Abstract:Objective To explore the value of perfusion CT quantitative analysis for predicting tumor regression grade (TRG) after chemoradiotherapy in patients with rectal cancer. Methods From June 2016 to June 2018, 94 rectal cancer patients diagnosed and treated in Cangzhou Central Hospital of Hebei Province were selected and were divided into reaction group (TRG 3-4) and nonreaction group (TRG 0-2) according to the results of surgical specimens. Perfusion CT was performed in both groups before treatment, and chemoradiotherapy and surgery were used. Baseline data and perfusion CT results including blood flow, blood volume, mean transit time (MTT), permeability surface (PS) were compared between the two groups, and receiver operating characteristic (ROC) curve was used to evaluate the predictive efficacy of perfusion CT indexes for chemoradiotherapy responsiveness. Results In this study, a total of 23 cases (24.47%) were responsive to chemoradiotherapy, and 71 cases (75.53%) were not responsive to chemoradiotherapy. Blood flow in reaction group ［(38.60±7.13) ml·100 g^-1·min^-1］ was significantly lower than that in non-reaction group ［(67.39±11.33) ml·100 g-1·min-1,t=3.273,P=0.001］. MTT in reaction group was significantly longer than that in non-reaction group ［(11.12±2.19) svs.(6.88±1.32) s,t=4.500,P＜0.001］. There was no significant difference in blood volume ［(4.62±0.73) ml/100 gvs.(5.01±1.04) ml/100 g］ and PS ［(13.72±3.82) ml·100 g^-1·min^-1vs.(11.40±2.59) ml·100 g^-1·min^-1］ between the two groups (t=0.818,P=0.415;t=0.409,P=0.683). The best cut-off points of blood flow and MTT for predicting chemoradiotherapy responsiveness were 50.89 ml·100 g^-1·min^-1and 8.99 s, the area under the curve (AUC) was 0.825 and 0.922, and the AUC of combined prediction of chemoradiotherapy responsiveness was 0.982, which was significantly better than that of single prediction (Z=2.868,P=0.004;Z=2.051,P=0.004). The accuracy (91.49%) and specificity (90.14%) of combined prediction of chemoradiotherapy responsiveness were significantly better than those of single prediction (blood flow: accuracy 75.53%, specificity 73.24%; MTT: accuracy 79.79%, specificity 78.87%), and the differences were statistically significant (χ²=8.800,P=0.012;χ²=6.766,P=0.034). Conclusion Blood flow and MTT in perfusion CT have great predictive value for chemoradiotherapy responsiveness in patients with rectal cancer.

Key words:Tomography,Rectal neoplasms,Neoplasm grading,Efficacy evaluation