Predicting Outcomes for Uterine Fibroid Embolization by using Deep Learning of Paired MRI Scans

项目总结 子宫肌瘤是女性良性肿瘤患病率最高的肿瘤，有各种各样的报道。 从4.5%上升到68.6%，对非洲裔美国女性有明显的偏见。据估计，经济上的 患有子宫肌瘤的有症状妇女给医疗系统带来的负担高达3400万美元。 目前，子宫肌瘤栓塞术被认为是一种高效的微创治疗方法。 成功率高达85%。然而，子宫切除术是最常见的手术， 占每年近600,000例，而每年仅进行14,000例UFE手术。一直以来 有充分的证据表明，少数族裔不太可能被转介进行微创手术，尽管 他们有全民保险。此外，社会经济地位较低的妇女，特别是 非洲裔美国人被不成比例地转诊为开放手术。因此，自动化工具，如 在这项提案中，能够提供不偏不倚的转介将是对抗这一点的显著优势 不幸的偏见。 该提案将具体探讨使用机器学习和深度学习方法来利用 新的回溯性数据集，汇编了从手术前和手术后配对提取的特征 磁共振成像(MRI)扫描了多达700名接受UFE的患者。这些型号将提供 UFE治疗效果评分，将提供客观和定量的衡量标准，以决定 患者是UFE的良好候选者。 这项提议的短期影响将是创建一个配对的UFE MRI扫描的精选数据库 已经对有关肌瘤位置和患者特征的各种指标进行了分析，这将使 临床社区开始提供定量方法来确定UFE候选者。长期的 这项提议的影响将在随后的临床试验中实现，这些临床试验将适当地验证这些人工智能模型以 预测哪些患者应该利用UFE作为治疗子宫肌瘤的非手术替代方案。

PROJECT SUMMARY Uterine fibroids represent the highest prevalence of benign tumors in women, with reports ranging anywhere from 4.5% to 68.6%, with a significant bias towards African American women. It is estimated that the economic burden on the healthcare system from symptomatic women with uterine fibroids is up to $34 million. Currently, uterine fibroid embolization (UFE) is considered a highly effective minimally invasive procedure with up to an 85% success rate. However, hysterectomies are the most commonly performed procedures, accounting for nearly 600,000 annually, while only 14,000 UFE procedures are performed annually. It has been well documented that minorities are less likely to be referred for minimally invasive procedures, even though there is universal insurance coverage for them. Furthermore, women in lower socio-economic status, particularly African Americans, have been disproportionately referred for open surgery. Therefore, automated tools, like the ones in this proposal, that can provide unbiased referrals will be significant advantage at combating this unfortunate bias. This proposal will specifically explore the use of machine learning and deep learning methods to leverage a novel retrospective dataset that compiles features extracted from paired pre-operative and post-operative magnetic resonance imaging (MRI) scans of up to 700 patients who underwent a UFE. These models will provide a UFE treatment effectiveness score that will provide an objective and quantitative metric to decide whether a patient is good candidate for UFE. The short term impact of this proposal will be the creation of a curated database of paired UFE MRI scans that have been analyzed for various metrics regarding fibroid positions and patient characteristics, that will allow the clinical community to begin providing quantitative methods to determine UFE candidates. The long-term impact of this proposal will be realized in subsequent clinical trials that validate these AI models properly to predict which patients should be leveraging UFE as a non-surgical alternative for treating fibroids.