Background: To address whether systematic lymph node dissection is always necessary in early lung cancer, we identified factors predicting nodal involvement in a screening series and applied them to nonscreening-detected cancers.
Methods: In the 97 patients with clinical T1-2N0M0 lung cancer (<3 cm), enrolled in the Continuous Observation of Smoking Subjects computed tomography (CT) screening study, who underwent curative resection with radical mediastinal lymph node dissection, we examined factors associated with hilar extrapulmonary and mediastinal nodal involvement. Nodule size plus positive/negative positron emission tomography (PET)-CT (usually as maximum standard uptake value [maxSUV]) were subsequently evaluated retrospectively for their ability to predict nodal involvement in 193 consecutive patients with nonscreening-detected clinical stage I lung cancer.
Results: Among Continuous Observation of Smoking Subjects patients, 91 (94%) were pN0, and six (6.2%) were pN+. All patients with maxSUV <2.0 (p = 0.08) or pathological nodule ≤10 mm (p = 0.027) were pN0 (62 cases). Nodal metastases occurred in 6 cases among the 29 (17%) patients with lung nodule >10 mm and maxSUV ≥2.0 (p = 0.002 versus the other 62 cases). In the nonscreening series, 42 of 43 cases with negative PET-CT (usually maxSUV <2.0) or nodule ≤10 mm were pN0; 33 of 149 (22%) cases with positive PET-CT (usually maxSUV ≥ 2.0) and nodule >10 mm were pN+ (p = 0.001 versus the 43 cases).
Conclusions: This limited experience suggests that in early-stage clinically N0 lung cancers with maxSUV <2.0 or pathological nodule size ≤10 mm, systematic nodal dissection can be avoided as the risk of nodal involvement is very low.