One of the major conclusions in the study by Ou and colleagues1 was that for patients with stage IB (T2N0M0) non-small cell lung cancer with a tumor size of > 3 cm, the T2P descriptor is a poor independent prognostic factor. The authors have acknowledged that a major limitation of their study was that VPI could not be separated from hilar atelectasis or obstructive pneumonitis. However, they presumed1 that the majority of patients within the T2P category had VPI. The effect of tumor size on the impact of VPI has rarely been demonstrated, and the results have remained controversial. Furthermore, the relationship between tumor size and frequency of VPI has rarely been investigated.2,3 Manac'h and colleagues2 have demonstrated that the frequency of VPI significantly increased as tumor size increased (tumors ≤ 3 cm, 10%; tumors > 3 to ≤ 5 cm, 19.6%; tumors > 5 cm, 33%). Shimizu and coworkers3 also reported that patients with tumors having a diameter of > 3 cm had a higher possibility of VPI. In the study by Ou and colleagues,1 the group of patients with T2 tumors > 3 cm with a T2P descriptor may consist of more patients with tumors of larger size than that of patients with T2 tumors > 3 cm alone. Therefore, the prognostic value of the T2P descriptor in T2 tumors > 3 cm may be partly contributed by larger tumor size. We wonder how Ou and colleagues1 handled the possible interaction of VPI and tumor size in their study.