Role of FDG-PET/CT in Identification of Histological Upgrade of Ductal Carcinoma in Situ (DCIS) in Needle Biopsy

Background: Accurate preoperative detection of the invasive components of ductal carcinoma in situ (DCIS) is essential for an appropriate treatment. 18F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) scan, which can indicate the metabolic activity and aggressiveness of breast cancer, may be used as one of the predictors of the invasive components of DCIS in needle biopsy. Objectives: To determine whether the FDG-PET/CT findings are associated with the histological upgrade of DCIS in biopsy. Patients and Methods: In this retrospective cohort, we reviewed 165 cases of DCIS in 162 patients, who underwent preoperative FDG-PET/CT examinations between April 2008 and September 2015. The clinicopathological characteristics and FDG-PET/CT findings of the patients were compared with respect to cancer invasion. The predictors of DCIS upgrade to invasive cancer were also examined. Moreover, the diagnostic performance of visual and semi-quantitative analyses of FDG-PET/CT in predicting invasion was compared. The semi-quantitative analyses were based on the maximum standardized uptake value (SUVmax), divided by the cutoff point in a receiver operating characteristic (ROC) curve analysis. Results: The final pathological findings indicated 119 cases of pure DCIS and 46 cases of DCIS with invasion. The optimal SUVmax threshold was 1.9 in the ROC curve analysis. Young age, high SUVmax, positivity in the visual analysis of FDG-PET/CT, and large pathological tumor size were significantly more frequent in the DCIS + invasion group. The significant predictors of DCIS histological upgrade were age (P = 0.011), SUVmax (P < 0.001), visual analysis of FDG-PET/CT (P = 0.004), and pathological tumor size (P = 0.003) in the univariate analysis. In the multivariate analysis, the SUVmax (odds ratio [OR] = 3.31, P = 0.003) and tumor size (OR = 1.20, P = 0.022) were significant when the model included the SUVmax, age, and size (model 1). On the other hand, age (OR = 0.96, P = 0.032), visual analysis (OR = 4.67, P = 0.006), and tumor size (OR = 1.25, P = 0.005) were significant predictors when the model included visual analysis, age, and size (model 2). The sensitivity was significantly higher in the visual analysis, whereas the specificity, positive predictive value (PPV), and accuracy were significantly higher in the semi-quantitative analysis. Conclusion: FDG-PET/CT is a potentially useful imaging tool to predict the upgrade of DCIS to invasive cancer.