PET/CT features discriminate risk of metastasis among single-bone FDG lesions detected in newly diagnosed non-small-cell lung cancer patients.

Abstract

We investigated the capacity of fluorodeoxyglucose (FDG) PET/CT features for stratifying probability of metastasis for single-bone FDG lesions in non-small-cell lung cancer (NSCLC). Subjects were 118 newly diagnosed NSCLC patients with a solitary bone FDG lesion and no evidence of other distant metastasis based on PET/CT, brain MRI, and contrast-enhanced chest CT. Bone lesion SUV_max and CT findings, primary tumor SUV_max, clinical T stage, and N stage were analyzed. The bone lesions were determined by biopsy, characteristic MRI findings and clinical follow-up to be metastatic in 33 (28.0%) and benign in 85 cases (72.0%). A cutoff bone SUV_max of 4.3 showed good diagnostic performance (81.8% sensitivity, 84.7% specificity, and 83.9% accuracy), but there was considerable overlap. Bone lesion PET/CT features of SUV_max ≤ 2, osteosclerotic rim or fracture correctly diagnosed 20/20 benign, while SUV_max > 10, soft-tissue mass or bone destruction correctly diagnosed 18/18 metastatic cases. In the remaining 80 cases, bone features of SUV_max > 4.3 and osteolytic change, and lung tumor features of SUV_max > 6.4, ≥ T2 stage (n = 70), and ≥ N1 stage (n = 43) favored metastasis. The presence of one or less of these features correctly diagnosed 38/38 benign, while the presence of four or more features correctly diagnosed 5/5 metastatic cases. The 37 cases with two or three features had either benign (n = 27) or metastatic bone disease (n = 10). Combining bone lesion and lung tumor PET/CT features can help stratify risk of bone metastasis in these patients. • In NSCLC with a single-bone FDG lesion, lesion SUV_maxis useful for differential diagnosis. • CT features of the single-bone FDG lesions provide additional diagnostic value. • High NSCLC SUV_max, greater T stage, and FDG positive nodes also favor metastasis.