Intensity Modulated Radiation Therapy: Integrated Models and Algorithms

该基金支持开发新的综合模型和算法，用于调强放射治疗（IMRT）治疗计划。IMRT被认为是治疗多种癌症最有效的放射疗法。现有的方法将处理计划问题分解为若干阶段，并依次求解相应的子问题。尽管这可能会导致较差的处理方案，但迄今为止，由于缺乏合适的模型公式来允许有效的求解方法，因此阻碍了子问题的整合。特别是，最受关注的子问题是确定用于照射癌症患者的每束光束的强度分布图或影响图的中心问题。然而，已知与治疗递送效率和质量有关的问题对IMRT治疗有重大影响。模型将充分考虑到治疗设备的潜力和局限性，以及治疗过程中患者解剖结构的变化，并将与算法一起开发，以解决由此产生的综合问题。所有开发的模型和算法都将在实际的临床患者数据上进行测试。每年，成千上万的病人通过放射疗法接受癌症治疗。该项目将为整合临床因素的IMRT治疗计划问题提供高效且稳健的综合模型，这些问题目前尚未得到令人满意的解决。在现实计算时间内解决由此产生的综合优化问题的有效方法的发展将为医生和临床医生提供实时决策支持，并将帮助他们找到治疗计划，而不需要在治愈或控制疾病的良好机会与潜在的生活质量之间做出选择。

This grant supports the development of new integrated models and algorithms for intensity modulated radiation therapy (IMRT) treatment planning. IMRT is considered to be the most effective radiation therapy for many forms of cancer. Current approaches decompose the treatment planning problem into several stages and the corresponding subproblems are solved sequentially. Although this may lead to inferior treatment plans, the lack of suitable model formulations that allow for efficient solution methods has so far prevented the integration of the subproblems. In particular, the subproblem that has received the most attention is the central problem of determining an intensity profile or fluence map for each of the beams used in irradiating a cancer patient. However, issues pertaining to treatment-delivery efficiency and quality are known to have a significant impact on IMRT treatments. Models that fully take into account both potential and limitations of the treatment equipment as well as changes in the patient anatomy over the course of the treatment will be developed along with algorithms to solve the resulting integrated problems. All of the developed models and algorithms will be tested on actual clinical patient data.Every year, hundreds of thousands of patients receive cancer treatment by radiation therapy. This project will provide highly effective and robust integrated models for IMRT treatment-planning problems that integrate clinical considerations that are currently not yet addressed satisfactorily. The development of efficient approaches to solve the resulting integrated optimization problems in realistic computational times will provide real-time decision support to physicians and clinicians and will help them to find treatment plans that do not require a choice between a good chance for a cure or control of disease and the potential quality of life.