3D Magnetic Resonance Imaging And Multimodal Radiomics In The Preoperative Diagnosis Study Of Rectal Cancer

PART Ⅰ Diagnostic study of 3D magnetic resonance imaging in preoperative evaluation of rectal cancerObjective: T2 weighted imaging is the most important magnetic resonance imaging sequence in the preoperative evaluation of rectal cancer.The imaging technology of 3D T2WI sequence is beneficial to achieve higher Inter-layer resolution,higher signal-to-noise ratio,and multiplanal reconstruction,and improve efficiency by simplifying the scanning process.The purpose of this study is to compare the diagnostic value of the 2D T2WI and 3D T2WI sequence for preoperative evaluation of patients with rectal cancer.Materials and Methods: This study was a prospective study that collected 243 rectal cancer patients who underwent MR examination at the First Hospital of Jilin University from January 2021 to October 2022.All enrolled patients underwent 2D T2WI and 3D T2WI rectal MR examinations and were evaluated for T stage,N stage,EMVI and MRF by two radiologists.The golden standard of T and N stages were based on the histopathological diagnostic results,and the golden standard of EMVI and MRF were based on the interpretation results by a senior radiologist.The diagnostic efficacy and consistency of diagnostic results for T stage,N stage,EMVI and MRF were assessed by two radiologists.Results: The average accuracy of 3D T2WI and 2D T2WI sequence for the diagnosis of T stage was 87.0% and 75.5%,respectively,and the diagnostic accuracy of early rectal cancer,low rectal cancer,high and middle rectal cancer in 3D T2WI were higher than 2D T2WI sequence(P value <0.05).The mean accuracy of 3D T2WI and 2D T2WI sequence for N staging was 87.3% and 79.6% respectively.The mean accuracy of 3D T2WI and 2D T2WI sequence for EMVI was 92.9% and 79.7% respectively.The mean accuracy of 3D T2WI and 2D T2WI sequence for MRF was 93.7% and 87.7% respectively.Conclusion: Compared with the traditional 2D T2WI sequence,the 3D T2WI sequence has higher diagnostic accuracy and evaluation consistency in the preoperative evaluation of T stage,N stage,EMVI and MRF,which has significantly improved its diagnostic efficacy for early rectal cancer and high rectal cancer.PART Ⅱ Diagnostic criteria for metastatic lymph nodes in rectal cancer based on3D-T2WI: a comparative study of MRI and histopathologyObjective:Some lymph nodes were found with or without "capsule sign" and "central spot sign".The main objective of this study was to evaluate whether the specific signs of 3D T2WI in lymph nodes have diagnostic value for metastatic lymph nodes and to frame a 3D T2WI sequence lymph node diagnostic model.Materials and Methods: In this study,39 rectal cancer patients were prospectively collected,and 145 lymph nodes(52 positive nodes and 93 negative nodes)were obtained by accurate matching of preoperative MR,postoperative MR and pathological results.The imaging features of accurately matched lymph nodes on 2D T2WI and 3D T2WI sequences were evaluated separately,and selected for independent risk or protective factors for metastatic lymph nodes by univariate and multivariate logisitic regression analysis.The lymph node diagnostic criteria based on 2D T2WI and 3D T2WI independent factors and existing guideline factors were constructed to compare the diagnostic efficiency of lymph nodes under different diagnostic criteria,and finally a lymph node diagnostic model based on 3D T2WI sequence was constructed.Results: The diagnostic criteria based on risk factors,protective factors or guideline factors from 3D T2WI sequence screening achieved high diagnostic accuracy,with 82.1%,92.4% and 87.6%,respectively,which were significantly higher than the diagnostic results from 2D T2WI sequence.The metastatic lymph node diagnostic model was constructed by combining the lymph node edge,internal signal,“capsule sign”,and “central spot sign”assessed in 3D T2WI sequence,the AUC value and 95%CI in the training dataset was0.981(0.963-1.000)and the AUC value and 95%CI in the validation dataset was 0.986(0.968,1.000).Conclusion: The capsule and central point sign assessed by 3D T2WI sequence facilitate the evaluation of metastatic lymph nodes,and the lymph node diagnostic model constructed by imaging features evaluated by 3D T2WI sequence can achieve high diagnostic accuracy.Part Ⅲ.Preoperative prediction of perineural invasion with multi-modality radiomics in rectal cancerObjective: Perineural invasion as a grossly underreported independent risk predictor in rectal cancer is important for patient treatment modality selection,which is hard to identify preoperatively.We aim to diagnosis PNI status in rectal cancer using multi-modality radiomics.Materials and Methods: Preoperative magnetic resonance images,computed tomography images,and medical records were analyzed in 94 consecutive patients with histopathological confirmed rectal cancer between June 2016 and October 2018.Of all the patients,70% were randomly assigned to the training dataset and the remaining 30% were selected for the testing dataset.Radiomics features were extracted and selected from the Volumes of Interests(VOIs)in the modality of T2-weighted imaging,diffusion-weighted imaging,and portal venous phase of contrast-enhanced CT,respectively.The radiomics score(Rad score),which named T2WI score,DWI score,and CT score,respectively,was calculated via a linear combination of the selected strongest features that weighted by their respective coefficients based on the corresponding modality.Then these scores were combined to generate MR score,concise score,and integrated score.Discrimination,calibration,and clinical benefit ability were used to evaluate the performance of the Rad scores in both training and testing datasets.Results: CT score and T2WI score were independent risk predictors for perineural invasion in rectal cancer patients[CT score,OR(95% CI)= 4.218(1.070–16.620);T2WI score,OR(95% CI)= 105.721(3.091–3615.790)].The concise score which combined CT score and T2WI score,showed the best performance [training dataset,AUC(95% CI)=0.906(0.833–0.979);testing dataset,AUC(95% CI)= 0.884(0.761–1.000)]Conclusions: The multi-modality Rad score could be used to preoperatively assess PNI status in rectal cancer,which provides valuable reference for clinicians in individual clinical decision-making system.