Background: Visceral pleural invasion (VPI) with PL1 or PL2 increases the T classification from T1 to T2 in non-small cell lung cancers (NSCLCs) ≤ 3 cm. We proposed a modified T classification based on VPI to guide adjuvant therapy.Research Question: Is it reasonable to upstage PL1-positive cases from T1 to T2 for NSCLCs ≤ 3 cm?Study Design and Methods: In total, 1,055 patients with resected NSCLC were retrospectively included. Tumor sections were restained with hematoxylin and eosin stain and Victoria blue elastic stain for the elastic layer. Disease-free survival (DFS) and overall survival (OS) were calculated by the Kaplan-Meier method. Subgroup analysis and a Cox proportional hazards model were used to further determine the impact of VPI on survival.Results: The extent of VPI was diagnosed as PL0 in 824 patients, PL1 in 133 patients, and PL2 in 98 patients. The 5-year DFS rates of patients with PL0, PL1, and PL2 were 62.6%, 60.2%, and 28.8% (P < .01), whereas the corresponding 5-year OS rates were 78.6%, 74.4%, and 50.0% (P < .01), respectively. As predicted, the DFS and OS of patients with PL2 were much worse than those of patients with PL0 (P < .01) and PL1 (P < .01). However, both the DFS and OS of patients with PL0 and PL1 were comparable (DFS: P = .198; OS: P = .150). For node-negative cases, the DFS and OS of patients with PL0 and PL1 were also comparable (DFS: P = .468; OS: P = .388), but patients with PL2 had much worse DFS and OS than patients with PL0 (P < .01) and PL1 (P < .01). Multivariable analyses suggested that PL2, together with node positivity and poor cell differentiation, was an independent adverse prognostic factor.Interpretation: In NSCLCs ≤ 3 cm, tumors with PL1 should remain defined as T1, not T2. Overtreatment by adjuvant chemotherapy in node-negative NSCLCs ≤ 3 cm might be avoided in PL1 cases. [ABSTRACT FROM AUTHOR]