Frequency and extent of uterine perfusion via ovarian arteries observed during uterine artery embolization for leiomyomas

OBJECTIVE. The objective of our study was to evaluate the frequency and extent of residual uterine perfusion via the ovarian arteries after bilateral uterine artery embolization (UAE) for the treatment of symptomatic uterine leiomyomas. MATERIALS AND METHODS. One hundred forty-five consecutive patients who underwent UAE were retrospectively evaluated for blood supply to the uterus via the ovarian arteries after UAE. After completion of UAE, uterine supply from the ovarian arteries was assessed by performing abdominal aortography in all patients. Selective ovarian arteriography, in addition, was performed in some patients. The criteria used to characterize ovarian artery perfusion as seen on the aortograms were vessel size compared with a 5-French catheter and visualization and extent of flow. When the ovarian arteries visibly supplied uterine tissue, a quantification system was applied as follows: the uterus was divided in 24 segments on the basis of a clock model that was superimposed over the uterine territory in the anteroposterior projection. Depending on its distance from the midpoint of the clock, perfusion segments were labeled as central, middle, distal, or peripheral for each hour of the clock resulting in a total of 24 (12 x 2) potential segments of residually perfused uterine tissue via the ovarian arteries. RESULTS. Two hundred ninety ovarian arteries were evaluated on aortography; of these, 202 (70%) were not seen during aortography. Of the visualized ovarian arteries (n = 88), 52% (46/88) were smaller than, 25% (22/88) were equal to, and 23% (20/88) were larger than the diameter of a 5-French catheter. The aortogram revealed that 61% (54/88) of the ovarian arteries extended into the pelvis, whereas 38% (33/88 [one missing data point]) did not. Selective injections were performed in 54 ovarian arteries. Of these, 69% (37/54) of the ovarian arteries had residual fibroid perfusion from the ovarian arteries after UAE (10 left-sided, 15 right-sided, six bilateral = 37 ovarian arteries). Residual fibroid perfusion was more likely in large ovarian arteries, particularly those with rapid flow visualized extending into the pelvis. The perfusion scores ranged from one to 18 segments (< 6 segments, n = 21 ovarian arteries; 6-12 segments, n = 12; > 12 segments, n = 4). Direct communication with the uterine arteries was seen in 20 ovarian arteries, 40% (8/20) of which did not show any uterine or fibroid perfusion, suggesting that fibroid flow had been occluded by UAE. CONCLUSION. Based on aortography, the presence of residual fibroid perfusion is more likely if the ovarian arteries are large, have rapid flow, or have flow that extends into the pelvis. Selective ovarian artery evaluation may be indicated in these cases to determine the extent of residual fibroid perfusion.