P-106. Exploration of a Desirability of Outcome Ranking (DOOR) Endpoint for Invasive Aspergillosis Using Two Registrational Trials

TL;DR

This study explores a new way to evaluate treatment outcomes for invasive aspergillosis by combining safety and efficacy into a single ranking system.

Contribution

The paper introduces a syndrome-specific DOOR endpoint for invasive aspergillosis using real clinical trial data.

Findings

Overall DOOR probabilities showed no significant differences between treatment arms in the two trials.

D-TRAE occurrences suggested a possible advantage for isavuconazole in one trial.

Sensitivity analyses confirmed no differences in treatment outcomes when using alternative definitions of clinical failure.

Abstract

By combining efficacy and safety outcomes into a single ordinal endpoint, desirability of outcome ranking (DOOR) may facilitate a holistic examination of patients’ experiences in clinical trials. We derived a syndrome-specific DOOR endpoint for invasive aspergillosis (IA) utilizing data from two registrational phase 3 clinical trials submitted to the FDA; Trial 1 compared isavuconazole to voriconazole, while Trial 2 compared posaconazole to voriconazole. Treatment duration in both trials was 6 weeks with possible extension to Day 84 [end of treatment (EOT)]; both utilized a primary endpoint of mortality at Day 42 [test of cure (TOC)]. A Data Review Committee (DRC) categorized clinical response at EOT in each trial as success (complete or partial) or failure (stable or progression).

Death (entire study period), absence of clinical response (ACR; death by Day 42), infectious complications (IC), serious adverse events (SAE) and discontinuations due to treatment-related adverse events (D-TRAE) were included as components in the DOOR endpoint with a 5-level ordinal scale. We applied the resultant IA-specific DOOR to patient-level data from the two trials to estimate the probability that a patient in one treatment arm would have a more desirable outcome than a patient in the other at both TOC and EOT. Sensitivity analyses were done with ACR encompassing (1) DRC-adjudicated clinical progression or (2) DRC-adjudicated clinical progression and stability rather than mortality.

There were no statistically significant differences in overall DOOR probabilities between the treatment arms for Trials 1 and 2 at TOC and EOT. The DOOR probability for D-TRAE nominally significantly favored the isavuconazole arm at EOT in Trial 1. No differences between treatment arms were seen in the sensitivity analyses when DRC-adjudicated definitions of clinical failure were used as ACR.

Though overall IA-specific DOOR probabilities between treatment arms in the two trials were similar, our work highlighted possible differences in the occurrences of D-TRAE. These results warrant further investigation and underscore the importance of key clinical data collection and consideration of improvements in the electronic case report form used in IA trials.

Vance G. Fowler, MD, MHS, Affinergy, Janssen, Contrafect: Advisor/Consultant|AstraZeneca; EDE; Basilea: Grant/Research Support|Debiopharm, GSK; Affinium, Basilea,: Advisor/Consultant|Destiny, Amphliphi, Armata, Akagera: Advisor/Consultant|Merck; Contrafect; Karius; Janssen: Grant/Research Support|UpToDate: Royalties|Valanbio: Stock options