events low, this survival advantage could not be proved. This can be seen in the large 95% CI of both the OS and the CSS results. The results show that immediate RC can lower cancer-specific mortality by as much as 57% or increase it threefold. The effect of immediate RC on OS in the present cohort was between lowering overall mortality by 60% to elevating it by 53%. Thus, a larger study with longer follow-up is needed to resolve this issue.

Two previous studies have shown that patients with clinical T1 disease upstaged to carcinoma invading bladder muscle at RC had lower recurrence-free survival and disease-specific survival than patients who were not upstaged at RC [15, 16]. A third study showed that patients with clinical T1 disease upstaged to pathological T2 at RC had worse recurrence-free survival than patients correctly staged as clinical T2 before RC. When patients without previous BCG treatment (about 36% of the cited study's cohort) were compared, patients treated with immediate RC for clinical T1 and upstaged to pathological T2 at RC had similar 3- and 5-year recurrence-free survival rates to patients accurately staged preoperatively (65% and 65% vs. 77% and 69%, respectively; P = 0.39) [17]. A major limitation of these three studies is that most patients included in them were not restaged with a second transurethral resection before RC, in contrast to patients included in the present cohort that were all restaged with a re-TURBT.

A recent report of an international cohort of 1136 patients treated with RC for clinical T1, of whom nearly 70% of patients were restaged with re-TURBT, showed higher recurrence and cancer-specific mortality rates among patients upstaged to carcinoma invading bladder muscle at RC [18].

Because a quarter of patients treated with immediate RC for clinical T1 in the present cohort (14/57 patients) were upstaged to carcinoma invading bladder muscle at RC, and because it has been shown, as described above, that such patients have worse recurrence-free survival and CSS when compared with patients not upstaged, and similar recurrence-free survival when compared with patients treated with RC for known muscle-invasive disease, these patients would have been under-treated had they not been treated with RC as this has been shown to be the treatment method of choice in patients with carcinoma invading bladder muscle. Moreover, patients treated with deferred RC had a higher rate of lymph node metastases at surgery (19% vs. 5%), possibly reflecting a loss of window of opportunity for cure in these patients.

We acknowledge the limitations of the present report: its retrospective nature restricts our ability to compare the two treatment groups; the cohort is relatively small and the number of events is limited; and a selection bias is clearly present because this cohort represents highly selected patients in a referral cancer centre. More than that, there is an obvious selection bias in the assignment of patients to the different treatment groups, as can be seen in the difference in baseline characteristics of the patients in the two treatment groups and as discussed above.

In conclusion, as patients with T1 bladder cancer and residual T1 disease on re-TURBT are at very high risk for progression to carcinoma invading bladder muscle, and in light of the fact that a quarter of these patients may harbour muscle-invasive disease not identified at re-TURBT, immediate RC should be considered in this group of patients. This is especially true in patients with unresectable disease in the re-TURBT. Patients treated with a bladder-preserving approach should be followed closely, as more than a quarter of them will eventually require RC and about 30% of patients treated with deferred RC will be found to have carcinoma invading bladder muscle at the time of RC and about 20% will harbour lymph node metastases at the time of surgery. Given the limitations of this study, however, a larger, preferably randomized, study with longer follow-up is needed to further elucidate the best treatment for patients with T1 bladder cancer and residual T1 disease on re-TURBT.

Acknowledgements

Supported by the Sidney Kimmel Center for Prostate and Urologic Cancers.

Conflict of Interest

None declared.

Abbreviations
OS

overall survival

CSS

cancer-specific survival

re-TURBT

restaging transurethral resection of the bladder tumour

RC

radical cystectomy

HR

hazard ratio

Ancillary