The telomere biomarker as a tool to inform decision-making for aggressive salvage therapy in men with rising PSA post prostatectomy

乳腺癌切除术后生化复发的男性接受挽救性放疗（RT）可能受益于额外的 抗雄激素治疗（AAT）通过降低其进展为远处转移和死亡的可能性。 然而，在RTOG 9601中，并非所有男性都受益。目前还没有预测性生物标志物来确定谁是 更可能或更不可能从积极挽救治疗（RT+AAT）中获益。为了满足这一未满足的需求， 精确的治疗决策，我们将评估端粒生物标志物作为预测生物标志物， 治疗反应在这种情况下。我们在概念上的创新假设是，端粒生物标志物- 癌细胞间端粒长度的变异性与基质细胞端粒长度的组合- 捕获有关肿瘤行为的信息，超出目前使用的指标，从而确定男性谁是 更可能或更不可能从积极的挽救治疗中获益。我们发现端粒 生物标志物是一个独立的预后标志物的致死性疾病在药物治疗的男子，确定3 预后分类：良好、中等和差。端粒生物标志物尚未被测试为预测 在任何情况下的治疗反应。我们将在两个互补的环境中解决这些目标，即试验和临床 实践中，共有839名男性和165起转移事件。在试验设置中，我们将使用RTOG 9601，其中男性 随机分配至RT+/-AAT组。在临床实践中，我们使用接受RT+/-AAT的队列， 约翰霍普金斯或波士顿医学中心，并有组织微阵列;在分析中，我们将加权 倾向评分，以尽量减少由于患者和肿瘤因素造成的偏倚。我们将评估这些目标， 端粒生物标志物：1.在标准化的环境中，测试是否有进展到转移和死亡的速度， 前列腺癌在RT+AAT和仅RT之间不同。2.在临床实践环境中，测试 RT+AAT和RT之间仅前列腺癌的转移进展和死亡不同。我们会弄脏 用于端粒和细胞类型特异性免疫荧光标记，并进行图像捕获和定量 图像分析，并得出每个人端粒生物标志物。我们将按生物标志物类别和用途进行分层 考克斯模型，以估计RT+AAT与进展之间的关联，并确定生物标志物是否增加 预测RT+AAT反应的能力，超过目前使用的生化后复发预后 指标在具有与中等预后相关的生物标志物类别的男性中，我们假设 与仅接受RT的男性相比，接受RT+AAT的男性的进展率较低。男性中 生物标志物类别与预后好或差相关，我们假设男性患者的进展率 接受RT+AAT治疗的男性中的死亡率与仅接受RT治疗的男性中的死亡率相似。在RTOG 9601中，RT+AAT更多 仅在某些亚组中比RT有效。为了优化决策，我们将确定生物标志物是否 在亚组中具有预测性。如果我们的假设得到证实，下一步将是前瞻性验证， 商业伙伴关系，以产生一个基于工具包的系统，用于自动化平台。

Men with biochemical recurrence after prostatectomy receiving salvage radiation (RT) may benefit from added anti-androgen therapy (AAT) by decreasing their likelihood of progressing to distant metastasis and death. However, in RTOG 9601, not all men benefitted. No predictive biomarker currently exists to identify who is more likely or less likely to benefit from aggressive salvage therapy (RT+AAT). To address this unmet need for precision treatment decision-making, we will evaluate the telomere biomarker as a predictive biomarker for treatment response in this setting. Our conceptually innovative hypothesis is that the telomere biomarker – the combination of cancer cell-to-cell variability in telomere length coupled with stromal cell telomere length – captures information about tumor behavior beyond currently used indicators and thus, identifies men who are more likely or less likely to benefit from aggressive salvage therapy. We discovered that the telomere biomarker is an independent prognostic marker for lethal disease in surgically-treated men, identifying 3 prognostic categories: good, intermediate, and poor. The telomere biomarker has not been tested as predictive of treatment response in any setting. We will address the aims in 2 complementary settings, trial and clinical practice, totaling 839 men and 165 metastatic events. In the trial setting, we will use RTOG 9601, in which men were randomized to RT+/-AAT. In the clinical practice setting, we use cohorts who received RT+/-AAT at Johns Hopkins or Boston Medical Center and have tissue microarrays; in the analysis, we will weight by a propensity score to minimize bias due to patient and tumor factors. We will evaluate these aims stratified by the telomere biomarker: 1. In the standardized setting, test if rate of progression to metastasis and death from prostate cancer differs between RT+AAT and RT only. 2. In the clinical practice setting, test if rate of progression to metastasis and death from prostate cancer differs between RT+AAT and RT only. We will stain for telomeres and cell-type specific immunofluorescence markers and perform image capture and quantitative image analysis, and derive each man’s telomere biomarker. We will stratify by biomarker category and use Cox models to estimate associations between RT+AAT and progression, and determine if the biomarker adds to predictive capability for response to RT+AAT beyond currently used post-biochemical recurrence prognostic indicators. In men with the biomarker category associated with intermediate prognosis, we hypothesize that the progression rate is lower in men who received RT+AAT compared to men who received RT only. In men with biomarker categories associated with good or poor prognosis, we hypothesize that the progression rate in men who received RT+AAT is similar to the rate in men who received RT only. In RTOG 9601, RT+AAT was more efficacious than RT only in some subgroups. For optimized decision-making, we will determine if the biomarker is predictive in subgroups. If our hypothesis is confirmed, next steps would be prospective validation and commercial partnership to generate a kit-based system for automated platforms.