Magnetic resonance imaging of rectum : Diagnostic and therapy related aspects

Sammanfattning: Papers I-II: The purpose of paper 1 was to assess the size and configuration of the perirectal fatty tissues (PF) using magnetic resonance imaging (MRI). In 25 subjects the volume and cross-sectional parameters based on the amount of PF to different sides of die rectum, and the total area occupied were retrospectively measured on MRI. There was a good correlation between anteroposterior diameter of the PF at four centimeter below S 1-2 (and the left-to-right diameter seven centimeter below S 1-2), and mesorectal volume (NW). Furthermore, the form of PF differed significantly between male and female subjects. In paper II, we analyzed the influence of MV on the accuracy of the first preoperative MRI. 267 patients with rectal cancer had their MV measured without knowledge of the prospective evaluations by the radiologist or the pathologist, and the discrepancies in the results were correlated to the MV and clinical data. T- or N-staging accuracy by MRI did not significantly correlate to MV. The difference between assessment by radiologist and pathologist did not differ based on MV. Finally patients with larger MV did not have fewer cases with involvement of mesorectal fascia (MF) or involvement of neighboring organs. Thus, MV does not appear to affect the locoregional prognostic factors, nor is it able to explain the difference in evaluation between the radiologist and pathologist. Papers II-IV: In paper III we tried to find out if the tumor size on MRI in patients without preoperative radiotherapy correlates to the corresponding pathologic findings. 18 patients were included. The tumor size was measured on MR and histopathologic specimen. Regression curves showed best correlations for area (r2=0.75) and volume (r2=0.65-0.82). With the formula proposed from this material, we assume that rectal tumors can be measured on MR images using a metric model, and then extrapolated to what we would expect from pathology, hence providing us with a tool where we could measure tumor response after neoadjuvant therapy. In paper IV, we used these tools to evaluate changes after radiotherapy and correlation between MRI and histopathology. RVs was defined as the residual pathologic tumor volume while RVm was similarly defined as the residual MRI tumor volume at 2 nd MRI. 25 patients with MRI before and after radiotherapy were included. The second MRI was not more accurate than the initial MRI for assessment of the T-stage or distance to circumferential resection margin (CRM). RVm showed significant correlation to RVs and pathologic T-stage. A 2 nd MRI alone after radiotherapy with delay before surgery has limited value in understanding the individual response to therapy, but followup volumetry can be helpful to understand which tumors have responded. Paper V: The aim of this study was to determine the sites of local recurrence (LR) following radical total mesorectal excision (TME) for rectal cancer in an effort to elucidate the reasons for recurrence. 33 CT and MRI of 37 patients with LR were examined. 29 LR were found in the lower two-thirds of the pelvis, with two appearing to originate from lateral pelvic lymph nodes (LN). Evidence of residual PF was identified in 15 patients. 14 of the LR originated from primary tumors in the upper rectum and 12 of them with evidence of residual PF. Lateral pelvic LN metastases are not a major cause of LR after TME. Partial mesorectal excision may be associated with an increased risk of LR from tumors in the upper rectum.