Pretreatment factors informing treatment selection between perioperative enfortumab vedotin plus pembrolizumab and neoadjuvant gemcitabine-cisplatin in cisplatin-eligible, resectable muscle-invasive bladder cancer: protocol for a systematic review of comparative effect modification and treatment-related toxicity

Protocol for a systematic review, registered with PROSPERO (Centre for Reviews and Dissemination, University of York). Registration number to be added to this record once assigned. AUTHORITATIVE VERSION. Version 1.9 (29 July 2026) comprises the full protocol version 1.7 (26 July 2026) as amended by the version 1.8 amendment and correction note (28 July 2026) and by the version 1.9 second amendment and correction note (29 July 2026). The three documents are deposited together and are cited together. Where they differ, the most recent document governs: the version 1.9 note governs over the version 1.8 note, which governs over the version 1.7 document. Version 1.9 supersedes all earlier drafts. CONTENTS OF THIS RECORD. Full protocol version 1.7 (26 July 2026) Amendment and correction note forming version 1.8 (28 July 2026) Second amendment and correction note forming version 1.9 (29 July 2026) Combined single-file document (cover page, version 1.7, and both amendment notes) as submitted to PROSPERO Completed PRISMA-P 2015 checklist Erratum recording typesetting defects in the version 1.7 document Plain-text electronic search strategies for MEDLINE via PubMed, Embase.com, CENTRAL, and Web of Science Core Collection, incorporating the amendments in version 1.8 note section 2.9 Archive of pre-registration development versions, with manifest and checksums REVIEW SCOPE. Among cisplatin-eligible adults with resectable muscle-invasive urothelial carcinoma of the bladder, which pretreatment characteristics modify the comparative effects of perioperative enfortumab vedotin plus pembrolizumab versus neoadjuvant gemcitabine plus cisplatin on event-free survival and on the prevalence of clinically significant peripheral neuropathy assessed at a landmark common to both randomized groups. Structured narrative synthesis (SWiM) is the default; pairwise meta-analysis is conditional on sufficient comparability. No network meta-analysis and no treatment ranking are planned. REGISTRATION STATUS. This is a results-known registration. The principal KEYNOTE-B15/EV-304 results entered the public domain on 17 December 2025 (sponsor topline release) and 26-28 February 2026 (ASCO Genitourinary Cancers Symposium, abstract LBA630), before protocol development commenced on 20 July 2026. The full peer-reviewed report was published online on 22 July 2026 (N Engl J Med 2026;395:338-348; doi:10.1056/NEJMoa2601486) and was available to the review team before the primary comparative harm estimand was specified in version 1.4, dated 26 July 2026. The registration is prospectively documented but not results-blind. Section 13.1 of the protocol states the full timeline and the three prespecified safeguards that follow, as extended by section 3 of each amendment note. SCOPE OF THE VERSION 1.9 AMENDMENT. The version 1.9 note amends the governance provisions only. It applies the conflict-of-interest exclusion rule already contained in section 14.3 of version 1.7 to the identity of the adjudicator, reconstitutes the review team accordingly, and completes the record of declared interests. No change is made to the review question, population, interventions or comparators, eligibility criteria, outcomes and their attribution rule, time origin, landmark, presentation hierarchy and denominator, search strategies, synthesis plan, or the decision not to conduct a network meta-analysis. No provision amended by the version 1.9 note was amended by the version 1.8 note. REVIEW STATUS AT DEPOSIT. No formal database searching, eligibility screening, standardized data extraction, risk-of-bias assessment, ICEMAN appraisal, or synthesis had begun.