Neuroendocrine transformation from EGFR/ALK-wild type or TKI-naïve non-small cell lung cancer: An under-recognized phenomenon

IntroductionNeuroendocrine transformation (NET) is a resistance mechanism for epidermal growth factor receptor (EGFR)/anaplastic lymphoma kinase (ALK) tyrosine kinase inhibitors (TKIs) in non-small cell lung cancer (NSCLC). We aimed to elucidate whether NET develops in TKI-naïve NSCLC by using molecular fingerprinting in paired pre- and post-NET tissues.Patients and methodsNET cases were identified based on the following criteria: the pre- and post-NET lesions must harbor mutual somatic mutations; neuroendocrine component should be absent in the sampled specimens of pre-NET lesions; and re-biopsy should be performed in either previously biopsied baseline lesions or newly developed lesions, but not in baseline-existing non-biopsied lesions, excluding synchronous neuroendocrine malignancy. p53 and Rb expression were evaluated via immunohistochemistry. Clinical characteristics, treatments, and outcomes were recorded and analyzed.ResultsFifteen NET cases were identified, including five EGFR/ALK wild-type, three EGFR-mutant TKI-naïve, and seven TKI-treated cases. All cases harbored mutual somatic mutations in paired pre- and post-NET lesions. Recurrent pre-NET mutations were detected in TP53 (44.4%), RB1 (33.3%), and PDGFRA (33.3%), but two of the three PDGFRA mutations were lost after NET, whereas pre-NET TP53 and RB1 mutations were retained in the corresponding post-NET lesions. Immunohistochemistry revealed inactivated p53/Rb in 90.9% and 72.7% of the pre-NET lesions, respectively.ConclusionsThis proof-of-concept study demonstrated that NET develops in NSCLCs without TKI targets or treatments. This phenomenon could be under-recognized, because re-biopsy was less frequently performed in these patients. Tissue re-biopsy should be preferred over liquid biopsy at the time of progression to account for histology transformation. p53/Rb IHC should be considered in addition to genomic TP53/RB1 evaluation for NET risk prediction.