To evaluate the treatment failure (TF) rate of leiomyoma after uterine artery embolization (UAE) for uterine leiomyomas in cases of the presence of anastomoses between the ovarian arteries (OA) and uterine arteries (UA).

The results of 163 consecutive UAE for uterine fibroid were reviewed. Mean patient age was 42.8 years (range, 25 to 57 years). TF was evaluated according to the anastomoses between OA and UA on pre-embolization angiography. Magnetic resonance images (MRIs) were obtained at 1~6 months or 12 months after UAE. MRIs were gadolinium (Gd)-enhanced images and/or T2-weighted images. MRIs exhibited focal enhancement portion on fibroid and evaluated the TF rate of the leiomyoma in patients of presence of anastomoses between OA and UA.

Fifty six patients had anastomoses between UA and OA on pre-embolization angiography (56/163, 34.4%). Angiographic subtypes were type Ia (n=19), type Ib (n=16), type II (n=11) and type III (n=10). Of all patients, 10 patients showed the focal enhancements of the leiomyomas on follow-up enhanced MRIs (10/163, 6.1%). Three treatments failed in patients demonstrated type Ia (3/19, 15,8%). One had type Ib (1/16, 6.3%). Other 6 had no anastomoses. There was no TF rate difference between patients with communication (4/56, 7.1%) and without communication (6/107, 5.7%). However, TF rate in patients with type Ia communication (15.8%) was higher than that without communication (5.7%; P<0.05).

Type Ia utero-ovarian anastomoses communication could be a contraindication for embolization treatment for leiomyoma.

We studied to determine the usefulness of dynamic magnetic resonance imaging(MRI) in the preoperative evaluation of invasion of colorectal cancer and to compare its usefulness with the conventional CT. To observe the enhancement pattern of colorectal wall after iv administration of Gd-DTPA between normal and cancerous wall.

Twenty patients with colon cancer and 8 patients with rectosigmoid cancer, who were diagnosed between October 1997 and June 1998 by barium enema, colonoscopic biopsy were evaluated. The patients population consisted of 16 men and 12 women, with ages ranging from 46 to 68 years(mean 61years). Preoperative staging was done by conventional CT and dynamic MRI. All MR images were performed with 1.5T superconducting magneting unit(Vision, Siemens, Erlangen, Germany). 2D-FLASH(Fast Low-Angle Shot) sequence was used for the dynamic and delayed images(TR/TE/NEX/FA=72.5-117.3/4.1/1/80°), and acquisition time of 13-15sec.

For the dynamic images, five MR images were obtained with a single breath hold. Precontrast images(axial, or sagittal or coronal) was obtained first, and then dynamic MR images were obtained at 30, 60, 90, 120sec after intravenous injection of 0.1mmol/kg Gd-DTPA. Ten to fifteen delayed images were also obtained with the interval of 4-5 minutes with a single breath hold. Preoperative staging of CT and MRI were decided with a consensus by two radiologists. Pathologic staging were done by TNM classification.

The dynamic MR image-determined stage of colorectal cancer correlated with histopathologic findings in 2 of 3 pT2 tumor(67%), 19 of 21 pT3(90%), and 4 of 4 pT4 tumors(100%). MRI correctily diagnosed tumor deposits of involved lymph nodes in 16 patients, overall accuracy was 57%(16/28%). And the signal intensities after IV Gd-DTPA administration between the cancerous wall and normal wall ws not significantly different at the 30, 60, and 90 seconds MR images with the indifferent Student t-test(p>0.05).

Dynamic MRI has a role for the preoperative assesment of colorectal carcinoma.