EURURO Vol. 72 No. 5

661 860

While it has been found that neoadjuvant chemotherapy
(NAC) before radical cystectomy (RC) improves survival for
muscle-invasive urothelial carcinoma of the bladder (UCB)
compared to RC alone , this survival benefit has been
primarily attributed to those patients without residual
disease after NAC (ypT0N0) . Unfortunately, however,
only 20–38% of patients achieve such a complete pathologic
response At the same time, limited data exist regarding
the prognosis for patients with residual UCB (rUCB) at RC
after chemotherapy.
Our objective was to characterize the outcomes for
patients with rUCB at RC following NAC and to determine
whether survival differs from pathologic stage–matched
controls treated with RC alone.
Following institutional review board approval, patients
who underwent NAC followed by RC for cT2–4N0M0 UCB
between 1980 and 2016 were identified using the Mayo
Clinic cystectomy registry. Non-urothelial histology, prior
UCB-directed radiotherapy, palliative cystectomy, concom-
itant nephroureterectomy or urethrectomy, and missing
match variables were exclusion criteria.
Patients who received NAC followed by RC for UCB were
matched with up to two control subjects as available (1:2,
n = 144; 1:1, n = 36) from patients who underwent RC alone,
with matching based on pathologic T and N stage (American
Joint Committee on Cancer TNM classification, 7th edition),
surgical margin status, and year of RC (Supplementary
Fig. 1). The final matched cohort included 180 patients
who underwent NAC + RC matched to 324 patients who
underwent RC alone (total n = 504).
With survival time measured from the date of RC, non-
urothelial recurrence-free survival (RFS; including local
recurrence and distant metastasis), cancer-specific survival
(CSS), and overall survival (OS) were estimated using the
Kaplan-Meier method and compared using the log-rank test.
Multivariable Cox regression analyses adjusted for age,
Charlson comorbidity index, and receipt of adjuvant chemo-
therapy were used to evaluate the associations between NAC
and outcomes. Analyses were stratified by the absence or
presence of rUCB at RC, with analysis of the subset without
rUCB at RC serving to provide confirmation of the validity of
the analytic approach. A 90-d conditional landmark analysis
was used for models including adjuvant chemotherapy.
Supplementary Table 1 summarizes the baseline char-
acteristics. The majority of patients in the NAC group
received cisplatin-based (83%) or carboplatin-based (9%)
combination chemotherapy (Supplementary Table 2). Me-
dian follow-up among survivors was 2.6 yr (interquartile
range 1.3–4.7), during which time 146 patients experienced
disease recurrence and 268 died, with 166 dying from UCB
(Supplementary Table 3).
For patients found to have no residual tumor at surgery
(ypT0N0 and pT0N0), there were no significant differences
in subsequent 5-yr RFS (90% vs 94%; p = 1), CSS (82% vs 93%;
p = 0.4), or OS (82% vs 82%; p = 0.5) between the NAC and
control groups ). Likewise, on Cox regression analysis,
the risks of recurrence (hazard ratio [HR] 1.00, 95%
confidence interval [CI] 0.22–4.56; p = 1), cancer-specific
mortality (HR 1.30, 95% CI 0.32–5.28; p = 0.7), and all-cause
mortality (HR 0.87, 95% CI 0.42–1.80; p = 0.7) did not
significantly differ between the NAC and control groups for
patients without residual UCB on RC ).
Among patients with rUCB at RC, those who had been
treated with NAC had significantly worse 5-yr RFS (50% vs
63%; p = 0.01), CSS (40% vs 59%; p = 0.003), and OS (33% vs
48%; p = 0.02) than pathologic stage-matched controls who
underwent RC alone . Moreover, on multivariable
analysis for patients with rUCB, NAC receipt was indepen-
dently associated with worse RFS (HR 1.84, 95% CI 1.28–
2.66; p = 0.001), CSS (HR 1.81, 95% CI 1.30–2.52; p < 0.001),
and OS (HR 1.57, 95% CI 1.18–2.08; p = 0.002; ).
We then further stratified patients by the specific
pathological stage of their residual cancer. On Kaplan-
Meier analysis for patients with <ypT2N0 rUCB at RC after
NAC (ie, those downstaged to organ-confined, non–muscle-
invasive disease) we found no difference when compared to
matched controls for 5-yr RFS (85% vs 82%; p = 0.8), CSS
(71% vs 83%; p = 0.5), or OS (64% vs 71%; p = 0.7;
Supplementary Figure 2). Likewise, on multivariable
analysis, NAC receipt was not associated with RFS (HR
0.95, 95% CI 0.45–1.99; p = 0.9), CSS (HR 1.05, 95% CI 0.51–
2.14; p = 0.9), or OS (HR 1.10, 95% CI 0.93–1.29; p = 0.3)
among these patients (Supplementary Table 4). Conversely,
patients with ypT2–4 or ypN+ rUCB at RC after NAC had
significantly worse 5-yr CSS (28% vs 46%; p = 0.03) and OS
(21% vs 36%; p = 0.04) compared to RC-only controls
Table 1 – Cox regression analyses of the association between receipt of neoadjuvant chemotherapy and oncologic outcomes following RC,
stratified by the presence or absence of rUCB at RC
Outcome rUCB (n = 401) ypT0N0/pT0N0 (n = 103)
HR (95% CI) p value HR (95% CI) p value
Recurrence 1.84 (1.28–2.66) 0.001 1.00 (0.22–4.56) 1
Cancer-specific mortality 1.81 (1.30–2.52) <0.001 1.30 (0.32–5.28) 0.7
All-cause mortality 1.57 (1.18–2.08) 0.002 0.87 (0.42–1.80) 0.7
HR = hazard ratio; CI = confidence interval; rUCB = residual urothelial carcinoma of the bladder; RC = radical cystectomy.
a Multivariable models adjusted for age, Charlson co-morbidity index. Models in the rUCB subgroup additionally adjusted for receipt of adjuvant chemotherapy
and employed a 90-day conditional landmark analysis.
b Includes local and distant non-urothelial recurrences (excludes urethral and upper tract urothelial tumors).
c In the analysis for ypT0N0 and pT0N0 patients, models for recurrence and cancer-specific mortality would be overspecified because of an insufficient number of
event (recurrence n = 8; cancer-specific deaths n = 10). Therefore, multivariable-adjusted estimates for these outcomes are not reported and univariable analyses
are reported instead.
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 6 6 0 – 6 6 4 661