Generalized Pairwise Comparisons Method to Assess Total Severe Oral Mucositis Burden and Interventional Benefit in HNC.

Severe oral mucositis (SOM, WHO Gr 3-4) is a common and debilitating side effect of intensity-modulated radiation therapy (IMRT) and concurrent cisplatin (CRT) in patients with head and neck cancer (HNC). Patient SOM burden is defined by multiple endpoints (incidence, duration, severity and time to onset) and it is critical to consider all in a "holistic" approach rather than a single endpoint. Generalized pairwise comparisons (GPC) statistical method allows combined assessment, while accounting for testing multiplicity. Two placebo (PBO)-controlled trials (phase 2b and phase 3 ROMAN) showed avasopasem manganese 90mg (AVA) significantly decreased individual endpoints of SOM incidence and duration, with nominal decrease in severity (WHO Gr 4 incidence) and delay in SOM onset. We sought to determine combined Net Treatment Benefit (NTB) of AVA vs. placebo (PBO) for both trials across these 4 key clinically relevant endpoints using GPC. Post-hoc meta-analysis of oral cavity and oropharynx patients in phase 2b (AVA = 76, PBO = 74) and ROMAN (AVA = 241, PBO = 166), stratified by cisplatin schedule (QW vs Q3W) and treatment setting (definitive vs post-op). GPC permitted simultaneous analysis of several prioritized outcomes (endpoints), comparing all possible pairs of 1 active (i.e., AVA) patient and 1 control (i.e., PBO) patient on first outcome, assigning Win, Loss or Tie to AVA per pair. Tied pairings moved on to the next outcome. Outcomes were prioritized as: 1) WHO grade 4 OM incidence (Severity), 2) SOM incidence, 3) days of SOM (Duration), 4) days to SOM onset; with 7 days difference defined as the clinical relevance threshold for (3) and (4). NTB = Wins - Losses on each outcome and combined; 1/NTB = Number Needed to Treat (NNT). Each trial was analyzed separately and integrated in a meta-analysis. Number of pairs analyzed: phase 2b = 1601, ROMAN = 13969. AVA showed significant NTB on each key outcome and all contributed meaningfully to combined NTB. Phase 2b GPC showed AVA combined NTB of 0.207 (P =0.025) and ROMAN GPC showed AVA combined NTB of 0.189 (P =0.001). Meta-analysis results (Table) show AVA NTB of 0.194 in favor of AVA (P <0.001), resulting in NNT of 1/0.194 = 5.15 patients. GPC method allows holistic analysis of SOM burden in HNC patients receiving concurrent cisplatin and IMRT. Phase 2b and ROMAN individual GPC analyses and integrated GPC meta-analysis show compelling evidence of AVA clinical benefit as a therapeutic for SOM. Funded by Galera Therapeutics, Inc; ClinicalTrials.gov NCT NCT02508389 & NCT03689712. [ABSTRACT FROM AUTHOR]