Clinical utility of novel biomarkers for prediction of early pregnancy failure

新型生物标志物预测早期妊娠失败的临床应用

• 批准号: 10563608
• 负责人: Kurt T Barnhart
• 金额: $ 71.84万
• 依托单位: UNIVERSITY OF PENNSYLVANIA
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-02-01 至 2028-01-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10563608
• 关键词: Affect; Area; Biological; Biological Markers; Caring; Cause of Death; Cessation of life; Characteristics; Clinical; Clinical Management; Clinical Trials; Complement; Consumption; Data; Development Plans; Diagnosis; Diagnostic; Discipline of obstetrics; Discrimination; Disease; Ectopic Pregnancy; Ethnic Origin; First Pregnancy Trimester; Funding; Gestational Age; Goals; Gynecologic; Hemorrhage; Human Chorionic Gonadotropin; Iatrogenesis; Interruption; Location; Machine Learning; Mammalian Oviducts; Medical; Medical Care Team; Methodology; Modeling; Morbidity - disease rate; National Institute of Child Health and Human Development; Pain; Pathway interactions; Patients; Performance; Physiology; Population; Predictive Value; Predictive Value of Tests; Pregnancy; Pregnancy Outcome; Procedures; Process; Productivity; Prospective Studies; Prospective cohort; Proteomics; Race; Recording of previous events; Research; Research Priority; Resources; Risk; Rupture; Sampling; Serum; Signs and Symptoms; Spontaneous abortion; Stress; Subgroup; Symptoms; System; Testing; Time; Uncertainty; United States National Institutes of Health; Uterus; Validation; Woman; Work; biomarker development; biomarker panel; biomarker performance; classification algorithm; clinical care; companion diagnostics; design; diagnostic tool; early pregnancy; early pregnancy loss; feature selection; health care delivery; improved; interest; machine learning algorithm; model building; novel; novel marker; pain symptom; population based; predictive marker; pregnancy failure; random forest; regression trees; screening; ultrasound

Abstract: The diagnosis and management for women who are at risk for ectopic pregnancy (EP) and spontaneous abortion (SAB), has not changed substantially in decades. While ultrasound can diagnose a significant portion of women at presentation, because of its limited accuracy in abnormal and early gestation, a large number of women need serial tests and procedures to determine the final location and viability of an early pregnancy. Misdiagnosis and iatrogenic complications during this time are still too common. This conundrum results in great stress and uncertainty for women undergoing clinical care and their health care teams. To date, we have demonstrated (and validated) that multiplexed biomarkers from divergent biological pathways can be used to minimize both false positive and false negative discrimination (rather than balancing the two errors with lower accuracy). In parallel, we have discovered and screened more than 54 novel biomarkers to obtain 11 candidates that can predict early pregnancy outcome. Using machine learning we have demonstrated that as few as 3 of these markers can predict the location of an early gestation (IUP vs EP or SAB) with 95% accuracy. Additionally, an overlapping group of 3 markers can predict the viability of a gestation (IUP vs SAB or EP) with 94% accuracy. When both tests are used in combination the accuracy is 96.9%. We now plan to externally validate these companion diagnostic(s) in a separate population-based prospective study and to determine the optimal conditions of use. We will optimize performance in women presenting with a pregnancy of unknown location as well as assess predictive values for all women presenting at risk for early pregnancy loss using an independent population-based prospective study (SA1). We will determine if accuracy can be improved with the combination of presenting signs, symptoms, and ultrasound findings (SA2). Moreover, we will validate our biomarker tests in two distinct populations of interest: women with high and low risk for EP: a) asymptomatic women prior to presentation for care (EAGeR Trial) and b) women diagnosed with a persistent pregnancy of unknown location (ACTorNOT Trial) (SA3). Developing novel biomarkers that distinguish normal physiology from the presence of gynecologic disorders is an NICHD research priority area. Our productive and established team is proposing rigorous cross disciplinary, state of the art, and novel research with great scientific impact that has the potential to produce a paradigm shift in clinical care models for the diagnosis and management of women in early pregnancy. Our detailed plan for biomarker development is iterative and nimble and, importantly, includes validation. Our development plan is informed by methodology from successful biomarker development, is designed to minimize known pitfalls, and is developed with FDA guidance.

摘要：有异位妊娠风险（EP）和 几十年来，自发流产（SAB）并未发生很大的变化。虽然超声可以诊断 出现大部分妇女，由于其在异常和妊娠中的准确性有限，所以 大量妇女需要串行测试和程序来确定早期的最终位置和可行性 怀孕。在这段时间里，误诊和医源性并发症仍然太普遍了。这个难题 对于接受临床护理及其卫生保健团队的妇女带来了极大的压力和不确定性。迄今为止， 我们已经证明（并得到验证）表明，来自不同生物学途径的多重生物标志物可以是 用于最大程度地减少假阳性和假阴性歧视（而不是平衡两个错误与 较低的精度）。同时，我们发现并筛选了54多个新型生物标志物以获得11 可以预测早期妊娠结局的候选人。使用机器学习，我们已经证明了这一点 这些标记中有3个可以预测95％的妊娠（IUP vs EP或SAB）的位置 准确性。此外，一个由3个标记的重叠组可以预测妊娠的生存能力（IUP与SAB 或EP）精度为94％。当两种测试结合使用时，精度为96.9％。我们现在计划 在基于人群的另一项前瞻性研究中，在外部验证这些伴侣诊断和 确定最佳使用条件。我们将优化出现怀孕的女性的表现 未知的位置以及评估所有出现早期怀孕风险的妇女的预测价值 使用基于人群的独立前瞻性研究（SA1）损失。我们将确定准确性是否可以 通过表现迹象，症状和超声检查结果（SA2）的结合来改善。而且，我们 将验证我们的生物标志物测试在两个截然不同的感兴趣种群中：EP高风险的妇女：A） 在介绍之前，无症状的妇女进行护理（急切的试验）和b）妇女被诊断出患有持续的妇女 未知位置的怀孕（Actornot试验）（SA3）。开发区分正常的新型生物标志物 存在妇科疾病的生理学是NICHD研究的优先领域。我们的生产力和 成熟的团队正在提出严格的交叉学科，最新技术和新颖的研究 科学影响有可能在诊断和 怀孕初期的妇女管理。我们的生物标志物开发计划的详细计划是迭代的， 敏捷，重要的是包括验证。我们的发展计划是通过方法来告知的 成功的生物标志物开发旨在最大程度地减少已知陷阱，并使用FDA开发 指导。