Acknowledging that studies by @MaxKates & others demonstrate BCG retreatment is effective in many selected patients with prior BCG exposure, I just don’t see how one can overcome the inherent selection bias in these retrospective studies to feel confident enough not to explore a subgroup analysis for RCTs enrolling such patients. And even if BCG monotherapy outcomes were comparable, that would not establish that the incremental benefit of adding PD-1/PD-L1 blockade is the same in previously treated and truly BCG-naïve patients. I would argue that in the absence of robust data demonstrating comparable outcomes after BCG retreatment versus first-ever BCG, prior BCG exposure should at minimum be a prespecified subgroup in randomized trials using BCG as the control arm. And if it wasn’t prespecified, the post-hoc analysis is still clinically important. POTOMAC and CREST should report it! • How many patients had prior BCG? • What was the HR in truly BCG-naïve subset? • What was the HR in remotely BCG-exposed patients? 5/n

The justification for including patients >2–3 years from prior BCG within “BCG-naïve” trials appears to come partly from this @IBCG_BladderCAposition paper, which states that clinical experience suggests their response to BCG retreatment is similar to that of truly BCG-naïve patients While this hasn’t really been my own clinical experience (especially for patients previously treated with maintenance BCG), anecdotal clinical experience is no substitute for objective comparative data, which is precisely why I think reporting on these subgroup analyses is so important. And regardless for a phase III trial with EFS/DFS as the endpoint, what matters is not simply initial response rates, it is the durability of disease control. 4/n @UroDocAsh , @UrogerliMD , @MaxKates ; @pcvblack @pjhensley11, @joanfundi

This effort lead by @DrAndrewKatims , former @UrologyMSK @SUO_YUO fellow now at @nymedcollege , first looked at 610 NMIBC patients treated with radical cystectomy without any intervening treatment after their pre-RC TURBT; across multiple NMIBC subsets (BCG-Naïve, BCG unresponsive, use of Narrow Band Imaging, etc) we found that ~20% of patients with CIS on pre-RC TURBT had no residual CIS within their cystectomy specimen & equally important was that ~50% without CIS on pre-RC TURBT had occult CIS within their cystectomy specimen. We then partnered with @MaxKates & @Sadra_Sepehri to look at a NMIBC cohort who had blue light TURBT before cystectomy without intervening treatment @HopkinsGBCI @brady_urology and found nearly identical results; ~20% eradication of CIS by TURBT alone & ~50% occult CIS on cystectomy n/3