| ObjectiveTo evaluate diffusion-weighted imaging(DWI)for assessment of pathological complete response(pCR)in locally advanced rectal cancer(LARC)after neoadjuvant chemoradiotherapy(CRT).Materials and methodsSeventy-three patients(median age,62 years;range,21-82 years;50 men and 23 women)who were diagnosed with LARC and underwent CRT with subsequent surgery and magnetic resonance imaging(MRI)in our hospital between January 2015 and December 2018 were prospectively enrolled in this study.According to postoperative pathology,they were divided into pCR group and non-pCR group.Clinical indexes before and after CRT(pre-/post-),including body mass index(BMI),carcinoembryonic antigen(CEA)and haemoglobin(HB),were recorded.Two radiologists,who were blinded to the histopathology results,independently reviewed pre-CRT MRI to evaluate tumor size,distance from the anal verge,circumferential resection margin(CRM),extramural vessel invasion(EMVI),T stage and N stage;placed three circular regions of interest(ROIs)of highest brightness signal intensity within the tumor on pre-and post-CRT DWI and copied the them to an apparent diffusion coefficient mapping to calculate tumor ADCs;drew manually on a section of pre-and post-CRT DWI to encompass the entire tumor lesion at the level where the largest diameter of the tumor was found to calculate the area of tumors(S);and further calculate their changes(?)and percentage changes(?%).Chi-square test,Student t test and Mann-Whitney U test were used to compare the variables between the pCR group and non-pCR group.Then,the factors with statistical significance were converted into categorical variables.Categorical variables indicated to be significantly related with pCR by multivariate logistic regression were incorporated into our scoring system to predict pCR.Diagnostic performance was evaluated by using the receiver operating characteristic(ROC)curves.ResultsTwenty(27.4%)patients were confirmed to have achieved pCR,while fifty-three(72.6%)did not.The patients in pCR group(median age,55 years;range,21-74 years)were younger than patients in non-pCR group(median age,63 years;range,29-64 years).There was no statistical significant difference in sex,BMI,CEA and HB between the two groups.The items assessed on pre-CRT MRI,including tumor size,distance from the anal verge,CRM,EMVI,T stage and N stage,were also not significantly different.The mean post-ADC(1.29×10-3 mm2/s),?ADC(0.35×10-3 mm2/s),AADC%(64.05%)and the median AS(110.65 mm2),?S%(76.43%)of pCR group was significantly higher compared to non-pCR group(1.14 ×10-3 mm2/s,0.35 ×10-3 mm2/s,44.92%,89.44 mm2,58.1%,respectively)(p<0.05).The median post-S of pCR group was significantly smaller compared to non-pCR group(38.18 vs 52.21 mm2,p=0.014).Conversely,the mean pre-ADC(0.79 vs 0.80 × 10-3 mm2/s,p=0.849)and pre-S(146.89 vs 147.22 mm2,p=0.504)of the pCR group did not differ significantly from that of the non-pCR group.Areas under the ROC curve(AUC)for identification of a pCR based on post-ADC,?ADC,?ADC%and post-S,AS,?S%,were 0.763,0.749,0.731and 0.688,0.660,0.718,respectively.Multivariate logistic regression indicated that age<57 years(OR=3.864,p=0.048),post-S<45 mm2(OR=8.416,p=0.003),post-ADC>1.20×10-3 mm2/s(OR=12.248,p=0.001)were independently associated with pCR,which were incorporated into our scoring modal to predict pCR.Score no less than 4 was a significant predictor for pCR.The AUC of this scoring system was 0.861(p<0.001)and the sensitivity,specificity,positive predictive value,negative predictive value and prediction accuracy were 70%,89%,68%,89%and 82%,respectively.Conclusion1.The patients with LARC from pCR group were younger than those from non-pCR group.Clinical indexes and items assessed on pre-CRT MRI are not reliable.2.The post-ADC,?ADC,?ADC%and post-S ?S,?S%are useful for differentiating pCR from non-pCR after CRT in patients with LARC.3.Age,post-S and post-ADC are significantly associated with pCR after CRT.A scoring modal based on these three factors may be valuable to predict pCR. |
PDF Full Text Request