EP790 - An Updated Report On Preoperative MRI Compared To Pathologic Nodal Status In Rectal Cancer Patients Who Did Not Undergo TNT

Purpose/Background: Rectal cancer affects more than 40,000 adults in the United States each year. Magnetic resonance imaging (MRI) is the mainstay for preoperative tumor staging. However, sensitivity and specificity for MRI reported lymph node positivity is < 80%, leading to some debate as to how this value should be considered when determining treatment stratification. Neoadjuvant chemotherapy and radiation, collectively referred to as TNT, are usually recommended for patients with lymph node positive disease discovered on preoperative MRI.

Methods/Interventions: The internal REDCap rectal cancer database at our single, large-volume, NAPRC-accredited cancer center was queried for patients diagnosed and/or treated at our institution. In these patients, preoperative lymph node positivity was determined by rectal cancer protocol MRI using the SAR Primary Rectal Cancer Staging Template, in which positivity is determined by a combination of nodal size, shape, borders, and signal intensity.

Results/Outcomes: Over the past five years, 224 patients had rectal cancer surgeries that included total mesorectal excisions (TME) at our institution. Of these patients, 59 had a preoperative rectal cancer protocol MRI and did not undergo TNT. 12 of these patients had node positive MRI scans, while 47 patients had node negative MRI scans. Of the 12 patients with node positive MRI scans, only 2 were node positive on final pathology. Of the 47 patients with node negative MRI scans, 36 were node negative on final pathology. In this limited cohort, sensitivity of MRI for detecting lymph node positive disease was 15%, while specificity was 78%. Positive predictive value (PPV) was only 17%, and negative predictive value (NPV) was 77%.

Conclusion/Discussion: The overall sensitivity and specificity of rectal MRI has been previously validated. However, there is a paucity of recent literature directly comparing nodal disease status between MRI reports and pathologic specimens in patients who never underwent TNT. Although a relatively small cohort, our retrospective data supports previous concerns about the validity and usefulness of using MRI synoptic reported preoperative lymph node status when deciding on treatment algorithms for patients with newly diagnosed rectal cancer. More investigation is warranted.

Table 1
MRI status vs Pathological Status of Mesorectal Lymph Nodes