Predictive value of magnetic resonance imaging characteristics before and after radiotherapy for the occurrence of severe radiation-induced late rectal injury in patients with rectal cancer / 中华胃肠外科杂志

ABSTRACT

Objective: Severe radiation-induced late rectal injury (sRLRI) directly affects the quality of life of patients with rectal cancer. Effective prediction of sRLRI before surgery may provide important information for the selection of surgical strategies and perioperative managements. The purpose of this study is to evaluate the feasibility of predicting sRLRI based on magnetic resonance imaging (MRI) features before and after radiotherapy for rectal cancer. Methods: This was a diagnostic study. Clinical and imaging data of 90 patients with rectal cancer receiving long-term radiotherapy from June 2013 to July 2018 in the Sixth Affiliated Hospital of Sun Yat-sen University were collected retrospectively. Case inclusion criteria: (1) rectal cancer was diagnosed by pathology and age of ≥ 18 years old; (2) patients received neoadjuvant chemoradiotherapy and anterior rectal resection; (3) follow up time ≥ 3 years; (4) patients had no history of other neoplasm. Exclusion criteria: (1) patients did not receive MRI examination in our hospital within 2 weeks before and/or 8 weeks after radiotherapy; (2) images were not good enough for evaluation; (3) medical records were incomplete; (4) patients had severe gastrointestinal diseases. According to the RTOG/EORTC classification criteria for radiation reactions, severe complications of grade 3-4 requiring surgical management were defined as sRLRI. T2WI and DWI images before and after radiotherapy were evaluated. The rectal wall thickness, bladder wall thickness, rectal sacral spacing and apparent diffusion coefficient (ADC) were measured. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value of the above indicators for sRLRI. Results: Among the 90 patients with rectal cancer, 34 (37.8%) developed sRLRI. Before radiotherapy, the median rectal wall thickness of sRLRI and non-sRLRI patients was 4.530 mm and 4.355 mm, respectively; the median bladder wall thickness was 3.962 mm and 3.868 mm, respectively; the median rectal sacral spacing was 15.557 mm and 12.433 mm, respectively; the median ADC value of rectal wall was 1.620 ×10(-3) mm(2)/s and 1.653 ×10(-3) mm(2)/s, respectively. There were no significant differences in above indicators between sRLRI and non-sRLRI patients (all P>0.05). After radiotherapy, compared with non-sRLRI patients, sRLRI patients had increased rectal wall thickness (median: 8.239 mm vs. 6.223 mm, Z=-3.512, P=0.001), rectal sacral spacing (median: 17.728 mm vs. 13.885 mm, Z=-2.247, P=0.025), and change of rectal wall thickness after radiotherapy (median: 98.106% vs. 49.584%, Z=-4.169, P<0.001). After radiotherapy, there were no significant differences in the bladder wall thickness and its change value, the ADC value of rectal wall and its change rate before and after radiotherapy between the two groups (all P>0.05). The area under the curve (AUC) of the change rates of rectal wall thickness after radiotherapy, rectal wall thickness and rectal sacral spacing after radiotherapy for predicting sRLRI was 0.763, 0.722 and 0.642, respectively, while the sensitivity was 85.3%, 70.6% and 76.5%, respectively, and the specificity was 64.3%, 71.4% and 57.1%, respectively. Conclusion: Based on MRI examinations, assessments of rectal wall thickness after radiotherapy, the change rate of rectal wall thickness after radiotherapy, and rectal sacral spacing after radiotherapy are helpful for evaluating the risk of sRLRI after radiotherapy for patients with rectal cancer.