Young Mi Jung, Seyeon Park, Youngbin Ahn, Haeryoung Kim, Eun Na Kim, Hye Eun Park, Sun Min Kim, Byoung Jae Kim, Jeesun Lee, Chan-Wook Park, Joong Shin Park, Jong Kwan Jun, Young-Gon Kim, Seung Mi Lee
{"title":"Identification of Preeclamptic Placenta in Whole Slide Images Using Artificial Intelligence Placenta Analysis.","authors":"Young Mi Jung, Seyeon Park, Youngbin Ahn, Haeryoung Kim, Eun Na Kim, Hye Eun Park, Sun Min Kim, Byoung Jae Kim, Jeesun Lee, Chan-Wook Park, Joong Shin Park, Jong Kwan Jun, Young-Gon Kim, Seung Mi Lee","doi":"10.3346/jkms.2024.39.e271","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Preeclampsia (PE) is a hypertensive pregnancy disorder linked to placental dysfunction, often involving pathological lesions like acute atherosis, decidual vasculopathy, accelerated villous maturation, and fibrinoid deposition. However, there is no gold standard for the pathological diagnosis of PE and this limits the ability of clinicians to distinguish between PE and non-PE pregnancies. Recent advances in computational pathology have provided the opportunity to automate pathological analysis for diagnosis, classification, prediction, and prediction of disease progression. In this study, we assessed whether computational pathology could be used to identify PE placentas.</p><p><strong>Methods: </strong>A total of 168 placental whole-slide images (WSIs) of patients from Seoul National University Hospital (comprising 84 PE cases and 84 normal controls) were used for model development and internal validation. For external validation of the model, 76 placental slides (including 38 PE cases and 38 normal controls) were obtained from the Boramae Medical Center (BMC). To establish standard criteria for diagnosing PE and distinguishing it from controls using placental WSIs, patch characteristics and quantification of terminal and intermediate villi were employed. In unsupervised learning, <i>K</i>-means clustering was conducted as a feature obtained through an Auto Encoder to extract the ratio of each cluster for each WSI. For supervised learning, quantitative assessments of the villi were obtained using a U-Net-based segmentation algorithm. The prediction model was developed using an ensemble method and was compared with a clinical feature model developed by using placental size features.</p><p><strong>Results: </strong>Using ensemble modeling, we developed a model to identify PE placentas. The model showed good performance (area under the precision-recall curve [AUPRC], 0.771; 95% confidence interval [CI], 0.752-0.790), with 77.3% of sensitivity and 71.1% of specificity, whereas the clinical feature model showed an AUPRC 0.713 (95% CI, 0.694-0.732) with 55.6% sensitivity and 86.8% specificity. External validation of the predictive model employing the BMC-derived set of placental slides also showed good discrimination (AUPRC, 0.725; 95% CI, 0.720-0.730).</p><p><strong>Conclusion: </strong>The proposed computational pathology model demonstrated a strong ability to identify preeclamptic placentas. Computational pathology has the potential to improve the identification of PE placentas.</p>","PeriodicalId":16249,"journal":{"name":"Journal of Korean Medical Science","volume":"39 39","pages":"e271"},"PeriodicalIF":3.0000,"publicationDate":"2024-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11473260/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Korean Medical Science","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.3346/jkms.2024.39.e271","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MEDICINE, GENERAL & INTERNAL","Score":null,"Total":0}
引用次数: 0
Abstract
Background: Preeclampsia (PE) is a hypertensive pregnancy disorder linked to placental dysfunction, often involving pathological lesions like acute atherosis, decidual vasculopathy, accelerated villous maturation, and fibrinoid deposition. However, there is no gold standard for the pathological diagnosis of PE and this limits the ability of clinicians to distinguish between PE and non-PE pregnancies. Recent advances in computational pathology have provided the opportunity to automate pathological analysis for diagnosis, classification, prediction, and prediction of disease progression. In this study, we assessed whether computational pathology could be used to identify PE placentas.
Methods: A total of 168 placental whole-slide images (WSIs) of patients from Seoul National University Hospital (comprising 84 PE cases and 84 normal controls) were used for model development and internal validation. For external validation of the model, 76 placental slides (including 38 PE cases and 38 normal controls) were obtained from the Boramae Medical Center (BMC). To establish standard criteria for diagnosing PE and distinguishing it from controls using placental WSIs, patch characteristics and quantification of terminal and intermediate villi were employed. In unsupervised learning, K-means clustering was conducted as a feature obtained through an Auto Encoder to extract the ratio of each cluster for each WSI. For supervised learning, quantitative assessments of the villi were obtained using a U-Net-based segmentation algorithm. The prediction model was developed using an ensemble method and was compared with a clinical feature model developed by using placental size features.
Results: Using ensemble modeling, we developed a model to identify PE placentas. The model showed good performance (area under the precision-recall curve [AUPRC], 0.771; 95% confidence interval [CI], 0.752-0.790), with 77.3% of sensitivity and 71.1% of specificity, whereas the clinical feature model showed an AUPRC 0.713 (95% CI, 0.694-0.732) with 55.6% sensitivity and 86.8% specificity. External validation of the predictive model employing the BMC-derived set of placental slides also showed good discrimination (AUPRC, 0.725; 95% CI, 0.720-0.730).
Conclusion: The proposed computational pathology model demonstrated a strong ability to identify preeclamptic placentas. Computational pathology has the potential to improve the identification of PE placentas.
背景:子痫前期(PE)是一种与胎盘功能障碍有关的妊娠高血压疾病,通常涉及急性动脉粥样硬化、蜕膜血管病变、绒毛成熟加速和纤维素沉积等病理病变。然而,目前尚无 PE 病理诊断的金标准,这限制了临床医生区分 PE 和非 PE 妊娠的能力。计算病理学的最新进展为诊断、分类、预测和疾病进展预测的病理分析自动化提供了机会。在本研究中,我们评估了计算病理学是否可用于识别 PE 胎盘:方法:共使用了 168 张首尔国立大学医院患者的胎盘全切片图像(WSI)(包括 84 例 PE 病例和 84 例正常对照)进行模型开发和内部验证。为了对模型进行外部验证,从 Boramae 医疗中心(BMC)获得了 76 张胎盘切片(包括 38 例 PE 病例和 38 例正常对照)。为了建立诊断 PE 的标准,并利用胎盘 WSI 将 PE 与对照组区分开来,我们采用了斑块特征以及末端绒毛和中间绒毛的定量方法。在无监督学习中,K均值聚类是通过自动编码器获得的特征,以提取每个WSI的每个聚类的比率。在监督学习中,使用基于 U-Net 的分割算法对绒毛进行定量评估。预测模型采用集合方法建立,并与利用胎盘大小特征建立的临床特征模型进行了比较:结果:利用集合建模法,我们建立了一个识别 PE 胎盘的模型。该模型显示出良好的性能(精确度-召回曲线下面积[AUPRC],0.771;95%置信区间[CI],0.752-0.790),灵敏度为77.3%,特异度为71.1%,而临床特征模型的精确度-召回曲线下面积[AUPRC]为0.713(95% CI,0.694-0.732),灵敏度为55.6%,特异度为86.8%。使用 BMC 导出的胎盘切片集对预测模型进行外部验证也显示出良好的区分度(AUPRC,0.725;95% CI,0.720-0.730):结论:所提出的计算病理学模型具有很强的识别子痫前期胎盘的能力。计算病理学有望改善对子痫前期胎盘的鉴定。
期刊介绍:
The Journal of Korean Medical Science (JKMS) is an international, peer-reviewed Open Access journal of medicine published weekly in English. The Journal’s publisher is the Korean Academy of Medical Sciences (KAMS), Korean Medical Association (KMA). JKMS aims to publish evidence-based, scientific research articles from various disciplines of the medical sciences. The Journal welcomes articles of general interest to medical researchers especially when they contain original information. Articles on the clinical evaluation of drugs and other therapies, epidemiologic studies of the general population, studies on pathogenic organisms and toxic materials, and the toxicities and adverse effects of therapeutics are welcome.