Effect of screening with a pan-TRK immunohistochemistry-based algorithm on NTRK fusion detection rates.

Abstract

e15135 Background: Rearrangements in the NTRK1, NTRK2, and NTRK3 genes are clinical actionable and predictive of response to TRK kinase inhibitors in solid tumors. Their detection in clinical practice is a priority, but their low prevalence (0.25-1%) poses a diagnostic challenge. In this study, we describe our immunohistochemical-based and cost-effective approach for NTRK rearrangements detection. Methods: We implemented a comprehensive NTRK testing approach at a reference center that combined clinical criteria (e.g., iodine therapy-resistant thyroid cancer); molecular biomarkers (e.g., non-small-cell lung carcinoma negative for driver alterations, colorectal carcinoma with deficient MMR and/or microsatellite instability); and histopathological criteria (identifying tumor subtypes associated with NTRK fusions, such as breast secretory carcinoma). Automated immunohistochemistry (Ventana BenchMark ULTRA IHC/ISH System) using the antibody pan-TRK (Ventana, EPR17341) was used as a screening tool. Positive results were confirmed through molecular techniques, including Next Generation Sequencing, NGS (Oncomine Focus Assay, Thermo Fisher, based on DNA and RNA) and/or FISH (NTRK3 break-apart probe, Zytovision). Results: We evaluated 172 patients (99 female/73 male), with a mean age of 52 years (SD 23.8). Eighteen (18) patients with neoplasms expressing TRK proteins were identified. Confirmatory studies using NGS identified fusions in 7 cases. In instances of poor RNA quality, FISH techniques (NTRK3) were employed, confirming the findings in an additional 3 patients. Overall prevalence of NTRK fusions was 5.8%. Histological types included the following tumor types: Thyroid carcinoma, 4 patients (one medullary and 3 papillary carcinomas, two of them in the pediatric population); soft tissue sarcomas, 3 patients (2 corresponding to infantile fibrosarcoma and a soft tissue provisional entity, "NTRK-rearranged spindle cell neoplasm”); secretory carcinoma of the breast, one patient; secretory carcinoma of the salivary gland, one patient; colorectal carcinoma, one patient, with a deficient MMR phenotype. NTRK3 gene fusions were found in 10 patients, (6 fusions with ETV6 detected by NGS and 3 detected by FISH), and NTRK1 fusion in 1 case (fusion with LMNA detected by NGS). The positive predictive value of pan-TRK immunohistochemistry as a screening tool was 0.55. Conclusions: In our experience, the comprehensive screening approach based on morphological and clinical/ molecular inclusion criteria, along with the use of immunohistochemical techniques (pan-TRK antibody), significantly enhances the detection rate of NTRK fusions (5.8% vs. 0.25-1%). Additionally, FISH provides a valuable tool in samples with poor RNA quality. This approach allows for a reduction in the impact of costly testing, such as NGS, on the healthcare system.