Here, RNA is attractive in that it provides information about not only single genes, but also fusions, amplifications, and signatures, as well as treatment-induced changes. “We know that genetic makeup of tumors undergoes significant dynamic changes under treatment pressure. So if we have a gene expression signature associated with sensitivity to chemotherapy, will this change over time if there is progression of disease?” Dr. Lenz said. “That will be the challenge in the future, to better understand the dynamics in order to then recommend second- and third-line treatments.”
Some issues must be overcome to use RNA sequencing in routine clinical care, he acknowledged. For example, this sequencing requires fresh frozen tumor tissue, and it must be largely free of any normal tissue.
“There is no doubt that comprehensive molecular characterization with DNA and RNA will lead to better treatment options and identifying patients who benefit most from very unique treatments,” Dr. Lenz concluded. “The key is, you need to do the testing in order to find and identify the most successful treatments for our patients.”
Study details
Most of the 150 COMPASS patients had metastatic disease (87%), and they were about evenly split between receiving mFOLFIRINOX versus gemcitabine plus nab-paclitaxel as first-line chemotherapy, Dr. Knox reported at the symposium, sponsored by the American Gastroenterological Association, the American Society for Clinical Oncology, the American Society for Radiation Oncology, and the Society of Surgical Oncology.
Genomic profiling revealed that roughly 40% overall had potentially actionable variants: alterations in the presence of KRAS wild type (8%), homologous recombination deficiency involving BRCA (6%), and others (25%) such as activating PIK3CA mutations or HER2 amplification.
In findings that Dr. Knox called “quite striking,” among patients given mFOLFIRINOX, median progression-free survival was 7.17 months for those with the RNA classic expression signature versus 2.50 months for those with the RNA basal-like signature (hazard ratio, 0.17; P less than .0001). In contrast, among patients given gemcitabine plus nab-paclitaxel, there was no significant difference by signature (5.65 vs. 4.93 months; P = .69).
GATA6 expression was strongly related to signature type, with classic tumors having high expression and basal-like tumors having low expression (P less than .0001).
Overall survival was similarly better for patients with classic versus basal-like tumors (8.8 vs. 5.2 months; HR, 0.53; P = .006) and for patients with GATA6 high- versus low-expression tumors (8.2 vs. 5.8 months; HR, 0.66; P = .08).
In the whole cohort, overall survival differed across subsets according to RNA signature and type of chemotherapy received (P = .0134). When analysis was restricted to patients given mFOLFIRINOX, median overall survival was 10.1 months with the classic signature but just 5.3 months with the basal-like signature (HR, 0.33; P = .0005).
A multivariate analysis among all chemotherapy-treated patients confirmed the survival benefit of classic signature over basal-like signature (HR, 0.55; P = .03).
Dr. Knox disclosed that she has a consulting or advisory role with Lilly, Merck, and Bristol-Myers Squibb; that she receives honoraria from Novartis; and that she receives research funding from AstraZeneca and Merck. The study is sponsored by University Health Network, Toronto.
SOURCE: O’Kane GM et al. GI Cancers Symposium, Abstract 188.