Histology determination of lung cancers: A report on genomic profiling of lung cancer of mixing histology.

Abstract

8570 Background: Histopathology, largely determined by morphology, plays a critical role in choosing appropriate treatment for lung cancer. The understanding of molecular determination of lung cancer histology is rudimentary. Our recently published data (Zhang, Science, 2014 and Liu, Nature Communications, 2016) have demonstrated that within the same patients with identical genetic background and identical exposure, tumor regions with different morphologic appearances may have very similar genomic profiles while tumors with the same morphology may have distinct genomic landscape. Methods: We collected 12 lung cancers of mixing histology with 2 to 4 histologic components within each tumor. In total, 26 tumor regions including 9 adenocarcinomas, 6 large-cell neuroendocrine carcinoma, 6 small cell carcinomas and 4 squamous cell carcinomas and one poorly differentiated lung carcinoma were microdissected and subjected to whole exome sequencing. Results: A substantial number of identical mutations were shared between different histologic components within the same tumor in all 12 patients. However, the proportion of shared mutations varies in different patients ranging from as little as 4% to as much as 99%. Mutation spectrum is also similar between different histologic components within the same tumors suggesting similar mutational process in place. Identical canonical cancer gene mutations including TP53, KRAS, PIK3CA, SOS1 and STK11 are generally shared between different histologic components within the same tumors. Canonical mutations in FBXW7 and MTHFR were detected in squamous component, but not small-cell component in one patient. Conclusions: Different histologic components of lung cancers of mixing histology are likely derived from the same progenitor cells, but the molecular timing of branch separation of subclones giving rise to different histologic components varies in different tumors. Although genomic aberrations may play a role in a subset of tumors, histologic features may not be determined at genomic level for most lung cancers. Gene expression and methylation analyses from these tumors are underway.