Purpose: The computed tomographic (CT) appearance of so-called ground glass components within lung adenocarcinomas correlate with noninvasive tumor histology, and solid radiographic components correlate with invasive histology. We hypothesized that T stage might be more accurately applied by considering the solid component nodule diameter rather than total nodule diameter.
Methods: We identified 74 patients with a solitary lung adenocarcinoma who underwent resection without receiving neoadjuvant therapy. Maximum total diameter and solid diameter of the nodules were measured on CT scans performed within 3 months of surgery. Cox proportional hazard modeling and Kaplan-Meier analyses were performed to determine whether total nodule diameter or solid component diameter was more predictive of overall survival.
Results: Thirty-three patients (45 %) had a solid nodule and 41 patients (55 %) had a part-solid nodule. Most patients were white (59 %) and female (69 %), and 42 % had never smoked. Seventy-four percent underwent lobectomy and 23 % sublobar resection. Sixty-six percent had pathologic stage I disease, 22 % stage II, and 12 % stage IIIA. Mean ± SD total and solid nodule diameters were 32.1 ± 17.5 and 24.8 ± 18.0 mm, respectively (p = 0.01). Among patients with part-solid nodules, multivariate modeling incorporating significant univariate predictors of survival (age, gender, procedure, N descriptor) revealed that maximum solid diameter was associated with overall survival (hazard ratio 1.4, p = 0.01), while maximum total diameter was not.
Conclusions: In a largely non-Asian cohort undergoing resection for adenocarcinoma, radiographic diameter of the solid component of a part-solid lesion on CT predicts overall survival better than total lesion diameter. These data provide further evidence to support altering the T descriptor for lung adenocarcinoma for part-solid nodules.