Preoperative evaluation of local staging and surgical retroperitoneal margin of colon tumors using multidetector computerized tomography

PURPOSE:This study aims to evaluate preoperative T, N stages and retroperitoneal surgical margin (RSM) in colon cancer by using multi-slice CT; furthermore, providing comparisons between diagnostic results of histopathologic examination as the gold standard.MATERIALS AND METHODS:In this retrospective study, 141 cases during the time period between January 2008 and January 2012, which are diagnosed as colorectal adenocarcinoma after histopathological analysis and also have multi-slice CT scans in the radiology archives of our institution, are evaluated in terms of tumor localizations, T-N stages of tumors and retroperitoneal surgical margin by two observers. Statistical analyses are provided by using gold standard pathological data.RESULTS:In determining extramural invasion, sensitivitiy, specificity, pozitive predictive value, negative predictive value, ratio of accurarcy was respectively, 81-87%, 50-75%, 95-97%, 26-27% and %81-84; besides a slight agreement between pathology and two observers. Moderate interobserver agreement was observed (?=0.425). In determining T stage, under, true ve over staging results -respectively- were 31.9%, 54.6%, 13.5% for the first observer and 23.4%, 51.1%, 25.5% for the second one. Determining lymph node metastasis sensitivitiy, specificity, pozitive predictive value, negative predictive value, ratio of accurarcy was respectively, 84%, %46-56, %60-65, %74-78, %64-70 with a slight agreement between pathology and two observers. Interobserver agreement was substantial (?=0,650).RSM involvement was present in 6 (4.7%.) cases. When only retroperitoneal colon segments are considered, 1.6% of the cases demonstrate RSM involvement. Four of the RSM positive six cases were located on sigmoid colon and one case was on transverse colon and caecum. Considering only retroperitoneal colon tumors and all colon tumors respectively, first observer?s sensitivity values were 0%- 33%, specificity values were 82.5%-81%, while same values for the second observer were %100-%50, %73-%80. A moderate interobserver agreement was obtained (?=0,518).CONCLUSION:In our knowledge our study is the the largest series to date of evaluating RSM with MDCT. In our study preoperative determining of clinically important extramural invasion is found 81-87% sensitivitity higher than other studies. Due to assuming lymph node cut-off in N staging as 5mm, our study results are observed to be more sensitive and with lower specificity value compared to the other studies in the literature. As in other studies, it is established that usage of oral contrast media has no effect on the sensitivitiy, specificity of T staging. As a novel observation, our study concludes that rectal contrast materials also has no effect in correctness of T stage diagnosis.Radiologic evaluation of the RCS is available in the literature only one study that contains lower number of patients compared to our study. RCS is evaluated in the retroperitoneal colon sections in the literature, however, there is no such a case series evaluating RSM involvement of the sigmoid colon. Colon may show anatomic and individual differences in patients. Taking this into consideration, not only retroperitoneal ascending and descending colon cancer, but also sigmoid and transverse colon RCS involvement may be present should be kept in mind.