MRI And Radiomics Values In Predicting Treatment Response And Outcomes Of Rectal Cancer Patients Treated By Neoadjuvant Chemoradiotherapy

Part 1 MRI Morphologic and Clinicopathologic Characteristics for Predicting Survival Outcomes in Patients with Locally Advanced Rectal CancerPurpose:The aim of this study was to analyze the relevance of more MRI morphologic and clinicopathologic factors with outcomes in patients with locally advanced rectal cancer(L ARC).Material and Methods:115 patients(39 females and 76 males;median age,54 years;age range,28-82 years)with LARC treated with neoadjuvant chemoradiotherapy(nCRT)were enrolled in this retrospective study.Pre-and post-nCRT MRI morphologic and clinicopathologic variabilities were evaluated in all patients.3-year DFS was estimated using Kaplan-Meier product-limit method,and Cox proportional hazards models were used to determine associations between morphologic or clinicopathologic variabilities and survival outcomes.Results:With a median follow-up of 48.0 months,the 3-year DFS was 79.0%for all patients.During follow-up,18 patients died,28 patients experienced relapse(26 distant,1 local and 1 both),and 69 patients were censored.MRI-detected extramural vascular invasion(mrEMVI)was the only significantly independent factor of long-time survival,while HR was 2.308(95%CI 1.151-4.629,P=0.018)on univariate and 2.495(95%CI 1.243-5.012,P=0.010)on multivariate analysis.The 3-year cumulati-ve survival rates in patients with mrEMVI negativity compared with positivity were 86.6%vs 65.0%(P =0.015),respectively.Conclusion:In conclusion,pre-nCRT mrEMVI status was the independent significant risk factor for long-time outcomes in LARC patients treated with nCRT.Precise evaluation of mrEMVI status facilitates individualized treatment and improves long-term outcomes of LARC patients.Part 2 To Compare the Predictive Value of the Radiomics Siganature Extrated from MRI Plain or Enhancement Imaging for the Survival of Rectal CancerPurpose:To compare the value of radiomics signature based on magnetic resonance imaging(MRI)plain and enhancement sequence for predicting the disease-free survival(DFS)of locally advanced rectal cancer(LARC)patients.Material and Methods:Fifty-one patients with rectal adenocarcinoma were retrospectively analyzed.All patients underwent neoadjuvant chemotherapy(nCRT)followed total mesorectal excision(TME).Follow-up time for the survival patients were more than 3 years.The image segmentation was performed on the oblique axial T2WI sequence of the high resolusion(HR)small FOV(Field of view,FOV)and the multi-phase enhancement sequence,respectively.Least absolute shrinkage and selection operator(LASSO)Cox regression was applied to select radiomics features and the imaging signature was constructed.According to the radiomics score of each patient,the patients were divided into the high risk group with shorter DFS and the low risk group with longer DFS.The differences in the DFS of the high and low risk groups were compared using Kaplan-Meier survival curves in the training and validation cohort,respectively.The predictive ability of the model was evaluated by concordance index(C index).Results:The DFS of the high and low risk groups on the enhancement sequence training and validation cohort were 0.0002 and 0,00913 respectively.The C index of the model were 0.904 and 0.700,respectively.The model has the better ability to predict survival.The DFS of the high and low risk groups on the plain sequence training cohort was statistically different,the P value was 0.005,the model C index was 0.711.There was no statistical difference on the plain sequence validation cohort,the P value was 0.767,the C index was only 0.500.Conclusion:Radiomics signature extracted from MRI venous phase enhancement sequence superior to plain sequence for predicting the DFS of LARC patients before nCRT.Part 3 A Novel Radiomic Signature as a Prognostic Biomarker for Locally Advanced Rectal CancerPurpose:To construct radiomics signature for predicting locally advanced rectal cancer(LARC)patients disease-free survival(DFS)and to compare the ability of a radiomics model to predict DFS with that of a clinicoradiologic risk model in individual patients with LARC.Method:108 consecutive patients with LARC treated with neoadjuvant chemoradiotherapy(nCRT)followed by total mesorectal excision(TME).The ITK-SNAP software was used for manual segmentation of three-dimensional pre-nCRT MR images.All manual tumor segmentations were performed by a gastrointestinal tract radiologist.A radiomic signature based on 485 three-dimensional features was generated using the least absolute shrinkage and selection operator(LASSO)Cox regression model.The association of the radiomic signature with DFS was investigated by Kaplan-Meier survival curves.Survival curves were compared by the log-rank test.Three models were built and assessed for their predictive values,using the Harrell concordance index and integrated time-dependent area under the curve.Results:The novel radiomic signature stratified patients into low-and high-risk groups for DFS in the training set(hazard ratio[HR]=6.83;P<0.001),and was successfully validated in the validation set(HR=2.92;P<0.001).The model combining the radiomic signature and clinicoradiologic has the best performance(C index=0.788,95%CI 0.72-0.86;iAUC of 0.837 at 3 years).Conclusion:The novel radiomic signature could be used as an independent biomarker to predict DFS in patients with LARC.Furthermore,combining this radiomic signature with clinicoradiologic features significantly improved the ability to estimate DFS,and may help guide individualized treatment in such patients.Part 4 MRI Texture Analysis in Predicting and Early assessment Treatment Response to Neoadjuvant Chemoradiotherapy in Locally Advanced Rectal Cancer PatientsPurpose:To evaluate the value of MRI texture analysis in prediction treatment response before and early neoadjuvant chemoradiotherapy(nCRT)in patients with locally advanced rectal cancer(LARC).Methods:This retrospective study comprised of 59 patients who underwent MRI examinations thrice,only the pre-CRT(before nCRT)and early CRT(10-15 fractions after nCRT initiation)were used for analysis.The tumoral texture parameters(including mean value,standard deviation(SD),skewness,kurtosis,uniformity,energy and entropy)were compared between pre-and early nCRT.Area Under receiver operating characteristic(ROC)Curves[AUCs]after computing ROC curves were used to compare the diagnostic performance of statistically significant difference parameters and logistic regression analysis predicted probabilities for discriminating responders and nonresponders.Results:Standard deviation,kurtosis,and homogeneity were statistically different pre-and early nCRT(P values were 0.0012,0.0001,and<0.0001,respectively).The AUC of pre-nCRT kurtosis for prediction tumor regression grading(TRG)was 0.67(95%Cl 0.54-0.79).The logistic variable 1 obtained from the multiparameter logistic regression analysis of three independent variables of homogeneity,energy,and entropy before nCRT,and the AUC of logistic variable 1 for prediction pathologic complete response(pCR)was 0.76(95%Cl 0.63-0.86).Logistic regression analysis of multivariate logistic regression analysis of the four variables of pre-nCRT standard deviation,homogeneity,energy,and entropy yielded an AUC of 0.78(95%CI 0.66-0.88).Conclusion:MRI texture parameters as imaging biomarkers have the potential to prediction tumoral treatment response to nCRT in patients with LARC.