Sarcopenia Weakens the Benefit of Bacillus Calmette‐Guérin in Non‐Muscle‐Invasive Bladder Cancer: A Computed Tomography‐Based Cohort Study

OBJECTIVES: Non-muscle-invasive bladder cancer (NMIBC) commonly recurs despite intravesical Bacillus Calmette-Guérin (BCG) therapy. Host factors that modify the effectiveness of BCG have not been completely defined. We evaluated whether sarcopenia is associated with outcomes after BCG in Japanese patients. METHODS: We conducted a retrospective single-center cohort study (2010-2023) including high-grade (HG) NMIBC patients treated with BCG (n = 139) and those not receiving BCG (n = 100). Sarcopenia was defined a priori using sex- and BMI-specific skeletal muscle index (SMI) and psoas muscle index (PMI) cutoffs derived from baseline L3 computed tomography and validated in Asian adults. The endpoints were intravesical recurrence-free survival (IVRFS), high-grade recurrence-free survival (HGRFS), progression-free survival (PFS), and overall survival (OS). Feature selection used LASSO-Cox with cross-validation; performance was assessed using time-dependent ROC. Within the BCG cohort, we conducted a 1:1 propensity score matching. RESULTS: In the BCG cohort, 93/139 (66.9%) and 69/139 (49.6%) patients were classified as sarcopenic according to SMI and PMI, respectively. One- and two-year apparent AUCs for intravesical recurrence were 0.75 and 0.79 (validation AUCs 0.68 and 0.71). Kaplan-Meier analyses revealed no differences in outcomes by sarcopenia in the non-BCG cohort but worse outcomes across endpoints in the BCG cohort. After matching, SMI-based sarcopenia was associated with shorter IVRFS, whereas PMI-based sarcopenia was associated with shorter IVRFS, HGRFS, PFS, and OS. CONCLUSIONS: Patients with sarcopenia were less likely to benefit from BCG. The incorporation of computed tomography-based muscle indices into pre-BCG decision-making may improve risk stratification and supportive care.