Rui Guo1, Bingqing Wang1, Yue Wang1, Yongbiao Zhang2, Tun Liu1, Qingguo Zhang1. 1. Department of Ear Reconstruction, Plastic Surgery Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Thirty-three Badachu Road, Shijingshan District. 2. Beijing Advanced Innovation Center for Big Data-Based Precision Medicine, Beihang University, Beijing, China.
Abstract
BACKGROUND: Microtia is a severe congenital malformation of the external ear. This study aimed to explore the epidemiologic characteristics and the possible risk factors in patients with severe microtia in China, and integrate significant variables into a predictive nomogram. METHODS: A total of 965 patients with microtia were included. This retrospective case study was conducted from July 2014 to July 2019 at Plastic Surgery Hospital in China. The detailed questionnaires concerning potential risk factors were completed and data were gathered. Chi-Square and Fisher tests were used to analyze the variables, and a multivariate logistic regression model was used to select variables related to severe microtia, and then construct a nomogram. The nomogram model was evaluated by the concordance index (C-index), calibration plot, and receiver operating characteristics (ROCs) curve. Bootstraps with 1000 resamples were applied to these analyses. RESULTS: Of the 965 microtia patients, 629 (65.2%) were male and 867 (89.8%) were sporadic. The cases were observed more commonly in unilateral (83.1%) and right-sided (52.0%). And multiple malformations were observed in 392 (40.6%) cases. Multivariate logistic regression analysis showed that maternal age, miscarriage frequency, virus infection, anemia, using progesterone, paternal alcohol intake, and topography of living areas were associated with a higher risk of severe microtia. All the significant variables were combined into a predictive nomogram (C-index = 0.755,95% CI = 0.703-0.807). Higher prediction accuracy (adjusted C-index = 0.749) was further verified via bootstrap validation. The calibration plot showed good performance, and the ROCs curve analysis demonstrated high sensitivity and specificity. CONCLUSIONS: Most microtia patients are male, sporadic, and accompanied by other malformations, which are similar to the phenotypic analysis results of other studies. A nomogram predicting severe microtia was constructed to provide scientific guidance for individualized prevention in clinical practice.
BACKGROUND: Microtia is a severe congenital malformation of the external ear. This study aimed to explore the epidemiologic characteristics and the possible risk factors in patients with severe microtia in China, and integrate significant variables into a predictive nomogram. METHODS: A total of 965 patients with microtia were included. This retrospective case study was conducted from July 2014 to July 2019 at Plastic Surgery Hospital in China. The detailed questionnaires concerning potential risk factors were completed and data were gathered. Chi-Square and Fisher tests were used to analyze the variables, and a multivariate logistic regression model was used to select variables related to severe microtia, and then construct a nomogram. The nomogram model was evaluated by the concordance index (C-index), calibration plot, and receiver operating characteristics (ROCs) curve. Bootstraps with 1000 resamples were applied to these analyses. RESULTS: Of the 965 microtia patients, 629 (65.2%) were male and 867 (89.8%) were sporadic. The cases were observed more commonly in unilateral (83.1%) and right-sided (52.0%). And multiple malformations were observed in 392 (40.6%) cases. Multivariate logistic regression analysis showed that maternal age, miscarriage frequency, virus infection, anemia, using progesterone, paternal alcohol intake, and topography of living areas were associated with a higher risk of severe microtia. All the significant variables were combined into a predictive nomogram (C-index = 0.755,95% CI = 0.703-0.807). Higher prediction accuracy (adjusted C-index = 0.749) was further verified via bootstrap validation. The calibration plot showed good performance, and the ROCs curve analysis demonstrated high sensitivity and specificity. CONCLUSIONS: Most microtia patients are male, sporadic, and accompanied by other malformations, which are similar to the phenotypic analysis results of other studies. A nomogram predicting severe microtia was constructed to provide scientific guidance for individualized prevention in clinical practice.