Predictors of non-neoplastic lesions in lung tumours showing ground-glass opacity on thin-section computed tomography based on a multi-institutional prospective study

OBJECTIVES: Peripheral small lung tumours (LTs) showing ground-glass opacity (GGO) tend to be treated without preoperative histological diagnosis due to difficulty in obtaining tissue samples. Exclusion of non-neoplastic lesions (NNLs) is essential when considering non-surgical treatment such as stereotactic radiotherapy. Here, we sought to determine preoperative factors associated with NNLs in peripheral LTs using data from a prospective study that investigated the efficacy of lesser pulmonary resection (JCOG0804/WJOG4507L). METHODS: The key eligibility criteria of the study were as follows: (i) peripherally located definitive or suspected LC with maximum diameter <= 2 cm and (ii) radiological non-invasive tumour with consolidation/tumour ratio (CTR) of <= 0.25 based on thin-section computed tomography (CT). Among all the resected LTs, incidences of NNL and precancerous lesions were examined. Also, logistic regression analysis was conducted to investigate the predictors of NNL using maximum tumour dimension (<= 1 cm/>1 cm) and CTR (0/>0) as an explanatory variable. RESULTS: Between May 2009 and April 2011, 333 patients were prospectively enrolled from 51 institutions into the study. Among 333 patients, 345 LTs were included in the analysis. There were 314 (91.0%) LCs, 17 (4.9%) precancerous lesions and 14 (4.1%) non-cancerous lesions. Maximum tumour dimension < 1 cm was identified as a significant predictor of NNLs with logistic regression analysis. There were 10 (8.6%) NNLs in 116 LT <= 1 cm, but 4 (1.7%) NNLs in 229 LTs >1 cm. CONCLUSIONS: NNLs were found in only 4.1% of peripheral LTs with GGO. However, when the tumour diameter was <= 1 cm, similar to 10% were NNLs, necessitating a histological diagnosis when non-surgical treatment was considered.