A Novel, Interpretable Machine Learning Model Predicts Furosemide Dosing After Congenital Cardiac Surgery - Takeaways - MDSpire

A Novel, Interpretable Machine Learning Model Predicts Furosemide Dosing After Congenital Cardiac Surgery

  • By

  • Daniel E. Ehrmann

  • Matthew W. Hodgman

  • Emily M. Wittrup

  • John R. Charpie

  • Gabe E. Owens

  • Ranjit Aiyagari

  • Kayvan Najarian

  • May 7, 2026

  • 0 min

Share

  • 1

    Fluid overload after congenital cardiac surgery is linked to significant morbidity and mortality, highlighting the need for improved diuretic titration methods.

  • 2

    Current diuretic dosing practices vary widely among clinicians and ICUs, potentially leading to harmful under- or over-diuresis in patients.

  • 3

    The TGFNN-R model is a novel, fully-interpretable neural network designed to predict furosemide dosage following congenital heart surgery.

  • 4

    TGFNN-R utilizes fuzzy sets to better mirror clinical decision-making, allowing for more nuanced interpretations of input variables.

  • 5

    The model aims to enhance clinical decision support systems for managing fluid overload in neonates post-cardiac surgery.

Original Source(s)

A Novel, Interpretable Machine Learning Model Predicts Furosemide Dosing After Congenital Cardiac Surgery

  • by Daniel E. Ehrmann, Matthew W. Hodgman, Emily M. Wittrup, John R. Charpie, Gabe E. Owens, Ranjit Aiyagari, Kayvan Najarian

  • May 7, 2026

Related Content

FDA OKs Linaclotide in Young Children

Phase 3 data expanded use of linaclotide 72 mcg in pediatric functional constipation.

Distribution and associated factors of neck and shoulder pain among primary school children in Shijiazhuang city, Hebei province, China

Cardiac and aerobic response to growth hormone therapy in children with short stature: a prospective study using VO2max and speckle tracking echocardiography