Multi-atlas and whole body radiomics approaches for image-guided treatment of gynecologic cancers

摘要 妇科癌症是全球女性癌症死亡的主要原因之一。这些 患者通常在社会经济上处于不利地位，很难获得筛查和疫苗接种。 因此，他们经常表现为局部晚期疾病，对于这种疾病，盆腔放疗(RT)和 同时使用顺铂(即放化疗)是治疗的标准。然而，这种治疗受到以下因素的限制 治疗失败率高。通过提供双重化疗强化治疗，或者 同时或作为放化疗后的辅助治疗，是一种很有前途的改善预后的策略。 然而，高强度化疗的实施因高胃肠道和高死亡率而变得复杂。 血液毒性。在提高放化疗疗效的同时减少毒性的策略是必要的。 标准的盆腔RT技术包括大量的正常组织，包括肠、骨 骨髓、骨骼、膀胱和直肠，导致可预防的辐射毒性。图像导引辐射 治疗(IGRT)可以改善目标定位和剂量学，优化目标剂量，同时将剂量降至最低 周围的正常组织。然而，IGRT可以是高度资源密集型的，并且相对有效 试验一直很欠缺。由于这个原因，关于IG-IMRT在这方面的用途存在相当大的争议 疾病。我们的研究小组一直处于开发新的、具有成本效益的IGRT方法的前沿 具有促进更好地提供同步和/或辅助化疗的广泛潜力。 以前我们已经发现，辐射导致的对造血活跃的骨髓的损伤是一种 对强化化疗耐受性的关键决定因素。使用机器学习方法，我们最近 开发了一种基于多图谱的IGRT方法，可以预测活跃骨髓的规范分布， 这可以在该技术被 买不到或买不起。拟议的新研究将研究基于多图谱的IGRT的能力 使用数据，与标准治疗相比，降低血液毒性并改善化疗效果 450名患者参加了一项随机III期试验(NRG-GY006)。此外，我们将使用序列整体 体部PET/CT研究放疗剂量和化疗强度对机体代偿能力的影响 ，并开发了新的全身放射组学生物标志物来量化 炎症状态，我们假设这会影响患者的预后和对化疗的耐受性。 这项新的研究将我们与当前R01拨款(1R01CA197059-01)相关的工作扩展到 开展与GY006试验相关的科学研究。这条研究路线的首要目标是 利用先进的影像引导放射治疗提高盆腔癌的放化疗的治愈率 技巧。如果成功，这项研究将极大地改变许多骨盆的治疗方法 以化疗和放射治疗为标准的恶性肿瘤。

ABSTRACT Gynecologic cancers are among the leading causes of cancer death in women worldwide. These patients typically are socioeconomically disadvantaged, with poor access to screening and vaccination. Consequently, they often present with locoregionally advanced disease, for which pelvic radiotherapy (RT) with concurrent cisplatin (i.e., chemoradiotherapy) is the standard of care. This treatment is limited, however, by high rates of treatment failure. Intensifying treatment through the delivery of chemotherapy doublets, either concurrently or as adjuvant therapy following chemoradiotherapy, is a promising strategy to improve outcomes. However, the delivery of intensive chemotherapy is complicated by high rates of gastrointestinal and hematologic toxicity. Strategies to reduce toxicity while increasing efficacy of chemoradiotherapy are needed. Standard pelvic RT techniques encompass large volumes of normal tissue including bowel, bone marrow, bone, bladder, and rectum, leading to preventable radiation-induced toxicity. Image-guided radiation therapy (IGRT) can improve target localization and dosimetry, optimizing target dose while minimizing dose to surrounding normal tissues. However, IGRT can be highly resource intensive, and comparative effectiveness trials have been lacking. For this reason, there is considerable controversy as to the utility of IG-IMRT in this disease. Our research group has been at the forefront of developing novel, cost-effective IGRT approaches with wide potential to facilitate better delivery of concurrent and/or adjuvant chemotherapy. Previously we have found that radiation-induced injury to hematopoietically active bone marrow is a critical determinant of tolerance to intensive chemotherapy. Using machine learning methods, we recently developed a multi-atlas-based IGRT method that can predict canonical distributions of active bone marrow, which can obviate the need for positron emission tomography (PET) in settings where this technology is unavailable or unaffordable. The proposed new research will study the ability of multi-atlas-based IGRT to reduce hematologic toxicity and improve chemotherapy delivery compared to standard treatment, using data from 450 patients enrolled to a randomized phase III trial (NRG-GY006). Furthermore, we will use serial whole body PET/CT to study the impact of radiation dose and chemotherapy intensity on the compensatory hematopoietic response, and have developed novel whole body radiomics biomarkers to quantify the inflammatory state, which we hypothesize can influence patients' outcomes and tolerance to chemotherapy. The new research extends our work associated with a current R01 grant (1R01CA197059-01) to conduct correlative science associated with the GY006 trial. The overarching goal of this research line is to augment the therapeutic ratio of chemoradiotherapy for pelvic cancers using advanced image-guided radiation techniques. If successful, this research would significantly alter the approach to the treatment of many pelvic malignancies for which chemoradiotherapy is standard.