Differential Diagnosis of Pulmonary Enteric Adenocarcinoma (PEAC) with the Assistance of Next Generation Sequencing.

Jie Zhang,Keke Yu,Han,Chan Xiang,Yuchen Han,Shaokun Chuai,Haohua Teng,Junyi Ye,Xiaojing Li,Jinchen Shao,Lei Zhu
DOI: https://doi.org/10.1200/jco.2018.36.15_suppl.e20537
IF: 45.3
2018-01-01
Journal of Clinical Oncology
Abstract:e20537 Background: The morphological features and immunohistochemical profiles of PEAC overlap with those of colonic adenocarcinomas; therefore, definitive diagnosis of PEAC has to be confirmed by a negative colonoscopy. Due to the rareness of PEAC, its molecular signature has not been comprehensively examined. In this study, we investigated the molecular signatures associated with PEAC and its histological counterparts, colorectal cancer (CRC) and metastatic colorectal carcinoma (MCC). We also developed a model to effectively distinguish PEAC from MCC/CRC. Methods: Four genetic variant groups clustered from 12 genes were utilized as input variables for model construction. In each round, the best model was selected by 10-fold cross-validation. In the end, 10 models were averaged to generate the parameters for the final model. Results: Capture-based targeted sequencing was performed on tumor tissue samples from 13 PEAC, 15 MCCs and 5 CRC patients using a panel consisting of 295 cancer-related genes. All PEAC patients harbored driver mutation of non-small cell lung cancer (NSCLC), expect for one, who had no mutation detected from this panel. No classic CRC mutation was detected in PEAC. In contrast, classic colon cancer mutations, such as APC and KRAS mutations, were the predominant mutations detected in CRC and MCC, accounting for 78% and 91% of all mutations detected, respectively. Furthermore, we constructed a two-level decision tree model based on 4 genetic variant groups clustered from 12 genes (NSCLC driver mutations, RAS family, APC and MMR genes) to accurately distinguish PEAC from its histological counterparts, achieving an accuracy of 97%. This model accurately classified every patient except for one, who had no mutation detected from this panel. Conclusions: Our data revealed that the molecular profile of PEAC, distinctive to CRC and MCC, resembles that of NSCLC; in contrast, the mutation spectrum of patients with MCC resembles that of CRC. We developed a two-level decision model, which can effectively distinguish PEAC from MCC/CRC, paving its way as routine diagnostic method in clinical settings.
What problem does this paper attempt to address?