According to WHO diagnostic criteria, both Small cell lung cancer (SCLC) and Large cell neuroendocrine (LCNEC) are categorized into high grade neuroendocrine carcinoma, based on morphological features and immunostaining of neuroendocrine markers including CD56, chromogranin A (CgA) and synaptophysin (Syn). However, the above markers have limited satisfactory significance of specificity and sensitivity. Insulinoma-associated protein 1 (INSM1) is highly expressed in SCLC cells as a zinc finger protein, and numerous studies have shown that INSM1 positive staining has a strong correlation with NE differentiation in SCLC and an alternative diagnostic biomarker for SCLC, but few studies are involved in LCNEC and its significance remains unclear in LCNEC. The aim of this study was to investigate the potential of INSM1 as a neuroendocrine marker by comparing the expression of INSM1 in SCLC and LCNEC and its clinicopathological significance. 347 surgically resected lung cancer including 290 pure SCLC and 57 pure LCNEC samples were retrieved in this retrospective study. INSM1 expression was detected by immunohistochemistry staining in SCLC and LCNEC tissue microarrays. The different expression of INSM1 between SCLC and LCNEC as well as the correlation with clinicopathological features including gender, age, smoking history, tumor location, AJCC 7th stage, reginal lymph node metastasis, pleural invasion and tumor thrombosis were analyzed by chi-square test. Of the included patients, male/female ratio was 2.33 (203/87) in SCLC and 10.4 (52/5) in LCNEC, the smoker rate was 64.5% (187/290) in SCLC and 43.9% (25/57) in NSCLC. According to the AJCC Cancer Staging Manual (7th edition), the SCLC patients were 29.0% (84/290) in stage I, 27.2% (79/290) in stage II, 40.7% (118/290) in stage III, and 3.1% (9/290) in stage IV, while the NSCLC patients were 49.1% (28/57) in stage I, 26.3% (15/57) in stage II, and 24.6% (14/57) in stage III, respectively. INSM1 was localized in the nucleus of lung cancer cells, a significant higher positive rate was detected in SCLC than that in LCNEC (271/290,93.4% in SCLC; 32/57,56.1% in LCNEC; p<0.05). INSM1 protein expression significantly correlated with smoking history (p=0.023), pleural invasion (p=0.010), and a positive trend of regional lymph node metastasis (p=0.052) in SCLC, but did not correlate with any of the clinicopathological features in LCNEC (p>0.05). INSM1 protein was found no prognostic significance neither in SCLC nor in LCNEC (p>0.05). INSM1 protein is highly expressed in SCLC and positively correlated with pleural invasion and smoking history. In contrast, INSM1 expression in LCNEC is significant lower and does not correlated with such clinicopathological features. Further research needs to be done on exploring the potential different underlying mechanisms and diagnostic significance of INSM1 in SCLC and LCNEC.