Report from ISMRM: Dark lumen MR colonography tops bright lumen imaging -- up to a point

Studies presented at the International Society for Magnetic Resonance in Medicine annual meeting extolled the virtues of dark lumen MR colonography. Although researchers found that 3D gradient-echo imaging is more effective and easier to perform than bright lumen imaging, they were reluctant to relegate bright lumen to the scrap heap.

Dr. Thomas C. Lauenstein of University Hospital in Essen, Germany, performed conventional colonoscopy and two MR protocols on 31 patients with suspected colorectal cancer. All subjects underwent standard bowel cleansing the night before imaging, and their colons were filled with tap water before MR imaging on a 1.5T scanner. Dual-array surface coils achieved complete abdominal coverage.

The dark lumen protocol used 3D gradient recovery echo imaging after intravenous administration of gadolinium contrast. Bright lumen imaging with TrueFISP did not require intravenous contrast, but this sequence had to be performed with the patient in both prone and supine positions to achieve reliable coverage of the colonic wall.

Dark lumen imaging accurately identified 16 of 20 polyps and all three carcinomas diagnosed with conventional colonoscopy. No false positives were reported. Bright lumen TrueFISP imaging detected 13 polyps and the three carcinomas, but five false positives weakened its performance, Lauenstein said.

Sensitivity and specificity rates of 83% and 100%, respectively, for dark lumen imaging compared with a sensitivity of 74% and specificity of 100% for the bright lumen technique.

An associated evaluation of the parenchymal organs identified hepatic metastases in two of the three patients with colorectal carcinoma. These findings were possible with both techniques, although the lesions were more conspicuous with the dark lumen technique.

Incidental findings led Lauenstein to conclude that TrueFISP is superior for identifying inflammatory bowel disease, warranting its use as a complementary imaging tool. His previous work popularized MR colonography as an alternative to CT colon fly-throughs.

3D GRE may have outperformed TrueFISP because the staff at University Hospital has more than two years of experience with the dark lumen technique.

"TrueFISP is a newer technique. Maybe over time it will show better results than it did in this study," he said.

A second study by Dr. Waleed Ajaj, also from University Hospital, produced promising results with a dark lumen technique using contrast-enhanced 3D VIBE. The approach tested on 50 patients produced results nearly identical to those generated by conventional colonoscopy for lesions larger than 5 mm. MR colonography uncovered 25 of 27 lesions found with conventional colonoscopy. MRI's sensitivity and specificity for lesions in this class were 93% and 100%, respectively.

Dark lumen MR was unable to identify any lesion smaller than 5 mm but did help diagnose two metachronous lesions not observed with conventional colonoscopy because of incomplete coverage. It correctly diagnosed 26 of 28 cases of colitis and four of five cases of diverticulitis.

A third study, presented by Dr. M.M.L deLen of the Academic Medical Center in Amsterdam, examined various types of fecal tagging and rectal filling strategies performed with bright lumen 2D HASTE and 3D Flash sequences. Oral gadolinium ingested the night before imaging in lieu of conventional colonic cleansing worked better than barium sulfate as a tagging agent, according to deLen.

Rectal filling with a water/gadolinium mixture helped uncover more polyps than water alone or air for colonic distension, she said. HASTE missed fewer polyps than 3D Flash, although Flash found some lesions that were not detected with HASTE.