CAREER: Stochastic Control for Adaptive Biologically Conformal Radiotherapy

这个教师早期职业发展（Career）项目的研究目标是在数学上开发一种新的癌症放疗模式，使治疗最佳地适应肿瘤生物学状况的时空演变。这将通过制定动态优化模型来实现，该模型考虑了肿瘤对辐射反应的不确定性，并允许治疗计划者根据肿瘤的状况调节辐射束强度，正如在每次治疗之前获得的功能图像所观察到的那样。目标将是在治疗过程结束时尽量减少肿瘤细胞的数量，同时限制辐射对附近健康组织的毒性作用。将设计有效的算法来解决由此产生的具有计算挑战性的随机优化问题。对于真正的患者特异性治疗，这些将与统计学习方法相结合，使用在治疗过程中收集的信息来估计反应不确定性。将建立计算机生成的测试用例来验证治疗策略，并从计算效率和治疗效果的角度对它们进行比较。如果成功，该项目将为一种新的个性化癌症治疗计划方法提供严格的数学基础，这种方法可以在正确的时间向正确的位置提供正确的辐射剂量，从而有可能改善健康状况。方法上的思想也将适用于其他疾病的治疗计划。PI将指导研究生，将研究成果纳入工程师预科的模拟讲习班，为代表性不足的本科生提供研究机会，并为卫生专业人员开发短期课程。结果将通过期刊出版物和科学会议传播。

The research objective of this Faculty Early Career Development (CAREER) project is to mathematically develop a novel cancer radiotherapy paradigm that optimally adapts treatment to the spatiotemporal evolution of a tumor's biological condition. This will be accomplished by formulating dynamic optimization models that account for uncertainty in tumor-response to radiation and allow treatment planners to modulate radiation beam intensities depending on the tumor's condition, as observed in functional images acquired prior to each treatment session. The goal will be to minimize the number of tumor cells remaining at the end of the treatment course, while limiting toxic effects of radiation on nearby healthy tissue. Efficient algorithms will be designed to solve the resulting computationally challenging stochastic optimization problems. For truly patient-specific treatment, these will be integrated with statistical learning methods that estimate response-uncertainty using information gathered over the treatment course. Computer-generated test cases will be built to validate treatment strategies and to compare them from computational efficiency and treatment efficacy perspectives.If successful, the project will result in a rigorous mathematical foundation for a new individualized cancer treatment planning method that delivers the right radiation dose to the right location at the right time - potentially leading to improved health outcomes. The methodological ideas will also be applicable while planning treatment for other diseases. The PI will mentor graduate students, incorporate research findings into simulation workshops for pre-engineers, provide research opportunities for underrepresented undergraduates, and develop a short course for health professionals. Results will be disseminated through journal publications and scientific conferences.