Negative hyperselection of patients with RAS wild-type metastatic colorectal cancer for panitumumab: A biomarker study of the phase III PARADIGM trial.

11 Background: The PARADIGM trial showed longer median overall survival (OS) with first-line mFOLFOX6 plus panitumumab (PAN) vs bevacizumab (BEV) in patients (pts) with RAS wild type (WT) and left-sided metastatic colorectal cancer (mCRC) (37.9 vs 34.3 months, respectively; hazard ratio [HR], 0.82; P=0.031) and similar OS in right-sided pts (HR 1.09; Yoshino T, et al. ASCO 2022 LBA1). We evaluated the usefulness of negative hyperselection by gene alterations in ctDNA related to primary resistance to anti-EGFR therapy. Methods: Baseline plasma ctDNA (>10 ng/mL and >10 nM DNA) from pts enrolled in the biomarker study (NCT02394834) was assessed using a custom panel (PlasmaSELECT-R 91, PGDx). Preplanned gene alterations for hyperselection included KRAS, NRAS, PTEN, and extracellular domain EGFR mutations, HER2 and MET amplifications, and ALK, RET, or NTRK1 fusions. Results: Among 802 pts in the full analysis set, 733 (91%) had evaluable pretreatment samples for ctDNA analysis. Of these pts, 204 (28%) had at least 1 gene alteration, including KRAS/NRAS (8%), BRAF V600E (11%), PTEN (5%), EGFR (3%), HER2 (5%), MET (1%), and fusions (1%). In 529 (72%) hyperselected pts without any gene alterations, OS tended to be longer with PAN vs BEV regardless of primary sidedness; the OS HRs ranged from 0.76 to 0.82, and the median OS gains ranged from 6.6 to 8.0 months (Table). Meanwhile, OS was similar or inferior with PAN vs BEV irrespective of the primary sidedness in pts with any of these gene alterations (Table). Conclusions: Negative hyperselection using ctDNA rather than tumor sidedness may identify appropriate pts for first-line PAN over BEV. These results warrant further validation in additional cohorts. Clinical trial information: NCT02394834 . [Table: see text]