To evaluate the diagnostic value of CT plain scan and enhanced scan combined with multi mode reconstruction in colon tumor.

The clinical data of 96 patients with suspected colonic lesions who underwent colonoscopy examination in Guang′an People′s Hospital were retrospectively analyzed. Seventy cases of colon cancer and 26 cases of colon polyps were confirmed by pathology. All patients were scanned and enhanced by CT scan, and the scan data were sent to the post-processing workstation for multiplanar reconstruction (MPR), transparency (Raysum) and colonic virtual endoscopy (CTVC) images. The CT findings of colon cancer were analyzed. The preoperative CT images of colon cancer underwent T and N staging, and compared with the postoperative pathological findings.

CT scan and enhanced scan combined with multi mode reconstruction technique showed the detection rate of colon cancer lesions and colon polyps detection rate of 100%. CT scan and enhanced scan combined with multimode reconstruction showed no significant difference in the diagnosis of serosa, adjacent organs and retroperitoneal lymph nodes (P>0.05). Preoperative CT scan and enhanced scan with multi model reconstruction in the diagnosis of T staging, less than T₂ period, T₃ period and T₄ period accuracy were 84.29%, 75%, 73.33%, 88.10%, respectively. The overall accuracy of the diagnosis of N staging, N₀ staging, N₁ staging, N₂ staging were 71.43%, 72.73%, 60%, and 77.27%, respectively. The preoperative CT scan and enhanced scan combined with multi mode reconstruction technique showed good agreement between the diagnosis of T and N staging and pathological diagnosis (Kappa=0.725, 0.547, P<0.05).

CT scan and enhanced scan combined with multi mode reconstruction technique have high accuracy in diagnosis of colonic tumor lesions, and can better display the intestinal cavity of colon cancer, which can also determine the extent of tumor invasion and retroperitoneal lymph node metastasis. It can be used for preoperative staging diagnosis, and has higher clinical value.