Locally advanced rectal cancer: The value of magnetic resonance imaging as a predictive tool.

Abstract

Listen

Translate article

e14666 Background: Magnetic Resonance Imaging (MRI) of pelvis stratifies patients (pts) with rectal cancer (ca) according to locally advanced disease (dx) & risk of local recurrence. By determining the extent of lymph node (LN) positivity, it enables the appropriate selection of pts for neoadjuvant therapy (NAT) prior to curative surgical resection (Sx). We assessed our institution’s experience of NAT versus upfront Sx in stage I-III rectal ca to ascertain the utility of MRI as a predictive tool in LN status evaluation. Methods: We retrospectively reviewed a prospectively maintained database for all pts with a diagnosis of stage I-III rectal ca from January 2006 to September 2012 in a specialist colorectal cancer centre. We analysed data with respect to preoperative MRI staging & definitive histopathological confirmation of LN stage. Results: A total of 210 pts were identified that had Sx for rectal ca with curative intent. Of these, 112 pts received NAT while 98 had upfront Sx. Of those who proceeded directly to Sx, there were 41 females & 57 males. Average age in this group was 73.4 years (range 33 -90). LN staging by MRI was accurate in 45.9% (n=45). LN status could not be evaluated (Nx) in 17.3% (n=17). In 19.4% (n=19) more advanced LN staging was apparent on histology while in 17.3% (n=17) LN stage was deemed more aggressive by MRI. Of those who had NAT, there were 75 males & 37 females. Average age was 62.3 years (range 26-82). In this cohort, 81.3% (n=91) had both a preoperative MRI & pathology report available for comparison. MRI in this group matched histology in 61.5% (n=56). Nx was recorded in 3.3 % (n=3). In 17.6% (n=16) more aggressive LN involvement was evident at histology while 17.6% (n= 16) had less advanced dx at tissue sampling. Overall in this study, MRI accurately predicted LN status in 53.4% (n=101). Conclusions: While MRI is a valuable tool in determining those pts with rectal ca that would benefit from NAT, its predictive value has limitations. With further analysis of the dataset at our institution, we aim to identify other factors involved & create a predictive nomogram for the rectal ca pt with locally advanced dx. We plan to validate this work by including data from other Irish cancer centres.