The impact of partial volume effect correction on the diagnostic performance of quantitative FDG-PET/CT parameters: A lesion based analysis of suspected lung malignancy.

Objectives The role of partial volume effect correction (PVC)in the accurate quantitation of PET/CT parameters has been emphasized in phantom studies, However, limited number of clinical studies have investigated PVC. The purpose of this prospective study was to evaluate the potential of PVC on quantitative FDG­-PET/CT parameters in benign malignant differentiation of lesions in patients with suspected lung cancer.

Methods One hundred suspected lung cancer lesions (60benign;40 malignant)with histopathological diagnosis were included in this study. All patients underwent FDG-­PET/CT imaging before surgery.Quantitative PET parameters such as SUVmax,SUVmean,pvcSUVmean,metabolic tumor volume(MTV),total lesion glycolysis(TLG=SUVmean*MTV) pvcTLG (pvcSUVmean*MTV)and SUVpeak were measured by using an adaptive contrast­ oriented thresholding segmentation & PVC algorithm. The diagnostic performance of parameters was compared by pairwise comparison of receiver operating characteristic (ROC) curves. We also performed a subgroup analysis for lesions with metabolic volume less than the median(MTV=6.6).

Results FDG-PET/CT parameters showed superior diagnostic performance over volumetric parameters while the area under curve(AUC)were 0.90, 0.87, 0.86, 0.84, 0.82, 0.75, 0.55, for pvcSUVmean, SUVmax,SUVmean, SUVpeak, pvcTLG, TLG and MTV, respectively. In pairwise comparison of ROC curves, pvcSUVmean showed superior performance over SUVmean(p=0.04),SUVpeak (p=0.02)but not over SUVmax(p=0.16).Similarly,PVC enhanced the diagnostic performance of TLG. In subgroup analysis of small lesions pvcSUVmean had the highest performance compared to all other parameters with statistical significance.

Conclusions In this study PVC improves the diagnostic performance of PET/CT in differentiating suspected lung lesions, particularly for small lesions and it could be considered as a potential source of quantification/diagnostic inaccuracy in future clinical studies.