External validation of machine learning models including newborn metabolomic markers for postnatal gestational age estimation in East and South-East Asian infants.

Journal: Gates open research
Published Date: Jun 21, 2021

Abstract

 Postnatal gestational age (GA) algorithms derived from newborn metabolic profiles have emerged as a novel method of acquiring population-level preterm birth estimates in low resource settings. To date, model development and validation have been carried out in North American settings. Validation outside of these settings is warranted.    This was a retrospective database study using data from newborn screening programs in Canada, the Philippines and China. ELASTICNET machine learning models were developed to estimate GA in a cohort of infants from Canada using sex, birth weight and metabolomic markers from newborn heel prick blood samples. Final models were internally validated in an independent sample of Canadian infants, and externally validated in infant cohorts from the Philippines and China.   Cohorts included 39,666 infants from Canada, 82,909 from the Philippines and 4,448 from China.  For the full model including sex, birth weight and metabolomic markers, GA estimates were within ±5 days of ultrasound values in the Canadian internal validation (mean absolute error (MAE) 0.71, 95% CI: 0.71, 0.72), and within ±6 days of ultrasound GA in both the Filipino (0.90 (0.90, 0.91)) and Chinese cohorts (0.89 (0.86, 0.92)). Despite the decreased accuracy in external settings, our models incorporating metabolomic markers performed better than the baseline model, which relied on sex and birth weight alone. In preterm and growth-restricted infants, the accuracy of metabolomic models was markedly higher than the baseline model.  Accuracy of metabolic GA algorithms was attenuated when applied in external settings.  Models including metabolomic markers demonstrated higher accuracy than models using sex and birth weight alone. As innovators look to take this work to scale, further investigation of modeling and data normalization techniques will be needed to improve robustness and generalizability of metabolomic GA estimates in low resource settings, where this could have the most clinical utility.

Authors

  • Steven Hawken
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
  • Malia S Q Murphy
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
  • Robin Ducharme
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
  • A Brianne Bota
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
  • Lindsay A Wilson
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
  • Wei Cheng
    Department of Dental Implantology, Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Institute of Stomatology, Nanjing University, Nanjing, China.
  • Ma-Am Joy Tumulak
    Newborn Screening Reference Centre, University of the Philippines Manila, Manila, Philippines.
  • Maria Melanie Liberty Alcausin
    Newborn Screening Reference Centre, University of the Philippines Manila, Manila, Philippines.
  • Ma Elouisa Reyes
    Newborn Screening Reference Centre, University of the Philippines Manila, Manila, Philippines.
  • Wenjuan Qiu
    Pediatric Endocrinology and Genetic Metabolism, XinHua Hospital, Shanghai, Shanghai, China.
  • Beth K Potter
    School of Epidemiology and Public Health, University of Ottawa, Ottawa, ON, Canada.
  • Julian Little
    School of Epidemiology and Public Health, University of Ottawa, Ottawa, ON, Canada.
  • Mark Walker
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
  • Lin Zhang
    Laboratory of Molecular Translational Medicine, Centre for Translational Medicine, Key Laboratory of Birth Defects and Related Diseases of Women and Children, Ministry of Education, Clinical Research Center for Birth Defects of Sichuan Province, West China Second Hospital, Sichuan University, Chengdu, Sichuan, 610041, China. Electronic address: zhanglin@scu.edu.cn.
  • Carmencita Padilla
    Department of Pediatrics, University of the Philippines Manila, Manilla, Philippines.
  • Pranesh Chakraborty
    Newborn Screening Ontario, Children's Hospital of Eastern Ontario, Ottawa, ON, Canada.
  • Kumanan Wilson
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada.

Keywords

No keywords available for this article.

External Resources

Popular Topics
Recent Journals