Assessment of venous thrombus time of progression by gray-scale median analysis

To determine if gray-scale median (GSM) analysis could differentiate acute and recent deep venous thrombosis (DVT).

Patients submitted to vascular ultrasound examination of lower extremities due to suspected DVT were evaluated. Patients with acute or recent femoropopliteal DVT were included, whereas those without DVT, with chronic or isolated calf DVT were excluded. Time of onset of DVT symptoms was recorded. A transverse image of the thrombosed vein and adjacent artery was obtained. Two sonographers determined a subjective impression of thrombus time of progression and classified it as acute or recent. Thrombus GSM was calculated with a software. ROC curve was used to determine GSM cut-off points. Fischer's exact and Student´s t tests were also used. P<0.05 indicated statistical significance.

128 veins of 63 extremities were studied. Thrombus GSM correlated with time of onset of DVT symptoms (P=0.005) and with subjective evaluation of thrombus time of progression (P<0.001). When DVT symptoms had begun up to two weeks before and thrombus was classified as acute, GSM was lower. Area under ROC curve for thrombus GSM was 0.76 (P<0.001). GSM 17.90, 23.03, and 40.02 cut-off points differentiated between acute and recent thrombi with 35.59% sensitivity and 91.30% specificity, 59.32% sensitivity and 84.06% specificity, 91.53% sensitivity and 31.88% specificity, respectively.

GSM was a reliable tool for objectively differentiating acute and recent DVT in most thrombus images. GSM 17.90 and 40.02 cut-off points showed high specificity and high sensitivity, respectively, for acute and recent DVT differentiation. In 17.90 to 40.02 interval, GSM cut-off point that best distinguished acute from recent DVT was 23.03.