Min Sun, MD, Ph.D., Oncologist, and Hematologist at UPMC Hillman Cancer Center, Clinical Assistant Professor at the University of Pittsburgh School of Medicine. In this video, he speaks about the ASCO Abstract - Testing of a machine learning (ML) model for ability to predict oxaliplatin and bevacizumab (bev) benefit in NRG Oncology/NSABP C-07 and C-08.
• C-07 was a Phase III clinical trial1 that proved the efficacy of combining oxaliplatin (Oxi) with fluorouracil plus leucovorin (FULV) in adjuvant therapy for patients with stage II and III colon cancer (CC).
• Oxi causes neuropathy in 90% of patients; de-escalation will improve patient treatment and quality of life.
• C-08 was a phase III trial2 that failed to demonstrate the efficacy of adding bevacizumab (Bev) to FOLFOX in the CC adjuvant context.
• An ML signature was established to predict responsiveness to Oxi-containing regimens by mining TCGA and public transcriptomes of CRC patients.
• Whole transcriptome data (DASL arrays) from C-07 and C-08 (N=1,284) were used in a double-blind method to test the ML signature.
• The FLOX arm from C-07 and the FOLFOX arm from OC-08 were combined to form a FLOX/FOLFOX group (N=644).
• Points were divided into two groups: Sig+ and Sig-. The cutoffs used to categorize Oxi and Bev benefit tests were different. The primary endpoint was recurrence-free survival.
• In FLOX/FOLFOX-trtd pts, the ML signature predicts the effectiveness of Oxi.
• Although the signature is connected with the Bev benefit, its prediction power is insignificant.
• Sig- detects patients who had the best outcomes when treated with FULV alone but did not benefit from Oxi.
De-escalation (no Oxi) will protect them from the negative consequences of Oxi.
• Because the Sig+ group benefits from Oxi, FOLFOX is the best option.
• The Sig+ group may benefit from adding Bev to FOLFOX, with a significant increase in result.