EP970 - Does pathologic complete response correlate with complete tumor response on MRI after total neoadjuvant therapy for locally advanced rectal cancer?

Image or Table (Do Not add text here):

Purpose/Background: Total neoadjuvant therapy (TNT) has quickly become the standard of care for patients with locally advanced rectal cancer (LARC). Initial and restaging MRIs are crucial in determining treatment pathways as more patients undergoing TNT achieve clinical complete response and are offered non-operative management (NOM). For these patients, it is difficult to assess the accuracy of restaging MRIs, as failure of NOM may be multifactorial and not solely related to poor prediction on MRI. However, the correlation between pathologic complete response found after surgery and tumor response on restaging MRI may serve as an important quality measure.

Methods/Interventions: Under the US Rectal Cancer Collaborative IRB, all patients who underwent TNT for LARC at our single institution were included in this retrospective study. For patients who underwent radical oncologic resection, we compared tumor response on restaging MRI following completion of TNT for patients who had pathologic complete response versus residual cancer. Descriptive and sensitivity and specificity analyses were performed.

Results/Outcomes: From January 2014 through August 2024, 121 patients with LARC underwent TNT at a single institution. The mean age of the patients was 57 (24–82 years) and the study population consisted of 73 males (60.3%) and 48 females (39.7%). Of the 121 patients, 32 (26.5%) had a clinical complete response, 27 (22.3%) underwent NOM, 92 (76%) had radical oncologic resection and one patient had incomplete response but refused surgery and one had clinical complete response on MRI but are awaiting discussion of the final treatment plan. Among those who underwent surgery, 15 (16.3%) had pathologic complete response while 77 (83.7%) had residual cancer on final pathology.

60 patients underwent surgery and had a restaging MRI after completion of TNT. Imaging was performed on average 4.8 weeks after completion of all treatments. Overall, 83.3% had incomplete response and 16.7% showed complete or near-complete response on the final restaging MRI. For the 10 patients who had pathologic complete response, only 1 (10%) had complete or near complete response on restaging MRI. For the 50 patients who had residual cancer, 40 (80%) had incomplete response on MRI. The sensitivity of MRI for predicting pathologic complete response was 10% with a positive predictive value (PPV) of 9.1%. For patients with incomplete pathologic response, the specificity of restaging MRI was 80% with a negative predictive value (NPV) of 81.6%.

Conclusion/Discussion: Our findings suggest that MRI is effective at predicting residual cancer when incomplete response is noted on restaging imaging, but is much less reliable at identifying pathologic complete response, even when imaging shows complete or near-complete tumor response. Our study is limited by the relatively large number of operative patients who did not receive a final restaging MRI and the early timing for restaging MRI which could underestimate pathologic complete response. The low sensitivity and PPV highlight the need for additional modalities, such as endoscopy and physical examination, to improve the accuracy of restaging. As complete clinical response rates increase in the setting of TNT, it is critical to determine the optimal time to restage patients and to utilize a multimodal approach to guide treatment decisions for patients with locally advanced rectal cancer.