Preliminary Study on Predicting Pathological Staging and Immunohistochemical Markers of Rectal Cancer Based on ADC
Keke Zhang1, You Zheng1, Hongliang Huang1
1First Clinical Medical College, Lanzhou University, Donggang West Road, Chengguan Dist, Lanzhou 730000, China; Department of Radiology, First Hospital of Lanzhou University, Lanzhou, China; Gansu Intelligent Imaging Medical Engineering Research Center, Lanzhou, China; Accurate Imaging Collaborative Innovation Gansu International Science and Technology Cooperation Base, Lanzhou, China.
Objective:
To analyze the correlation between histogram parameters of ADC and pathological staging of rectal cancer and CD31, CD2-40, S-100, and to explore its predictive value.
Materials And Methods:
MRI findings of 60 patients with surgically and pathologically proved rectal cancer were analyzed retrospectively. Patients were divided into pT1-2, pT3-4, pN0, pN1-2, G1-2 and G3 groups according to TNM staging of UICC tumors (2019) and WHO classification of digestive system tumors (2019). Cases were divided into CD31 (+) and CD31 (-), CD2-40 (+) and CD4-20 (-), S-100 (+) and S-100 (-) groups according to the expression of immunohistochemical markers. The ROI was delineated layer by layer on the ADC images by Firevoxel software, and the histogram parameters were extracted. The histogram parameters (ADC mean, ADC minimum, ADC maximum, ADC mode, ADC quartile), skewness, kurtosis and entropy were compared between each group. The bivariate logistic regression model was used to predict the tumor staging and immunohistochemical results.
Results:
1. ADC10th, ADC mean and Entropy were higher than pT3-4, ADC mean was higher than pT1-2, Entropy was lower than pT1-2, ADC10th, ADC25th, ADC50th, ADC mean and ADC mode were lower than pT3-40 (-) in CD2-40 (+) group, and the difference was statistically significant (p < 0.05); 2. The lower area of the curve (AUC) of rectal cancer pT, pN and CD2-40 (+) is 0.952 (0.892-1.000), 0.882 (0.791-0.972), 0.913 (0.840-0.985); 3. In the logistic regression model, higher ADC, Ropy and higher pN stages are independent predictors of tumor pT stages (OR = 1.156, 1.144,111.528); p = 0.045, 0.048, 0.002); higher Ropy and lower pT stages are independent predictors of tumor pN stages in the model (OR = 73.939, 0.024; p = 0.019, 0.001); higher ADC and lower differentiation are independent predictors of tumor CD2-40 stages in the model (ADC = 1.17, 0.048, 0.011); and higher Ropy and lower pT stages are independent predictors of tumor CD2-40 stages in the model (ADC = 1.096, 0.094, 0.044).
Conclusion:
Histogram analysis based on ADC images has potential value in predicting the pathologic stage and immunohistochemical markers of rectal cancer, and logistic regression model has better diagnostic efficacy than single parameter.
More Related Videos
07:13Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
15:17Colorectal Cancer Cell Surface Protein Profiling Using an Antibody Microarray and Fluorescence Multiplexing
Published on: September 25, 2011