Does histological assessment accurately distinguish separate primary lung adenocarcinomas from intrapulmonary metastases? A study of paired resected lung nodules in 32 patients using a routine next-generation sequencing panel for driver mutations

Abstract

AimVarious approaches have been reported for distinguishing separate primary lung adenocarcinomas from intrapulmonary metastases in patients with two lung nodules. The aim of this study was to determine whether histological assessment is reliable and accurate in distinguishing separate primary lung adenocarcinomas from intrapulmonary metastases using routine molecular findings as an adjunct.MethodsWe studied resected tumour pairs from 32 patients with lung adenocarcinomas in different lobes. In 15 of 32 tumour pairs, next-generation sequencing (NGS) for common driver mutations was performed on both nodules. The remainder of tumour pairs underwent limited NGS, or EGFR genotyping. Tumour pairs with different drivers (or one driver/one wild-type) were classified as molecularly unrelated, while those with identical low-frequency drivers were classified as related. Three pathologists independently and blinded to the molecular results categorised tumour pairs as related or unrelated based on histological assessment.ResultsOf 32 pairs, 15 were classified as related by histological assessment, and 17 as unrelated. Of 15 classified as related by histology, 6 were classified as related by molecular analysis, 4 were unrelated and 5 were indeterminate. Of 17 classified as unrelated by histology, 14 were classified as unrelated by molecular analysis, none was related and 3 were indeterminate. Histological assessment of relatedness was inaccurate in 4/32 (12.5%) tumour pairs.ConclusionsA small but significant subset of two-nodule adenocarcinoma pairs is inaccurately judged as related by histological assessment, and can be proven to be unrelated by molecular analysis (driver gene mutations), leading to significant downstaging.