A novel NGS-based diagnostic algorithm for classifying multifocal lung adenocarcinomas in pN0M0 patients

Abstract

The classification of multifocal lung adenocarcinomas (MLAs), including multiple primary lung adenocarcinomas (MPLAs) and intrapulmonary metastases (IPMs), has great clinical significance in staging and treatment determination. However, the application of molecular approaches in pN0M0 MLA diagnosis has not been well investigated. Here, we performed next-generation sequencing (NGS) analysis in 45 pN0M0 MLA patients (101 lesion pairs) who were initially diagnosed as having MPLA by comprehensive histologic assessment (CHA). Five additional patients with intrathoracic metastases were used as positive controls, while 197 patients with unifocal lung adenocarcinomas (425 random lesion pairs) were used as negative controls. By utilizing a predefined NGS criterion, all IPMs in the positive control group could be accurately classified, whereas 13 lesion pairs (3.1%) in the negative control cohort were misdiagnosed as IPMs. Additionally, 14 IPM lesion pairs were diagnosed in the study group, with at least 7 misdiagnoses. We thus developed a refined algorithm, incorporating both NGS and histologic results, that could correctly diagnose all the known MPLAs and IPMs. In particular, all IPMs identified by the refined algorithm were diagnosed to be IPMs or suspected IPMs by CHA reassessment. The refined algorithm-diagnosed MPLAs patients also had significantly better progression-free survival than the refined algorithm-diagnosed IPMs (p < 0.0001), which is superior to conventional NGS or CHA diagnoses. Overall, we developed an NGS-based algorithm that could accurately distinguish IPMs from MPLAs in MLA patients. Our results demonstrate a promising clinical utility of NGS to complement traditional CHA-based MLA diagnosis and help determine patient staging and treatment.