Diagnostic performance of magnetic resonance to assess treatment response after neoadjuvant therapy in patients with locally advanced rectal cancer

Purpose Our study aimed to evaluate the diagnostic performance of rectal magnetic resonance imaging (MRI) for local restaging in patients with non-metastatic locally advanced rectal cancer (LARC) after neoadjuvant chemoradiotherapy (CRT) using surgical histopathology of total mesorectal excision as the reference standard. Methods Ninety-five patients with LARC who underwent rectal MRI after CRT between January 2014 and December 2016 were included. Accuracy, sensitivity, specificity, positive, and negative predictive value for local staging regarding T-stage, N-stage, circumferential resection margin, and MRI tumor regression grade (ymriTRG) were calculated, and inter-test agreements were assessed. Results 22/95 (23.2%) patients had radiological complete response (rCR), whereas 20/95 (21.1%) had pathological complete response (pCR). Among the patients with pCR, 11/20 (55%) had rCR. Fair agreement was demonstrated between ymriTRG and pathological TRG (ypTRG) (kappa = 0.255). The sensitivity and specificity for detection of pCR were 61.1% (95% CI 35.7-82.7) and 89.6% (95% CI 80.6-95.4). For the detection of ypTRG grades 1 and 2, the corresponding values were 67.2% (95% CI 54.3-78.4) and 51.6 (95% CI 33.1-69.8). The accuracy of ymriTRG was 24.2% (95% CI 15.6-32.8). Inter-test agreement in TRG between MRI and pathology was overall fair (kappa = 0.255) and slight (kappa = 0.179), if TRG 1 + 2. ] Conclusion Qualitative assessment on MRI for diagnosing pCR showed moderate sensitivity and high specificity, whereas the diagnosis of TRG had moderate sensitivity and low specificity with slight to fair inter-test agreement when compared with pathological specimens.