SBIR TOPIC 417 Multiple Brain Metastases Stereotactic Radiosurgery Management Platform - Detection, Planning and Follow-up

据估计，美国每年约有17万人被诊断为脑转移瘤(BMS)，而且大多数患者都有多发性脑转移瘤(MBMS)的临床表现。立体定向放射外科(SRS)已被广泛应用于MBMS的治疗，其毒副作用远低于传统的全脑放射治疗。目前的SRS平台是针对有限的靶点进行单分数治疗的，因此使用现有平台的MBMS SRS的临床工作流程很繁琐：1)劳动密集型的手动MBMS识别和分割；2)繁琐的人工规划MBMS到不同的治疗阶段；以及3)繁琐的跟踪MBMS治疗随访中的变化。我们的团队一直在积极开发(R01支持)基于人工智能(AI)的MBMS管理平台，即AimBMS，以优化SRS MBMS患者的护理。这份SBIR提案旨在翻译、验证AimBMS平台并将其商业化。该方案包括三个目标：1)开发私有云，将AimBMS平台整合到临床实践环境中；2)通过现场增量学习，提炼和优化特定地点实践的AimBMS平台性能；3)在5个合作机构验证AimBMS平台性能。该项目的成功将使AimBMS平台从单一的学术机构使用变为广泛的SRS社区实践，使大量MBMS人口受益。

It was estimated that ~170,000 patients diagnosed with Brain metastases (BMs) in United States each year and most patients have clinical manifestation of multiple BMs (mBMs). Stereotactic radiosurgery (SRS) has been used frequently for treating mBMs and demonstrated much less toxicity than conventional whole brain radiotherapy. Current SRS platform was designed for single-fraction treatment on limited number of targets, thus clinical workflow of mBMs SRS using extant platform is cumbersome: 1) labor-intensive manual mBMs identification and segmentation; 2) cumbersome manual planning mBMs into different treatment sessions; and 3) tedious tracking mBMs changes in treatment follow-up. Our team have been actively developing (R01-supported) an Artificial intelligence (AI)-based mBMs management platform, namely AimBMs, to optimize SRS mBMs patient care. This SBIR proposal targets to translate, validate and commercialize the AimBMs platform. The proposal consists of three aims: 1) develop a private cloud to integrate AimBMs platform into clinical practice environment; 2) refine and optimize AimBMs platform performance for site-specific practice using on-site incremental learning; and 3) validate AimBMs platform performance in 5 collaborating institutes. The success of this project will lead to bringing the AimBMs platform from a single academic institution use to a broad SRS community practice, benefitting a large mBMs population.