Could tumor characteristics identified by colonoscopy predict the locally advanced rectal carcinoma? / 中华医学杂志(英文版)

ABSTRACT

<p><b>BACKGROUND</b>Neoadjuvant chemoradiation is now considered the standard care for locally advanced rectal carcinoma (T3-4 or/and N1-2 lesions), but the accuracy of staging examinations including endorectal ultrasonography (ERUS) and MRI is far from excellent. In addition, the above staging equipment or professionals who perform the examinations may not be available in some hospitals, while preoperative colonoscopy and biopsy are usually obtainable in most hospitals. The objective of the present study was to investigate the clinical and pathological characteristics of locally advanced rectal carcinoma and identify candidates for neoadjuvant chemoradiation.</p><p><b>METHODS</b>This was a retrospective study. Patients who were treated for rectal cancer at Changhai Hospital from January 1999 to July 2008 were identified from our prospectively collected database. Statistical analysis was performed using SPSS Software System (version 15.0). The Mann-Whitney test, chi-square test and multivariate Logistic regression analysis were performed.</p><p><b>RESULTS</b>A total of 1005 cases were included in this research, of which 761 cases were identified as locally advanced rectal carcinoma depending on postoperative TNM staging. The results of multivariate Logistic regression analysis indicated seven independent risk factors that could be used to predict a locally advanced rectal carcinoma independently a high grade (including poor differentiation and undifferentiation) (OR 3.856; 95% CI 2.064 to 7.204; P = 0.000); large tumor size (OR 2.455; 95% CI 1.755 to 3.436; P = 0.000); elevated preoperative serum CEA level (OR 1.823; 95% CI 1.309 to 2.537; P = 0.000); non-polypoid tumor type (OR 1.758; 95% CI 1.273 to 2.427; P = 0.001); the absence of synchronous polyps (OR 1.602; 95% CI 1.103 to 2.327; P = 0.013); the absence of blood in stool (OR 1.659; 95% CI 1.049 to 2.624; P = 0.030); and a greater circumferential tumor extent (OR 1.813; 95% CI 1.055 to 3.113; P = 0.031). Based on these findings, a Logistic equation was established, the accuracy of which was 64% according to the information of the additional 50 cases.</p><p><b>CONCLUSIONS</b>Some independent risk factors related with locally advanced rectal carcinoma were identified, based on which it is possible to establish a Logistic equation as a tool to predict candidates of neoadjuvant chemoradiation. Further research about optimization of the equation is warranted.</p>