RETURN TO ABSTRACT LISTING


Objective:
To assess the role of post Gadolinium-Based Contrast Agent (GBCA) MRI in identifying patients with pathologic complete response (pCR) in mucinous rectal cancer (MRC) using dynamic contrast-enhanced (DCE) MRI and static images.

Materials and Methods:
This initial retrospective study included 42 consecutive patients (M/F 23/19, mean age 57 y.) with MRC who had both baseline and post-chemoradiotherapy (CRT) MRI prior to surgical resection. Sagittal DCE MRI with adequate image quality was identified in 35 patients. Axial contrast-enhanced (CE) (pre- and post-contrast) 3D T1w images were identified in 42 patients. One radiologist segmented rectal tumors on the sagittal DCE images and 2 radiologists segmented tumors on the axial T1W pre- and post-contrast images. All patients had surgery and pathologic response was retrieved from pathology reports and divided into pCR and non pCR. From DCE images, pharmacokinetic parameters Ktrans and kep were estimated by fitting DCE data to Toft’s perfusion model using a population-based arterial input function. On axial contrast-enhanced images, mean enhancement (enh) was estimated. Relative change in Ktrans (?Ktrans), kep (?kep), and enh (?enh) after CRT were calculated. Pre-CRT, post-CRT, and the relative change of all parameters were compared to pathological response. The area under the ROC Curve (AUC) was used to measure overall diagnostic accuracy.

Results:
In the 35 patients who underwent DCE, histopathology showed pCR: n=9 (26%), and non-pCR: n=26. In distinguishing patients with pCR, pre-CRT Ktrans, pre-CRT kep, and ?kep were statistically significant (p-values 0.047, 0.007, and 0.033, respectively) yielding AUCs of 0.73, 0.81, and 0.74, respectively. In the 42 patients who underwent CE, pCR: n=8 (19%), and non-pCR: n=34 . In distinguishing patients with pCR, for reader 1, enh(pre-CRT), enh(post-CRT), and ?enh (p-values=0.045, 0.039, and 0.0041, respectively) were statistically significant. For reader 2, enh(pre-CRT), and ?enh (p-values=0.02 and 0.006, respectively) were statistically significant. For ?enh, the AUC was 0.83 for reader 1, and 0.82 for reader 2

Conclusion:
Using GBCA, DCE and CE MRI shows promise in the prediction of treatment response in MRC. Pre-CRT Ktrans, pre-CRT kep, and ?kep were statistically significant between patients with pCR and non-pCR. For both readers, ?enh was significantly different between patients with pCR and non-pCR.