Contemporary use of perioperative cisplatin-based chemotherapy in patients with muscle-invasive bladder cancer
Level I evidence indicates that neoadjuvant cisplatin-based chemotherapy, in combination with radical cystectomy (RC), is associated with a significant survival advantage for patients with muscle-invasive bladder cancer. Despite this, neoadjuvant chemotherapy is not uniformly used. Our objective was to determine the patterns of utilization of neoadjuvant chemotherapy in patients undergoing RC for muscle invasive bladder cancer in a contemporary cohort in a tertiary care center.
A retrospective review was performed of patients with bladder cancer who underwent RC between 2003 and 2008 at our institution. Clinical stage, pathologic stage, renal function, and perioperative chemotherapy treatments were tabulated. Primary outcome measures were the type and use of neoadjuvant chemotherapy among eligible patients. Secondary measures were the utilization patterns of adjuvant chemotherapy, renal function, pathologic outcomes, and disease specific and overall survival. Reasons for nonutilization of chemotherapy were also examined.
Among 238 patients who underwent RC for bladder cancer, 145 had a preoperative clinical stage ≥T2. Only 17% (25 of 145) of these patients received cisplatin-based neoadjuvant chemotherapy. The renal function was adequate (CrCl > 60 ml/min) in 97 (67%) of these patients. Patients who received neoadjuvant chemotherapy had higher p0 rates (29% vs 8%) than patients who did not receive neoadjuvant therapy. Advanced patient age, comorbidities, concerns over toxicity of chemotherapy, and the modest nature of benefit from neoadjuvant chemotherapy may explain why this treatment is not often used.
Despite level I evidence, neoadjuvant cisplatin-based chemotherapies continue to be underutilized in the management of bladder cancer, even at a high-volume tertiary center. A prospective evaluation of management choices, including the patient and physician factors involved in the use of perioperative cisplatin-based chemotherapy in bladder cancer, is indicated. Cancer 2011. © 2010 American Cancer Society.
With 68,000 new cases diagnosed each year, cancer of the bladder is the fourth most common cancer, and causes more than 13,000 deaths in the United States alone.1, 2 Patients with muscle-invasive bladder cancer comprise less than a third of reported cases, but have a worse prognosis because of higher risk of progression, development of metastasis, and death.
While radical cystectomy represents the primary therapeutic modality for the treatment of muscle-invasive bladder cancer, several studies have shown that a combined approach of neoadjuvant cisplatin-based chemotherapy and surgery is associated with improved survival in comparison to surgery alone.3-5 In 2003, a phase III clinical trial (SWOG-8710) showed a survival benefit in patients who received both neoadjuvant cisplatin-based chemotherapy (3 cycles of MVAC: methotrexate, vinblastine, doxorubicin, and cisplatin) and radical cystectomy, compared with those who underwent radical cystectomy alone (77 months vs 46 months). Several analyses have confirmed that neoadjuvant cisplatin chemotherapy in combination with surgery has a modest survival benefit (5%), with those patients achieving a complete pathological response after chemotherapy having the best outcome.6, 7
The data from SWOG-8710 and subsequent meta-analyses are compelling, with a number needed to treat (NNT) estimated at 20. Similar or higher NNT outcomes in breast cancer have led to the institution and widespread administration of neoadjuvant and adjuvant treatments as standard of care.8 However, despite such evidence, neoadjuvant chemotherapy is not uniformly used for patients with muscle-invasive bladder cancer.9
The use of perioperative chemotherapy in a contemporary cohort (after the publication of SWOG-8710) was examined, including an evaluation of the factors involved in such decision-making. We report an underutilization of perioperative chemotherapy, positing that a prospective evaluation of the factors involved in decision-making will help elucidate why patients who could benefit from neoadjuvant chemotherapy do not receive this.
MATERIALS AND METHODS
Approval was obtained through the UT Southwestern Medical Center Institutional Review Board for this study. A retrospective review was performed of all patients who underwent radical cystectomy at our institution from 2003 to 2008. Demographics for all patients, as well as clinical and pathologic stage, were determined from reviewing medical records. Clinical stage was derived from cystoscopy with transurethral resection of bladder tumor, examination under anesthesia, and preoperative imaging. Preoperative imaging was routinely obtained within 4 weeks of the diagnosis of muscle-invasive bladder cancer.
Study Inclusion and Exclusion Criteria
Patients were included if information on clinical stage, pathologic stage, renal function, and use or nonuse of neoadjuvant or adjuvant chemotherapy was available. Patients with nontransitional cell histologies on transurethral resection pathology were excluded.
Intraoperative Technique and Postoperative Surveillance
All patients underwent an open radical cystectomy with bilateral pelvic lymph dissection. Lymph node dissection routinely included the obturator, internal iliac, external iliac, common iliac, presacral, and distal preaortic and precaval lymph nodes, with distal extent of dissection to the bifurcation of the aorta. Patients were observed postoperatively on average every 3 months for the first year and then every 6 months after that, with basic laboratory, cross sectional imaging, and physical exam. Patients with pathologic T3 or T4 disease or positive lymph nodes were considered eligible for adjuvant chemotherapy.
Estimation of Renal Function
Glomerular filtration rates (GFR) prior to and 3 months after surgery were estimated respectively by the MDRD equation, as described below:
GFR (mL/min/1.73 m2) = 175 × (Scr)−1.154 × (Age)−0.203 × (0.742 if female) × (1.210 if African American).
Similar analyses were performed using the Cockroft-Gault equation, shown below, to evaluate creatinine clearance.
GFR (mL/min) = (140-age) * (Wt in kg) * (0.85 if female) / (72 * Cr)
An estimated GFR or creatinine clearance cutoff of ≥60 ml/min was used to evaluate eligibility for cisplatin-based chemotherapy. Overall findings using both the MDRD and the Cockroft-Gault equation indicate similar conclusions. For the sake of brevity and clarity, only the results from the MDRD equation are tabulated.
Use of Perioperative Chemotherapy
We investigated whether patients undergoing radical cystectomy for a clinical stage ≥T2 received neoadjuvant chemotherapy (eligibility criteria for SWOG-8710).
In addition, the use of adjuvant chemotherapy for patients with pathologic stage ≥T3 or positive lymph nodes on the radical cystectomy specimen was examined.
Kaplan-Meier curves were created for disease-specific and overall survival. A Cox regression analysis using age, sex, extravesicular disease (T3 or T4), lymph node status, lymphovascular invasion, use of neoadjuvant chemotherapy, and preoperative renal function was performed to determine prognostic indicators. A Pearson chi-square test or Fisher exact probability test was used to compare non-continuous variables. Data were analyzed using SPSS 17.0 statistical software (SPSS Inc., Chicago, Ill).
From 2003 to 2008, there were 238 patients who underwent RC at our institution for transitional cell bladder cancer (Table 1). The mean follow-up was 21.3 months (range 0.2 to 69.7, median 16.2). A majority of patients (70%) undergoing RC had estimated renal function adequate for cisplatin-based therapy (Table 2).
Table 1. Demographic Data for Entire Cohort of Patients
| 18||2 (0.8)|
| 40||7 (2.9)|
| 50||51 (21.4)|
| 60||73 (30.7)|
| 70||75 (31.5)|
| 80||30 (12.6)|
| Men||180 (76.0)|
| Women||58 (24.0)|
| Caucasian||216 (90.8)|
| African American||12 (5.0)|
| Asian||5 (2.1)|
| Hispanic||5 (2.1)|
Table 2. Estimated Preoperative GFR in mL/min for Entire Cohort of Patients by MDRD Formula
|0||>120||5 (2.1)||4 (2.8)|
|1||90-120||41 (17.2)||21 (14.5)|
|2||60-90||121 (50.8)||72 (49.7)|
|3||30-60||64 (26.9)||45 (31.0)|
|4||15-30||6 (2.5)||2 (1.4)|
|5||<15||1 (0.4)||1 (0.7)|
Table 3. Clinical Stages and Pathologic Stages of Patients Receiving No or Any Neoadjuvant Chemotherapy
|TIS, Ta, T1||93 (45.1%)||0|
|T2||106 (51.5%)||22 (68.8%)|
|T3||6 (2.9%)||3 (9.4%)|
|T4||1 (0.5%)||7 (21.9%)|
|B. Radical Cystectomy Specimens From Patients Receiving No or Any Neoadjuvant Chemotherapy Pathologic Stage|
|T0||16 (7.8%)||10 (28.6%)|
|TIS, Ta, T1||54 (26.2%)||6 (17.2%)|
|T2||40 (19.4%)||4 (11.4%)|
|T3||75 (36.4%)||11 (31.4%)|
|T4||21 (10.2%)||4 (11.4%)|
|N+||56 (27.2%)||8 (25.0%)|
|HG||189 (99.5%)||23 (100%)|
Eligibility for and Use of Neoadjuvant Chemotherapy
Among these 238 patients, 145 had a preoperative clinical stage ≥T2. According to the SWOG-8710 trial, these persons could have benefited from neoadjuvant chemotherapy. Of these 145 patients, only 32 (22%) received neoadjuvant chemotherapy. These 32 patients included 7 who received non-cisplatin based chemotherapy. Overall, only 25 of 145 (17%) patients with preoperative clinical stage ≥T2 received neoadjuvant cisplatin-based chemotherapy. These include 22 of 128 (17%) patients with clinical stage T2, 3 of 9 (33%) patients with clinical stage T3, and 7 of 8 (88%) patients with clinical stage T4 (Table 3a). Excluding T4 patients, those who actually received optimal cisplatin-based neoadjuvant chemotherapy for clinical stage T2 or T3 drops further to 13% (18 of 137).
Reasons for not giving neoadjuvant chemotherapy are limited by the retrospective nature of the review, but appear to include concerns over patient age and/or comorbidities in 29, patient preference in 9, toxicity of chemotherapy in 12, symptoms or active bleeding in 6 patients, and/or clinically localized or resectable cancer in 19. In 66 patients, a satisfactory reason for excluding neoadjuvant chemotherapy could not be established.
Cisplatin can cause renal toxicity, and compromised renal function is often cited as the reason for avoiding cisplatin-based therapy. We examined if eligibility for cisplatin-based therapy could explain the reason for its low utilization. Of the 145 patients with a preoperative clinical stage ≥T2, 97 (70%) were eligible for cisplatin-based chemotherapy on the basis of estimated GFR >60 ml/min by the MDRD equation (Table 2). Similar proportions were observed with the usage of creatinine clearance cutoffs of 60ml/min/1.73m2 by the Cockroft-Gault equation (69%) or the Jelliffe equation (66%).
Only 14% (14 of 97) patients eligible by MDRD criteria received neoadjuvant cisplatin chemotherapy. At our institution, measured creatinine clearances were used as well to determine eligibility for cisplatin-based therapy. An additional 11 patients with a MDRD calculated GFR <60ml/min were found to have either hydrated or measured creatinine clearances >45ml/min/1.73m2, and subsequently received cisplatin-based neoadjuvant chemotherapy. Overall, only 25 of 145 (17%) patients with preoperative clinical stage ≥T2 received cisplatin-based neoadjuvant chemotherapy. All of the patients referred to a medical oncologist received neoadjuvant chemotherapy.
Eligibility for and Use of Adjuvant Chemotherapy
Following RC, 124 patients had a pathologic stage ≥T3109 and/or positive lymph nodes,65 and by criteria were at high risk for recurrence (Table 3b). Of these 124 patients at high risk for recurrence after surgery alone, 47 (38%) received adjuvant chemotherapy. Of the patients with pathologic stage ≥T3 and negative nodes (n=59), 10% (6 of 59) received adjuvant chemotherapy. In contrast, 63% (41 of 65) of patients with nodal positive disease received adjuvant chemotherapy. The primary reasons for not giving adjuvant chemotherapy included age and/or comorbidities in 17, patient preference in 20, pathologic stageT3a with negative surgical margins and no nodal involvement in 17, prior neoadjuvant chemotherapy in 3, and death before follow-up in 6; 14 patients were not available. More than a third of the patients (25 of 72) referred to medical oncology for adjuvant chemotherapy did not receive it.
Of the patients receiving adjuvant chemotherapy, 43 (92%) received cisplatin-based therapy and 4 (9%) received carboplatin or taxol-based therapy. Among the 124 patients at high risk for recurrence, 86 (69%) were eligible for cisplatin-based chemotherapy based on GFR >60 ml/min. Thirty-eight of 86 (44%) high-risk patients eligible by MDRD criteria received adjuvant chemotherapy, including 35 (41%) who received cisplatin-based chemotherapy. Again, an additional 8 patients with an MDRD estimated GFR <60ml/min were found to have hydrated or measured creatinine clearances ≥45ml/min and received cisplatin-based adjuvant chemotherapy.
Overall, of the cohort of 238 patients undergoing radical cystectomy at our institution, only 25 preoperative clinical stage T2 or greater and 35 high-risk postoperative patients received optimal neoadjuvant and adjuvant cisplatin-based chemotherapy, respectively.
Survival, Recurrence, and Metastasis Outcomes
The mean follow-up for our cohort of patients was 21.3 months (range 0.2 to 69.9, median 16.2). By Kaplan-Meier survival analysis, the mean overall and disease-specific survival times were 43.9 and 46.5 months, respectively, with 67 patients dying of bladder cancer at a mean time of 15.8 months (range 0.2-57.5, median 10.9). In our follow-up period, 51 patients developed metastasis at a mean time of 12.4 months (range 1.2-54.6, median 8.2), with 39 of those patients dying of bladder cancer. In addition, 56 patients developed a recurrence at a mean time of 12.7 months (range 1.2-62.0, median 8.4), with 39 of those patients dying of bladder cancer.
Effect of Neoadjuvant Chemotherapy
Patients who received neoadjuvant chemotherapy had a higher probability of pathologic T0 disease than patients who did not receive neoadjuvant chemotherapy (28% vs 8%, P = .002, 2-tailed Fisher exact probability test). There was no statistically significant difference in survival for patients who did and did not receive neoadjuvant chemotherapy (P > 0.8, log rank test). In a multivariate analysis of the 145 patients eligible for neoadjuvant chemotherapy, neoadjuvant chemotherapy was not an independent predictor of increased disease-specific or overall survival (Table 4).
Table 4. Multivariate Analysis
|Age >65 y||1.52 (0.80-2.90)||1.43 (0.79-2.59)||1.34 (0.75-2.58)||1.39 (0.78-2.46)|
|Sex||0.70 (0.37-1.30)||0.73 (0.41-1.31)||0.91 (0.50-1.66)||0.86 (0.49-1.51)|
|Lymph node+||1.92 (0.97-3.76)||2.21 (1.16-4.19)||2.30 (1.16-4.55)||2.64 (1.34-5.03)|
|≥T3||2.35 (1.08-5.10)||2.51 (1.21-5.17)||1.97 (0.76-5.15)||2.00 (0.84-4.83)|
|LVI||2.53 (1.12-5.65)||2.59 (1.21-5.52)||2.10 (1.04-4.22)||2.20 (1.13-4.28)|
|Neoadjuvant chemotherapy||1.49 (0.71-3.11)||1.41 (0.70-2.83)||n/a||n/a|
|Adjuvant chemotherapy||NA||NA||0.51 (0.26-0.995)||0.51 (0.28-0.94)|
Effect of Adjuvant Chemotherapy
In a multivariate analysis of the 124 patients eligible for adjuvant chemotherapy, adjuvant chemotherapy was a significant independent predictor of increased overall and disease-specific survival, with hazard ratios of 0.507 (95% CI, 0.275-0.937) and 0.512 (95% CI, 0.263-0.995), respectively (Table 4). Of the 59 patients with extravesical disease and negative nodes, adjuvant chemotherapy was given to 10% of patients; however, 36% developed metastasis and 34% died from bladder cancer. Among patients with a pathologic stage of T3a with negative surgical margins and negative nodes (n=36), only 1 patient received adjuvant chemotherapy. Yet 39% developed metastasis and 31% died of bladder cancer in our follow-up period.
Bladder cancer is a chemosensitive malignancy. Cisplatin is the single most active chemotherapeutic agent against bladder cancer. Carboplatin is often substituted for cisplatin in the management of patients with metastatic bladder cancer, especially in the setting of renal insufficiency.10 However, several trials have indicated that carboplatin is inferior to cisplatin in the management of patients with bladder cancer.11, 12 In addition, there are only a limited number of small non-randomized trials on the utility of carboplatin in the perioperative setting.13, 14
Since its introduction, cisplatin-based combination chemotherapy has been the standard of care in treating patients with locally advanced and metastatic transitional cell carcinoma in neoadjuvant, adjuvant, and metastatic settings.15-20 Data from 2 large randomized trials of cisplatin-based chemotherapy and a meta-analysis composed of 3005 patients indicate that neoadjuvant cisplatin-based chemotherapy is associated with a significant survival benefit in the 30%-40% patients who achieve a complete pathological response.6, 7 Our study did not show a survival difference between patients who did and did not receive neoadjuvant chemotherapy, but there was a significant difference between pathologic T0 rates with nearly 30% of the neoadjuvant patients achieving pathologic T0. The absence of demonstrable survival benefit of neoadjuvant chemotherapy in this cohort may be related to short follow-up and small cohort size.
The true proportion of patients deemed ineligible for cisplatin-based chemotherapy is not known. In the International Bladder Cancer study, 45.3% of the patients (106 of 234) with clinical stage T2-4a, Nx, M0 bladder cancer were deemed ineligible for cisplatin-based chemotherapy.21 Patients deemed ineligible for cisplatin-based chemotherapy had worse overall and cancer-specific survival outcomes (31%, 3-year overall survival) than patients who received cisplatin-based chemotherapy (62%, 3-year overall survival). However, the relative contribution of chemotherapy and comorbidities (other than renal dysfunction) that affect survival was not known.
With the increasing use of mathematic formulas to calculate CrCl to determine renal eligibility, a calculated CrCl cutoff of 60 mL/min or GFR <60 mL/min/1.73 m2 has been used to determine renal eligibility for cisplatin-based chemotherapy. Dash et al have shown that, in the Memorial Sloan-Kettering experience, the overall proportion of patients deemed ineligible for cisplatin-based chemotherapy was 28% using the CG formula.22 Our findings agree with this prior report, with 32%-33% of patients with clinical stage T2 or higher not meeting the cutoff either by the Cockroft-Gault or MDRD formulas. By present standards, these patients either would have been given no treatment or inferior carboplatin-based regimens, and may have had worse survival outcomes.23
Previously, we had shown that formulas to calculate creatinine clearances are unreliable and that up to 39% of patients with calculated CrCl <60 mL/min would have been eligible for cisplatin-based chemotherapy by evaluating measured CrCl using the same cutoff of 60 mL/min.23 We had also shown that ineligibility from calculated CrCls, using the cutoff of 60 mL/min, was not significantly associated with the ability to receive at least 3 cycles of chemotherapy. At our institution, we re-evaluate all patients with an estimated CrCl <60 and obtain a hydrated CrCl or measured CrCl using a 24-hour urine collection. Consequently, 17 patients who would not have received cisplatin-based chemotherapy on the basis of arbitrary CrCl cutoffs received and tolerated chemotherapy.
In our study, neoadjuvant chemotherapy was associated with p0 rate of 27%, similar to the 30% rate reported in the SWOG trial. Importantly, only patients receiving cisplatin-based therapy—and none of the patients receiving carboplatin-based therapies— achieved p0. Since most of our patients in this cohort received the doublet of gemcitabine and cisplatin, these data indicate that similar rates of pathologic downstaging may be seen with gemcitabine and cisplatin as reported for MVAC.3, 7 Although neoadjuvant chemotherapy in our cohort was not associated with a survival advantage, this may reflect the higher stage of patients receiving neoadjuvant chemotherapy and a lack of randomization. The lower p0 rates and lack of a survival benefit may also in part be attributed to the agents used (GC vs MVAC).
Reasons why neoadjuvant chemotherapy was not offered in this cohort included patient symptoms, patient preference, patient age and comorbidities, physician preference, and no clinical signs of disease progression (ie, lymphadenopathy). Furthermore, surgeons may not want to delay surgical treatment when there is bulky disease. Some of these reasons may reflect a lack of confidence in the published data to date, concerns over the toxicity and tolerability of chemotherapy, the modest nature of the benefit from neoadjuvant cisplatin-based therapy, and the desire to reserve chemotherapy for an adjuvant or salvage setting. Additional variables, such as the presence of lymphovascular invasion in the TURBT specimen and preoperative hydronephrosis, may influence the decision for chemotherapy in some patients. However, given the retrospective nature of this study, we are not able to discern the number of patients who may have received chemotherapy because of these confounders. The primary limitation is that most patients were referred to our center after having had a TURBT, and the presence or absence of lymphovascular invasion in a TURBT specimen is not uniformly noted. As a result of these concerns, the NCCN and the EAU guidelines both indicate that neoadjuvant chemotherapy only be “considered” (not mandated) in patients with muscle-invasive bladder cancer. Several ongoing studies are attempting to identify subsets of patients who will receive the most benefit from neoadjuvant chemotherapy using tissue biomarkers, prognostic models such as nomograms, and pathologic characteristics such as lymphovascular invasion.
Our data indicate that perioperative chemotherapy is underused, even in a high-volume tertiary carereferral center committed to multimodality therapy. In an analysis of 7161 patients with stage III bladder cancer from the National Cancer Database, perioperative chemotherapy was only administered to 11.6% of patients, with 10.4% receiving adjuvant chemotherapy and 1.2% receiving neoadjuvant chemotherapy.9 Our utilization patterns are also weighted towards adjuvant chemotherapy, with 34% of patients receiving adjuvant chemotherapy and only 20% receiving neoadjuvant chemotherapy. It is interesting that these utilization patterns are not supported by Level I evidence, which endorses only neoadjuvant chemotherapy.7, 24, 25 Numerous studies indicate that fewer patients undergo adjuvant chemotherapy, partly because of delayed recovery and complications following surgery, patient refusal, poor performance status, and psychological distress.24-28 Despite a clear lack of evidence supporting adjuvant chemotherapy, its utilization in our tertiary center and the National Cancer Database may reflect the desire of clinicians to decide on chemotherapy on the basis of pathologic rather than clinical staging parameters.26
Analysis of our small cohort shows a survival advantage for patients who receive adjuvant chemotherapy. While these data are not adequately powered, randomized, or prospective, they may support the rationale for delaying chemotherapy until after RC. Interestingly, our data indicate that patients with pathologic stage T3a, negative lymph nodes, and negative surgical margins are still at significant risk for recurrence and mortality from bladder cancer. These data suggest that patients with pathologic stage ≥T3 or positive nodes should be strongly considered for adjuvant chemotherapy.
Due to the retrospective nature of this study, the attributable factors suffer from a recall bias and are not analyzable. Moreover, as this study only examines patients undergoing RC, these data may underestimate the true number of patients with clinical stage >T2 referred for or who underwent chemotherapy. To further evaluate the reasons for failure of urologists in tertiary referral centers to prescribe neoadjuvant chemotherapy, a prospective evaluation of treatment decisions needs to be performed and is underway. We hope to understand institutional and personal preferences that affect the use of neoadjuvant chemotherapy in bladder cancer.
Bladder cancer is chemosensitive, and current evidence suggests neoadjuvant chemotherapy can improve outcomes. However, neoadjuvant cisplatin-based chemotherapies continue to be underutilized in the management of bladder cancer, even in a contemporary cohort at a high-volume tertiary center. Adjuvant cisplatin-based chemotherapy also appears to be underutilized, even in patients at high risk for recurrence. Compromised renal function does not explain the low use of cisplatin-based chemotherapies. A prospective evaluation of the management choices and patient/physician factors involved in the use of perioperative cisplatin-based chemotherapy in bladder cancer is indicated.
CONFLICT OF INTEREST DISCLOSURES
Supported by a grant from the Dorothy and James Cleo Thompson Jr Foundation (to GVR).