Prognostic value of the frequency of vascular invasion in stage I non-small cell lung cancer

Objectives There is no standard pathological method for determining vessel invasion in lung cancer. Herein, we examine whether vessel invasion can be accurately assessed using hematoxylin–eosin staining alone, and investigate the prognostic impact of the presence and frequency of vessel invasion in lung cancer. Methods Vessel invasion was assessed by hematoxylin–eosin, Victoria blue, and D2-40 in 251 completely resected stage I non-small cell lung cancer patients. Vessel invasion was classified into 3 grades according to the number of invaded vessels. Results Using hematoxylin–eosin and Victoria blue, vascular invasion was detected in 27 (10.8 %) and 75 (29.9 %) of patients, respectively. Lymphatic permeation was detected in 126 (50.2 %) and 70 (27.9 %) of patients using hematoxylin–eosin and D2-40 staining. Hematoxylin–eosin staining did not accurately detect a high frequency of vessel invasion; only 5 and 21.7 % of high-frequency vascular invasion and lymphatic permeation cases diagnosed with Victoria blue and D2-40 were detected. Multivariate analysis based on elastic stain and immunostaining indicated that plural invasion, a high frequency of vascular invasion (hazard ratio 4.00), and a high frequency of lymphatic permeation (hazard ratio 2.30) were independent predictors of cancer recurrence within 3 years. Likewise, an age ≥70 years, male, and a high frequency of vascular invasion (hazard ratio 3.41) were independent predictors of overall survival. Conclusions Vascular invasion should be confirmed by elastic stains, and the frequency, not but the presence, of vascular invasion is a powerful independent prognostic factor in completely resected stage I non-small cell lung cancer patients.