Detection of deep vein thrombosis using FDG-PET/CT volumetric parameters: A diagnostic performance study.

Objectives It has been shown that quantitative FDG-PET/CT parameters are able to detect inflammation in the vascular structures throughout the body and recently promising results towards detection of deep venous thrombosis (DVT) have been achieved. The aim of this study was to prospectively evaluate the ability of FDG-PET/CT volumetric indices in diagnosis of DVT.

Methods 23 patients (9 DVT+, 14 DVT-) with suspected DVT underwent compression ultrasound (CUS) and FDG-PET/CT scans. To measure the amount of FDG uptake in the lower extremities, mean standard uptake value (SUVmean), maximum standard uptake value (SUVmax), metabolic active volume (MAV) and total lesion glycolysis (TLG=MAV*SUVmean) were assessed by an expert physician blinded to the clinical and CUS results of the patient. An adaptive contrast-oriented thresholding algorithm was used to delineate thrombosed veins. Contralateral non-DVT lower limb (with lower FDG uptake) was used as internal control to calculate ratio, difference and relative ratio of FDG uptake parameters. Receiver operating curve (ROC) analyses were used for assessment of sensitivity and specificity of the volumetric parameters of FDG-PET/CT.

Results MAV, TLG and SUVmax relative ratios were the strongest in detection of DVT. The areas under the ROC curves were 0.91 for MAV (p=0.0001), 0.83 for TLG (p=0.0003), and 0.72 for SUVmax (p=0.04). For detection of DVT, relative ratio of MAV > 102% was 88.89% sensitive and 85.71% specific, relative ratio of TLG > 243% was 67% sensitive and 93% specific, while relative ratio of SUVmax > 11% showed 77% sensitivity and 57% specificity.

Conclusions In this patient population, volumetric FDG-PET/CT parameters seem to be promising diagnostic biomarkers for diagnosis of DVT. However, larger scale analyses are needed for further clarification of the role of these parameters in DVT patients.