Tomas Ripolles MD, Presenter: Nothing to Disclose

Maria Jesus Martinez Perez, Abstract Co-Author: Nothing to Disclose

Claudia P. Fernandez Ruiz, Abstract Co-Author: Nothing to Disclose

Jose Vizuete, Abstract Co-Author: Nothing to Disclose

Diana Patricia Gomez Valencia MD, Abstract Co-Author: Nothing to Disclose

Gregorio Martin-Benitez, Abstract Co-Author: Nothing to Disclose

To evaluate the usefulness of ultrasound as the initial diagnostic method for differentiating diverticulitis from colon cancer in patients with sigmoid colon stenosis, especially in the emergency setting.

Fifty-two patients with sigmoid stenosis were examined by US and CT during the period February 2006 - January 2013. Immediately after US or CT scans each stenosis was classified as malignant or benign. Off-site, two readers, who were unaware of the proven diagnosis, independently and retrospectively analyzed 13 different morphological ultrasound criteria retrieved from a literature review to differentiate between benign and malignant strictures. The two readers were asked to give a diagnosis of malignant, benign or indeterminate stenosis. Sensitivity, specificity and accuracy were calculated by considering the pathological analysis or by clinical follow up of at least one year. The interobserver agreement was calculated by the kappa statistics.

There were 22 sigmoid carcinomas and 30 diverticulitis. The on-site US results were 93% sensitivity, 96% specificity and 95% accuracy for the colon carcinoma diagnosis; CT sensitivity was 87%. The strongest sensitive morphological features for cancer were loss of normal layer structure (87%), length <10 cm (82%) and abrupt edges (82%). Adjacent lymph nodes (100%), wall thickness >15 mm (93%) and absence of diverticula (88%) were the most specific findings for carcinoma. For diverticulitis, the most sensitive and specific criteria were preserved mucosal folds and conservation of the inner layer (90 and 95,5% respectively). Pericolic fat infiltration or abscess were not good criteria for differentiating them. The agreement on morphologic features oscillated between 0,441 (length <10 cm) and 0,903 (lymph nodes), being >0,8 in 5 out of 13 features. Off-site US diagnosis, excluding 4 indeterminate cases, oscillated between 94-98% of accuracy or 95-100% of sensitivity. The interobserver agreement was 0,782, coinciding in the diagnosis of malignant or benign stenosis in 46 out of 52 cases.

Our experience suggests that diverticulitis can often be differentiated from colon carcinoma on the basis of some US findings described in the literature.

It is not possible to perform colonoscopy or CT-colonography to exclude carcinoma in patients with diverticulitis subjected to conservative management until inflammatory changes have subsided.

Ripolles, T, Martinez Perez, M, Fernandez Ruiz, C, Vizuete, J, Gomez Valencia, D, Martin-Benitez, G, Sigmoid Stenosis Caused by Diverticulitis versus Carcinoma: Can They be Differentiated by Ultrasound?.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13013177.html