Recurrence of disease in patients with muscle-invasive urothelial carcinoma of the bladder (UCB) is characterized by distant metastases. Evidence supports the use of neoadjuvant, cisplatin-based combination chemotherapy for resectable clinical (c)T2-T4aN0M0 stage UCB1 preceding radical cystectomy (RC). However, the adjuvant chemotherapy (AC) literature is plagued with underpowered and incomplete trials. A recent, large, retrospective cohort study suggests a benefit in patients who have tumors classified as pathologic stage ≥T3 and lymph node involvement.2 Despite such data, neoadjuvant chemotherapy (NC) seldom has been used and, indeed, has been used less than AC. A recent presentation demonstrated that only 12% of patients with T2-T4aN0M0 disease received NC.3 Surprisingly, a higher proportion of patients (22%) received AC. A mere 9% of all patients received cisplatin-based NC. Although the reasons for this practice pattern are unclear, renal dysfunction, poor performance status, comorbidities, and age may be responsible. Moreover, data demonstrate the high prevalence of renal dysfunction in the perioperative setting, and the probability of renal dysfunction increases with age to >40% of patients aged >70 years.4
Conversely, RC permits pathologic risk stratification, whereas current clinical staging frequently under stages patients. Comprehensive pathologic information may obviate unnecessary perioperative chemotherapy in patients who have a high probability of achieving a cure with RC alone, and AC may be reserved for those with a high risk of recurrence, ie, extravesical or lymph node-positive UCB. For example, patients with cT2N0 disease who are down staged to less than pT2N0 status at RC and who exhibit low-grade histology and a lack of lymphovascular invasion enjoy outstanding long-term outcomes and 5-year cancer-specific survival rates of approximately 95% without perioperative chemotherapy.5 NC may over treat such patients and expose them to the adverse effects of cisplatin chemotherapy, some of which may be longstanding, eg, peripheral neuropathy and renal dysfunction. Another issue that complicates decision-making is that approximately 33% of patients may not be candidates for AC within 90 days of RC because of complications.6 Conversely, neoadjuvant methotrexate, vinblastine, doxorubicin, and cisplatin (MVAC) does not appear to render RC infeasible or increase postoperative complications in a cisplatin-eligible population.1 One randomized trial compared AC with a combination of NC and AC and reported similar outcomes that did not suggest any inferior feasibility of AC.7 These data and the observation that trials of AC have accrued poorly suggest that, with the exception of selected patients who have an excellent performance status, NC may be more practical than AC. Thus, a rationale may be made for NC in most patients except cisplatin-ineligible patients and cisplatin-eligible patients with an excellent performance status and clinical T2N0 disease.
In this issue of Cancer, Scosyrev and colleagues report a retrospective cohort study of neoadjuvant gemcitabine plus cisplatin (GC) for cT2-T4Nany UCB performed at a single institution.8 Their study corroborates the finding that only a small proportion of patients receive NC (25 of 160 patients), and a higher proportion receives AC (42 of 135 patients). Patients who received neoadjuvant GC were younger and had fewer comorbidities. Moreover, patients who received neoadjuvant GC were more likely to have clinical extravesical or lymph node disease, probably reflecting a preference of some urologists to reserve NC for those with more advanced UCB. Neoadjuvant GC yielded a pT0 rate of 20% compared with the approximately 30% pT0 rate observed in a prospective trial of MVAC in patients with cT2-T4aN0 disease.1 This is likely a result of including more advanced and lymph node positive disease; indeed, patients with lymph node involvement are included in metastatic disease trials. Another explanation is that, in patients with metastases (including lymph node disease), up to 6 cycles of GC are planned; whereas, in the study by Scosyrev et al, the median number of cycles delivered was 4. Nevertheless, it is disappointing that neoadjuvant GC did not eradicate lymph node disease in those with clinical lymph node disease. Other data demonstrate that NC followed by RC and extended lymphadenectomy may yield durable remissions in patients who had baseline pelvic lymphadenopathy.9 Conversely, outcomes appear dismal in patients who have pathologic lymph node-positive disease at RC after NC, and separate studies have reported a 2-year recurrence-free survival rate of 13.5% and a median survival of 2.4 years.10, 11 Therefore, it may be advisable to administer 6 cycles of chemotherapy to patients who have clinical lymph node involvement and to proceed to RC only if a clinical complete response occurs with chemotherapy. In addition, those with residual lymph node disease after RC should be considered for trials of noncross-resistant therapy. Another research goal could be to avoid RC in patients who are pathologically clear of all disease after NC. Unfortunately, current staging is grossly inadequate to identify residual pathologic disease before surgery.12 The value of RC after chemotherapy as opposed to chemotherapy alone for patients with clinical lymph node disease warrants study. Regarding the regimen used, optimally administered GC has appeared similar to MVAC in a retrospective cohort comparison.13 Notably, the study by Scosyrev et al included patients who received cisplatin as a split-weekly dose in addition to the traditional bolus dose once every 3 to 4 weeks. Weekly cisplatin in combination with gemcitabine appears to demonstrate similar activity and an excellent nephrotoxicity profile in patients with baseline renal dysfunction (creatinine clearance ≥40 mL/minute) and advanced urothelial carcinoma.14 Therefore, a weekly GC regimen may be worth exploring in the neoadjuvant setting for patients with renal dysfunction.
Novel treatment strategies are necessary to overcome the poor applicability of cisplatin-based chemotherapy. Although both the quality of surgery and pathologic down-staging to pT0 or <pT2 appear to be associated with improved survival, downstaging to pT0 may be the more optimal pathologic endpoint, as suggested by a trend toward better long-term outcomes compared with the presence of residual nonmuscle-invasive disease (median survival, 13.6 years vs 10.6 years; hazard ratio, 2.05; P = .054).11 Notably, ongoing nonrandomized phase 2 trials are evaluating dose-dense (DD)-MVAC alone or with bevacizumab and GC in combination with sunitinib or bevacizumab. Some of those trials include UCB as well as upper tract urothelial carcinoma (UTUC). However, the pT0 rate with neoadjuvant chemotherapy for UTUC may be lower than that for UCB, perhaps because of differences in tumor biology. In a recent report, only 14% of patients with UTUC exhibited pT0, and the majority had received cisplatin-based chemotherapy.15 Therefore, UTUC may continue to warrant separate, dedicated trials. It is also unclear whether the combination of chemotherapy with biologic agents will increase the pT0 rate, even if long-term outcomes are enhanced, or vice versa. Therefore, the pathologic response endpoints require validation in the setting of chemobiologic combinations but, with further validation, may expedite drug development. Data from the breast cancer setting suggests that a relatively large increment in pathologic complete response may be necessary to translate to a survival benefit. Because DD-MVAC appeared to yield better outcomes in 1 randomized trial for advanced urothelial carcinoma, the comparison of neoadjuvant DD-MVAC and GC is appealing. In addition, neoadjuvant weekly cisplatin or carboplatin in combination with gemcitabine for those with renal dysfunction may warrant further investigation. The lack of a uniformly applicable chemotherapy template to populations with and without renal dysfunction has impeded their accrual into trials. Potentially, the regimen of weekly GC may be broadly applicable with a creatinine clearance ≥40 mL/minute and may provide a uniform template to develop neoadjuvant chemobiologic combinations. Baseline clinical staging procedures also vary, with physicians variably using a combination of bimanual examination and imaging. Although trials of NC have required pathologic demonstration of muscle invasion, a clinical stage ≥T3 and/or hydronephrosis coupled with lack of demonstration of muscle-invasion on biopsy may occur. A consensus to include such patients in NC trials may be justified.
Two-year and 3-year progression-free survival appear to be associated with 5-year survival in the setting of RC alone or with AC, and further validation of this endpoint may complement pathologic endpoints and assist in accelerating the development of novel perioperative regimens.16 Patients who are ineligible for neoadjuvant cisplatin or who elect an initial RC approach could be offered trials evaluating brief neoadjuvant biologic agents to assess biologic activity in tumor tissue. One issue to consider when enrolling patients on brief neoadjuvant therapy trials, especially those with T2N0 disease, is that approximately 40% of these patients may be down staged to pT0 by the initial transurethral resection of the bladder tumor alone, thereby not permitting a analysis of post-therapy tumor tissue.5 Hence, those with small-volume T2N0 tumors for whom the probability of achieving complete transurethral resection of the bladder tumor is high may not be optimal candidates for such trials. In addition, patients who have residual disease at the time of RC after NC should be offered trials evaluating novel biologic agents in the adjuvant setting. Moreover, those who proceed to initial RC and have extravesical or lymph node-positive disease probably should be offered AC trials evaluating combination cisplatin-based chemotherapy alone or with added biologic agents. This is based on the assumption that, despite the lack of definitive data with regard to the efficacy of AC, cisplatin-based AC is reasonable in high-risk patients given the circumstantial evidence to support it. Those who undergo initial RC and are not cisplatin candidates may be offered adjuvant clinical trials evaluating tolerable biologic agents or non-nephrotoxic chemotherapy, as exemplified by an ongoing European randomized trial evaluating gemcitabine monotherapy.
A layer of complexity is conferred by the role that molecular factors will play in predicting disease recurrence (prognostic factors) and benefit from chemotherapy (predictive factors). Gene expression profiling may assist in selecting patients.17, 18 Prospective trials need to incorporate molecular studies to eventually deliver systemic therapy not simply to those at high risk of disease recurrence but also to those likely to benefit. The advent of high-throughput technologies is allowing comprehensive identification of molecular markers. The challenge remains to optimize measurement of these targets and to translate molecular markers into utility in clinical management. The science with respect to systemic therapy for urothelial carcinoma is in flux with no clear direction because of a poor understanding of the molecular mechanisms underlying progression. Commitment to trials and a combination of pragmatic trial designs and better knowledge of biology are imperative to yield advances. A multidisciplinary approach and better collaboration is indispensable.