“In the absence of established or predefined biomarkers for regorafenib, we performed a broad exploratory biomarker analyses at the DNA, RNA, and protein level that represents a much more comprehensive approach than previous studies of regorafenib or sorafenib,” wrote Michael Teufel, PhD, of Bayer Healthcare Pharmaceuticals in Whippany, N.J., and his associates. The preplanned, retrospective analysis of data from the phase 3 RESOURCE trial was reported in.
The randomized trial included 567 patients whose hepatocellular carcinoma had progressed on sorafenib. Regorafenib significantly outperformed placebo with regard to overall survival (OS). Dr. Teufel and his associates performed next-generation sequencing on 17 archived tumor samples containing sufficient tissue (all from regorafenib recipients). They also performed immune profiling on 46 tumor samples (32 from regorafenib recipients and 14 from placebo recipients), protein analysis on 499 plasma samples (332 from regorafenib recipients and 167 from placebo recipients), and microRNA analysis on 343 plasma samples (234 regorafenib recipients and 109 placebo recipients).
Among 266 proteins tested, decreased levels of 5 proteins correlated with significantly longer OS on regorafenib therapy. These proteins are involved in inflammation or hepatocellular carcinogenesis, the researchers noted. Importantly, none were associated with survival independent of treatment. These five proteins included angiopoietin 1 (hazard ratio for OS, 0.53; 95% confidence interval, 0.38-0.73), cystatin B (hazard ratio, 0.47; 95% CI, 0.34-0.64); the latency-associated peptide of transforming growth factor beta (HR, 0.46; 95% CI, 0.33-0.64), oxidized low-density lipoprotein receptor 1 (HR, 0.54; 95% CI, 0.41-0.72), and C-C motif chemokine ligand 3 (HR, 0.54; 95% CI, 0.39-0.74).
Additionally, baseline concentrations of 47 of the 266 proteins correlated with a time to progression (TPP) benefit on regorafenib therapy (adjusted P less than or equal to .05 for each). The 47 proteins included all 5 that predicted an OS benefit. All but two proteins (calbindin and gelsolin) showed the same directional effect as for OS (that is, low expression predicted response).
Nine plasma microRNA’s levels correlated with improved OS on regorafenib (adjusted P less than or equal to .05): MIR30A, MIR122, MIR125B, MIR200A, MIR374B, MIR15B, MIR107, MIR320, and MIR645. Notably, expression was linked to longer OS specifically among patients with the Hoshida S3 subtype of hepatocellular carcinoma. Next-generation sequencing of tumor samples also identified 49 variants in 27 oncogenes or tumor-suppressor genes. Mutations in CTNNB1 were found in 3 of 10 patients who progressed on regorafenib, and VEGFA amplification was found in 1 of 7 regorafenib responders.
“Thus far, rational biomarker selection has been unsuccessful in identifying predictive markers for regorafenib in colorectal cancer and gastrointestinal stromal tumors,” the researchers commented. “The broader approach used in this study is not only biologically warranted considering the heterogeneity of hepatocellular carcinoma tumors, but is also needed due to the multiple targets and pathways affected by MKIs such as regorafenib. Levels of these circulating biomarkers and genetic features of tumors might be used to identify patients with hepatocellular carcinoma most likely to respond to regorafenib.”
Bayer funded the study, provided the study drug, and was involved in all aspects of the study. Dr. Teufel and three coinvestigators are Bayer employees. Dr. Teufel and two coinvestigators own stock in Bayer. Three other coinvestigators disclosed ties to Bayer and other pharmaceutical companies.
SOURCE: Teufel M et al. Gastroenterology. 2019 Jan 30. .