Quantitative semi-automatic 3D assessment of uterine fibroids after intra-arterial embolization - a feasibility study


      Uterine artery embolization (UAE) is an effective alternative to surgical treatment. While contrast-enhanced MRI (ceMRI) continues to be the gold standard to assess anatomic response to treatment, no standardized criteria exist to date. Most methods measure diameters on a representative slice and evaluate enhancement based on individual visual perception. We propose to introduce a reproducible, quantitative approach calculating the exact total lesion volume (TLV) and lesion viability expressed as enhancing lesion volume (ELV).

      Materials and Methods

      This study included 22 patients with uterine fibroids, which were treated with UAE and underwent pre- and post-procedural ceMRI. TLV and ELV were calculated as follows: 1) A semi-automatic 3D segmentation of the dominant lesion was performed on the ceMRI scan to calculate TLV 2) The pre-contrast MRI scan was subtracted from the ceMRI scan. 3) After subtraction, the 3D segmented volume was applied on the difference image and statistics were calculated in ROIs representing normal soft tissue. 4) ELV was defined as areas of TLV where the enhancement exceeded the value of the ROI. ELV was expressed in % of TLV and in cm3. Patients were classified as responders, if symptoms (e.g. menorrhagia, pain) showed improvement at a 6-month follow-up. Statistics were calculated using the paired t-test and the Mann-Whitney test.


      The mean pre-procedural TLV was 130.2cm3 [5.35-522.1cm3] and the mean ELV was 99.2% [43.3-100%]. UAE achieved a reduction of TLV in 21 (95%) of the patients with a mean post-procedural TLV of 76.6cm3 [0.5-340.9cm3] (p<.001) and a reduction of ELV in all patients with a mean ELV of 37.8% [4.6-97.6%] (p<.001). All patients with a residual ELV of <10cm3 (N=10) were responders, while 50% (N=6) of patients with a residual ELV >10cm3 (N=12) were non-responders. Non-responders (N=6) had a significantly higher TLV (p=.021) as compared to responders (N=16) on baseline MRI.


      Our method is feasible; it eliminates subjectivity in assessing lesion response after UAE and provides an accurate method to quantify anatomic response to treatment. The implied correlation with clinical outcome deserves further investigation in a larger cohort.