Using radiomics feature to distinguish malignant from benign solitary pulmonary nodules on dual time point PET/CT images

Objectives: Lung cancer is the most common cancer worldwide, which usually present as solid pulmonary nodules (SPNs) on early diagnostic images. Classification of malignant disease at this early timepoint is critical for improving success of surgical resection and increasing 5 year survival. This study aimed to determine whether radiomics feature(Cluster prominence) derived from dual time point PET images (DTPI) can help to differentiate between malignant and benign solitary pulmonary nodules (SPNs).

Methods: We retrospectively analyzed data from patients with solid pulmonary nodules (SPNs) who had DTPI 18F-FDG PET/CT between 2003 and 2014 at a single institution. Patient without pathology confirmed diagnosis or those receiving surveillance follow-up imaging less than 1 year were removed from this study. The early PET scans and delayed PET scans were performed in the same session at 60min and 180min post injection, respectively. Nodules were identified and segmented on each scan by two experienced nuclear medicine physicians. The metabolic volumes (MV) on early PET scans were calculated, and MV smaller than 5cc were removed from this study. For each lesion MV, SUVmax and radiomics features (Cluster prominence) were extracted on early PET images and delayed images. of each lesion were calculated on both early PET and delayed PET images. A 5 scaled visual interpretation score was given for each lesion base on the dual time point PET/CT images by two experienced nuclear medical physicians. ROC analysis was performed on each radiomics feature, SUVmax and visual interpretation scores individually. The areas under curve (AUC) were calculated.

Results: Eighty five SPNs were included. AUCs for discriminating between benign and malignant SPNs of SUVmax and visual interpretation were 0.78 and 0.77, respectively. Cluster prominence extracted from early and delayed PET had larger AUCs(0.81 and 0.82, respectively) than SUVmax and visual interpretation.

Conclusions: Radiomics feature cluster prominence on FDG DTPI images were useful for discriminating benign from malignant nodules in larger SPNs. Compared to SUVmax or visual interpretation, cluster prominence extracted from PET images improved discrimination between benign and malignant SPNs.