Head-to-head comparison of maximum standardized uptake value and three-dimensional consolidation-to-tumor ratio for preoperative prediction of International Association for the Study of Lung Cancer grade 3 in clinical stage IA invasive lung adenocarcinoma.

内容摘要

Biological aggressiveness varies substantially even within clinical stage IA invasive lung adenocarcinoma. The International Association for the Study of Lung Cancer (IASLC) grading system identifies grade 3 tumors as a high-risk subset associated with adverse outcomes and aggressive pathological features. Preoperative identification of grade 3 tumors may therefore be useful for surgical planning and nodal assessment. Both maximum standardized uptake value (SUVmax) on fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT) and three-dimensional consolidation-to-tumor ratio (3D-CTR) on CT are candidate preoperative markers, but direct comparison of their discriminative ability in the same cohort has been limited. We therefore compared the diagnostic performance of SUVmax and 3D-CTR for predicting IASLC grade 3 in clinical stage IA invasive lung adenocarcinoma. We retrospectively analyzed 224 patients with clinical stage IA invasive lung adenocarcinoma who underwent curative resection at Osaka Medical and Pharmaceutical University Hospital between January 2019 and October 2025. The primary endpoint was IASLC grade 3 determined on resected specimens. SUVmax and 3D-CTR were evaluated preoperatively. Receiver operating characteristic (ROC) analysis was used to calculate the area under the curve (AUC), and AUCs were compared using DeLong's test. Secondary analyses included correlation analysis, logistic regression analysis, and exploratory ROC analyses for occult lymph node metastasis (OLNM), spread through air spaces (STAS), lymphatic vessel invasion (LVI), and pleural invasion (PL). Of the 224 patients, 59 (26.3%) had IASLC grade 3 tumors. Compared with grade 1-2 tumors, grade 3 tumors were more frequently associated with smoking history, a pure-solid appearance, and significantly higher CTR, 3D-CTR, and SUVmax. In ROC analysis, SUVmax had a significantly higher AUC than 3D-CTR for predicting grade 3 (0.792 vs. 0.694, DeLong P=0.02). SUVmax and 3D-CTR showed a moderate positive correlation (Spearman's &#x3c1;=0.486, P<0.001). In multivariable logistic regression analysis, SUVmax remained independently associated with grade 3 [odds ratio (OR) 1.25, 95% confidence interval (CI): 1.13-1.38, P<0.001], whereas 3D-CTR showed only borderline significance (OR 1.01, 95% CI: 1.00-1.02, P=0.050). In exploratory ROC analyses, SUVmax also outperformed 3D-CTR for predicting OLNM, whereas no significant differences were observed for STAS, LVI, or PL. In clinical stage IA invasive lung adenocarcinoma, SUVmax showed stronger discriminative ability than 3D-CTR for preoperative prediction of IASLC grade 3 in this cohort. These findings should be interpreted cautiously and require validation in independent cohorts before clinical application.