Strategies for discerning chemotherapy response and resistance in ovarian cancer

项目总结/摘要 卵巢癌是女性癌症死亡的第五大原因。大多数卵巢癌的病例都很高- 分级浆液性卵巢癌（HGSOC），其占大多数卵巢癌死亡率。尽管有 随着对这种肿瘤类型认识的增加，使用“旧的”DNA损伤化疗药物仍然是一个问题。 卵巢癌患者的首选。HGSOC的标准治疗包括减积手术， 以铂为基础的化疗，目的是消除所有剩余的微转移。不幸的是 大量HGSOC患者对这些治疗没有反应或仅部分反应。患者 对铂类化疗无反应的患者必须接受同样的治疗， 毒性，无明显益处。尽管癌症基因组学和精准医学在世界范围内取得了迅速进展， 在过去的十年里，仍然令人惊讶的是，没有强大的分子生物标志物可以预测患者的反应， 已经在卵巢癌中确定了铂剂，指出需要掺入功能性的 指标和早期分子变化纳入预测框架。我们假设早期 药物暴露后的分子和功能改变，这些改变是肿瘤细胞如何协调反应的基础 对药物治疗的影响决定了长期的治疗反应，因此对药物治疗的发展至关重要。 预测模型我们进一步假设，患者的药物敏感性和生物能量依赖性 肿瘤是这种预测的重要功能读数。我们建议整合早期的时间 基因组、转录组、代谢表型、药物敏感性和组织学信息输入机器 学习模型，以识别对铂的患者特异性反应最具预测性的一组指标 HGSOC中的化疗。这是通过将新鲜制备的精确切割的肿瘤 切片培养模型，一种创新的单细胞芯片上代谢细胞术分析，和一种新的原位空间 在活组织上的代谢谱测定。成功完成后，这项研究可以提供一个预测 用于鉴定对铂化疗具有较高耐药性风险的HGSOC患者的诊断测定， 以及在治疗开始之前可能受益于这种治疗的患者。这些结果将提供 与肿瘤遗传学和其他临床因素互补的创新分子和功能信息， 更多信息的临床决策，并帮助重新指导铂难治性患者接受其他治疗 在治疗前进行新疗法的方法或临床试验。

Project Summary/Abstract Ovarian cancer is the fifth leading cause of cancer death among females. Most cases of ovarian cancer are high- grade serous ovarian cancer (HGSOC), which accounts for most ovarian cancer mortality. In spite of the increasing knowledge of this tumor type, the use of “old” DNA-damaging chemotherapeutic drugs remains the first option for ovarian cancer patients. The standard treatment of HGSOC involves de-bulking surgery followed by platinum-based chemotherapy with the aim to eliminate all the remaining micro-metastases. Unfortunately, a significant number of HGSOC patients do not respond or only partially respond to these treatments. A patient who does not respond to platinum chemotherapies has to go through the same treatment and experience the toxicity without meaningful benefit. Despite the rapid advances in cancer genomics and precision medicine in the last decade, what remains surprising is that no robust molecular biomarker that can foretell patient response to platinum agents has been identified in ovarian cancer, pointing to the need for the incorporation of functional metrics and early-stage molecular changes into the predictive framework. We hypothesize that early-stage molecular and functional alterations upon drug exposure that underlie how tumor cells orchestrate a response to the drug treatment are dictating longer-term therapy responses and thus critical to the development of predictive models. We further hypothesize that the drug susceptibility and bioenergetic dependency of the patient tumors are important functional readouts for such a prediction. We propose to integrate early-stage temporal genome, transcriptome, metabolic phenotypes, drug susceptibility, and histology information into machine learning models to discern a set of metrics most predictive for patient-specific responses to platinum chemotherapy in HGSOC. This is enabled by a strategic integration of a freshly prepared, precision-cut tumor slice culture model, an innovative single-cell on-chip metabolic cytometry assay, and a novel in situ spatial metabolic profiling assay on live tissues. Upon successful completion, this study could deliver a predictive diagnostic assay for identifying HGSOC patients with a higher risk of resistance to platinum chemotherapy, as well as patients who may benefit from such treatments, prior to the onset of therapy. These results will provide innovative molecular and functional information complementary to tumor genetics and other clinical factors for more informative clinical decision-making and help re-direct platinum-refractory patients to other therapeutic approaches or clinical trials of novel therapies before the treatment.