Submit your email address to sign up for the Visionary,
or monthly newsletter.

X

Search the Radiation Oncology Institute

X

Submit your email address to sign up for the Visionary,
or monthly newsletter.

X

Search the Radiation Oncology Institute

X

Hesham Elhalawani, MD, MSc

    Identifying an Imaging Biomarker for Cerebral Radiation Necrosis

    Dr. Greer

    Hesham Elhalawani, MD, MSc, and his team at Brigham and Women’s Hospital are using radiomics to help predict which patients being treated with immunotherapy and stereotactic radiosurgery (SRS) for brain metastases are at greatest risk of developing radiation necrosis (RN). Up to 40% of patients with cancer are diagnosed with brain metastases, and treatment with SRS is common and on the rise. If the brain tissue in the treated area is injured, RN can be a long-term side effect and results in symptoms such as headaches, cognitive impairment and seizures. Additionally, immunotherapy seems to increase the risk of developing RN and is frequently used to treat cancers that have a tendency to spread to the brain. Distinguishing RN from a return of the cancer is difficult using conventional MRI, and neurosurgery to perform a biopsy is often required for an accurate diagnosis. Dr. Elhalawani aims to develop a “virtual biopsy” using radiomics, which is the analysis of data from medical images, and would be a non-invasive method to diagnose RN.

    With support from a James D. Cox Research Award from the ROI, Dr. Elhalawani is leveraging the power of artificial intelligence (AI) to conduct a longitudinal analysis of MRIs performed before and after SRS to identify imaging biomarkers to predict which patients are most likely to develop RN. The project is a key component of his clinical fellowship in radiation oncology and central nervous system oncology and is being conducted under the mentorship of Ayal Aizer, MD, MHS.

    In the first study of its kind, Dr. Elhalawani’s team is:

    • Analyzing retrospective data including anatomic and multi-parametric MRIs and planning target volumes from patients treated with SRS and immunotherapy to identify radiomic features associated with RN.
    • Developing an AI decision-making tool that can predict which patients are most likely to develop RN based on the radiomic features of MRIs conducted as part of the standard clinical workflow.
    • Validating the “virtual biopsy” decision-making tool in a cohort of patients enrolled in an ongoing prospective clinical trial and comparing the model’s performance to a diagnosis based on surgical biopsy, the current standard of care.

    A non-invasive method to predict which patients are most likely to develop RN that relies on routine imaging would be a significant clinical advance that could transform practice. Many patients stand to benefit from this innovation by allowing for earlier, more targeted treatment of RN or tumor recurrence, reducing the need for additional neurosurgery and improving their quality of life.