The traditional histologic classification of lung cancer is not satisfactory to describe the morphologic characteristics of individual tumors, because it does not fully cover cytologic features. This paper describes a novel typing system using morphometric profiling that covers a variety of morphologic features including histologic architecture, cell type, cytoplasmic color and internal structure, nuclear outline, chromatin pattern, nucleoli count and remarkableness, and average and deviation of nuclear size and circularity. In all, 201 cases of lung tumors (whose sizes are <20 mm) were examined. Results of a hierarchical clustering analysis were used to draw a dendrogram. We here tentatively focused on 8 morphometric clusters and analyzed their potential association with a variety of clinicopathologic and molecular genetic features. Significant differences in postoperative recurrent risk, growth activity, oncogenic mutation (EGFR or KRAS), impairments of tumor suppressors (p53 and p16), sex predisposition, and smoking status were found among the 8 clusters. The system has the potential to improve histopathologic diagnosis and our understanding of carcinogenesis in the lung.