- Email: [email protected]
Optimization of Patient Selection for Neoadjuvant Chemotherapy in Muscle-invasive Urothelial Carcinoma of the Bladder
Original Study
Optimization of Patient Selection for Neoadjuvant Chemotherapy in Muscle-invasive Urothelial Carcinoma of the Bladder Patrick J. Hensley,1 Jeffrey Goodwin,1 Daniel L. Davenport,2 Stephen E. Strup,1 Andrew James1 Abstract Few predictors are known for determining the response to neoadjuvant chemotherapy (NAC) for bladder cancer. We investigated the clinicopathologic predictors of extravesical disease in cohorts receiving NAC and upfront cystectomy. The present study demonstrated an inferior response to alternative NAC regimens, implicating cardiovascular comorbidities and nutritional status in the tolerability and response to NAC. Background: Radical cystectomy (RC) is delayed in a subset of patients who respond poorly to neoadjuvant chemotherapy (NAC). The present study investigated the clinicopathologic characteristics predicting extravesical disease at RC and the factors associated with NAC tolerability to improve patient selection and the sequence of definitive therapy. Materials and Methods: Patients with cT2 urothelial carcinoma of the bladder who underwent NAC were stratified by the final pathologic stage: complete (ypT0N0), partial ( pT2), and nonresponse (> pT2 and/or Nþ). Patients treated with upfront cystectomy were divided into those with organ-confined ( pT2) and those with extravesical disease (> pT2 and/or Nþ). Results: Of 145 patients, 89 received NAC and 56 underwent upfront RC. The univariate predictors of extravesical disease in the patients treated with upfront RC included increased age (P ¼ .021), higher Eastern Cooperative Oncology Group performance status (P < .001), hydronephrosis (P ¼ .021), and cardiovascular risk factors. The complete, partial, and nonresponse rates to NAC were 25.8%, 39.3%, and 34.8%, respectively. The multivariate predictors of pathologic progression on NAC included low serum albumin (P ¼ .005), hydronephrosis (P ¼ .040), incomplete NAC (P ¼ .014), and alternative NAC (non-gemcitabine/cisplatin or MVAC, P ¼ .022). Significant multivariate predictors of incomplete NAC included increased age, coronary artery disease (P ¼ .027), and Eastern Cooperative Oncology Group performance status. Conclusion: Redundant clinicopathologic features predicted adverse cystectomy pathology in patients treated with both NAC and upfront RC. The results of the present study demonstrated an inferior pathologic response to alternative NAC regimens in clinically organ-confined disease and implicated cardiovascular comorbidities and nutritional status in the tolerability and response to NAC. Our findings predicate the importance of using patient-specific factors to guide the sequence of definitive treatment toward timely, centralized care to improve clinical outcomes. Clinical Genitourinary Cancer, Vol. -, No. -, --- ª 2018 Elsevier Inc. All rights reserved. Keywords: Bladder cancer, Cisplatin, Cystectomy, Pathologic response, Performance status
Introduction Neoadjuvant chemotherapy (NAC), followed by radical cystectomy (RC), is the reference standard treatment for muscle1
Department of Urology Department of Surgery, University of Kentucky College of Medicine, Lexington, KY
2
Submitted: Dec 8, 2017; Revised: Feb 10, 2018; Accepted: Feb 18, 2018 Address for correspondence: Patrick J. Hensley, MD, Department of Urology, University of Kentucky, 800 Rose Street, Lexington, KY 40536 E-mail contact: [email protected]
1558-7673/$ - see frontmatter ª 2018 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.clgc.2018.02.007
invasive bladder cancer. RC provides local control and NAC treats micrometastatic disease, conferring a 5% to 10% survival benefit compared with RC alone. Patients receiving NAC exhibit a greater rate of pathologic T0 (pT0) at RC, providing a plausible explanation of the survival benefit.1,2 Both RC and NAC, however, are associated with considerable morbidity. Several series have shown a 28% to 64% complication rate, 26% to 35% rehospitalization rate, and up to 8% 90-day mortality after RC.3 In prospective studies, NAC with MVAC (methotrexate, vinblastine, doxorubicin, cisplatin) was associated
Clinical Genitourinary Cancer Month 2018
-1
Optimal Patient Selection for NAC with a 35% to 37% rate of grade 3/4 toxicity,1,2 with these toxicities associated with decreased overall survival (hazard ratio, 2.71; P < .001).4 Gemcitabine/cisplatin, although better tolerated than MVAC, has been associated with a 2% to 27% rate of grade 3/4 hematologic toxicity.5 In this subset of patients, NAC might delay the time to definitive extirpative therapy and result in additional morbidity. These patients may benefit from foregoing NAC for upfront RC. The identification of clinical and pathologic variables that could predict a favorable response to NAC would allow for a more selective approach to NAC usage. Additionally, the identification of factors that could predict for a patient’s ability to tolerate and complete NAC may improve patient selection. We investigated these variables in subsets of patients treated with NAC followed by RC and in patients who received upfront RC to identify clinicopathologic characteristics that may be used to improve patient-specific clinical decision making and, ultimately, reduce morbidity and improve the outcomes for patients with muscle-invasive bladder cancer.
Materials and Methods
2
-
An institutional review board-approved retrospective review of patients was performed. A waiver of informed consent was obtained. The inclusion criteria included clinical stage T2 urothelial carcinoma of the bladder with muscularis propria invasion found on transurethral resection pathologic examination, cross-sectional imaging depicting no evidence of extravesical disease, and negative findings with bimanual examination under anesthesia. All patients underwent RC, pelvic lymph node dissection, and urinary diversion at the University of Kentucky from January 2010 to April 2017. Tumor staging was determined using the 2015 American Joint Committee on Cancer staging system. Patients either underwent upfront RC or NAC (regimen described), followed by RC. Patients with history of initial radiation therapy for bladder cancer, those who had undergone partial cystectomy, and patients with pure adenocarcinoma, squamous cell carcinoma, or neuroendocrine carcinoma were excluded. The patients who underwent NAC were stratified into 3 groups according to the final pathologic stage: complete response (ypT0N0), partial response (final pathologic stage pT2), and nonresponse (final pathologic stage > pT2 and/or Nþ). The patients who had undergone upfront cystectomy were divided into those with organ-confined disease (final pathologic stage pT2) and those with extravesical disease (final pathologic stage > pT2 and/or Nþ). A review was performed of preoperative clinical variables and the reports pathology reports from transurethral resection specimens and cystectomy specimens. An incomplete chemotherapy regimen was defined as any break, delay, or dose reduction in the prescribed chemotherapy regimen. An alternative chemotherapy regimen was defined as any variant from gemcitabine/cisplatin or MVAC. All outside pathology slides were reviewed by the pathology department of the University of Kentucky to confirm the stage, grade, and presence or absence of variant histologic features. Any level of variant histologic features as described on the transurethral resection pathology report was considered variant. Data were statistically analyzed using analysis of variance or c2 tests of univariate differences and logistic regression for multivariable analysis after combining those with partial responses and no
Clinical Genitourinary Cancer Month 2018
responses. Significance was set at P < .05. All statistical calculations were performed using SPSS, version 22 (IBM Corp, Armonk, NY).
Results Comparison of Cohorts: Upfront RC Versus NAC Of the 145 patients who met the inclusion criteria, 89 underwent NAC followed by RC and 56 underwent upfront RC. The clinicopathologic and demographic variables for these 2 groups of patients are listed in Table 1. The patients who had undergone upfront RC were, on average, older and had a greater baseline serum creatinine than did patients treated with NAC. NAC was prescribed more commonly for patients with variant histologic features. Although not statistically significant, a trend was found toward a greater Eastern Cooperative Oncology Group (ECOG) performance status in patients who had undergone upfront RC. Patients treated with NAC experienced a mean delay in the interval from diagnosis to definitive surgery that was 16 weeks longer than that for the upfront surgery cohort. No statistically significant socioeconomic differences were found between the 2 groups (data not shown).
Clinical and Pathologic Predictors of NAC Response The complete pathologic response rate to NAC was 26% and the nonresponse rate was 35% (Table 1). Statistically significant univariate predictors of a poor response to NAC included increased age, low serum albumin, increased ECOG performance status, and the presence of hydronephrosis (unilateral or bilateral). Cardiovascular comorbidities, including the presence of hyperlipidemia, hypertension, coronary artery disease, or a history of myocardial infarction, were univariate predictors of a poor response to NAC. Although more common in the nonresponse group, the pooled incidence of variant histologic features was not significant among the groups. Further statistical analysis of the specific variant histologic features was not performed, given the low incidence. All 4 patients with micropapillary differentiation on transurethral resection were nonresponders. No patients with micropapillary, sarcomatoid, or clear cell differentiation were in the complete response group. The nonresponders began their first chemotherapy cycle after the diagnosis of muscle-invasive disease, on average, nearly 3 weeks later than the complete responders (Table 1). The deviation from a standard chemotherapeutic regimen as prescribed was associated with a poor pathologic response to NAC. Patients who received alternative regimens from the standard gemcitabine/cisplatin or MVAC (most commonly gemcitabine/carboplatin) responded poorly to NAC, as did the patients who required a break, delay, or dose reduction in NAC cycles (defined as incomplete NAC). These findings were confirmed as independent predictors of disease progression with NAC on multivariate regression analysis (Table 2).
Clinical Predictors of Incomplete NAC We investigated the clinical predictors of patients receiving an incomplete NAC regimen. Increased age and ECOG performance status were the most significant predictors of a patient’s ability to tolerate chemotherapy as prescribed on univariate analysis (Table 3). These were also statistically significant predictors of a poor pathologic response to NAC. Although chemotherapy regimens prescribed and administered at our institution was not a statistically
Table 1 Univariate Predictors of Pathologic Response to NAC and Extravesical Disease After RC NAC Variable Patients, n (%) Female gender, % Age, y 2
BMI, kg/m
Upfront RC
NR
PR
CR
All Patients
31 (34.8)
35 (39.3)
23 (25.8)
89 (100)
22.6
17.1
21.7
20.2
66.3 1.7
65.0 1.5
58.8 1.6
63.8 1.0
26.7 0.8
28.6 1.0
28.1 1.4
27.8 0.6
P Value
OCD
EVD
All Patients
22 (39)
34 (61)
56
a
18
18
18
a
65 2
71 2
69 1
a
28.4 1.2
26.7 0.8
27.4 0.7
.925
.003 .393
P Value (NAC vs. Upfront RC)
1.000a
.830a
a
.005a
a
.625a
a
.021 .218
80.6
82.9
65.2
77.5
.215
82
88
86
.698
.282a
3.1 0.8
3.5 0.4
3.6 0.5
3.4 0.6
.004a
3.5 0.1
3.4 0.1
3.5 0.1
.408a
.550a
Pre-NAC/Sx serum creatinine, mg/dL
1.03 0.06
0.98 0.05
0.98 0.05
1.00 0.03
a
.499
1.3 0.2
1.5 0.2
1.4 0.1
a
.589
.010a
Post-NAC serum creatinine, mg/dL
1.11 0.06
1.04 0.07
0.96 0.04
1.04 0.04
.108a
NA
NA
NA
NA
NA
Tobacco use, % Albumin, g/dL (n ¼ 83)
a
P Value
.037a
ECOG PS, % 0
45.1
51.5
73.9
55.0
82
35
53
1
45.2
37.1
26.1
37.1
18
35
29
2
9.7
11.4
0.0
7.9
0
24
14
3
0
.159a
0
0
0
6
4
48.4
17.1%
17.4
28.1
.008a
14
44
32
.021a
.709a
Hyperlipidemia
51.6
34.3
17.4
36.0
.010a
23
56
43
.026a
.484a
Hypertension
90.3
60.0
65.2
71.9
.028a
50
85
71
.007a
1.000a
18.0
a
23
a
.524a
a
1.000a
a
.073a
a
.013a
Hydronephrosis, %
0
<.001a
Cardiovascular comorbidities, %
Coronary artery disease Diabetes mellitus Stroke Myocardial infarction
32.3 22.6 3.2
14.3 17.1 0.0
4.3 21.7 0
20.2 1.1
.007
a
.897
a
.239
9 9 5
32 29 9
21 7
.056 .099
1.000
a
0
0
0
1.000
a
NA
NA
NA
NA
5.7
0
10.1
.005
Interval from diagnosis to NAC, d
62 7
53 5
42 4
53 3
Interval from NAC end to Sx, d
91 18
62 6
76 20
76 9
.483
NA
NA
NA
Interval from diagnosis to Sx, d
223 20
191 9
203 20
205 9
.416a
64 10
113 22
93 14
.026
a
NA
NA NA
.060a
<.001a
Biopsy pathologic findings Carcinoma in situ
16.1
25.7
17.4
20.2
.834a
23
29
27
.759a
.418a
Variant histologic features (any)
29.0
17.1
21.7
22.5
.472a
27.3
47.1
39.3
.169a
.039a
Squamous cell
9.7
5.7
17.4
10.1
18.2
29.4
25.0
Glandular
3.2
8.6
8.7
6.7
4.5
2.9
3.6
Sarcomatoid
3.2
0
0
1.1
0
2.9
1.8
Micropapillary
12.9
0
0
4.5
4.5
8.8
7.1
Patrick J. Hensley et al
Clinical Genitourinary Cancer Month 2018
22.6
-3
Final Nþ
Data presented as n (%), mean standard error, or %. Abbreviations: BMI ¼ body mass index; CR ¼ complete response; ECOG ¼ Eastern Cooperative Oncology Group; EVD ¼ extravesical disease; OCD ¼ organ-confined disease; NAC ¼ neoadjuvant chemotherapy; NR ¼ no response; PR ¼ partial response; PS ¼ performance status; Sx ¼ surgery. a Statistically significant.
.524a <.001a 38 0 <.001a 0
Clinical Genitourinary Cancer Month 2018
52
0
18
<.001a 50 13 50 Incomplete NAC
77
9 15 Alternative NAC
NAC at our institution
36
65 54
21
.052a 0 2.9 Plasmacytoid
NAC regimen, %
Clear cell
0
39
CR PR
52
1.1
All Patients
-
Variable
NR
NAC Table 1 Continued
4
.013a
2.9 0
23
1.8
0 0
0
P Value (NAC vs. Upfront RC) All Patients EVD OCD P Value
Upfront RC
P Value
Optimal Patient Selection for NAC Table 2 Multivariate Predictors of Pathologic Progression During NAC Variable Mean albumin, g/dL (n ¼ 83) Hydronephrosis Incomplete NAC Alternative NAC Myocardial infarction
OR (95% CI) 0.218 3.601 4.499 4.926 12.091
(0.076-0.627) (1.057-12.266) (1.359-14.891) (1.262-19.231) (1.736-84.219)
P Value .005 .040 .014 .022 .012
Abbreviations: CI ¼ confidence interval; NAC ¼ neoadjuvant chemotherapy; OR ¼ odds ratio.
significant predictor of overall pathologic response (Table 1), patients who received chemotherapy at the University of Kentucky were more likely to complete NAC relative to those who underwent treatment at outlying institutions. Additionally, the interval from diagnosis to the initiation of NAC correlated inversely with a patient’s ability to tolerate a complete treatment course. On multivariate regression analysis, greater ECOG performance status, the presence of coronary artery disease, and older age were all independent predictors of patients receiving an incomplete NAC regimen (Table 4).
Clinical and Pathologic Predictors of Extravesical Disease in Patients Who Underwent Upfront RC Patients who received upfront RC had a shorter interval from the diagnosis of muscle-invasive disease to definitive extirpative surgery. The univariate clinical and pathologic predictors of extravesical disease at RC for cT2 urothelial carcinoma of the bladder in patients who underwent upfront RC included increased age and ECOG performance status, the presence of hydronephrosis, and a history of hypertension or hyperlipidemia (Table 1). Ten patients had an ECOG performance status of 2 to 3, none of whom had organconfined disease at RC. The incidence of variant histologic features was 20% greater in the extravesical disease group, with a predominance of sarcomatoid and micropapillary differentiation found in this group. ECOG performance status was the strongest predictor of extravesical disease at RC on multivariate analysis. In addition to ECOG, only coronary artery disease was an additional independent predictor of extravesical disease on multivariate analysis (odds ratio, 7.0; 95% confidence interval, 1.2-41.9; P ¼ .033). An ECOG performance status of 2 to 3 had a complete association with extravesical disease, precluding multivariate statistical analysis.
Discussion The optimal patient selection for NAC before RC for patients with muscle invasive urothelial cell carcinoma of the bladder is often challenging owing to the poor predictors of both disease response and patient tolerability of chemotherapy. The ability to reliably predict which patients will benefit from and/or tolerate NAC to completion could improve the outcomes for patients with muscle-invasive bladder cancer. The commonly cited reasons for forgoing NAC include the delay to definitive surgery and the potential toxicity of chemotherapeutic regimens.6 The present study identified clinicopathologic characteristics predicting pathologic response to NAC and patient-specific factors associated with patients’ ability to complete
Patrick J. Hensley et al Table 3 Univariate Predictors of Incomplete NAC Variable
Table 4 Multivariate Predictors of Incomplete NAC
Full Course Incomplete NAC NAC P Value
Patients, n
44
44
Female gender, %
18.2
22.7
60 y
59.1
25.0
61-70 y
27.3
36.4
71
13.6
38.6
Tobacco use, %
75.0
79.5
.800
Married, %
65.9
47.7
.131
Lives with family, %
75.0
63.6
.355
0
75.0
36.4
1
20.5
52.3
2
4.5
11.4
Age group, %
.792 .003
ECOG PS, %
Coronary artery disease Myocardial infarction
.001
59.1
84.1
.017
6.8
29.5
.011
4.5
15.9
.157
Carcinoma in situ
13.6
25.0
.280
Mean pre-NAC serum creatinine >1.2 mg/dL, %
12.5
23.8
.255
Interval from diagnosis to NAC, d
44 3
63 6
.008
65.9
38.6
.018
NAC at our institution, %
ECOG PS 1 versus 0 ECOG PS 2 versus 0 Coronary artery disease Age 60-70 versus <60 y Age >70 versus <60 y
OR (95% CI) 4.746 3.085 5.183 3.363 4.549
(1.639-13.742) (0.441-21.591 (1.205-22.292) (1.070-10.567) (1.232-16.798)
P Value .004 .256 .027 .038 .023
Abbreviations: CI ¼ confidence interval; ECOG ¼ Eastern Cooperative Oncology Group; NAC ¼ neoadjuvant chemotherapy; OR ¼ odds ratio; PS ¼ performance status.
Cardiovascular comorbidities, % Hypertension
Variable
Data presented as n (%), mean standard error, or %. Abbreviations: ECOG ¼ Eastern Cooperative Oncology Group; NAC ¼ neoadjuvant chemotherapy; PS ¼ performance status.
NAC for clinical stage T2 urothelial carcinoma of the bladder. The complete pathologic response rate to NAC in our cohort of 26% was comparable to that reported in previous studies at 23% to 28%.7,8 Previous studies have sought to define the categories of pathologic response for prognostic purposes, including a complete pathologic response (ypT0N0) and a spectrum of stable, organ-confined disease and progressive extravesical disease during NAC.7,8 Although a complete pathologic response to NAC has been intuitively associated with improved clinical outcomes relative to patients with residual disease, the clinical implications for survival are less clear. Gandhi et al8 studied the survival outcomes for 150 patients with muscle-invasive bladder cancer stratified by pathologic response. The 5-year cancer-specific survival (CSS) was improved in responders (ypT1) compared with nonresponders (ypT2; 91% vs. 27%; P < .01). No statistically significant difference was found in survival on the subgroup analysis between patients with ypT0 and those with residual, nonemuscle-invasive disease.8 The rate of pT0 disease at RC from transurethral resection alone ranges from 7% to 20%.9,10 Kassouf et al11 found similar CSS rates among patients without residual disease who had received NAC, followed by RC, or RC alone (92% vs. 86%, respectively; P ¼ .65), suggesting that the treatment of micrometastatic disease by NAC might be of less importance than its effect on the primary tumor. The present study was limited to patients who had undergone RC for clinical stage T2 urothelial carcinoma, because we believe
this clinical scenario provides the greatest room for interpretation in the value of NAC. Although hydronephrosis has been identified as a risk factor for extravesical disease, we chose to consider patients without local tumor extension or lymphadenopathy as having clinical stage T2, because w35% of patients with hydronephrosis have been shown to have organ-confined disease at RC.12 The preponderance of patients undergoing RC for urothelial carcinoma do so for clinical stage T2 disease.7-11,13 Pokuri et al7 completed a retrospective review of 50 patients who had undergone NAC, followed by robotic cystectomy for cT2 urothelial carcinoma. Pokuri et al7 divided patients according disease status (ypT0 and non-ypT0 disease), with variant histologic features predicting for an incomplete response. Their study was limited by its small cohort size and the inclusion of patients with clinically extravesical disease. In our cohort, micropapillary and sarcomatoid variants were exclusively found in the nonresponse group. In a retrospective study of patients with variant histologic features from the National Cancer Database, NAC was associated with improved rates of organ-confined disease at RC relative to patients treated with upfront RC.14 In our cohort, alternative NAC regimens primarily consisted of gemcitabine/carboplatin or conversion to Taxol-based regimens because of intolerance. In locally advanced or metastatic disease, gemcitabine/cisplatin exhibited an improved pathologic response,15 time to progression, and overall survival, with comparable acceptable toxicity, relative to gemcitabine/carboplatin.5,16 Consistent with other studies, we have demonstrated a significantly inferior pathologic response to alternative NAC in those with clinically organ-confined disease.12,17 Alternative regimens were associated with a nearly fivefold increase in pathologic progression during NAC compared with standard gemcitabine/cisplatin or MVAC (odds ratio, 4.926; 95% confidence interval, 1.262-19.231; P ¼ .022). In a retrospective study of patients with noneorgan-confined disease, pathologic downstaging was exhibited in 59% of patients treated with neoadjuvant gemcitabine/cisplatin compared with 7% of patients treated with gemcitabine/carboplatin; however, the difference was not statistically significant (P ¼ .444).18 Chappidi et al4 divided patients with cT2N0 disease into those with stable (ypT2N0M0) and progressive (>ypT2 or Nþ) disease after NAC. Patients with stable disease during NAC had better overall survival (OS) than did those with progression during NAC, and OS comparable to that of patients with cT2 disease who had undergone upfront RC. Nonresponders to NAC demonstrated worse CSS compared with that of the entire upfront surgery cohort, further emphasizing the benefit of predicting the chemotherapy response.4,8 Their study also correlated an increased tumor size with
Clinical Genitourinary Cancer Month 2018
-5
Optimal Patient Selection for NAC poor pathologic response. In our study, the presence of hydronephrosis, used as a surrogate for tumor burden and depth of invasion, correlated similarly with nonresponse. Although the investigators initially reported a correlation between NAC intolerance and poor pathologic response, this was not substantiated in their subsequent work with the same cohort or the National Cancer Database.4,8 In our study, incomplete NAC regimens were nearly 6 times more prevalent in nonresponders relative to complete responders (77% vs. 13%, respectively). Chappidi et al4 also found a lower Charlson comorbidity index was predictive of the pathologic response. Our study also identified an association of better ECOG performance status with an improved pathologic response. No patients with ECOG score of 2 had a complete pathologic response in our study. Similar to the findings from Chappidi et al,4 a delay in the initiation of NAC was associated with an inferior pathologic response. In addition to ECOG performance status, we found that the presence of cardiovascular comorbidities and decreased nutritional status were predictive of a poor pathologic response to NAC and the patient’s ability to tolerate the full NAC course. This might be related, in part, to the ability of diseased vasculature to deliver concentrated chemotherapeutics to target tissues, because cardiovascular disease is considered a field effect. Furthermore, advanced age was predictive of both a poor pathologic response and the inability of patients to tolerate NAC. Previous work has shown that older age at RC is associated with pathologic upstaging, worse CSS, and greater rates of recurrent disease.19 Considered together, patient age, functional status, nutritional status, and comorbidities should be considered when selecting candidates for both NAC and RC. As a tertiary referral center for a large rural patient population, many of our patients undergo staging transurethral resection at outlying institutions and more than one half received their NAC outside of our institution. Our data suggest that patients who received NAC at our academic institution relative to the community setting were both more likely to receive full-course therapy and to exhibit a favorable pathologic response. Additionally, an increased time from the diagnosis of muscle-invasive disease to the initiation of NAC was associated with a poor pathologic response and the ability to tolerate NAC. After the diagnosis of muscle-invasive disease, patients in the nonresponse group started NAC nearly 3 weeks later than did those patients in the complete response group. The delay in NAC could be attributable to delays in referral from outside institutions, delays in obtaining staging imaging, or lengthy medical comorbidity optimization, because these patients were, overall, less healthy than were the NAC responders. Historically there are improved clinical and oncologic outcomes for high-volume centers,20 and our data further support the utility of centralized, coordinated, and timely care in a tertiary referral center. Furthermore, the interval from diagnosis to RC was, on average, 16 weeks longer in the NAC group than for the patients who received upfront RC. This delay might explain the inferior outcomes for nonresponders to NAC compared with upfront RC and further implicates the importance of predicting pathologic response/progression during NAC.4 Our study additionally investigated clinical predictors of extravesical disease for patients with clinical stage 2 urothelial carcinoma
6
-
Clinical Genitourinary Cancer Month 2018
who underwent upfront RC. Pathologic upstaging occurs in w40% of patients treated with RC alone, with 32% to 36% of patients with cT2 disease having extravesical disease found on final pathologic evaluation.21,22 The predictors of upstaging at upfront RC in a large retrospective cohort of >600 patients include high-grade tumors, lymphovascular invasion, female gender, and histologic variants.22 Culp et al13 identified high-risk patients who received RC alone as those with hydronephrosis, extravesical disease on clinical stage, and/or the presence of variant histologic features. These patients experienced decreased CSS (64% vs. 84%; P < .001) and OS (47% vs. 65%; P < .001).13 A high degree of redundancy was found in the clinicopathologic predictors of advanced disease between our NAC and upfront RC cohorts. The factors associated with extravesical disease in cT2 patients who received upfront RC included advanced age, greater ECOG performance status, and the presence of cardiovascular risk factors. None of our patients with organ-confined disease had an ECOG score >1. Similar to Culp et al,13 hydronephrosis and variant histology predicted extravesical disease, with nearly double the incidence of variant histologic features in our cohort with extravesical disease. Evidence supports the role for upfront RC for the appropriate patient. Bhindi et al23 case-matched patients treated with NAC followed by RC with patients treated with upfront RC. They found that patients with residual disease after NAC followed by RC exhibited significantly worse CSS and OS compared with their stage-matched controls who had undergone upfront RC, especially ypT2-T4 or Nþ disease.23 The identification of factors that predict response to NAC would further optimize treatment selection on a patient-specific basis. Future optimization of treatment selection will implicate genomic sequencing information and DNA expression profiles. A recent study by Plimack et al24 revealed DNA alterations in nearly 300 cancer-related genes in patients treated with NAC, including 3 particular mutations in DNA repair genes (ATM, RB1, FANCC ), which predicted for OS and the pathologic response to NAC. Choi et al25 characterized the expression profiles of 3 subtypes of urothelial carcinoma: basal, luminal, and p53-like. These distinct, wellcharacterized subtypes predicted for the response to cisplatin-based NAC.25 Our group is currently evaluating the role of epithelialemesenchymal transition markers in predicting chemotherapeutic resistance. A multidisciplinary approach to selecting candidates with cT2 disease for NAC has been recommended. Patients who received NAC were younger and had fewer cardiovascular risk factors than those who received upfront RC. We found these factors to be significant predictors of a poor NAC response. Our data suggest that candidates should not only be evaluated for NAC according to the clinical indications but also using patient-specific factors such as age, ECOG performance status, and the presence of cardiovascular comorbidities, because these are predictive of the tolerability of chemotherapy. An ECOG performance status of 1 was a significant predictor of incomplete NAC compared with an ECOG performance status of 0. The small number of patients with an ECOG performance status of 2 limited the multivariate statistical analysis (Table 4). A high degree of redundancy was found in the clinicopathologic predictors of NAC tolerability with the predictors of adverse pathologic features found at
Patrick J. Hensley et al RC in both the NAC and the upfront surgery cohorts, suggesting these are not mutually exclusive. Rather, to a degree, patients with poor overall health, deconditioning, and adverse pathologic features respond poorly, regardless of the treatment modality. Nonetheless, emerging data suggest inferior oncologic outcomes in patients with residual muscle-invasive disease after NAC compared with patients who received upfront RC.23 Therefore, further pursuit of clinicopathologic and molecular markers that predict the final pathologic stage is warranted. The present study was limited by its retrospective nature with short follow-up times, precluding an outcomes analysis. However, in cohorts of patients undergoing RC for muscle-invasive disease, the final pathologic stage and pathologic response to NAC are validated surrogate endpoints for cancer-specific clinical outcomes.8,9,11,21 Although we found clinical utility in the subclassification of NAC response into 3 groups (complete, partial, and nonresponse), the small number of patients in each group limited the statistical analysis. Additionally, the results of the present study were limited to those patients with clinical stage T2 urothelial carcinoma of the bladder and should not be applied to patients with locally advanced, extravesical, or metastatic disease. Lymphovascular invasion status on pathologic examination of transurethral resection was unknown for a large subgroup of patients and that group was excluded from analysis. Additionally, variant histologic features were not quantified in all cases. Records regarding outside facility chemotherapy administration and reasons for incomplete regimens were limited, and this should remain a focus for ongoing efforts to predict patient tolerability of NAC regimens and the pathologic and clinical response to therapy.
Recent evidence has suggested that patients with pathologic
progression during NAC will have worse CSS than patients who undergo upfront surgery. The goal of the present study was to investigate the clinicopathologic predictors of the pathologic response in patients who receive NAC followed by RC and to determine factors predicting extravesical disease in patients treated with upfront RC. We demonstrated an inferior pathologic response to alternative NAC regimens (not gemcitabine/cisplatin or MVAC) in clinically organ-confined disease. Additionally, cardiovascular comorbidities and nutritional status were associated with lower tolerability and response to NAC. Incomplete chemotherapy regimens, defined as those with a break, delay, or dose reduction, predicted for pathologic progression on NAC. The patient characteristics associated with incomplete NAC regimens included increased age and poor performance status. The present study identified several clinical and pathologic characteristics predictive of tolerance and the response to NAC, improving patient selection for NAC and sequencing of definitive therapy. The importance of multidisciplinary, centralized care of patients with muscle-invasive disease was evident by the improved outcomes for our patients who received their chemotherapy within our institution and had undergone definitive treatment in a timely manner.
Disclosure The authors declare that they have no competing interests.
Conclusion The present study identified several clinical and pathologic characteristics predictive of tolerance and response to NAC in patients with muscle-invasive urothelial cell carcinoma. Patients who received alternative NAC regimens, and those who poorly tolerated their NAC, were more likely to exhibit pathologic progression. ECOG performance status, cardiovascular comorbidities, and nutritional status are associated with pathologic response and an ability to tolerate NAC to completion. Consideration of these patient-specific factors might allow for more appropriate treatment selection and reduce the delay of definitive surgical treatment in the substantial subset of patients who cannot tolerate or respond to NAC. This demonstrates the importance of multidisciplinary, centralized care for these patients. In the absence of novel or better tolerated chemotherapeutic agents, ongoing efforts should be focused on tailoring sequencing of therapy to individual patients to improve the clinical outcomes of those with muscle-invasive bladder cancer.
Clinical Practice Points NAC for muscle-invasive urothelial carcinoma of the bladder
results in a complete pathologic response in 23% to 38% of patients. The survival advantage for patients with a pathologic response to cisplatin-based NAC compared with those with pathologic progression is well-documented.
References 1. Grossman HB, Natale RB, Tangen CM, et al. Neoadjuvant chemotherapy plus cystectomy compared with cystectomy alone for locally advanced bladder cancer. N Engl J Med 2003; 349:859-66. 2. International Collaboration of Trialists; Medical Research Council Advanced Bladder Cancer Working Party (now the National Cancer Research Institute Bladder Cancer Clinical Studies Group); European Organisation for Research and Treatment of Cancer Genito-Urinary Tract Cancer Group; Australian Bladder Cancer Study Group; National Cancer Institute of Canada Clinical Trials Group; Finnbladder; Norwegian Bladder Cancer Study Group; Club Urologico Espanol de Tratamiento Oncologico Group, Griffiths G, Hall R, Sylvester R, et al. International phase III trial assessing neoadjuvant cisplatin, methotrexate, and vinblastine chemotherapy for muscle-invasive bladder cancer: long-term results of the BA06 30894 trial. J Clin Oncol 2011; 29:2171-7. 3. Tan WS, Lamb BW, Kelly JD. Complications of radical cystectomy and orthotopic reconstruction. Adv Urol 2015; 2015:323157. 4. Chappidi MR, Kates M, Brant A, et al. Assessing cancer progression and stable disease after neoadjuvant chemotherapy for organ-confined muscle-invasive bladder cancer. Urology 2017; 102:148-58. 5. Dogliotti L, Cartenì G, Siena S, et al. Gemcitabine plus cisplatin versus gemcitabine plus carboplatin as first-line chemotherapy in advanced transitional cell carcinoma of the urothelium: results of a randomized phase 2 trial. Eur Urol 2007; 52:134-41. 6. David KA, Milowsky MI, Ritchey J, Carroll PR, Nanus DM. Low incidence of perioperative chemotherapy for stage III bladder cancer 1998 to 2003: a report from the National Cancer Data Base. J Urol 2007; 178:451-4. 7. Pokuri VK, Syed JR, Yang Z, et al. Predictors of complete pathologic response (pT0) to neoadjuvant chemotherapy in muscle-invasive bladder carcinoma. Clin Genitourin Cancer 2016; 14:e59-65. 8. Gandhi NM, Baras A, Munari E, et al. Gemcitabine and cisplatin neoadjuvant chemotherapy for muscle-invasive urothelial carcinoma: predicting response and assessing outcomes. Urol Oncol 2015; 33:204.e1-7. 9. Palapattu GS, Shariat SF, Karakiewicz PI, et al. Bladder Cancer Research Consortium. Cancer specific outcomes in patients with pT0 disease following radical cystectomy. J Urol 2006; 175:1645-9. 10. Volkmer BG, Kuefer R, Bartsch G Jr, et al. Effect of a pT0 cystectomy specimen without neoadjuvant therapy on survival. Cancer 2005; 104:2384-91.
Clinical Genitourinary Cancer Month 2018
-7
Optimal Patient Selection for NAC 11. Kassouf W, Spiess PE, Brown GA, et al. P0 stage at radical cystectomy for bladder cancer is associated with improved outcome independent of traditional clinical risk factors. Eur Urol 2007; 52:769-74. 12. Green DA, Rink M, Hansen J, et al. Accurate preoperative prediction of non-organ confined bladder urothelial carcinoma at cystectomy. BJU Int 2013; 111:404-11. 13. Culp SH, Dickstein RJ, Grossman HB, et al. Refining patient selection for neoadjuvant chemotherapy before radical cystectomy. J Urol 2014; 191:40-7. 14. Vetterlein MW, Wankowicz SAM, Seisen T, et al. Neoadjuvant chemotherapy prior to radical cystectomy for muscle-invasive bladder cancer with variant histology. Cancer 2017; 123:4346-55. 15. Mossanen M, Lee F, Cheng H, et al. Nonresponse to neoadjuvant chemotherapy for muscle-invasive urothelial cell carcinoma of the bladder. Clin Genitourin Cancer 2014; 12:210-3. 16. Galsky MD, Chen GJ, Oh WK, et al. Comparative effectiveness of cisplatin-based and carboplatin-based chemotherapy for treatment of advanced urothelial carcinoma. Ann Oncol 2012; 23:406-10. 17. Stimson CJ, Cookson MS, Barocas DA, et al. Preoperative hydronephrosis predicts extravesical and node positive disease in patients undergoing cystectomy for bladder cancer. J Urol 2010; 183:1732-7. 18. Mertens LS, Meijer RP, Meinhardt W, et al. Occult lymph node metastases in patients with carcinoma invading bladder muscle: incidence after neoadjuvant chemotherapy and cystectomy vs after cystectomy alone. BJU Int 2014; 114:67-74.
8
-
Clinical Genitourinary Cancer Month 2018
19. Nielsen ME, Shariat SF, Karakiewicz PI, et al. Bladder Cancer Research Consortium (BCRC). Advanced age is associated with poorer bladder cancerspecific survival in patients treated with radical cystectomy. Eur Urol 2007; 51:699-706. 20. Afshar M, Goodfellow H, Jackson-Spence F, et al. Centralisation of radical cystectomies for bladder cancer in England, a decade on from the “Improving Outcomes Guidance”: the case for super centralisation. BJU Int 2018; 121: 217-24. 21. Shariat SF, Palapattu GS, Karakiewicz PI, et al. Discrepancy between clinical and pathologic stage: impact on prognosis after radical cystectomy. Eur Urol 2007; 51: 137-49. 22. Turker P, Bostrom PJ, Wroclawski ML, et al. Upstaging of urothelial cancer at the time of radical cystectomy: factors associated with upstaging and its effect on outcome. BJU Int 2012; 110:804-11. 23. Bhindi B, Frank I, Mason RJ, et al. Oncologic outcomes for patients with residual cancer at cystectomy following neoadjuvant chemotherapy: a pathologic stagematched analysis. Eur Urol 2017; 17:30406-12. 24. Plimack ER, Dunbrack RL, Brennan TA, et al. Defects in DNA repair genes predict response to neoadjuvant cisplatin-based chemotherapy in muscle-invasive bladder cancer. Eur Urol 2015; 68:959-67. 25. Choi W, Porten S, Kim S, et al. Identification of distinct basal and luminal subtypes of muscle-invasive bladder cancer with different sensitivities to frontline chemotherapy. Cancer Cell 2014; 25:152-65.
Optimization of Patient Selection for Neoadjuvant Chemotherapy in Muscle-invasive Urothelial Carcinoma of the Bladder Patrick J. Hensley,1 Jeffrey Goodwin,1 Daniel L. Davenport,2 Stephen E. Strup,1 Andrew James1 Abstract Few predictors are known for determining the response to neoadjuvant chemotherapy (NAC) for bladder cancer. We investigated the clinicopathologic predictors of extravesical disease in cohorts receiving NAC and upfront cystectomy. The present study demonstrated an inferior response to alternative NAC regimens, implicating cardiovascular comorbidities and nutritional status in the tolerability and response to NAC. Background: Radical cystectomy (RC) is delayed in a subset of patients who respond poorly to neoadjuvant chemotherapy (NAC). The present study investigated the clinicopathologic characteristics predicting extravesical disease at RC and the factors associated with NAC tolerability to improve patient selection and the sequence of definitive therapy. Materials and Methods: Patients with cT2 urothelial carcinoma of the bladder who underwent NAC were stratified by the final pathologic stage: complete (ypT0N0), partial ( pT2), and nonresponse (> pT2 and/or Nþ). Patients treated with upfront cystectomy were divided into those with organ-confined ( pT2) and those with extravesical disease (> pT2 and/or Nþ). Results: Of 145 patients, 89 received NAC and 56 underwent upfront RC. The univariate predictors of extravesical disease in the patients treated with upfront RC included increased age (P ¼ .021), higher Eastern Cooperative Oncology Group performance status (P < .001), hydronephrosis (P ¼ .021), and cardiovascular risk factors. The complete, partial, and nonresponse rates to NAC were 25.8%, 39.3%, and 34.8%, respectively. The multivariate predictors of pathologic progression on NAC included low serum albumin (P ¼ .005), hydronephrosis (P ¼ .040), incomplete NAC (P ¼ .014), and alternative NAC (non-gemcitabine/cisplatin or MVAC, P ¼ .022). Significant multivariate predictors of incomplete NAC included increased age, coronary artery disease (P ¼ .027), and Eastern Cooperative Oncology Group performance status. Conclusion: Redundant clinicopathologic features predicted adverse cystectomy pathology in patients treated with both NAC and upfront RC. The results of the present study demonstrated an inferior pathologic response to alternative NAC regimens in clinically organ-confined disease and implicated cardiovascular comorbidities and nutritional status in the tolerability and response to NAC. Our findings predicate the importance of using patient-specific factors to guide the sequence of definitive treatment toward timely, centralized care to improve clinical outcomes. Clinical Genitourinary Cancer, Vol. -, No. -, --- ª 2018 Elsevier Inc. All rights reserved. Keywords: Bladder cancer, Cisplatin, Cystectomy, Pathologic response, Performance status
Introduction Neoadjuvant chemotherapy (NAC), followed by radical cystectomy (RC), is the reference standard treatment for muscle1
Department of Urology Department of Surgery, University of Kentucky College of Medicine, Lexington, KY
2
Submitted: Dec 8, 2017; Revised: Feb 10, 2018; Accepted: Feb 18, 2018 Address for correspondence: Patrick J. Hensley, MD, Department of Urology, University of Kentucky, 800 Rose Street, Lexington, KY 40536 E-mail contact: [email protected]
1558-7673/$ - see frontmatter ª 2018 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.clgc.2018.02.007
invasive bladder cancer. RC provides local control and NAC treats micrometastatic disease, conferring a 5% to 10% survival benefit compared with RC alone. Patients receiving NAC exhibit a greater rate of pathologic T0 (pT0) at RC, providing a plausible explanation of the survival benefit.1,2 Both RC and NAC, however, are associated with considerable morbidity. Several series have shown a 28% to 64% complication rate, 26% to 35% rehospitalization rate, and up to 8% 90-day mortality after RC.3 In prospective studies, NAC with MVAC (methotrexate, vinblastine, doxorubicin, cisplatin) was associated
Clinical Genitourinary Cancer Month 2018
-1
Optimal Patient Selection for NAC with a 35% to 37% rate of grade 3/4 toxicity,1,2 with these toxicities associated with decreased overall survival (hazard ratio, 2.71; P < .001).4 Gemcitabine/cisplatin, although better tolerated than MVAC, has been associated with a 2% to 27% rate of grade 3/4 hematologic toxicity.5 In this subset of patients, NAC might delay the time to definitive extirpative therapy and result in additional morbidity. These patients may benefit from foregoing NAC for upfront RC. The identification of clinical and pathologic variables that could predict a favorable response to NAC would allow for a more selective approach to NAC usage. Additionally, the identification of factors that could predict for a patient’s ability to tolerate and complete NAC may improve patient selection. We investigated these variables in subsets of patients treated with NAC followed by RC and in patients who received upfront RC to identify clinicopathologic characteristics that may be used to improve patient-specific clinical decision making and, ultimately, reduce morbidity and improve the outcomes for patients with muscle-invasive bladder cancer.
Materials and Methods
2
-
An institutional review board-approved retrospective review of patients was performed. A waiver of informed consent was obtained. The inclusion criteria included clinical stage T2 urothelial carcinoma of the bladder with muscularis propria invasion found on transurethral resection pathologic examination, cross-sectional imaging depicting no evidence of extravesical disease, and negative findings with bimanual examination under anesthesia. All patients underwent RC, pelvic lymph node dissection, and urinary diversion at the University of Kentucky from January 2010 to April 2017. Tumor staging was determined using the 2015 American Joint Committee on Cancer staging system. Patients either underwent upfront RC or NAC (regimen described), followed by RC. Patients with history of initial radiation therapy for bladder cancer, those who had undergone partial cystectomy, and patients with pure adenocarcinoma, squamous cell carcinoma, or neuroendocrine carcinoma were excluded. The patients who underwent NAC were stratified into 3 groups according to the final pathologic stage: complete response (ypT0N0), partial response (final pathologic stage pT2), and nonresponse (final pathologic stage > pT2 and/or Nþ). The patients who had undergone upfront cystectomy were divided into those with organ-confined disease (final pathologic stage pT2) and those with extravesical disease (final pathologic stage > pT2 and/or Nþ). A review was performed of preoperative clinical variables and the reports pathology reports from transurethral resection specimens and cystectomy specimens. An incomplete chemotherapy regimen was defined as any break, delay, or dose reduction in the prescribed chemotherapy regimen. An alternative chemotherapy regimen was defined as any variant from gemcitabine/cisplatin or MVAC. All outside pathology slides were reviewed by the pathology department of the University of Kentucky to confirm the stage, grade, and presence or absence of variant histologic features. Any level of variant histologic features as described on the transurethral resection pathology report was considered variant. Data were statistically analyzed using analysis of variance or c2 tests of univariate differences and logistic regression for multivariable analysis after combining those with partial responses and no
Clinical Genitourinary Cancer Month 2018
responses. Significance was set at P < .05. All statistical calculations were performed using SPSS, version 22 (IBM Corp, Armonk, NY).
Results Comparison of Cohorts: Upfront RC Versus NAC Of the 145 patients who met the inclusion criteria, 89 underwent NAC followed by RC and 56 underwent upfront RC. The clinicopathologic and demographic variables for these 2 groups of patients are listed in Table 1. The patients who had undergone upfront RC were, on average, older and had a greater baseline serum creatinine than did patients treated with NAC. NAC was prescribed more commonly for patients with variant histologic features. Although not statistically significant, a trend was found toward a greater Eastern Cooperative Oncology Group (ECOG) performance status in patients who had undergone upfront RC. Patients treated with NAC experienced a mean delay in the interval from diagnosis to definitive surgery that was 16 weeks longer than that for the upfront surgery cohort. No statistically significant socioeconomic differences were found between the 2 groups (data not shown).
Clinical and Pathologic Predictors of NAC Response The complete pathologic response rate to NAC was 26% and the nonresponse rate was 35% (Table 1). Statistically significant univariate predictors of a poor response to NAC included increased age, low serum albumin, increased ECOG performance status, and the presence of hydronephrosis (unilateral or bilateral). Cardiovascular comorbidities, including the presence of hyperlipidemia, hypertension, coronary artery disease, or a history of myocardial infarction, were univariate predictors of a poor response to NAC. Although more common in the nonresponse group, the pooled incidence of variant histologic features was not significant among the groups. Further statistical analysis of the specific variant histologic features was not performed, given the low incidence. All 4 patients with micropapillary differentiation on transurethral resection were nonresponders. No patients with micropapillary, sarcomatoid, or clear cell differentiation were in the complete response group. The nonresponders began their first chemotherapy cycle after the diagnosis of muscle-invasive disease, on average, nearly 3 weeks later than the complete responders (Table 1). The deviation from a standard chemotherapeutic regimen as prescribed was associated with a poor pathologic response to NAC. Patients who received alternative regimens from the standard gemcitabine/cisplatin or MVAC (most commonly gemcitabine/carboplatin) responded poorly to NAC, as did the patients who required a break, delay, or dose reduction in NAC cycles (defined as incomplete NAC). These findings were confirmed as independent predictors of disease progression with NAC on multivariate regression analysis (Table 2).
Clinical Predictors of Incomplete NAC We investigated the clinical predictors of patients receiving an incomplete NAC regimen. Increased age and ECOG performance status were the most significant predictors of a patient’s ability to tolerate chemotherapy as prescribed on univariate analysis (Table 3). These were also statistically significant predictors of a poor pathologic response to NAC. Although chemotherapy regimens prescribed and administered at our institution was not a statistically
Table 1 Univariate Predictors of Pathologic Response to NAC and Extravesical Disease After RC NAC Variable Patients, n (%) Female gender, % Age, y 2
BMI, kg/m
Upfront RC
NR
PR
CR
All Patients
31 (34.8)
35 (39.3)
23 (25.8)
89 (100)
22.6
17.1
21.7
20.2
66.3 1.7
65.0 1.5
58.8 1.6
63.8 1.0
26.7 0.8
28.6 1.0
28.1 1.4
27.8 0.6
P Value
OCD
EVD
All Patients
22 (39)
34 (61)
56
a
18
18
18
a
65 2
71 2
69 1
a
28.4 1.2
26.7 0.8
27.4 0.7
.925
.003 .393
P Value (NAC vs. Upfront RC)
1.000a
.830a
a
.005a
a
.625a
a
.021 .218
80.6
82.9
65.2
77.5
.215
82
88
86
.698
.282a
3.1 0.8
3.5 0.4
3.6 0.5
3.4 0.6
.004a
3.5 0.1
3.4 0.1
3.5 0.1
.408a
.550a
Pre-NAC/Sx serum creatinine, mg/dL
1.03 0.06
0.98 0.05
0.98 0.05
1.00 0.03
a
.499
1.3 0.2
1.5 0.2
1.4 0.1
a
.589
.010a
Post-NAC serum creatinine, mg/dL
1.11 0.06
1.04 0.07
0.96 0.04
1.04 0.04
.108a
NA
NA
NA
NA
NA
Tobacco use, % Albumin, g/dL (n ¼ 83)
a
P Value
.037a
ECOG PS, % 0
45.1
51.5
73.9
55.0
82
35
53
1
45.2
37.1
26.1
37.1
18
35
29
2
9.7
11.4
0.0
7.9
0
24
14
3
0
.159a
0
0
0
6
4
48.4
17.1%
17.4
28.1
.008a
14
44
32
.021a
.709a
Hyperlipidemia
51.6
34.3
17.4
36.0
.010a
23
56
43
.026a
.484a
Hypertension
90.3
60.0
65.2
71.9
.028a
50
85
71
.007a
1.000a
18.0
a
23
a
.524a
a
1.000a
a
.073a
a
.013a
Hydronephrosis, %
0
<.001a
Cardiovascular comorbidities, %
Coronary artery disease Diabetes mellitus Stroke Myocardial infarction
32.3 22.6 3.2
14.3 17.1 0.0
4.3 21.7 0
20.2 1.1
.007
a
.897
a
.239
9 9 5
32 29 9
21 7
.056 .099
1.000
a
0
0
0
1.000
a
NA
NA
NA
NA
5.7
0
10.1
.005
Interval from diagnosis to NAC, d
62 7
53 5
42 4
53 3
Interval from NAC end to Sx, d
91 18
62 6
76 20
76 9
.483
NA
NA
NA
Interval from diagnosis to Sx, d
223 20
191 9
203 20
205 9
.416a
64 10
113 22
93 14
.026
a
NA
NA NA
.060a
<.001a
Biopsy pathologic findings Carcinoma in situ
16.1
25.7
17.4
20.2
.834a
23
29
27
.759a
.418a
Variant histologic features (any)
29.0
17.1
21.7
22.5
.472a
27.3
47.1
39.3
.169a
.039a
Squamous cell
9.7
5.7
17.4
10.1
18.2
29.4
25.0
Glandular
3.2
8.6
8.7
6.7
4.5
2.9
3.6
Sarcomatoid
3.2
0
0
1.1
0
2.9
1.8
Micropapillary
12.9
0
0
4.5
4.5
8.8
7.1
Patrick J. Hensley et al
Clinical Genitourinary Cancer Month 2018
22.6
-3
Final Nþ
Data presented as n (%), mean standard error, or %. Abbreviations: BMI ¼ body mass index; CR ¼ complete response; ECOG ¼ Eastern Cooperative Oncology Group; EVD ¼ extravesical disease; OCD ¼ organ-confined disease; NAC ¼ neoadjuvant chemotherapy; NR ¼ no response; PR ¼ partial response; PS ¼ performance status; Sx ¼ surgery. a Statistically significant.
.524a <.001a 38 0 <.001a 0
Clinical Genitourinary Cancer Month 2018
52
0
18
<.001a 50 13 50 Incomplete NAC
77
9 15 Alternative NAC
NAC at our institution
36
65 54
21
.052a 0 2.9 Plasmacytoid
NAC regimen, %
Clear cell
0
39
CR PR
52
1.1
All Patients
-
Variable
NR
NAC Table 1 Continued
4
.013a
2.9 0
23
1.8
0 0
0
P Value (NAC vs. Upfront RC) All Patients EVD OCD P Value
Upfront RC
P Value
Optimal Patient Selection for NAC Table 2 Multivariate Predictors of Pathologic Progression During NAC Variable Mean albumin, g/dL (n ¼ 83) Hydronephrosis Incomplete NAC Alternative NAC Myocardial infarction
OR (95% CI) 0.218 3.601 4.499 4.926 12.091
(0.076-0.627) (1.057-12.266) (1.359-14.891) (1.262-19.231) (1.736-84.219)
P Value .005 .040 .014 .022 .012
Abbreviations: CI ¼ confidence interval; NAC ¼ neoadjuvant chemotherapy; OR ¼ odds ratio.
significant predictor of overall pathologic response (Table 1), patients who received chemotherapy at the University of Kentucky were more likely to complete NAC relative to those who underwent treatment at outlying institutions. Additionally, the interval from diagnosis to the initiation of NAC correlated inversely with a patient’s ability to tolerate a complete treatment course. On multivariate regression analysis, greater ECOG performance status, the presence of coronary artery disease, and older age were all independent predictors of patients receiving an incomplete NAC regimen (Table 4).
Clinical and Pathologic Predictors of Extravesical Disease in Patients Who Underwent Upfront RC Patients who received upfront RC had a shorter interval from the diagnosis of muscle-invasive disease to definitive extirpative surgery. The univariate clinical and pathologic predictors of extravesical disease at RC for cT2 urothelial carcinoma of the bladder in patients who underwent upfront RC included increased age and ECOG performance status, the presence of hydronephrosis, and a history of hypertension or hyperlipidemia (Table 1). Ten patients had an ECOG performance status of 2 to 3, none of whom had organconfined disease at RC. The incidence of variant histologic features was 20% greater in the extravesical disease group, with a predominance of sarcomatoid and micropapillary differentiation found in this group. ECOG performance status was the strongest predictor of extravesical disease at RC on multivariate analysis. In addition to ECOG, only coronary artery disease was an additional independent predictor of extravesical disease on multivariate analysis (odds ratio, 7.0; 95% confidence interval, 1.2-41.9; P ¼ .033). An ECOG performance status of 2 to 3 had a complete association with extravesical disease, precluding multivariate statistical analysis.
Discussion The optimal patient selection for NAC before RC for patients with muscle invasive urothelial cell carcinoma of the bladder is often challenging owing to the poor predictors of both disease response and patient tolerability of chemotherapy. The ability to reliably predict which patients will benefit from and/or tolerate NAC to completion could improve the outcomes for patients with muscle-invasive bladder cancer. The commonly cited reasons for forgoing NAC include the delay to definitive surgery and the potential toxicity of chemotherapeutic regimens.6 The present study identified clinicopathologic characteristics predicting pathologic response to NAC and patient-specific factors associated with patients’ ability to complete
Patrick J. Hensley et al Table 3 Univariate Predictors of Incomplete NAC Variable
Table 4 Multivariate Predictors of Incomplete NAC
Full Course Incomplete NAC NAC P Value
Patients, n
44
44
Female gender, %
18.2
22.7
60 y
59.1
25.0
61-70 y
27.3
36.4
71
13.6
38.6
Tobacco use, %
75.0
79.5
.800
Married, %
65.9
47.7
.131
Lives with family, %
75.0
63.6
.355
0
75.0
36.4
1
20.5
52.3
2
4.5
11.4
Age group, %
.792 .003
ECOG PS, %
Coronary artery disease Myocardial infarction
.001
59.1
84.1
.017
6.8
29.5
.011
4.5
15.9
.157
Carcinoma in situ
13.6
25.0
.280
Mean pre-NAC serum creatinine >1.2 mg/dL, %
12.5
23.8
.255
Interval from diagnosis to NAC, d
44 3
63 6
.008
65.9
38.6
.018
NAC at our institution, %
ECOG PS 1 versus 0 ECOG PS 2 versus 0 Coronary artery disease Age 60-70 versus <60 y Age >70 versus <60 y
OR (95% CI) 4.746 3.085 5.183 3.363 4.549
(1.639-13.742) (0.441-21.591 (1.205-22.292) (1.070-10.567) (1.232-16.798)
P Value .004 .256 .027 .038 .023
Abbreviations: CI ¼ confidence interval; ECOG ¼ Eastern Cooperative Oncology Group; NAC ¼ neoadjuvant chemotherapy; OR ¼ odds ratio; PS ¼ performance status.
Cardiovascular comorbidities, % Hypertension
Variable
Data presented as n (%), mean standard error, or %. Abbreviations: ECOG ¼ Eastern Cooperative Oncology Group; NAC ¼ neoadjuvant chemotherapy; PS ¼ performance status.
NAC for clinical stage T2 urothelial carcinoma of the bladder. The complete pathologic response rate to NAC in our cohort of 26% was comparable to that reported in previous studies at 23% to 28%.7,8 Previous studies have sought to define the categories of pathologic response for prognostic purposes, including a complete pathologic response (ypT0N0) and a spectrum of stable, organ-confined disease and progressive extravesical disease during NAC.7,8 Although a complete pathologic response to NAC has been intuitively associated with improved clinical outcomes relative to patients with residual disease, the clinical implications for survival are less clear. Gandhi et al8 studied the survival outcomes for 150 patients with muscle-invasive bladder cancer stratified by pathologic response. The 5-year cancer-specific survival (CSS) was improved in responders (ypT1) compared with nonresponders (ypT2; 91% vs. 27%; P < .01). No statistically significant difference was found in survival on the subgroup analysis between patients with ypT0 and those with residual, nonemuscle-invasive disease.8 The rate of pT0 disease at RC from transurethral resection alone ranges from 7% to 20%.9,10 Kassouf et al11 found similar CSS rates among patients without residual disease who had received NAC, followed by RC, or RC alone (92% vs. 86%, respectively; P ¼ .65), suggesting that the treatment of micrometastatic disease by NAC might be of less importance than its effect on the primary tumor. The present study was limited to patients who had undergone RC for clinical stage T2 urothelial carcinoma, because we believe
this clinical scenario provides the greatest room for interpretation in the value of NAC. Although hydronephrosis has been identified as a risk factor for extravesical disease, we chose to consider patients without local tumor extension or lymphadenopathy as having clinical stage T2, because w35% of patients with hydronephrosis have been shown to have organ-confined disease at RC.12 The preponderance of patients undergoing RC for urothelial carcinoma do so for clinical stage T2 disease.7-11,13 Pokuri et al7 completed a retrospective review of 50 patients who had undergone NAC, followed by robotic cystectomy for cT2 urothelial carcinoma. Pokuri et al7 divided patients according disease status (ypT0 and non-ypT0 disease), with variant histologic features predicting for an incomplete response. Their study was limited by its small cohort size and the inclusion of patients with clinically extravesical disease. In our cohort, micropapillary and sarcomatoid variants were exclusively found in the nonresponse group. In a retrospective study of patients with variant histologic features from the National Cancer Database, NAC was associated with improved rates of organ-confined disease at RC relative to patients treated with upfront RC.14 In our cohort, alternative NAC regimens primarily consisted of gemcitabine/carboplatin or conversion to Taxol-based regimens because of intolerance. In locally advanced or metastatic disease, gemcitabine/cisplatin exhibited an improved pathologic response,15 time to progression, and overall survival, with comparable acceptable toxicity, relative to gemcitabine/carboplatin.5,16 Consistent with other studies, we have demonstrated a significantly inferior pathologic response to alternative NAC in those with clinically organ-confined disease.12,17 Alternative regimens were associated with a nearly fivefold increase in pathologic progression during NAC compared with standard gemcitabine/cisplatin or MVAC (odds ratio, 4.926; 95% confidence interval, 1.262-19.231; P ¼ .022). In a retrospective study of patients with noneorgan-confined disease, pathologic downstaging was exhibited in 59% of patients treated with neoadjuvant gemcitabine/cisplatin compared with 7% of patients treated with gemcitabine/carboplatin; however, the difference was not statistically significant (P ¼ .444).18 Chappidi et al4 divided patients with cT2N0 disease into those with stable (ypT2N0M0) and progressive (>ypT2 or Nþ) disease after NAC. Patients with stable disease during NAC had better overall survival (OS) than did those with progression during NAC, and OS comparable to that of patients with cT2 disease who had undergone upfront RC. Nonresponders to NAC demonstrated worse CSS compared with that of the entire upfront surgery cohort, further emphasizing the benefit of predicting the chemotherapy response.4,8 Their study also correlated an increased tumor size with
Clinical Genitourinary Cancer Month 2018
-5
Optimal Patient Selection for NAC poor pathologic response. In our study, the presence of hydronephrosis, used as a surrogate for tumor burden and depth of invasion, correlated similarly with nonresponse. Although the investigators initially reported a correlation between NAC intolerance and poor pathologic response, this was not substantiated in their subsequent work with the same cohort or the National Cancer Database.4,8 In our study, incomplete NAC regimens were nearly 6 times more prevalent in nonresponders relative to complete responders (77% vs. 13%, respectively). Chappidi et al4 also found a lower Charlson comorbidity index was predictive of the pathologic response. Our study also identified an association of better ECOG performance status with an improved pathologic response. No patients with ECOG score of 2 had a complete pathologic response in our study. Similar to the findings from Chappidi et al,4 a delay in the initiation of NAC was associated with an inferior pathologic response. In addition to ECOG performance status, we found that the presence of cardiovascular comorbidities and decreased nutritional status were predictive of a poor pathologic response to NAC and the patient’s ability to tolerate the full NAC course. This might be related, in part, to the ability of diseased vasculature to deliver concentrated chemotherapeutics to target tissues, because cardiovascular disease is considered a field effect. Furthermore, advanced age was predictive of both a poor pathologic response and the inability of patients to tolerate NAC. Previous work has shown that older age at RC is associated with pathologic upstaging, worse CSS, and greater rates of recurrent disease.19 Considered together, patient age, functional status, nutritional status, and comorbidities should be considered when selecting candidates for both NAC and RC. As a tertiary referral center for a large rural patient population, many of our patients undergo staging transurethral resection at outlying institutions and more than one half received their NAC outside of our institution. Our data suggest that patients who received NAC at our academic institution relative to the community setting were both more likely to receive full-course therapy and to exhibit a favorable pathologic response. Additionally, an increased time from the diagnosis of muscle-invasive disease to the initiation of NAC was associated with a poor pathologic response and the ability to tolerate NAC. After the diagnosis of muscle-invasive disease, patients in the nonresponse group started NAC nearly 3 weeks later than did those patients in the complete response group. The delay in NAC could be attributable to delays in referral from outside institutions, delays in obtaining staging imaging, or lengthy medical comorbidity optimization, because these patients were, overall, less healthy than were the NAC responders. Historically there are improved clinical and oncologic outcomes for high-volume centers,20 and our data further support the utility of centralized, coordinated, and timely care in a tertiary referral center. Furthermore, the interval from diagnosis to RC was, on average, 16 weeks longer in the NAC group than for the patients who received upfront RC. This delay might explain the inferior outcomes for nonresponders to NAC compared with upfront RC and further implicates the importance of predicting pathologic response/progression during NAC.4 Our study additionally investigated clinical predictors of extravesical disease for patients with clinical stage 2 urothelial carcinoma
6
-
Clinical Genitourinary Cancer Month 2018
who underwent upfront RC. Pathologic upstaging occurs in w40% of patients treated with RC alone, with 32% to 36% of patients with cT2 disease having extravesical disease found on final pathologic evaluation.21,22 The predictors of upstaging at upfront RC in a large retrospective cohort of >600 patients include high-grade tumors, lymphovascular invasion, female gender, and histologic variants.22 Culp et al13 identified high-risk patients who received RC alone as those with hydronephrosis, extravesical disease on clinical stage, and/or the presence of variant histologic features. These patients experienced decreased CSS (64% vs. 84%; P < .001) and OS (47% vs. 65%; P < .001).13 A high degree of redundancy was found in the clinicopathologic predictors of advanced disease between our NAC and upfront RC cohorts. The factors associated with extravesical disease in cT2 patients who received upfront RC included advanced age, greater ECOG performance status, and the presence of cardiovascular risk factors. None of our patients with organ-confined disease had an ECOG score >1. Similar to Culp et al,13 hydronephrosis and variant histology predicted extravesical disease, with nearly double the incidence of variant histologic features in our cohort with extravesical disease. Evidence supports the role for upfront RC for the appropriate patient. Bhindi et al23 case-matched patients treated with NAC followed by RC with patients treated with upfront RC. They found that patients with residual disease after NAC followed by RC exhibited significantly worse CSS and OS compared with their stage-matched controls who had undergone upfront RC, especially ypT2-T4 or Nþ disease.23 The identification of factors that predict response to NAC would further optimize treatment selection on a patient-specific basis. Future optimization of treatment selection will implicate genomic sequencing information and DNA expression profiles. A recent study by Plimack et al24 revealed DNA alterations in nearly 300 cancer-related genes in patients treated with NAC, including 3 particular mutations in DNA repair genes (ATM, RB1, FANCC ), which predicted for OS and the pathologic response to NAC. Choi et al25 characterized the expression profiles of 3 subtypes of urothelial carcinoma: basal, luminal, and p53-like. These distinct, wellcharacterized subtypes predicted for the response to cisplatin-based NAC.25 Our group is currently evaluating the role of epithelialemesenchymal transition markers in predicting chemotherapeutic resistance. A multidisciplinary approach to selecting candidates with cT2 disease for NAC has been recommended. Patients who received NAC were younger and had fewer cardiovascular risk factors than those who received upfront RC. We found these factors to be significant predictors of a poor NAC response. Our data suggest that candidates should not only be evaluated for NAC according to the clinical indications but also using patient-specific factors such as age, ECOG performance status, and the presence of cardiovascular comorbidities, because these are predictive of the tolerability of chemotherapy. An ECOG performance status of 1 was a significant predictor of incomplete NAC compared with an ECOG performance status of 0. The small number of patients with an ECOG performance status of 2 limited the multivariate statistical analysis (Table 4). A high degree of redundancy was found in the clinicopathologic predictors of NAC tolerability with the predictors of adverse pathologic features found at
Patrick J. Hensley et al RC in both the NAC and the upfront surgery cohorts, suggesting these are not mutually exclusive. Rather, to a degree, patients with poor overall health, deconditioning, and adverse pathologic features respond poorly, regardless of the treatment modality. Nonetheless, emerging data suggest inferior oncologic outcomes in patients with residual muscle-invasive disease after NAC compared with patients who received upfront RC.23 Therefore, further pursuit of clinicopathologic and molecular markers that predict the final pathologic stage is warranted. The present study was limited by its retrospective nature with short follow-up times, precluding an outcomes analysis. However, in cohorts of patients undergoing RC for muscle-invasive disease, the final pathologic stage and pathologic response to NAC are validated surrogate endpoints for cancer-specific clinical outcomes.8,9,11,21 Although we found clinical utility in the subclassification of NAC response into 3 groups (complete, partial, and nonresponse), the small number of patients in each group limited the statistical analysis. Additionally, the results of the present study were limited to those patients with clinical stage T2 urothelial carcinoma of the bladder and should not be applied to patients with locally advanced, extravesical, or metastatic disease. Lymphovascular invasion status on pathologic examination of transurethral resection was unknown for a large subgroup of patients and that group was excluded from analysis. Additionally, variant histologic features were not quantified in all cases. Records regarding outside facility chemotherapy administration and reasons for incomplete regimens were limited, and this should remain a focus for ongoing efforts to predict patient tolerability of NAC regimens and the pathologic and clinical response to therapy.
Recent evidence has suggested that patients with pathologic
progression during NAC will have worse CSS than patients who undergo upfront surgery. The goal of the present study was to investigate the clinicopathologic predictors of the pathologic response in patients who receive NAC followed by RC and to determine factors predicting extravesical disease in patients treated with upfront RC. We demonstrated an inferior pathologic response to alternative NAC regimens (not gemcitabine/cisplatin or MVAC) in clinically organ-confined disease. Additionally, cardiovascular comorbidities and nutritional status were associated with lower tolerability and response to NAC. Incomplete chemotherapy regimens, defined as those with a break, delay, or dose reduction, predicted for pathologic progression on NAC. The patient characteristics associated with incomplete NAC regimens included increased age and poor performance status. The present study identified several clinical and pathologic characteristics predictive of tolerance and the response to NAC, improving patient selection for NAC and sequencing of definitive therapy. The importance of multidisciplinary, centralized care of patients with muscle-invasive disease was evident by the improved outcomes for our patients who received their chemotherapy within our institution and had undergone definitive treatment in a timely manner.
Disclosure The authors declare that they have no competing interests.
Conclusion The present study identified several clinical and pathologic characteristics predictive of tolerance and response to NAC in patients with muscle-invasive urothelial cell carcinoma. Patients who received alternative NAC regimens, and those who poorly tolerated their NAC, were more likely to exhibit pathologic progression. ECOG performance status, cardiovascular comorbidities, and nutritional status are associated with pathologic response and an ability to tolerate NAC to completion. Consideration of these patient-specific factors might allow for more appropriate treatment selection and reduce the delay of definitive surgical treatment in the substantial subset of patients who cannot tolerate or respond to NAC. This demonstrates the importance of multidisciplinary, centralized care for these patients. In the absence of novel or better tolerated chemotherapeutic agents, ongoing efforts should be focused on tailoring sequencing of therapy to individual patients to improve the clinical outcomes of those with muscle-invasive bladder cancer.
Clinical Practice Points NAC for muscle-invasive urothelial carcinoma of the bladder
results in a complete pathologic response in 23% to 38% of patients. The survival advantage for patients with a pathologic response to cisplatin-based NAC compared with those with pathologic progression is well-documented.
References 1. Grossman HB, Natale RB, Tangen CM, et al. Neoadjuvant chemotherapy plus cystectomy compared with cystectomy alone for locally advanced bladder cancer. N Engl J Med 2003; 349:859-66. 2. International Collaboration of Trialists; Medical Research Council Advanced Bladder Cancer Working Party (now the National Cancer Research Institute Bladder Cancer Clinical Studies Group); European Organisation for Research and Treatment of Cancer Genito-Urinary Tract Cancer Group; Australian Bladder Cancer Study Group; National Cancer Institute of Canada Clinical Trials Group; Finnbladder; Norwegian Bladder Cancer Study Group; Club Urologico Espanol de Tratamiento Oncologico Group, Griffiths G, Hall R, Sylvester R, et al. International phase III trial assessing neoadjuvant cisplatin, methotrexate, and vinblastine chemotherapy for muscle-invasive bladder cancer: long-term results of the BA06 30894 trial. J Clin Oncol 2011; 29:2171-7. 3. Tan WS, Lamb BW, Kelly JD. Complications of radical cystectomy and orthotopic reconstruction. Adv Urol 2015; 2015:323157. 4. Chappidi MR, Kates M, Brant A, et al. Assessing cancer progression and stable disease after neoadjuvant chemotherapy for organ-confined muscle-invasive bladder cancer. Urology 2017; 102:148-58. 5. Dogliotti L, Cartenì G, Siena S, et al. Gemcitabine plus cisplatin versus gemcitabine plus carboplatin as first-line chemotherapy in advanced transitional cell carcinoma of the urothelium: results of a randomized phase 2 trial. Eur Urol 2007; 52:134-41. 6. David KA, Milowsky MI, Ritchey J, Carroll PR, Nanus DM. Low incidence of perioperative chemotherapy for stage III bladder cancer 1998 to 2003: a report from the National Cancer Data Base. J Urol 2007; 178:451-4. 7. Pokuri VK, Syed JR, Yang Z, et al. Predictors of complete pathologic response (pT0) to neoadjuvant chemotherapy in muscle-invasive bladder carcinoma. Clin Genitourin Cancer 2016; 14:e59-65. 8. Gandhi NM, Baras A, Munari E, et al. Gemcitabine and cisplatin neoadjuvant chemotherapy for muscle-invasive urothelial carcinoma: predicting response and assessing outcomes. Urol Oncol 2015; 33:204.e1-7. 9. Palapattu GS, Shariat SF, Karakiewicz PI, et al. Bladder Cancer Research Consortium. Cancer specific outcomes in patients with pT0 disease following radical cystectomy. J Urol 2006; 175:1645-9. 10. Volkmer BG, Kuefer R, Bartsch G Jr, et al. Effect of a pT0 cystectomy specimen without neoadjuvant therapy on survival. Cancer 2005; 104:2384-91.
Clinical Genitourinary Cancer Month 2018
-7
Optimal Patient Selection for NAC 11. Kassouf W, Spiess PE, Brown GA, et al. P0 stage at radical cystectomy for bladder cancer is associated with improved outcome independent of traditional clinical risk factors. Eur Urol 2007; 52:769-74. 12. Green DA, Rink M, Hansen J, et al. Accurate preoperative prediction of non-organ confined bladder urothelial carcinoma at cystectomy. BJU Int 2013; 111:404-11. 13. Culp SH, Dickstein RJ, Grossman HB, et al. Refining patient selection for neoadjuvant chemotherapy before radical cystectomy. J Urol 2014; 191:40-7. 14. Vetterlein MW, Wankowicz SAM, Seisen T, et al. Neoadjuvant chemotherapy prior to radical cystectomy for muscle-invasive bladder cancer with variant histology. Cancer 2017; 123:4346-55. 15. Mossanen M, Lee F, Cheng H, et al. Nonresponse to neoadjuvant chemotherapy for muscle-invasive urothelial cell carcinoma of the bladder. Clin Genitourin Cancer 2014; 12:210-3. 16. Galsky MD, Chen GJ, Oh WK, et al. Comparative effectiveness of cisplatin-based and carboplatin-based chemotherapy for treatment of advanced urothelial carcinoma. Ann Oncol 2012; 23:406-10. 17. Stimson CJ, Cookson MS, Barocas DA, et al. Preoperative hydronephrosis predicts extravesical and node positive disease in patients undergoing cystectomy for bladder cancer. J Urol 2010; 183:1732-7. 18. Mertens LS, Meijer RP, Meinhardt W, et al. Occult lymph node metastases in patients with carcinoma invading bladder muscle: incidence after neoadjuvant chemotherapy and cystectomy vs after cystectomy alone. BJU Int 2014; 114:67-74.
8
-
Clinical Genitourinary Cancer Month 2018
19. Nielsen ME, Shariat SF, Karakiewicz PI, et al. Bladder Cancer Research Consortium (BCRC). Advanced age is associated with poorer bladder cancerspecific survival in patients treated with radical cystectomy. Eur Urol 2007; 51:699-706. 20. Afshar M, Goodfellow H, Jackson-Spence F, et al. Centralisation of radical cystectomies for bladder cancer in England, a decade on from the “Improving Outcomes Guidance”: the case for super centralisation. BJU Int 2018; 121: 217-24. 21. Shariat SF, Palapattu GS, Karakiewicz PI, et al. Discrepancy between clinical and pathologic stage: impact on prognosis after radical cystectomy. Eur Urol 2007; 51: 137-49. 22. Turker P, Bostrom PJ, Wroclawski ML, et al. Upstaging of urothelial cancer at the time of radical cystectomy: factors associated with upstaging and its effect on outcome. BJU Int 2012; 110:804-11. 23. Bhindi B, Frank I, Mason RJ, et al. Oncologic outcomes for patients with residual cancer at cystectomy following neoadjuvant chemotherapy: a pathologic stagematched analysis. Eur Urol 2017; 17:30406-12. 24. Plimack ER, Dunbrack RL, Brennan TA, et al. Defects in DNA repair genes predict response to neoadjuvant cisplatin-based chemotherapy in muscle-invasive bladder cancer. Eur Urol 2015; 68:959-67. 25. Choi W, Porten S, Kim S, et al. Identification of distinct basal and luminal subtypes of muscle-invasive bladder cancer with different sensitivities to frontline chemotherapy. Cancer Cell 2014; 25:152-65.