Introduction

Cisplatin-based neoadjuvant chemotherapy (C-NAC) is recommended for eligible patients with muscle-invasive bladder cancer (MIBC) [1, 2] based on randomized trials [3,4,5] and meta-analyses [6]. In locally advanced disease, chemotherapy optimizes disease control and helps selecting patients for locoregional treatment. Common regimens are ddMVAC (dose-dense methotrexate, vinblastine, doxorubicin, cisplatin) and GC (gemcitabine–cisplatin).

Median age at bladder cancer diagnosis in the USA is 73. However, in pivotal C-NAC trials, median age of participants is 63–64 [3, 5, 7]; less than 20–25% are aged ≥ 70. In routine practice, patients aged ≥ 70 may be offered C-NAC thrice less frequently [8]. Only small retrospective studies have reported outcomes in patients aged ≥ 70 treated with GC or gemcitabine–carboplatin [9, 10]; 3 cycles of neoadjuvant ddMVAC was reported feasible in 43 patients aged ≥ 75 [11].

Here, we report on 56 well-characterized MIBC patients aged ≥ 75 treated with C-NAC. We report applicability, safety, and efficacy of standard-of-care C-NAC in this understudied population.

Patients and methods

Patients and treatment

This single-center retrospective study included patients aged ≥ 75 with cT2-4NanyM0 MIBC who underwent ≥ 1 cycle of ddMVAC (methotrexate 30 mg/m2, vinblastine 3 mg/m2, doxorubicin 30 mg/m2, and cisplatin 70 mg/m2 every 2 weeks for up to 6 cycles) or GC (gemcitabine 1250 mg/m2 on day 1 and day 8 and cisplatin 70 mg/m2 on day 1 every 3 weeks for up to 4 cycles) from 2011 to 2021. The treating oncologist determined C-NAC regimen with on-demand decisional help from oncogeriatricians. Locoregional management was determined in multidisciplinary tumor boards.

Before each cycle, renal function was estimated using Cockroft–Gault formula creatinine clearance (CGF CrCl) and body surface area-indexed estimated glomerular filtration rate (CKD-EPI eGFR). Policy for cisplatin administration was as follows: 70 mg/m2 if CrCl ≥ 60 mL/min; 50 mg/m2 if 50–59 mL/min; 40 mg/m2 if 40–49 mL/min; and end of treatment if < 40 mL/min. In the occurrence of grade ≥ 3 toxicities, a 20% dose reduction was done for all drugs.

Outcomes

Primary outcome was overall survival (OS: time from chemotherapy initiation to death from any cause). Secondary outcomes were deliverance of ≥ 4 cycles, pathological downstaging, and safety. Adverse events (AEs) were described using CTCAE v5.0. Subgroup analyses focused on baseline characteristics including age-adjusted Charlson comorbidity index (aaCCI), polypharmacy (≥ 5 medications daily), and nutritional status. Undernutrition was defined as BMI < 21 kg/m2 and/or significant weight loss (> 5% in 1 month and/or > 10% in 6 months). Skeletal muscle index at third lumbar vertebra level (L3SMI) was assessed on pre-chemotherapy CT scans; previously reported thresholds were used to define sarcopenia (< 41 cm2/m2 for women, < 43 cm2/m2 for non-overweight men, < 53 cm2/m2 for overweight/obese men) [13].

Statistical analysis

Continuous variables were described using median (range) and dichotomous variable using absolute number and/or proportions. Survival was described using Kaplan–Meier estimates and compared using log rank test. Association between clinical factors and OS was explored using univariate and multivariate Cox regression models. Dichotomous variables were compared using Fisher’s exact test.

Results

Patients

Fifty-six patients were included Table 1 and Supplementary Table 1. Median age was 79. Most patients had good performance status and low aaCCI. Eight patients had BMI < 21 kg/m2; 6 had serum albumin level < 35 g/L. In contrast, 42 of 52 evaluable patients had CT-scan sarcopenia.

Table 1 Baseline characteristics of patients
Full size table

Treatments

Supplementary Table 2 summarizes C-NAC and locoregional management.

Thirty-one patients received ddMVAC; 15 received GC. GC was more frequently used with increasing age (p value: 0.095 by Student’s t test; Supplementary Table 1). Dose reduction was performed in 11% of patients at first cycle (gemcitabine dose reduction: 3; doxorubicin dose reduction: 1; cisplatin dose reduction: 2 ddMVAC patients with impaired CGF CrCl) and in 54% at any time. Median number of cycles was 4; 73% of patients underwent ≥ 4 cycles. Premature C-NAC discontinuation (before 6 ddMVAC cycles or 4 GC cycles) occurred in 66% of patients, due to toxicity (28 patients), locoregional treatment requirements (6), intercurrent infection (1), or patient’s death (1). Toxicities prompting discontinuation (possibly concomitant) included renal (11 patients), hematological (10), general/asthenia (10), and digestive (6).

Forty-three patients (77%) underwent curative-intent locoregional treatment (radical cystectomy: 54%; chemoradiotherapy: 23%). Median time from last C-NAC cycle to cystectomy was 49 days (range 32–254).

Five patients (9%) underwent active endoscopic surveillance. Lack of local management in the remaining 8 patients was due to intercurrent affection (lymphoma: 1; SARS-CoV-2 infection: 2), patient’s choice (2), disease progression (2), or death (1).

Safety

Table 2 summarizes AEs from C-NAC. All patients experienced AEs from chemotherapy; 55% experienced grade 3–4 AEs (ddMVAC: 66%; GC: 27%), most often hematological toxicity from ddMVAC. Red blood cell transfusion was performed in 29% of patients and platelet transfusion in 7%. Febrile neutropenia rate was 7%. One patient died shortly after the first C-NAC cycle (cause unknown). Immediate surgical outcomes are described in Supplementary Table 2; 1 patient died from cystectomy complications.

Table 2 Adverse events from chemotherapy
Full size table

Downstaging

Seventeen of 30 cystectomy patients (57%) had downstaging to ypT < 2N0/x (pT0N0: 14; ypTisN0: 1; ypT1N0: 1; ypT0Nx: 1).

Seven of 15 patients (47%) who underwent post-chemotherapy TURBT and no cystectomy had no residual disease (ycT0) and 2 patients (13%) had non-muscle-invasive residual disease (ycT1).

Survival

Median follow-up from chemotherapy initiation was 36 months (47 months in patients alive at last follow-up).

Twenty-one patients had died; causes were disease progression (14), treatment related (during C-NAC: 1; postoperative: 1), intercurrent disease (3), and unknown (2).

Three-year OS was 65%; 5-year OS was 60%; median OS was not reached (Fig. 1A). OS was non-significantly better with ddMVAC than with GC; we found no association with locoregional management (data not shown).

Fig. 1
figure 1

Overall survival. A Overall survival in all patients. B Overall survival by age-adjusted Charlson comorbidity index (aaCCI). HR 4.45 (95% CI 1.61–12.30), p = 0.0002 (univariate). C Overall survival by polypharmacy status. HR 4.87 (95% CI 1.72–13.80), p = 0.0001 (univariate)

Full size image

Impact of baseline parameters

Polypharmacy and aaCCI ≥ 8 were non-significantly associated with decreased chemotherapy feasibility and increased toxicity (Table 3).

Table 3 Univariate analysis of secondary outcomes according to age and comorbidity-related parameters
Full size table

In univariate analysis, OS was significantly worse in patients with polypharmacy or aaCCI ≥ 8 (Fig. 1B, C, Table 4). In a multivariate analysis of 56 patients, polypharmacy retained a significant prognostic value (Table 4).

Table 4 Univariate and multivariate analyses of overall survival
Full size table

Undernutrition was associated with worse OS in 43 evaluable patients in univariate analysis (Supplementary Fig. 1). Incomplete data and lesser reliability of retrospective assessment precluded integration in the multivariate analysis.

Discussion

In this study, 73% of MIBC patients aged ≥ 75 completed 4 C-NAC cycles and 77% underwent cystectomy or chemoradiotherapy. Nine patients underwent 5 ddMVAC cycles, and 9 other 6 cycles. Grade 3–4 toxicity (including febrile neutropenia), downstaging, and OS compared aptly with those observed in patients aged 15 years younger in median in prospective trials [7, 12]. Interestingly, downstaging in cystectomy patients (ypT < 2N0: 57%; ypT0N0: 47%) was higher than in the recent study with 3 ddMVAC cycles (24%) [11], suggesting a benefit of longer-course ddMVAC.

Clinicians may systematically refrain from offering C-NAC to elderly patients because of perceived frailty. Although age alone is not a criterion, cisplatin eligibility decreases with age solely because of inadequate renal function defined by CGF CrCl < 60 mL/min [13]. Twenty-two patients in our study conformed to this definition: only 5 of them (23%) prematurely discontinued chemotherapy for renal toxicity. Comparatively, this was the case for 6 of 34 patients with initial CGF CrCl ≥ 60 mL/min (18%). CGF CrCl accuracy decreases with age, and guidelines endorse CKD-EPI eGFR to accurately estimate renal function in elderly patients [14]. Only 6 of our patients had CKD-EPI eGFR < 60 mL/min, still with no more frequent discontinuation for renal toxicity. Data from VESPER showed that CGF CrCl ≥ 50 mL/min was sufficient to safely administer C-NAC [15]. To mirror this, we postulate that CKD-EPI eGFR ≥ 50 mL/min may be a reasonable cutoff for cisplatin eligibility in elderly MIBC patients. External validation focusing on long-term renal outcomes is warranted.

Polypharmacy and aaCCI were significantly associated with OS and non-significantly with C-NAC feasibility and toxicity. This emphasizes the need for careful pretreatment evaluation of elderly MIBC patients with comorbidities. Several studies described the prognostic value of aaCCI in cancer: for instance, aaCCI is associated with mortality in non-metastatic patients aged ≥ 70 with diverse primaries [16], or with lower completion of adjuvant chemotherapy for pancreatic cancer [17]. Our observations echo these findings.

Seventeen patients underwent a complete prospective geriatric assessment. Median age was 79 (76–90). Thirteen had ECOG PS 0–1. Twelve had positive polypharmacy status. Median aaCCI was 7 (5–9). Based on a comprehensive assessment [18], 8 patients had undernutrition (moderate: 6; severe: 2). Six of 16 evaluable patients had ≥ 1 functional criterion for impaired muscle function (history of falls, unheld monopodal support, timed get-up-and-go test > 20 s, walking speed < 1 m/s); all 6 had CT-scan sarcopenia. Most had fully preserved autonomy (ADL 6/6: 14; IADL 4/4: 11). This prospectively characterized subset of our study population illustrates its fitness, especially when considering the bias in referring apparently frailer patients for geriatric assessment.

Despite being generally fit, 81% of our elderly patients had CT-scan sarcopenia. Interestingly, in a study of 146 MIBC patients, prevalence of similarly-defined sarcopenia was 45.9% [19] and mean age in sarcopenic patient was 73, as opposed to 66 in non-sarcopenic patients. Thus, conventional thresholds defining CT-scan sarcopenia seem inadequate for elderly patients. Functional assessment is warranted.

Few patients died from intercurrent illnesses, illustrating that the course of cancer defines prognosis in elderly MIBC patients, with negligible competitive mortality. Pathological response to C-NAC retained major prognostic value. Indeed, the life expectancy for a cancer-free 80-year old in France is currently 9–11 years (depending on gender), justifying all feasible curative-intent treatments in eligible elderly MIBC patients, even octogenarians.

Our study has limitations, first its retrospective nature. Outcomes such as number of C-NAC cycles, downstaging and OS are little impacted by memory bias, but safety data should be considered with more caution. Few patients were lost to follow-up. Findings regarding prognostic value of aaCCI and polypharmacy are hypothesis-generating and should be prospectively assessed in a validation cohort. Lack of systematic complete nutritional assessment including muscle strength is a limitation, given the high prevalence and dramatic prognostic value of undernutrition in elderly cancer patients [20, 21]. Treatment selection bias in the decision to administer C-NAC is also a concern; unfortunately we are not able to describe the selection processes that took place at both our institution and referring centers that led to exclusion of patients from C-NAC.

Still, although the number of patients remains small, this is the most-documented series to date of MIBC patients aged ≥ 75 treated with C-NAC. Our patients were probably more fit than the whole of elderly MIBC patients, so we recommend that clinicians remain attentive to baseline general health status and comorbidities before offering C-NAC, and refer them to oncogeriatricians for pretreatment evaluation.

In conclusion, C-NAC (including ddMVAC) is feasible and effective in selected MIBC patients aged ≥ 75, who should not be systematically barred from standard-of-care chemotherapy based on age alone. Evaluation of comorbidities and geriatric parameters as well as of tumor characteristic and cisplatin eligibility criteria informs patient selection.