To determine the gross tumor volume (GTV) to clinical target volume margin for non-small-cell lung cancer treatment planning.

A total of 35 patients with Stage T1N0 adenocarcinoma underwent wedge resection plus immediate lobectomy. The gross tumor size and microscopic extension distance beyond the gross tumor were measured. The nuclear grade and percentage of bronchoalveolar features were analyzed for association with microscopic extension. The gross tumor dimensions were measured on a computed tomography (CT) scan (lung and mediastinal windows) and compared with the pathologic dimensions. The potential coverage of microscopic extension for two different lung stereotactic radiotherapy regimens was evaluated.

The mean microscopic extension distance beyond the gross tumor was 7.2 mm and varied according to grade (10.1, 7.0, and 3.5 mm for Grade 1 to 3, respectively, p < 0.01). The 90th percentile for microscopic extension was 12.0 mm (13.0, 9.7, and 4.4 mm for Grade 1 to 3, respectively). The CT lung windows correlated better with the pathologic size than did the mediastinal windows (gross pathologic size overestimated by a mean of 5.8 mm; composite size [gross plus microscopic extension] underestimated by a mean of 1.2 mm). For a GTV contoured on the CT lung windows, the margin required to cover microscopic extension for 90% of the cases would be 9 mm (9, 7, and 4 mm for Grade 1 to 3, respectively). The potential microscopic extension dosimetric coverage (55 Gy) varied substantially between the stereotactic radiotherapy schedules.

For lung adenocarcinomas, the GTV should be contoured using CT lung windows. Although a GTV based on the CT lung windows would underestimate the gross tumor size plus microscopic extension by only 1.2 mm for the average case, the clinical target volume expansion required to cover the microscopic extension in 90% of cases could be as large as 9 mm, although considerably smaller for high-grade tumors. Fractionation significantly affects the dosimetric coverage of microscopic extension.