Association of accelerated approval (AA) status with National Comprehensive Cancer Network (NCCN) guideline recommendation for cancer drug indications: A cross-sectional study.

Abstract

1599 Background: Many oncology drugs are approved via FDA’s accelerated approval pathway based on trials showing changes to surrogate measures only ‘reasonably likely’ to predict clinical benefit. Despite this preliminary evidence, accelerated approval drugs are incorporated into clinical practice and included in guidelines. We assessed NCCN evidence and preference ratings for drugs with accelerated vs traditional approval. Methods: Accelerated approval oncology drugs were identified using an FDA list as of June 30 2022. Drug labelling revealed other oncology indications of these drugs. Accelerated approvals were stratified if they had been converted to traditional approval after confirmatory studies or withdrawn. Indications with no NCCN guidelines or tumor-agnostic approvals were excluded. Relevant NCCN guidelines were reviewed as of Oct 19 2022. Each indication was assessed for ratings of level of evidence and treatment preference. Levels of evidence included category 1 (consensus, high-level evidence), 2A (consensus, lower-level evidence), 2B (lower-level evidence without consensus), or 3 (major disagreement). Treatment preference rankings included preferred, alternative preferred, other recommended, or useful in certain circumstances. Results: Among 315 indications for 100 drugs, 156 (50%) had traditional approval and 159 (50%) had current or previous accelerated approval. Among accelerated approval indications, 78 (49%) had been converted to traditional approval and 21 (13%) withdrawn. Among all indications, 105 (33%) were rated as category 1 evidence, 185 (59%) as 2A, 6 (2%) as 2B, and 2 (0.6%) as 3. The treatment preference for 153 (49%) was preferred, for 55 (17%) was recommended, for 5 (2%) was alternative preferred, and for 38 (12%) was useful in certain circumstances. Compared to traditional approval, accelerated approval indications less often had category 1 evidence (3% v 47%), and were less often listed as preferred (40% v 58%). Converted accelerated approval indications had mid-range ratings for evidence (38% category 1) and treatment preference (47% preferred). 11 withdrawn accelerated approval indications remained in NCCN guidelines. Conclusions: In most cases, NCCN guidelines identified accelerated approval indications as having lower evidence than traditionally approved cancer drugs. Many traditional approvals and conversions from accelerated approvals had only category 2A evidence, suggesting better evidence is needed for traditional approval of oncology drugs. But in about half of accelerated approval indications withdrawn by the drug’s manufacturer, NCCN continued recommending the drug. Professional organizations like NCCN should ensure their guidelines reflect that accelerated approval drugs are being marketed based on preliminary evidence.