Evaluation des diagnostischen Stellenwertes der kontrastmittelverstärkten dynamischen MR-Angiographie des Unterschenkels

The peripheral arterial occlusive disease (PAOD) belongs to the spectrum of cardiovascular disorders, which is the most common reason of death in Germany. The prevalence of PAOD rises significantly with increasing age. About 20 % of men older than 65 years are affected. Therefore an early and correct diagnosis is vital for therapy and further outcome of the patient. Diagnostic methods are assessing medical history, physical examination, especially the evaluation of the ABI, and advanced medical imaging. Until today conventional digital subtraction angiography (DSA) is considered to be the gold standard among the imaging diagnostic methods. Because of different risks (radiation exposure, contrast media reactions, nephrotoxicity) and the complications associated with invasive methods (bleeding, infection) an alternative imaging technique is desired. Because of the technological progress 3D CE-MRA has shown to be one of the most promising choices. Aim of the study was to compare time-resolved 4D CE-MRA-sequences (TWIST) with high spatial resolution 3D CE-MRA-sequences (MIP) to the gold standard DSA by detecting clinical relevant stenosis in the calf. The evaluated vessel segments were the tibiofibular trunk, the anterior and posterior tibial artery and the fibular artery. Sensitivity and specificity were evaluated. Also, the reader agreement and the agreement between the two sequences (MIP, TWIST) of the corresponding reader has been evaluated. Furthermore image quality was recorded and evaluated. In this study, 100 cases from a 17 months time period were taken into account. The gender ratio was 60 % men and 40 % women, while the average age was 72.2 years. Evaluating the degree of stenosis by DSA, about 65 % of the cases showed clinical relevant stenosis, which means a stenosis of more than 51 %. 30 % of the cases did not show any stenosis in DSA and about 5 % showed no clinical relevant stenosis (stenosis less than 50 %). Concerning the evaluation of image quality it became apparent that image quality of MIP has been lower than the ones of DSA and TWIST. Analysing the statistics, it became apparent that high spatial resolution MIP-sequences showed inferior results than time-resolved TWIST sequences. The average sensitivity of MIP was 62.5 % and the average specificity was 48 %. In comparison, the average sensitivity of TWIST was 82 % and the average specificity was 59.5 %. The investigator comparison it became noticeable that the less experienced investigator surveyed a lower quality of MIP results. To verify the impact of image quality on statistical results, the data was overhauled. Only cases with very good and good image quality were included and were reevaluated. This led to overall better indices, increasing the average sensitivity of MIP to 84.5 % and of TWIST to 91 %. The average specificity of MIP and TWIST increased to 71.5 %. The reader agreement for MIP sequences was moderate (ĸ = 0.49) while it was substantial (ĸ = 0.65) for TWIST sequences. After overhauling the data reader agreement improved also. For MIP it became substantial (ĸ = 0.78) and for TWIST it was almost perfect (ĸ = 0.81). Investigator-independent method can be assumed because of the improved reader agreement. This was noticeable at both MRA sequences, whereat time-resolved TWIST showed a higher investigator independence than high spatial resolution MIP. The sequence agreement showed a significant better evaluation (p < 0.05) for clinical relevant diseased cases in DSA for TWIST sequences compared to MIP sequences. There was no significant difference in evaluation for healthy cases in DSA for MIP and TWIST. After overhauling the data there was no difference in evaluation for both healthy and diseased cases in DSA for MIP and TWIST. In conclusion the already used 3D CE-MRA is a potential alternative for DSA for diagnostics and therapy planning. Especially the additional acquisition of time-resolved TWIST sequences yields useful image material. If images of good quality are available MIP reconstructions are sufficient for evaluation. Because of the mixed image quality, in clinical routine a two step approach is suggested. First, a 3D CE-MRA of the complete lower extremity should be performed. This should be followed by an immediate analysis of image quality of the calf region. If image quality is considered suboptimal, an additional time-resolved TWIST sequence has to be acquired.