Zhang TY, Lin Y, Fan R, Hu SR, Cheng MM, Zhang MC, Hong LW, Zhou XL, Wang ZT, Zhong J. Potential model for differential diagnosis between Crohn's disease and primary intestinal lymphoma.
World J Gastroenterol 2016;
22:9411-9418. [PMID:
27895429 PMCID:
PMC5107705 DOI:
10.3748/wjg.v22.i42.9411]
[Citation(s) in RCA: 19] [Impact Index Per Article: 2.1] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 06/15/2016] [Revised: 07/28/2016] [Accepted: 08/19/2016] [Indexed: 02/06/2023] Open
Abstract
AIM
To evaluate the usefulness of different parameters to differentiate Crohn’s disease (CD) from primary intestinal lymphoma (PIL).
METHODS
The medical records of 85 patients with CD and 56 patients with PIL were reviewed retrospectively. Demographic, clinical, laboratory, endoscopic, and computed tomographic enterography (CTE) parameters were collected. The univariate value of each parameter was analyzed. A differentiation model was established by pooling all the valuable parameters. Diagnostic efficacy was analyzed, and a receiver operating characteristic (ROC) curve was plotted.
RESULTS
The demographic and clinical parameters that showed significant values for differentiating CD from PIL included age of onset, symptom duration, presence of diarrhea, abdominal mass, and perianal lesions (P < 0.05). Elevated lactate dehydrogenase and serum β2-microglobulin levels suggested a PIL diagnosis (P < 0.05). The endoscopic parameters that showed significant values for differentiating CD from PIL included multiple-site lesions, longitudinal ulcer, irregular ulcer, and intraluminal proliferative mass (P < 0.05). The CTE parameters that were useful in the identification of the two conditions included involvement of ≤ 3 segments, circular thickening of the bowel wall, wall thickness > 8 mm, aneurysmal dilation, stricture with proximal dilation, “comb sign”, mass showing the “sandwich sign”, and intussusceptions (P < 0.05). The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of the differentiation model were 91.8%, 96.4%, 93.6%, 97.5%, and 88.5%, respectively. The cutoff value was 0.5. The area under the ROC curve was 0.989.
CONCLUSION
The differentiation model that integrated the various parameters together may yield a high diagnostic efficacy in the differential diagnosis between CD and PIL.
Collapse