(1) Background: Radiomics use high-throughput mining of medical imaging data to extract unique information and predict tumor behavior. Currently available clinical prediction models poorly predict treatment outcomes in pancreatic adenocarcinoma. Therefore, we used radiomic features of primary pancreatic tumors to develop outcome prediction models and compared them to traditional clinical models. (2) Methods: We extracted and analyzed radiomic data from pre-radiation contrast-enhanced CTs of 74 pancreatic cancer patients undergoing stereotactic body radiotherapy. A panel of over 800 radiomic features was screened to create overall survival and local-regional recurrence prediction models, which were compared to clinical prediction models and models combining radiomic and clinical information. (3) Results: A 6-feature radiomic signature was identified that achieved better overall survival prediction performance than the clinical model (mean concordance index: 0.66 vs. 0.54 on resampled cross-validation test sets), and the combined model improved the performance slightly further to 0.68. Similarly, a 7-feature radiomic signature better predicted recurrence than the clinical model (mean AUC of 0.78 vs. 0.66). (4) Conclusion: Overall survival and recurrence can be better predicted with models based on radiomic features than with those based on clinical features for pancreatic cancer.
3/125 and 0/30 features with CoV<0.05 respectively). Mean signal intensity ratio F1/SIM in GTV correlates strongly with corresponding changes in ROI_tis (Pearson r Z 0.88, p Z 0.0006), with no significant trend in either (t-test p Z 0.28 and p Z 0.46). Histogram width features in tumor show an increase SIM-F1 and a drop F1-F5 (e.g. Interquartile range (IQR): p Z 0.042, 1.3AE0.4 vs. 0.92AE0.19 F1/SIM vs. F5/F1 ratios, Standard Deviation (SD): p Z 0.08, 1.3AE0.5 vs. 0.93AE0.15). The only patient with pathological complete response at surgery was also only one of 10 to show SIM-F1 decrease (ratios: IQR 0.96, SD 0.92). Normalization by median ROI_tis signal further strengthened the trend (IQR p Z 0.01, SD p Z 0.005). No trends were seen in ROI_tis (IQR p Z 0.50, SD p Z 0.48). Conclusion: These results present early application for the radiomic and histogram analysis for MRgRT image quantification. PDAC may be difficult to define in TRUFI MRgRT images, highlighting the value of observed high spatial robustness of features. Tumor specific changes in histogram width were observed. These metrics, associated with tumor internal heterogeneity, show promise to elucidate clinically relevant information from the images. Preliminary link with histological response was established. Efficient normalization method was found to improve sensitivity. More work is required to understand the biological basis of the observed trends.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.