Biomarker-Based Tools for Treatment Response Decision Support of Bladder Cancer

基于生物标志物的膀胱癌治疗反应决策支持工具

• 批准号: 10617683
• 负责人: Ajjai Shivaram Alva
• 金额: $ 8.42万
• 依托单位: UNIVERSITY OF MICHIGAN AT ANN ARBOR
• 依托单位国家: 美国
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-05-06 至 2025-04-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10617683
• 关键词: Adjuvant Chemotherapy; Biological Markers; Bladder; Bladder Neoplasm; Cancer Etiology; Cessation of life; Characteristics; Clinical; Clinical Treatment; Clinical Trials; Clinical assessments; Collaborations; Computers; Data; Decision Support Systems; Diagnosis; Disease; Ensure; Evaluation; Feedback; Future; Goals; Image; Image Analysis; Interobserver Variability; Laboratory Research; Machine Learning; Malignant neoplasm of urinary bladder; Monitor; Morbidity - disease rate; Neoadjuvant Therapy; Oncologist; Outcome; Output; Patient Care; Patients; Performance; Process; Quality of life; Radical Cystectomy; Reproducibility; Site; System; Techniques; Testing; Time; Training; Tumor Volume; United States; Woman; alternative treatment; bench to bedside; cancer imaging; cancer type; chemotherapy; clinical care; clinical decision support; clinical development; clinical research site; design; effectiveness validation; experience; improved; innovation; men; mortality; outcome prediction; predictive modeling; preservation; prospective; quantitative imaging; radiologist; radiomics; responders and non-responders; response; support tools; tool; treatment response; tumor

PROJECT SUMMARY / ABSTRACT Bladder cancer is a common type of cancer that can cause substantial morbidity and mortality among both men and women. Bladder cancer causes over 16,870 deaths per year in the United States with 79,030 new bladder cancer cases diagnosed in 2017. A reliable assessment of the response to neoadjuvant therapy at an early stage is vital for identifying tumors that do not respond and allowing the patient a chance of alternative treatment. We have successfully developed a computer decision support system (CDSS-T) for monitoring of bladder cancer treatment response. A quantitative image analysis tool for bladder cancer (QIBC) that quantifies the bladder gross tumor volume (GTV) and image characteristics is an important component of CDSS-T. Advanced machine learning techniques are used to merge the GTV and radiomic biomarkers into an effective predictive model. The goal of this project is to validate the effectiveness of CDSS-T as an aid to the radiologists and the oncologists in assessment of bladder cancer change as a result of treatment through pilot clinical trials. We will (1) perform a preparatory clinical trial with the clinicians at UM, which will simulate the real prospective clinical trial with high quality retrospective data, (2) deploy the QIBC and CDSS-T tools at the three collaborating clinical sites, (3) use the QIBC and CDSS-T tools at the different clinical sites in the prospective pilot clinical trial (standard clinical workflow) utilizing the decision support in parallel to the standard clinical patient care, and (4) analyze the impact of the QIBC and CDSS-T tools on the clinicians' performance in the pilot clinical trial and assess the potential impact on clinical outcome. We hypothesize that this innovative approach can improve clinicians' accuracy, consistency and efficiency in bladder GTV estimation and assessment of treatment response. To test these hypotheses, we will perform the following specific tasks: (1) to evaluate the performance of the QIBC and CDSS-T tools on cases not previously used, new to the system, for both the prediction accuracy and the automatic standalone functionality, refine and optimize the design of the user interface based on clinicians' feedback after their hands-on experience with the system to ensure its practicality and robustness, familiarize clinicians with the performance of the CDSS-T tools and the interpretation of the CDSS-T outputs as a part of user training for the prospective pilot clinical trial, (2) to optimize the QIBC and CDSS-T tools for the clinical workflow at the different clinical sites based on the site clinicians' feedback in order to operate efficiently and in a standalone mode by clinicians, (3) to record the clinicians' predicted outcomes without and with the use of the tools during the pilot clinical trials, estimate the accuracy of assessing response to neoadjuvant chemotherapy in the current clinical treatment paradigm by comparing the estimated response to the histopathologically determined response after radical cystectomy, and (4) to statistically analyze the impact of the QIBC and CDSS-T tools on the performance of the clinicians in the pilot clinical trial and statistically assess the potential impact on clinical outcome.

项目摘要 /摘要 膀胱癌是一种常见的癌症类型 和女人。膀胱癌在美国每年造成16,870人死亡，有79,030个新膀胱 2017年诊断出的癌症病例。对早期对新辅助治疗的反应的可靠评估 阶段对于识别不反应并允许患者有其他治疗的肿瘤至关重要。 我们已经成功地开发了一个计算机决策支持系统（CDSS-T）来监视膀胱 癌症治疗反应。膀胱癌（QIBC）的定量图像分析工具，可量化 膀胱总肿瘤体积（GTV）和图像特性是CDSS-T的重要组成部分。先进的 机器学习技术用于将GTV和放射线生物标志物合并为有效的预测 模型。该项目的目的是验证CDSS-T作为放射科医生和 通过试验临床试验治疗，肿瘤学家评估膀胱癌的变化。我们将 （1）与UM的临床医生一起进行预备临床试验，该试验将模拟真正的前瞻性临床 使用高质量回顾性数据的试验，（2）在三个合作的临床上部署QIBC和CDSS-T工具 站点，（3）在潜在的临床临床试验中使用QIBC和CDSS-T工具 （标准临床工作流程）通过与标准临床患者护理并行利用决策支持，（4） 分析QIBC和CDSS-T工具对临床医生在试验临床试验中表现的影响 评估对临床结果的潜在影响。我们假设这种创新的方法可以改善 临床医生在膀胱GTV估计中的准确性，一致性和效率 回复。为了检验这些假设，我们将执行以下特定任务：（1）评估性能 QIBC和CDSS-T工具在以前未使用的情况下，新的是系统的新型预测准确性 以及自动独立功能，完善并优化用户界面的设计 临床医生对系统的实践经验后，反馈以确保其实用性和鲁棒性， 熟悉CDSS-T工具的性能以及CDSS-T输出的解释为 预期试验临床试验的用户培训的一部分，（2）优化QIBC和CDSS-T工具 基于现场临床医生的反馈，在不同临床站点的临床工作流程有效运行 （3）以独立的模式，以记录临床医生的预测，没有使用 试验临床试验期间的工具，估计评估对新辅助反应的准确性 通过比较当前临床治疗范式的化学疗法，通过比较 自由基膀胱切除术后组织病理学确定的反应，（4）统计分析 QIBC和CDSS-T工具涉及临床医生在试点临床试验中的性能和统计评估 对临床结果的潜在影响。