Search

Abstract Microsatellite instability high (MSI-H) and mismatch repair deficient (dMMR) tumor status have been demonstrated to predict patient response to immunotherapies. We developed and validated a next-generation sequencing (NGS)-based companion diagnostic (CDx) to detect MSI-H solid tumors via a comprehensive genomic profiling (CGP) assay, FoundationOne®CDx (F1CDx). To determine MSI status, F1CDx calculates the fraction of unstable microsatellite loci across >2000 loci using a fraction-based (FB) analysis. Across solid tumor types, F1CDx demonstrated a high analytical concordance with both PCR (n = 264) and IHC (n = 279) with an overall percent agreement (OPA) of 97.7% and 97.8%, respectively. As part of a retrospective bridging clinical study from KEYNOTE-158 Cohort K and KEYNOTE-164, patients with MSI-H tumors as determined by F1CDx demonstrated an objective response rate (ORR) of 43.0% to pembrolizumab. In real-world cancer patients from a deidentified clinicogenomic database, F1CDx was at least equivalent in assessing clinical outcome following immunotherapy compared with MMR IHC. Demonstrated analytical and clinical performance of F1CDx led to the pan-tumor FDA approval in 2022 of F1CDx to identify MSI-H solid tumor patients for treatment with pembrolizumab. F1CDx is an accurate, reliable, and FDA-approved method for the identification of MSI-H tumors for treatment with pembrolizumab.

Abstract Treatment of non-small cell lung cancer is increasingly biomarker driven with multiple genomic alterations, including those in the epidermal growth factor receptor (EGFR) gene, that benefit from targeted therapies. We developed a set of algorithms to assess EGFR status and morphology using a real-world advanced lung adenocarcinoma cohort of 2099 patients with hematoxylin and eosin (H&E) images exhibiting high morphological diversity and low tumor content relative to public datasets. The best performing EGFR algorithm was attention-based and achieved an area under the curve (AUC) of 0.870, a negative predictive value (NPV) of 0.954 and a positive predictive value (PPV) of 0.410 in a validation cohort reflecting the 15% prevalence of EGFR mutations in lung adenocarcinoma. The attention model outperformed a heuristic-based model focused exclusively on tumor regions, and we show that although the attention model also extracts signal primarily from tumor morphology, it extracts additional signal from non-tumor tissue regions. Further analysis of high-attention regions by pathologists showed associations of predicted EGFR negativity with solid growth patterns and higher peritumoral immune presence. This algorithm highlights the potential of deep learning tools to provide instantaneous rule-out screening for biomarker alterations and may help prioritize the use of scarce tissue for biomarker testing.

Abstract In this study, we report the clinicopathologic and genomic profiles of 891 patients with RET fusion driven advanced solid tumors. All patient samples were tested using a tissue-based DNA hybrid capture next generation sequencing (NGS) assay and a subset of the samples were liquid biopsies tested using a liquid-based hybrid capture NGS assay. RET fusions were found in 523 patients with NSCLC and in 368 patients with other solid tumors. The two tumor types with the highest number of RET fusion were lung adenocarcinoma and thyroid papillary carcinoma, and they had a prevalence rate 1.14% (455/39,922) and 9.09% (109/1199), respectively. A total of 61 novel fusions were discovered in this pan-tumor cohort. The concordance of RET fusion detection across tumor types among tissue and liquid-based NGS was 100% (8/8) in patients with greater than 1% composite tumor fraction (cTF). Herein, we present the clinicopathologic and genomic landscape of a large cohort of RET fusion positive tumors and we observed that liquid biopsy-based NGS is highly sensitive for RET fusions at cTF ≥1%.