Quantitative PET Imaging for Oncologic Immune Response Prediction

摘要 免疫检查点抑制剂显著提高了一些癌症的总存活率，这些癌症 反过来又引发了巨大的科学和财政投资，以进一步扩大这种治疗方法 范例。然而，目前，免疫治疗的好处只在少数患者中实现。 使问题进一步复杂化的是，许多免疫疗法存在严重不良免疫事件的风险，而且 解剖分期、18F-FDG PET显像等检测疗效的方法存在混淆 通过潜在的免疫渗透的存在。这些入侵的免疫细胞可能引起潜在的反应 肿瘤体积增大和18F-FDG消耗增加，这使它们与进展期难以区分 恶性肿瘤。由于目前缺乏诊断能力，许多患者接受治疗的唯一选择 免疫治疗必须确定他们是否有反应是总体生存，这是一个长期和潜在的 确定治疗效果的危险方法。此外，考虑到越来越多的药物 以及正在进行临床试验的组合，在早期阶段监测治疗效果的能力将 潜在地有助于将新的治疗方法更快地提交审批。目前，还没有批准的生物标记物 治疗前判断疗效，并对PD-L1等肿瘤标志物进行活检分析 结果只带来了结果的适度改善。因此，预测反应的生物标记物将允许 在临床前和临床研究方面都取得了重大进展。 颗粒酶B，由T效应细胞在激活后分泌，是一种有效的诱导剂 细胞凋亡，是免疫治疗反应的有力预测指标。我开发了一种新颖的、有选择性的PET 检测颗粒酶B的分泌和活性形式的成像肽，允许区分 对免疫治疗的积极反应和无反应，即“耗尽”含有颗粒酶B的T细胞 可能存在，但不会主动分泌这种酶。使用颗粒酶B肽进行PET成像允许 在肿瘤体积改变前对免疫治疗反应的高度敏感和特异性预测 小鼠同基因癌症模型。这种表型并不局限于老鼠，因为人类样本分析了 抗体和My多肽显示应答时颗粒酶B水平显著高于非应答时 有反应的病人。因此，颗粒酶B PET成像提供了对早期反应的独特见解，而不是 目前可以使用任何其他技术。当前的方法不能准确地定义响应之前 破坏性抽样，作为一种反应被定义为缺乏进展。考虑到这些限制，我建议 使用颗粒酶B PET成像根据颗粒酶B水平对小鼠进行分层，然后进行生化和 有反应和无反应肿瘤的遗传学分析。PET成像的非侵入性不会 仅在任何解剖变化之前提供准确的反应区分，但也将允许 根据初步的PET成像结果进行二次治疗。

Abstract Immune checkpoint inhibitors have markedly improved overall survival in a number of cancers, which has in turn sparked tremendous scientific and financial investment into further expansion of this treatment paradigm. Currently, however, the benefits of immunotherapy have only been realized in a minority of patients. Further complicating the issue, many immunotherapies carry risks of severe adverse immune events, and methods to detect therapeutic efficacy such as anatomical staging and 18F-FDG PET imaging are confounded by the potential presence of immune infiltrate. These invading immune cells can cause potentially responding tumors to increase in size and in 18F-FDG consumption, which make them indistinguishable from progressing malignancies. Because of the lack of current diagnostic capabilities, the only option many patients undergoing immunotherapy have to determine if they are responding is overall survival, which is a long and potentially dangerous approach to determining therapeutic efficacy. Additionally, given the increasing number of drugs and combinations being clinically trialed, the ability to monitor therapeutic efficacy at an earlier stage would potentially help bring new treatments to approval much faster. Currently, there is no approved biomarker for determining therapeutic efficacy, and biopsy analysis of tumor markers such as PD-L1 prior to treatment have only resulted in modest improvements of outcome. Thus a biomarker that predicted response would permit significant advances in both the pre-clinical and clinical investigations. Granzyme B, which is secreted by T effector cells following activation and acts as a potent inducer of apoptosis, is a strong predictor of immunotherapy response. I have developed a novel and selective PET imaging peptide that detects the secreted and active form of granzyme B, permitting differentiation between active response to immunotherapy and non-response in which “exhausted” T cells that contain granzyme B may be present but are not actively secreting the enzyme. PET imaging with the granzyme B peptide permits highly sensitive and specific prediction of response to immunotherapy prior to changes in tumor volume in murine syngeneic models of cancer. This phenotype is not limited to mice, as human samples analyzed both by antibody and my peptide show significantly higher levels of granzyme B in responding versus non- responding patients. Thus, granzyme B PET imaging offers a unique insight into early response that is not currently possible using any other technique. Current methods cannot accurately define a response prior to destructive sampling, as a response is defined as lack of progression. Given these limitations, I am proposing to use granzyme B PET imaging to stratify mice based on granzyme B levels, followed by biochemical and genetic analysis of responding and non-responding tumors. The non-invasive nature of PET imaging will not only provide accurate differentiation of response prior to any anatomic changes, but will also allow for secondary therapeutic manipulations based on initial PET imaging results.