Discovery of gene fusions in driver-negative NCI-MATCH screening samples.

person Stefan T Kaluziak Massachusetts General Hospital/Harvard Medical School, Boston, MA info_outline Stefan T Kaluziak, A. JOHN Iafrate, Chris Alan Karlovich, P. Mickey Williams, Jeffrey Sklar, Lisa McShane, Elizabeth Codd, Rashi Purohit, Grace Kirkpatrick

Authors person Stefan T Kaluziak Massachusetts General Hospital/Harvard Medical School, Boston, MA info_outline Stefan T Kaluziak, A. JOHN Iafrate, Chris Alan Karlovich, P. Mickey Williams, Jeffrey Sklar, Lisa McShane, Elizabeth Codd, Rashi Purohit, Grace Kirkpatrick Organizations Massachusetts General Hospital/Harvard Medical School, Boston, MA, Frederick National Laboratory for Cancer Research, Frederick, MD, Yale University, New Haven, CT, Biometric Research Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute, Bethesda, MD, Massachusetts General Hospital, Boston, MA Abstract Disclosures Research Funding U.S. National Institutes of Health U.S. National Institutes of Health Background: NCI-MATCH aimed to determine the treatment efficacy of identifying actionable genetic alterations, irrespective of cancer histology, and to tailor treatment to the putative molecular drivers of tumors. Leveraging next-generation sequencing (NGS) and ThermoFisher’s Oncomine assay, this phase II trial genotyped 5,540 patient samples during active screening. Reviewing the genotyping statistics, we noted that NCI-MATCH identified fusions in 164 tumors, only 2.96%. This result is substantially lower than expected, given that clinical assays used at MGH’s and Yale’s diagnostic laboratories identified fusions in 8.5% of the 7,500 sequenced cases, 2.5% of which were novel. Notably, the Oncomine assay, based on multiplexed PCR, is unable to detect driver fusion events when either of the gene partners are unknown, or a novel breakpoint is involved. Cross-referencing identified fusions from MGH’s cohort with Oncomine’s primers, 5-30% of these fusions were potentially missed, suggesting that we could detect up to 30% more fusions than originally identified. Developed at MGH’s Molecular Diagnostics Laboratory, Anchored Multiplex PCR (AMP) interrogates 138 known fusion partners, enabling unbiased fusion detection when only one partner is known. With a high likelihood of identifying novel fusion drivers using AMP technology, we sequenced 663 cases that had no identified mutations of interest (MOI) from the NCI-MATCH cohort. Methods: A total of 200ng of RNA per sample was shipped from ECOG-ACRIN’s repository at MD Anderson Cancer Center to MGH (n = 319) and Yale (n = 349). Samples were processed with ArcherDX’s FusionPlex Pan Solid Tumor v2 kit and sequenced on the NextSeq500 with 150 base pair (bp) Paired-End (PE) kits. Data analysis was completed with ArcherDX’s analytics platform in conjunction with custom Python and R code. Reported fusions were manually verified on a case-by-case basis. All statistical tests were conducted in R. Results: Of the cases sequenced, most (663/693, 95.7%) were successfully processed with AMP. We identified 142 cases with at least one gene fusion out of 663 MOI-Negative cases (21%; 95% CI [18%, 25%]), with 90 as unique fusions, significantly more than the 8.5% observed at MGH (P < .001). Of the 142 fusion-positive cases, 36 have potentially actionable driver events, 27 contain novel fusions not previously described, and in 63 cases no drugs are available to target either component of the fusion. Additionally, we identified fusions in (12/34, 35%) Cholangiocarcinoma and (38/101, 37.6%) Sarcoma cases, 9 and 3 of which were potentially actionable respectively. Conclusions: As hypothesized, we identified fusions that the NCI-MATCH screening had missed, detailing sequencing technology improvement that could be broadly applicable to many clinical laboratories sequencing tumors. Clinical trial information: NCT02465060.

Clinical