MP54-01: Discordance Between Clinical and Pathologic Carcinoma In Situ for TURBT and Radical Cystectomy in Patients With Non-Muscle Invasive Bladder Cancer: Implications for BCG Unresponsive Clinical Trial Design and Interpretation

Manuel R. de Jesus Escano*, Carissa Chu, Song Jiang, Wesley Yip, New York, NY, Nima Almassi, Cleveland, OH, Peter Reisz, Andrew Lenis, Nicole Benfante, Eugene Cha, Timothy F Donahue, Guido Dalbagni, Alvin Goh, Bernard Bochner, Eugene Pietzak, New York, NY

Introduction: The FDA recommends non-muscle invasive bladder cancer (NMIBC) trials stratify patients into two groups: carcinoma in situ (CIS) with or without papillary disease (CIS±Ta/T1) and papillary only disease (Ta/T1). Complete response rate is used as indicator of drug efficacy for approval in Phase II BCG unresponsive trials, yet limited data supports this classification. Moreover, effects of TURBT on response rate is not well stablished in NMIBC trials. To address this, we analyzed the concordance of CIS between TURBT and corresponding radical cystectomy (RC) specimens in NMIBC patients.

Methods: Patients treated with immediate RC at our center for high grade NMIBC between 2005-2015. Pre-RC TURBT/EUA and RC were performed at our center for study inclusion. Enhanced cystoscopy techniques during TURBTs were recorded. BCG unresponsive was defined using FDA criteria. BCG relapsing was defined as prior BCG treatment, but not meeting BCG unresponsive criteria. Clinical CIS (cCIS) was defined as histologic CIS in TURBT specimen, and pathologic CIS (pCIS) as histologic CIS in RC specimen.

Results: Of 610, 302 (49%) were BCG naïve and 308 (51%) had prior BCG. 358 (59%) had positive cCIS, 419 (69%) had positive pCIS, and 486 (80%) had any CIS (either +cCIS or +pCIS). In those with +cCIS, 289 (81%) had residual pathologic CIS (+cCIS/+pCIS), while eradication of CIS (+cCIS/-pCIS) was seen in 69 (19%). For those without cCIS at TURBT, concordance with RC (-cCIS/-pCIS) was seen in 124 (49%), while occult pathologic CIS (-cCIS/+pCIS) was detected in 128 (51%). In those BCG unresponsive (n=99), eradication of CIS by TURBT was seen in 16% (8/51) who would have been included in CIS±Ta/T1 cohorts, while occult pathologic CIS was seen in 52% (25/48). A similar frequency of CIS eradication (18%) and occult pathologic CIS (46%) was seen in the BCG relapsing subgroup.

Conclusions: We report the potential utility of CIS as a marker of residual disease. Based on our findings, >15% of NMIBC patients in single-arm FDA registration trials might experience complete responses from the effects of TURBT alone, raising concerns about the interpretation of these trials. Furthermore, occult pathologic CIS was found in half of NMIBC patients who would incorrectly be placed into the papillary Ta/T1 only cohort of current NMIBC clinical trial stratification schema. Better measures of residual CIS after TURBT are needed to improve patient stratification and drug efficacy evaluations in NMIBC trials.

Source of Funding: SKC for Prostate and Urologic Cancers, NIH/NCI to MSKCC through CCSG award #P30 CA008748, MJ and HRK CMO, MSKCC SPORE in BC P50-CA221745, MSKCC SPORE CE A., MSKCC Dept. of Surgery FR A., Bochner-Fleisher Scholars in BC Award, NIH/NCATS #UL1-TR-002384, NIH/NCI K12 Paul Calabresi CD A. for C Onc (K12 CA184746), Wofchuck Family A.