Didier Bielen

Imaging detection and diagnosis of hepatocellular carcinomas (HCCs) and dysplastic nodules (DNs) in cirrhotic patients is important because the number, size, and type of focal lesions strongly influence patient management. Focal lesions detected by imaging examinations during pretransplantation evaluation were correlated with focal lesions detected during detailed pathological examination of 49 cirrhotic explant livers. Within 6 months before transplantation, color Doppler ultrasonography (US), contrast-enhanced computed tomography (CT), and magnetic resonance (MR) imaging were performed in 94%, 33%, and 55% of patients, respectively. In 2% to 8% of patients, different types of benign focal lesions were present, and a considerable proportion was interpreted as (pre)malignant on imaging examination. US detected only the largest HCCs (patient sensitivity, 40%; specificity, 100%) and no DNs. On a per-patient basis, contrast-enhanced CT and MR imaging had poor sensitivity (20% and 27%, respectively) and good specificity (100% and 94%, respectively) for DNs. Patient sensitivity and specificity of both techniques for HCC were reasonable (50% for CT, 70% for MR imaging) and good (79% for CT, 82% for MR imaging), respectively. Neither technique was able to detect smaller (pre)malignant lesions. As a consequence, 10% of patients underwent transplantation, although they exceeded the tumor number limit. Currently used imaging techniques cannot correctly determine the exact tumor burden in some cirrhotic patients. Regular contrast-enhanced MR examination of cirrhotic patients waiting for liver transplantation is the best tool for the early detection of (pre)malignant lesions.

To evaluate the feasibility of MR colonography (MRC) with air using two-dimensional (2D) T1-weighted fast spin-echo (T1wFSE) in patients scheduled for conventional colonoscopy (CC) after classic bowel preparation, and assess the ability of the technique to detect colonic lesions. The distention was sufficient for diagnosis, and the technique provided adequate delineation of the wall in the majority of segments. Residual fluid obscured the wall in different segments, especially in the ascending and descending colon (supine position) and in the cecum, transverse, and sigmoid colon (prone position). These findings were consistent with CT colonography. MRC visualized three lesions, missed one lesion >10 mm, visualized none of four lesions <5 mm, and yielded one false-positive lesion (5-10 mm). Missed lesions can be due to inconsistency in the slice positions between consecutive breath-holds, which is inherent to the multishot technique. Residual fluid may have obscured the smaller lesions. The shortcomings of the technique are limited coverage and signal drop-off at the borders of the field of view (FOV). Before multishot 2D T1wFSE colonography can become a valid screening method, improved patient preparation and a more practical technique are needed.

To compare thin-slice multidetector computed tomography colonography (CTC) that uses stool tagging with colonoscopy. One hundred fifty patients scheduled for colonoscopy underwent high-resolution CTC. An iodinated contrast agent was added to the preparation to tag the residual colonic fluid and stool. The effect of fluid tagging was assessed first. Sensitivity and specificity were calculated for two independent readers. In addition, values were recalculated separately for the first and last 75 patients. Tagging was optimal in 95.3% of the cases, and reader confidence was high. Sensitivities were 64.1%-66.7% (for the 2 readers) for 5- to 9-mm polyps and 91.7% for larger polyps. The overall specificity was 94.2% and 95%. Sensitivity improved during the study for both 5- to 9-mm polyps (from 54.2%-58.3% to 80%) and polyps larger than 9 mm (from 50% to 100%). Specificity changed nonuniformly. The combination of fluid tagging and high-resolution scanning in CTC showed high sensitivity and specificity, especially concerning sensitivity for polyps of 10 mm and larger.

We evaluated an automated polyp size measurement tool in computed tomographic colonography for its accuracy and value for patient risk stratification. A simulation program generated a raw data phantom with sessile and pedunculated polyps of known sizes using 120 to 140 kV and 50, 40, 20, 15, and 10 mAs. All polyps were measured by clicking on the polyp surface. Comparison of the calculated size with the known polyp sizes allowed calculation of reproducibility and accuracy. For patients with proven polyps, we also compared automated measurements with manual and endoscopic measurements to evaluate the effect on patient risk stratification. The automated measurement tool allowed accurate measurements. In the patient study, assignment to the correct size group was not significantly different from the radiologist's results. However, it slightly improved patient risk stratification by reducing both failed and unnecessary colonoscopy referral. An automated tool for polyp measurement in patients facilitates patient risk stratification.

The aim of this study was to evaluate the discomfort associated with CT colonography compared with colonoscopy and bowel purgation cleansing, and to evaluate patient preference between CT colonography and colonoscopy. In a total of 124 patients, scheduled for multidetector virtual CT colonography and diagnostic colonoscopy, patient acceptance and future preference were assessed during the different steps of the procedure (colon preparation, CT examination, and conventional colonoscopy). Patients who described contradictory findings between the degree of discomfort and their preference regarding follow-up examinations were retrospectively reinterviewed regarding the reason for this discrepancy. Colonoscopy was graded slightly more uncomfortable than virtual CT colonography, but the preparation was clearly the most uncomfortable part of the procedure. Concerning their preference regarding follow-up examinations, 71% of the patients preferred virtual CT colonography, 24% preferred colonoscopy, and 5% had no preference. Twenty-eight percent of the patients preferred virtual CT colonography despite that they thought it was equally or even more uncomfortable than colonoscopy. This was mainly due to the faster procedure (17 patients), the lower physical challenge (14 patients), and the lack of sedation (12 patients) of virtual CT colonography. Factors other than the discomfort related to the examinations play an important role in the patient's preference for virtual CT colonography, namely the faster procedure, the lower physical challenge, and the lack of sedation. Since the preparation plays a major decisive factor in the patient acceptance of virtual CT colonography, more attention should be given to fecal tagging.

The purpose of this study was to evaluate the accuracy of diffusion-weighted magnetic resonance imaging (DW-MRI) in differentiating HCC from benign cirrhotic lesions compared with conventional dynamic contrast-enhanced MRI. Fifty-five patients with cirrhosis underwent conventional and DW-MRI at 1.5 Tesla. Signal intensity ratios (SI(ratio)) of solid liver lesions to adjacent hepatic parenchyma were measured for b0, b100, b600 and b1000, and the apparent diffusion coefficients (ADC) were calculated. In 27 patients, imaging results were compared to histopathology, and in 28 patients, to imaging follow-up. Based on predetermined thresholds, sensitivity and specificity of DW-MRI and conventional MRI were compared. SI(ratio) was significantly different between malignant and benign lesions at all b-values (P < 0.0001). No significant difference in ADC was seen (P = 0.47). For detection of malignant lesions, DW-MRI with b600-SI(ratio) yielded a sensitivity of 95.2% compared to 80.6% for conventional MRI (P = 0.023) and a specificity of 82.7% compared to 65.4% (P = 0.064). The improved accuracy was most beneficial for differentiating malignant lesions smaller than 2 cm. DW-MRI with b600-SI(ratio) improved the detection of small HCC and the differentiation of pseudotumoral lesions compared with conventional MRI.

Perianal fistulization is the result of a chronic inflammation of the perianal tissues. A wide spectrum of clinical manifestations, ranging from simple to complex fistulas, can be seen, the latter especially in patients with Crohn disease. Failure to detect secondary tracks and hidden abscesses may lead to therapeutic failure, such as insufficient response to medical treatment and relapse after surgery. Currently, magnetic resonance (MR) imaging is the preferred technique for evaluating perianal fistulas and associated complications. Initially used most often in the preoperative setting, MR imaging now also plays an important role in evaluating the response to medical therapy.