Radiomic model predicts radiotherapy outcomes for patients with brain metastases
Researchers have developed a new radiomic-clinical model that can predict treatment outcomes of whole-brain radiotherapy based on data from MRI exams.
Professor LI Hai and WANG Hongzhi from Hefei Institutes of Physical Science of Chinese Academy of Sciences (CAS) led the research efforts, which were detailed recently in European Radiology. Experts combined radiomics and the game theory-based SHapley Additive exPlanations (SHAP) method to predict how patients with brain metastases would respond to treatment in the hopes that their results could be used in the management and treatment of such a consequential diagnosis.
To build the model, the experts used radiomics features (extracted from pre-treatment MRI scans) and clinical features of 228 patients with brain metastases (184 in a training cohort and 44 for validation). Patients were categorized as nonresponding and responding.
In terms of accuracy, the radiomic-clinical model yielded AUCs of .928 and .851 for predicting radiotherapy treatment outcomes in the training and validation cohorts, respectively.
While the model did perform well in assessing treatment responses, the experts explained that one of the most beneficial aspects of their model was that its results are interpretable in a “clinician-friendly way.”
“SHAP summary plot illustrated the feature’s value affected the feature’s impact attributed to model, and SHAP force plot showed the integration of features’ impact attributed to individual response,” Hai, Honzhi and co-authors wrote.
The experts said this, in turn, helps avoid the “black box” effect that is common with many machine learning algorithms and enables providers to understand the model’s processes that lead to its conclusions.
In the future, the authors believe that their model could play a role in personalizing radiation therapy treatments for patients with brain metastases.
View the research here.
Reference:
Wang, Y., Lang, J., Zuo, J.Z. et al. The radiomic-clinical model using the SHAP method for assessing the treatment response of whole-brain radiotherapy: a multicentric study. Eur Radiol (2022).
More on radiomics:
DBT-based radiomics nomogram predicts lymphovascular invasion in breast cancer patients
CT-based radiomics features can help diagnose COPD earlier than ever before
Radiomics-clinical model accurately predicts osteoporotic spinal fracture timeline on CT images
CT-based radiomics nomogram accurately predicts colorectal cancer prognosis
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She joined Innovate Healthcare in 2021 and has since put her unique expertise to use in her editorial role with Health Imaging.