Advanced diffusion MRI for evaluating early response to radiation treatment in cervical cancer

Project Summary Despite the advent of aggressive cervical cancer screening programs, cervical cancer remains one of the most common cancers affecting women under age 35, and the fourth most common cause of cancer death worldwide. The standard of care for early stage (≥IB) cervical cancer is hysterectomy or radiation. Unfortunately, the

consequences of radical treatment include fertility loss, nerve injury causing bladder and bowel dysfunction, and pelvic pain. There is a critical need to reduce cervical cancer mortality, while minimizing the potential morbidities of treatment. To achieve this end requires refined approaches for diagnosis and evaluation of response to

treatment using noninvasive biomarkers to differentiate indolent from clinically significant disease at the earliest possible time-point. PET/CT is currently the mainstay in evaluating response to treatment and is highly confounded by post-treatment changes such as edema. Magnetic resonance imaging (MRI) with advanced

diffusion-weighted imaging may offer an alternative approach to evaluate treatment response, with additional advantages of being a radiation-free and contrast media-free exam. The overall objective in this application is to develop and evaluate a robust advanced diffusion-weighted imaging technique that provides a highly sensitive

and specific reflection of cervical cancer tumor burden and treatment response at the earliest possible time point. Our hypothesis is that restriction spectrum imaging (RSI), an advanced diffusion imaging technique, is as sensitive and specific as standard of care post-treatment PET/CT for evaluation of treatment efficacy of cervical

cancer and can be performed 3 months earlier than standard of care PET/CT. The aims of this proposal are 1) Determine the RSI model for cervical cancer evaluation, 2) Develop and validate a cervical cancer classification algorithm from multi-parametric MRI based on the Aim 1 biophysical model using established machine learning

techniques, 3) Prospectively validate RSI-MRI compared to PET/CT in evaluating response to radiation treatment in cervical cancer patients (≥ stage IB) in a pilot study0. The main significance of this study is the development of a radiation-free and non-contrast imaging technique for evaluating response to treatment three

months earlier than the current standard of care PET/CT. This will allow appropriate treatment earlier preventing unnecessary progression of disease. The innovation proposed involves developing a diffusion model specific for cervix imaging within the RSI framework based on the biophysical characteristics of healthy and malignant

cervical tissue. We will then apply this quantitative technique prospectively on a preliminary cohort of patients before and after treatment and compare to the standard of care PET/CT imaging. At the completion of the study, a new tool for evaluating response to treatment in cervical cancer that is contrast and radiation free will be

available. This is directly translatable to the clinical setting to benefit cervical cancer patients. Cervical cancer patients will be better served, particularly patients not responding to treatment, as they can be directed to appropriate treatment at an earlier time point.