Muscle-invasive bladder cancer is an aggressive disease, with a high rate of metastases. One third of bladder cancers are advanced at presentation and another 15% to 30% of highgrade superficial tumors progress to muscle-invasive disease, usually within 5 years. Up to 50% of patients with infiltrating disease will eventually succumb to their illness, most likely due to occult metastatic disease that was present at the time of initial diagnosis. Survival is dependent upon grade, pathologic stage, and nodal status. Extensive lymph node dissection has both prognostic and therapeutic significance. Effective chemotherapy is potentially curative in selected patients, as up to 70% of patients with metastatic disease have tumors that are chemosensitive.

For patients with muscle-invasive bladder cancer, chemotherapy is administered either before, in the “neoadjuvant” setting, or after cystectomy in the “adjuvant” setting, to eradicate micrometastatic disease. Whether or not chemotherapy can impact in a significant manner upon survival is still uncertain, based on the data derived from randomized clinical trials. The majority of studies evaluating neoadjuvant and adjuvant chemotherapy have unfortunately been underpowered, have utilized ineffective chemotherapy, or have had methodological flaws (1). Two sufficiently powered neoadjuvant studies as well as meta-analyses provide some insight. The effectiveness of adjuvant chemotherapy has still not been clearly ascertained from randomized trials.

Neoadjuvant chemotherapy has been designed for patients with operable clinical stage T2 to T4a muscle-invasive disease. Chemotherapy prior to cystectomy or, in some cases, full-dose radiation therapy (RT) is geared toward treating micrometastatic disease present at the time of diagnosis. The rationale for giving upfront chemotherapy is based on the intent to improve survival by treating micrometastatic disease, present but not detected at the time of initial staging. Neoadjuvant chemotherapy has also been proposed in some programs with the intention of bladder preservation (2).

There are several advantages to giving chemotherapy in the neoadjuvant setting. Systemic therapy is delivered early when the burden of metastatic disease is minimal, and the therapy is usually well tolerated prior to surgery or RT. The toxicity is presumably lower than in patients with metastatic disease, as subjects with localized disease usually have a better performance status. Patients are often able to tolerate a higher dose intensity and more cycles of chemotherapy preoperatively than postoperatively (3). Drug delivery is also in all probability better as the vascular bed is intact. Neoadjuvant chemotherapy allows in vivo drug sensitivity testing that may provide useful information for subsequent therapy. The primary tumor can be evaluated for response, which also has major prognostic significance. In addition, preoperative chemotherapy may downstage tumors, potentially permitting technically easier surgery (4). It is not known whether three or four cycles are needed as no study has ever systematically evaluated this question.

Disadvantages of neoadjuvant chemotherapy include uncertainties and unnecessary treatment of misstaged lowrisk tumors. Discrepancies between clinical and pathological staging have been reported in some 30% of cases and can complicate the evaluation of response to neoadjuvant chemotherapy (5,6). The major disadvantage of neoadjuvant chemotherapy, however, is the delay in definitive local therapy in patients who do not respond and whose disease progresses. An interval longer than 12 weeks between the diagnosis of muscle invasion and cystectomy has been associated with a poorer outcome (7,8).

Another theoretical disadvantage of neoadjuvant chemotherapy is the potential increased incidence of perioperative morbidity. This has actually been evaluated in very few studies (9,10). In a comparative study of neoadjuvant and adjuvant chemotherapy, neoadjuvant chemotherapy did not increase perioperative morbidity (9). Additionally, randomized studies comparing patients with and without neoadjuvant treatment demonstrate a similar postoperative complication risk. In the United States and in the majority of European countries, radical cystectomy is the preferred treatment for fit patients who have a good performance status.

Many randomized trials have evaluated the impact of neoadjuvant chemotherapy on survival. These trials have shown either a trend toward better survival or no benefit. The majority of trials have, however, been underpowered. Randomized neoadjuvant chemotherapy trials are summarized in Table 26.1 (3,11,12,13,14,15,16,17,18,19,20,21,22).

The largest trial of neoadjuvant chemotherapy was performed by the European Organization for Research and Treatment of Cancer (EORTC) and the Medical Research Council (MRC). Nine hundred and seventy-six patients were accrued over a 5½-year period in 106 institutions and were randomized to receive CMV (methotrexate, vinblastine, and cisplatin) neoadjuvant chemotherapy or no chemotherapy. Cystectomy and/or RT were permitted as definitive management of the primary tumor. The study aimed to detect an absolute improvement in survival of 10% (from 50% to 60%) with a power of 90% and a type I error of 5%.

When first published, there was a nonsignificant trend toward improvement in survival in patients treated with neoadjuvant CMV chemotherapy (22). In an update presented at the ASCO meeting in 2002 (but still not published), with longer follow-up of 7.4 years, the data have just reached statistical significance (p = 0.048). There was a 5.5% benefit with neoadjuvant CMV (14). Survival at 5 years was 50% compared to 44%, and at 8 years was 43% versus 37%. Following neoadjuvant chemotherapy, no residual tumor or pathologic complete response (pCR) was found in 32.5% of cystectomy specimens. Chemotherapy-related mortality was 1%. Chemotherapy did not increase the rate of postoperative complications.

Although the study demonstrated an improvement with CMV chemotherapy, the improvement was less than what was originally sought in this trial. A number of factors may explain why the chemotherapy arm in this study achieved somewhat lower survival than anticipated. The CMV regimen used in this study is not considered as standard of care, and has never been compared to MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin) in a randomized trial. Definitive therapy was comprised of either cystectomy or RT (or both), with 42% of patients receiving RT alone as local therapy in the neoadjuvant chemotherapy group and 43% in the no chemotherapy group. RT is believed to be equivalent to cystectomy in Europe, but no prospective randomized trial has ever proven this.

The Southwest Oncology Group (SWOG) Intergroup trial attempted to detect an increase in survival with MVAC chemotherapy. Three hundred and seventeen patients with clinical T2-T4a bladder cancer were randomized to three cycles of neoadjuvant MVAC chemotherapy followed by cystectomy or to cystectomy alone (15). Enrolment took place over 11 years at 126 institutions. Patients were stratified prior to randomization, age (<65 or ≥65 years), and clinical stage (cT2 vs. cT3 or cT4a). The statistical plan was to perform one-sided testing, as the investigators hypothesized that medical practice would change only if there was an improvement with chemotherapy.

The median survival was 77 months for patients who received neoadjuvant MVAC chemotherapy as compared to 46 months for patients randomized to surgery alone (p = 0.06; two-sided stratified log rank test). Due to the very long accrual period and follow-up, these results achieved borderline statistical significance, and the estimated risk of death was reduced by 25% in the group assigned to receive MVAC and cystectomy, as compared with the group assigned to undergo cystectomy alone. In both groups, improved survival was associated with the absence of residual cancer in the cystectomy specimen. Significantly more patients in the combination-therapy group had no residual disease than patients in the cystectomy group (38% vs.15%, p < 0.001). Patients with locally advanced (cT3 or T4a) disease had the greatest survival benefit from neoadjuvant chemotherapy (65 months vs. 24 months). In this trial, there were no deaths attributable to MVAC.

Many oncologists prefer to use newer agents such as gemcitabine or the taxanes, which have not been evaluated in randomized studies. Gemcitabine plus cisplatin (GC) has similar efficacy and less toxicity than MVAC in metastatic disease but relatively is untested in the neoadjuvant setting. Four cycles of neoadjuvant GC were administered before radical cystectomy to 42 patients at Memorial Sloan-Kettering Cancer Center (MSKCC) with comparisons made to historical controls treated with neoadjuvant MVAC (23). No residual disease (pT0) was obtained in 26% and no residual muscle-invasive disease (<pT2) in 36%. This was compared to 54 historical MVAC-treated patients, in whom pT0 was achieved in 28% and <pT2 in 35%.

In another study from The Cleveland clinic these results with neoadjuvant GC were not confirmed (24). In this report, there were 117 patients who underwent radical cystectomy of whom 29 (25%) received neoadjuvant chemotherapy; 20 (69%) patients received GC, 4 (14%) received MVAC, and 5 (17%) received other regimens. Few patients achieved a pCR with neoadjuvant chemotherapy and most experienced rapid disease progression. The authors hypothesized that the poor outcomes may have been related to the use of non-MVAC-based regimens or excessive delay in performing cystectomy. In the absence of supportive data for GC in the neoadjuvant setting, MVAC is still the preferred regimen. Excessive delays in performing RC may negate the benefit of neoadjuvant chemotherapy (25). Although the results with GC may appear similar to MVAC, further confirmative studies are required before its efficacy in metastatic disease can be extrapolated to the neoadjuvant setting.

The majority of neoadjuvant randomized trials have failed to show an unequivocal survival benefit in favor of chemotherapy. However, the majority of the studies have been compromised
by inadequate sample sizes, suboptimal chemotherapy, premature closure, or inadequate follow-up (1). For this reason, meta-analyses have attempted to explain and interpret these data (26,27).

The advanced bladder cancer (ABC) meta-analysis included data on 3,005 patients from 11 randomized neoadjuvant chemotherapy trials (27). In patients who received single-agent cisplatin chemotherapy, neoadjuvant chemotherapy appeared detrimental, but the numbers were too small to support definitive conclusions (p = 0.264). The overall analysis was, however, in favor of chemotherapy HR = 0.89 (95% CI 0.81-0.98), p = 0.022. This is primarily based upon the subset of patients who received cisplatin-based combination neoadjuvant chemotherapy. The meta-analysis showed a reduction in the risk of death of 14% (HR = 0.86, 95% CI 0.77-0.95, p = 0.003) translating into a 5% absolute survival benefit at 5 years (95% CI: +2% to +9%). The results are similar to the results obtained in the EORTC and MRC trial, which was the largest trial in the analysis (14). The authors noted no difference in the relative risk reduction in relation to the type of local treatment (cystectomy, RT, or the combination) or in relation to patient characteristics. However, apart from this study, neoadjuvant chemotherapy prior to RT has not been shown to improve survival (28).

Because of the inaccuracies of clinical staging, the potential morbidity related to MVAC chemotherapy, a 70% cure rate in pT2No disease with surgery alone, and only a modest 5% improvement in absolute overall survival with combined therapy, many favor limiting chemotherapy to patients with a pathologic stage of pT3 or greater or node-positive disease (29).