Background: This study was undertaken to assess the reliability of duplex scanning, as compared with digital subtraction angiography, in diagnosing peripheral arterial disease of the lower limbs.
Patients and methods: 60 legs were studied. Each leg was divided in 5 arterial segments, from aortoiliac to popliteal. Duplex scanning and digital subtraction angiography were performed. The disease in each segment was assessed as significant or insignificant or classified into five categories concerning the grade of stenosis (patent vessel or no diameter reduction, mild stenosis, i.e. 1-19% diameter reduction, moderate and severe stenosis, i.e. 20-49% and 50-99% diameter reduction, respectively, and occluded vessel). Three statistical approaches were applied: calculating sensitivity and specificity, Kappa statistics, and weighted Kappa value. Median grades of arterial stenosis across different segments, according to angiography measurements, were also calculated.
Results: The duplex sensitivity in detecting significant lesions ranged from 0.46 to 0.88. The Kappa values of agreement between duplex and angiography ranged from 0.35 to 0.64. Weighted Kappa values ranged from 0.45 to 0.72. Median grades of arterial lesions ranged from 2 (interquartile range 1-2) to 4 (3-4) and differed significantly (Kruskal-Wallis H test, p < 0.001).
Conclusions: The sensitivities and specificities suggested various duplex reliabilities in detecting significant arterial disease across different lower limbs segments. The Kappa values, in general, testified duplex's insufficient accuracy in grading the severity of stenosis. However, weighted Kappa values confirmed duplex's better ability to approximate the grade of stenosis. The arterial segments themselves were unevenly affected with the peripheral arterial disease.