H. Zhong

SU-F-R-33: Can CT and CBCT Be Used Simultaneously for Radiomics Analysis

PURPOSE To investigate whether CBCT and CT can be used in radiomics analysis simultaneously. To establish a batch correction method for radiomics in two similar image modalities. METHODS Four sites including rectum, bladder, femoral head and lung were considered as region of interest (ROI) in this study. For each site, 10 treatment planning CT images were collected. And 10 CBCT images which came from same site of same patient were acquired at first radiotherapy fraction. 253 radiomics features, which were selected by our test-retest study at rectum cancer CT (ICC>0.8), were calculated for both CBCT and CT images in MATLAB. Simple scaling (z-score) and nonlinear correction methods were applied to the CBCT radiomics features. The Pearson Correlation Coefficient was calculated to analyze the correlation between radiomics features of CT and CBCT images before and after correction. Cluster analysis of mixed data (for each site, 5 CT and 5 CBCT data are randomly selected) was implemented to validate the feasibility to merge radiomics data from CBCT and CT. The consistency of clustering result and site grouping was verified by a chi-square test for different datasets respectively. RESULTS For simple scaling, 234 of the 253 features have correlation coefficient ρ>0.8 among which 154 features haveρ>0.9 . For radiomics data after nonlinear correction, 240 of the 253 features have ρ>0.8 among which 220 features have ρ>0.9. Cluster analysis of mixed data shows that data of four sites was almost precisely separated for simple scaling(p=1.29 * 10-7 , χ2 test) and nonlinear correction (p=5.98 * 10-7 , χ2 test), which is similar to the cluster result of CT data (p=4.52 * 10-8 , χ2 test). CONCLUSION Radiomics data from CBCT can be merged with those from CT by simple scaling or nonlinear correction for radiomics analysis.

SU-E-T-739: The Logistic Regression and Cox Regression Model for Predicting Local Recurrence, Distant Metastases, and Overall Survival of Rectal Cancer Patients

Purpose: The purpose was to develop predicting models for the local recurrence, distant metastases, and overall survival for rectal cancer patients treated with chemoradiotherapy. These models were based on different statistical methods for indicating various relationships between clinical features and follow-up results, which may provide support on decision-making in treatment. Methods: Models were developed based on logistic regression and cox proportional hazards model using patient data (N=277) from Fudan University Shanghai Cancer Center. The patient data set was randomly split into two proportions as the training and validation datasets. The clinical features used as variables include sex, age, tumor location, RT dose, adjuvant chemotherapy, surgery procedure, clinical tumor stage, and pTNM stage. The Bootstrapping was used for constructing confidence intervals; and the model performance was evaluated by the concordance index (c-index). Results: The logistic regression model is easy to use because the conditional independence assumption is unnecessary. The Cox regression model can dispose data with time factors so follow-up over a certain time period can be predicted. For the logistic regression model, the c-index for validation are 0.73 (LR), 0.77 (DM), and 0.76(OS), while numerical values for the cox regression model are 0.72 (LR), 0.77 (DM), and 0.70(OS). Both models show the probability to predict follow-up events based on c-index. Clinical tumor stage and pathologic stage are crucial features for both models (p<0.05). Conclusion: The Logistic Regression and Cox Regression Model can be used for Predicting Local Recurrence, Distant Metastases, and Overall Survival of Rectal Cancer Patients. The performance of models can be further improved with additional parameters provided from different facilities.