Integrated blood and radiomic subtyping to guide immunotherapy treatment selection and early response assessment in metastatic non-small cell lung cancer

摘要 免疫检查点抑制剂(ICIS)改善了转移性非小细胞肺癌(NSCLC)的预后， 提供者现在可以在多种基于ICI的一线方案之间进行选择，包括ICI单一疗法和 ICI接受化疗。然而，这种选择的增加使临床管理变得复杂，几乎没有 用于指导前期ICI治疗选择的生物标志物，以及用于早期治疗评估的不完整指标 对ICI治疗的反应。因此，迫切需要新的分析工具来优化和个性化 免疫治疗的治疗策略。虽然之前的生物标记物工作主要集中在基于组织的 分子图谱，这些已经显示出有限的预测能力，并且由于 获取治疗前和治疗中组织的实际限制。相比之下，成像和基于血液的分析提供了 一种独特的非侵入性机制，通过这种机制，肿瘤的生物学和治疗上的变化可以 研究和建模。因此，我们提出了一种综合的放射-血液分析来开发前和后的预测因子。 指导非小细胞肺癌的临床治疗。我们的首要目标是发展放射学- 利用我们在数据科学方面的专业知识，为晚期非小细胞肺癌的精确免疫治疗提供血液签名， 胸部肿瘤学、癌症基因组学、计算肿瘤学、临床化验发展，并建立 研究合作。我们的初步数据表明，我们成功地利用了多参数分析 循环肿瘤DNA以确定与ICI结果和疾病复发相关的分子表型 并在开发具有卓越结果预测和证明的新的放射组学亚型技术中 与潜在肺癌生物学的关系。因此，我们假设辐射和血液的结合 指标可以非侵入性地为非小细胞肺癌管理的治疗决策提供信息，同时促进我们的 了解非小细胞肺癌生物学。为了推进这一假设，我们汇集了一组独特的 转移性非小细胞肺癌患者接受ICI方案治疗，并获得高质量的放射扫描、血样和 分子和临床数据：我们内部的肺癌数据库(Gemini，n=5000)；我们的验证数据集 与马萨诸塞州总医院(MGH)癌症中心(MGH，n=600)合作；多中心 协作性STAND UP 2癌症/马克基金会队列(SU2C，n=400)，以及一项预期的III期ICI试验 (Lonestar，n=300)。我们的建议建立在这些独特的队列和我们有希望的初步数据的基础上 构建预测模型以指导前期ICI治疗选择和改善治疗反应 评估，而补充研究将揭示这些临床预测因素背后的生物学。一个 我们提案的主要优势是我们的跨学科团队在开发、验证和翻译方面的专业知识 这些创新的预测模型针对高度相关的临床问题。一体化的发展 基于血液和成像的放射性基因组生物标记物将有助于改善患者的临床管理 转移性非小细胞肺癌，同时帮助该领域迈向非侵入性精密免疫肿瘤学的新时代。

ABSTRACT Immune checkpoint inhibitors (ICIs) have improved outcomes in metastatic non-small cell lung cancer (NSCLC), and providers may now choose between multiple first-line ICI-based regimens including ICI monotherapy and ICI with chemotherapy. However, this increase in options has complicated clinical management, with few biomarkers to guide upfront ICI treatment selection, and incomplete metrics for early on-treatment assessment of response to ICI therapy. Hence, there is an urgent need for novel analytics tools to optimize and personalize immunotherapy treatment strategies. While prior biomarker efforts have focused largely on tissue-based molecular profiling, these have demonstrated limited predictive power and are difficult to implement due to practical limitations in acquiring pre- and on-treatment tissue. In contrast, imaging and blood-based assays offer a unique and non-invasive mechanism by which the biology of the tumor and the changes on treatment can be studied and modeled. Thus, we propose an integrated radiomic-blood analysis to develop predictors of pre- and on-treatment response to guide the clinical management of NSCLC. Our primary goal is to develop radiomic- blood signatures for precision immunotherapy in advanced NSCLC by leveraging our expertise in data science, thoracic oncology, cancer genomics, computational oncology, clinical assay development, and established research collaborations. Our preliminary data demonstrates our success in utilizing multi-parametric profiling of circulating tumor DNA to identify molecular phenotypes associated with ICI outcome and disease recurrence, and in developing novel radiomic subtyping techniques with superior outcome prediction and demonstrated association with underlying lung cancer biology. Hence, we hypothesize that coupling radiomic and blood-based metrics can non-invasively inform therapeutic decision-making in NSCLC management while advancing our understanding of NSCLC biology. To advance this hypothesis, we have assembled a unique set of cohorts of metastatic NSCLC patients treated with ICI regimens with high-quality radiographic scans, blood samples, and molecular and clinical data: our in-house lung cancer database (GEMINI, n=5000); a validation dataset from our collaboration with the Massachusetts General Hospital (MGH) Cancer Center (MGH, n=600); the multicenter collaborative Stand Up 2 Cancer/Mark Foundation cohort (SU2C, n=400), and a prospective phase III ICI trial (LONESTAR, n=300). Our proposal builds on these unique cohorts and our promising preliminary data to construct predictive models to guide up-front ICI therapy selection and improve on-treatment response assessment, while complementary investigations will uncover the biology underlying these clinical predictors. A major strength of our proposal is our interdisciplinary team’s expertise in developing, validating, and translating these innovative predictive models toward highly relevant clinical questions. The development of integrative blood- and imaging-based radio-genomic biomarkers will help improve the clinical management of patients with metastatic NSCLC while helping progress the field toward a new era of non-invasive precision immunooncology.