Manreet K Kanwar1, Mardi Gomberg-Maitland2, Marius Hoeper3, Christine Pausch4, David Pittrow5, Geoff Strange6, James J Anderson7, Carol Zhao8, Jacqueline V Scott9, Marek J Druzdzel10, Jidapa Kraisangka11, Lisa Lohmueller12, James Antaki13, Raymond L Benza14. 1. Cardiovascular Institute at Allegheny Health Network, Pittsburgh, PA, USA. 2. George Washington School of Medicine and Health Sciences, Washington, DC, USA. 3. Dept of Respiratory Medicine, Hannover Medical School, German Center for Lung Research (DZL), Hannover, Germany. 4. GWT-TUD GmbH, Dresden, Germany. 5. Faculty of Institute for Clinical Pharmacology, Technical University, Dresden, Germany. 6. School of Medicine, University of Notre Dame, Fremantle, Australia. 7. Respiratory Dept, Sunshine Coast University Hospital, Nambour, Australia. 8. Actelion Pharmaceuticals US, A Janssen Pharmaceutical Company of Johnson & Johnson, San Francisco, CA, USA. 9. School of Biomedical Engineering, Carnegie Mellon University, Pittsburgh, PA, USA. 10. Faculty of Computer Science, Bialystok University of Technology, Bialystok, Poland. 11. Faculty of Information and Communication Technology, Mahidol University, Salaya, Thailand. 12. Language Technologies Institute, Carnegie Mellon University, Pittsburgh, PA, USA. 13. Meinig School of Biomedical Engineering, Cornell University, Ithaca, NY, USA. 14. Ohio State Medical Center, Columbus, OH, USA raymond.benza@osumc.edu.
Abstract
BACKGROUND: Current risk stratification tools in pulmonary arterial hypertension (PAH) are limited in their discriminatory abilities, partly due to the assumption that prognostic clinical variables have an independent and linear relationship to clinical outcomes. We sought to demonstrate the utility of Bayesian network-based machine learning in enhancing the predictive ability of an existing state-of-the-art risk stratification tool, REVEAL 2.0. METHODS: We derived a tree-augmented naïve Bayes model (titled PHORA) to predict 1-year survival in PAH patients included in the REVEAL registry, using the same variables and cut-points found in REVEAL 2.0. PHORA models were validated internally (within the REVEAL registry) and externally (in the COMPERA and PHSANZ registries). Patients were classified as low-, intermediate- and high-risk (<5%, 5-20% and >10% 12-month mortality, respectively) based on the 2015 European Society of Cardiology/European Respiratory Society guidelines. RESULTS: PHORA had an area under the curve (AUC) of 0.80 for predicting 1-year survival, which was an improvement over REVEAL 2.0 (AUC 0.76). When validated in the COMPERA and PHSANZ registries, PHORA demonstrated an AUC of 0.74 and 0.80, respectively. 1-year survival rates predicted by PHORA were greater for patients with lower risk scores and poorer for those with higher risk scores (p<0.001), with excellent separation between low-, intermediate- and high-risk groups in all three registries. CONCLUSION: Our Bayesian network-derived risk prediction model, PHORA, demonstrated an improvement in discrimination over existing models. This is reflective of the ability of Bayesian network-based models to account for the interrelationships between clinical variables on outcome, and tolerance to missing data elements when calculating predictions.
BACKGROUND: Current risk stratification tools in pulmonary arterial hypertension (PAH) are limited in their discriminatory abilities, partly due to the assumption that prognostic clinical variables have an independent and linear relationship to clinical outcomes. We sought to demonstrate the utility of Bayesian network-based machine learning in enhancing the predictive ability of an existing state-of-the-art risk stratification tool, REVEAL 2.0. METHODS: We derived a tree-augmented naïve Bayes model (titled PHORA) to predict 1-year survival in PAH patients included in the REVEAL registry, using the same variables and cut-points found in REVEAL 2.0. PHORA models were validated internally (within the REVEAL registry) and externally (in the COMPERA and PHSANZ registries). Patients were classified as low-, intermediate- and high-risk (<5%, 5-20% and >10% 12-month mortality, respectively) based on the 2015 European Society of Cardiology/European Respiratory Society guidelines. RESULTS: PHORA had an area under the curve (AUC) of 0.80 for predicting 1-year survival, which was an improvement over REVEAL 2.0 (AUC 0.76). When validated in the COMPERA and PHSANZ registries, PHORA demonstrated an AUC of 0.74 and 0.80, respectively. 1-year survival rates predicted by PHORA were greater for patients with lower risk scores and poorer for those with higher risk scores (p<0.001), with excellent separation between low-, intermediate- and high-risk groups in all three registries. CONCLUSION: Our Bayesian network-derived risk prediction model, PHORA, demonstrated an improvement in discrimination over existing models. This is reflective of the ability of Bayesian network-based models to account for the interrelationships between clinical variables on outcome, and tolerance to missing data elements when calculating predictions.
Authors: Marius M Hoeper; Tilmann Kramer; Zixuan Pan; Christina A Eichstaedt; Jens Spiesshoefer; Nicola Benjamin; Karen M Olsson; Katrin Meyer; Carmine Dario Vizza; Anton Vonk-Noordegraaf; Oliver Distler; Christian Opitz; J Simon R Gibbs; Marion Delcroix; H Ardeschir Ghofrani; Doerte Huscher; David Pittrow; Stephan Rosenkranz; Ekkehard Grünig Journal: Eur Respir J Date: 2017-08-03 Impact factor: 16.671
Authors: Nazzareno Galiè; Marc Humbert; Jean-Luc Vachiery; Simon Gibbs; Irene Lang; Adam Torbicki; Gérald Simonneau; Andrew Peacock; Anton Vonk Noordegraaf; Maurice Beghetti; Ardeschir Ghofrani; Miguel Angel Gomez Sanchez; Georg Hansmann; Walter Klepetko; Patrizio Lancellotti; Marco Matucci; Theresa McDonagh; Luc A Pierard; Pedro T Trindade; Maurizio Zompatori; Marius Hoeper Journal: Eur Respir J Date: 2015-08-29 Impact factor: 16.671
Authors: Raymond L Benza; Dave P Miller; Robyn J Barst; David B Badesch; Adaani E Frost; Michael D McGoon Journal: Chest Date: 2012-08 Impact factor: 9.410
Authors: G E D'Alonzo; R J Barst; S M Ayres; E H Bergofsky; B H Brundage; K M Detre; A P Fishman; R M Goldring; B M Groves; J T Kernis Journal: Ann Intern Med Date: 1991-09-01 Impact factor: 25.391
Authors: Geoff Strange; Edmund M Lau; Eleni Giannoulatou; Carolyn Corrigan; Eugene Kotlyar; Fiona Kermeen; Trevor Williams; David S Celermajer; Nathan Dwyer; Helen Whitford; Jeremy P Wrobel; John Feenstra; Melanie Lavender; Kenneth Whyte; Nicholas Collins; Peter Steele; Susanna Proudman; Vivek Thakkar; Dominic Keating; Anne Keogh Journal: Heart Lung Circ Date: 2017-09-20 Impact factor: 2.975
Authors: Raymond L Benza; Manreet K Kanwar; Amresh Raina; Jacqueline V Scott; Carol L Zhao; Mona Selej; C Greg Elliott; Harrison W Farber Journal: Chest Date: 2020-09-01 Impact factor: 9.410
Authors: Francois Haddad; Kevin Contrepois; Myriam Amsallem; Andre Y Denault; Roberto J Bernardo; Alokkumar Jha; Shalina Taylor; Jennifer Arthur Ataam; Olaf Mercier; Tatiana Kuznetsova; Anton Vonk Noordegraaf; Roham T Zamanian; Andrew J Sweatt Journal: Chest Date: 2021-11-11 Impact factor: 10.262