Y J Zhao1, H J Zhang2, C X Li3, T Wu4, X M Shen1, J Zhang5. 1. Ministry of Education - Shanghai Key Laboratory of Children's Environmental Health, Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China. 2. Departments of Pathology and Bio-Bank, The International Peace Maternity and Child Health Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China. 3. School of Mathematics and Computational Science, Sun Yat-sen University, Guangzhou, China. 4. Ministry of Education - Shanghai Key Laboratory of Children's Environmental Health, Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China; Chengdu Women's and Children's Central Hospital, Chengdu, Sichuan Province, China. 5. Ministry of Education - Shanghai Key Laboratory of Children's Environmental Health, Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China. Electronic address: junjimzhang@gmail.com.
Abstract
INTRODUCTION: Placental morphological and histopathological measures can be numerous and it is very time-consuming to collect all the information. When planning a large birth cohort study, researchers often face the dilemma as to whether and what information to collect in a placenta. The aim of this study was (1) to systematically select placental measures that have significant clinical implications, and (2) to assess the accuracy of these measures. METHODS: We used placental pathology information from the Collaborative Perinatal Project (CPP), in which 45,785 births had comprehensive information on placental morphology and pathology. We chose 20 major childhood diseases as outcomes. The statistical method of LASSO (least absolute shrinkage and selection operator) was used to select important placental measures that would have better predictability for outcomes. RESULTS: LASSO selected 81 measures as candidates. After having consulted placental pathologists, we further narrowed down to the 38 and 23 most important measures to form two shortened evaluation systems that could be used in clinical practice and research. The sensitivity and specificity of these measures for composite child diseases were 68% and 34%, respectively, for the 38 measures and 48% and 53%, respectively, for the 23 measures. CONCLUSIONS: We presented a potentially useful system for pathological characterization of the placenta. The use of these relatively simple and accessible characteristics as biomarkers may be considered in large birth cohort studies.
INTRODUCTION: Placental morphological and histopathological measures can be numerous and it is very time-consuming to collect all the information. When planning a large birth cohort study, researchers often face the dilemma as to whether and what information to collect in a placenta. The aim of this study was (1) to systematically select placental measures that have significant clinical implications, and (2) to assess the accuracy of these measures. METHODS: We used placental pathology information from the Collaborative Perinatal Project (CPP), in which 45,785 births had comprehensive information on placental morphology and pathology. We chose 20 major childhood diseases as outcomes. The statistical method of LASSO (least absolute shrinkage and selection operator) was used to select important placental measures that would have better predictability for outcomes. RESULTS: LASSO selected 81 measures as candidates. After having consulted placental pathologists, we further narrowed down to the 38 and 23 most important measures to form two shortened evaluation systems that could be used in clinical practice and research. The sensitivity and specificity of these measures for composite child diseases were 68% and 34%, respectively, for the 38 measures and 48% and 53%, respectively, for the 23 measures. CONCLUSIONS: We presented a potentially useful system for pathological characterization of the placenta. The use of these relatively simple and accessible characteristics as biomarkers may be considered in large birth cohort studies.