Artificial Intelligence in Healthcare The Cleveland Clinic Way Aziz - PowerPoint PPT Presentation

Artificial Intelligence in Healthcare “The Cleveland Clinic Way” Aziz Nazha, MD Director, Center for Clinical Artificial Intelligence Associate Medical Director, Enterprise Analytics Assistant Professor of Medicine Lerner College of Medicine / CWRU Taussig Cancer Institute Cleveland Clinic @AzizNazhaMD

Terminology Artificial Intelligence Machine Learning Deep Learning Supervised Learning Unsupervised Learning

Terminology Input Output Variables Classification Image Regression Text Algorithm

Democratizing AI

Myelodysplastic Syndromes Current Paradigm Stage Treatment Diagnosis Higher Risk ü Transplant ü Azacitdine (6 months) ü Decitabine (6 months) Lower Risk ü Supportive care ü Growth factors ü Others

Myelodysplastic Syndromes Challenges Stage Treatment Diagnosis ü Subjective ü Predicted outcome Uncertainty in = predicting ü 30-40% is changed Actual outcome response or resistance ü Dysplasia not ü Significant difference enough to call it MDS

Diagnosis

Computer Vision and Medicine A Esteva et al. Nature 1–4 (2017) doi:10.1038/nature21056

MDS (Diagnosis) Normal Normal MDS MDS AML AML

MDS (Diagnosis), Do we Need a Biopsy? ü 1 ü 2 + Clinical Algorithm Variable Importance Final Model + Mutational

MDS vs. Other Disorders

Staging / Prognosis

MDS (Staging) Survival IPSS-R

New Model Building Data Random Survival Forest Important Variables Demographic Training CC + MLL Clinical ü 1 X ü 1 Genomics Validation Moffitt

Web Application Final Model

Predicting Response/Resistance to Therapy

Response / Resistance to Chemotherapy 6 Azacitidine 30 - 40% months Decitabine FDA approved Response Rate Treatment duration

Association Rules (Recommender System) ? Mentor 1 ? Mentor 2

Association Rules (Recommender System) ? Mentor 1 Mentor 2

Association Rules (Recommender System) Response ASXL1 TET2 RUNX1 Patient 1 (HMA) Resistance TP53 RCOR SRSF2 Patient 2 (HMA)

Association Rules Validation Rules Association Rules (Resistance) ASXL1 , NF1 29% pts ASXL1 , EZH2 , TET2 > 3 mutations/sample ASXL1 , EZH2 , RUNX1 EZH2 , SRSF2 , TET2 Accuracy: 93% in the ASXL1 , EZH2 , SRSF2 Validation cohort ASXL1 , RUNX1 , SRSF2 ASXL1 , TET2 , SRSF2 ASXL1, BCOR, RUNX1

Lab Validation 1 .0 T E T 2 M T /S R S F 2 M T /R U N X 1 M T CRISPR/cas9 V iability 0 .5 T E T 2 M T 0 .0 0 1 2 3 4 5 6 D e cita b in e , µ M In the Lab Multiplex CRISPR/cas9

CCAI

Mission & Vision Become the world’s largest To harness the power of and most collaborative AI to advance medicine center for innovation and and transform healthcare advancement of AI in healthcare

Projects Medicine Respiratory Others Cancer ICU Diagnosis Hospital Omics q Project 1 Operation q Project 22 q Project 17 q Project 2 q Project 18 q Project 13 q Project 3 q Project 14 q Project 15 Prognosis q Project 16 q Project 5 Liver/GI q Project 6 Asthma q Project 7 q Project 23 q Project 19 q Project 8 Predicting Response to Pediatric Cardiology Therapy q Project 20 q Project 24 q Project 9 q Project 21 q Project 10 q Project 11 q Project 12

Collaborations

Education AI for healthcare providers course Cleveland Clinic June 10-26 Medical School AI track Lerner college of medicine Computer Science School Healthcare course Case Western Reserve University

Conclusion “ The electric light did not come from the continuous improvement of candles” -Oren Harari

Acknowledgments Cleveland Clinic Mikkael Sekeres, MD, MS Jaroslaw Maciejewski, MD, PhD Brian Bolwell, MD Matt Kalaycio, MD Anjali Advani, MD Sudipto Mukherjee, MD, PhD Dana Farber Aaron Gerds, MD, MS MLL Hetty Carrawy, MD, MBA David Steensma, MD Betty Hamilton, MD Benjamin Ebert, MD Ronald Sobecks, MD Torsten Haferlach, MD Coleman Lindsley, MD, PhD Navneet Majhail, MD Claudia Haferlach, MD Manja Megendorfer, MBA Maciejewski’s lab Wencke Walter, PhD Babl Jha’s Lab Stephan Hutter, PhD Cornell Leukemia nurses and staff Gail Roboz, MD Johns Hopkins France Amy DeZern, MD Eric Solary, MD Rafael Itzykson, MD, PhD Moffitt Cancer Center Alan List, MD MD Anderson Rami Komrokji, MD Eric Padron, MD Guillermo Garcia-Manero, MD Nazha Lab David Sallman, MD Hagop Kantarjian, MD University of Florence Orkun Balglo, MD Valeria Santini, MD Ping Chao, MD Jacob Shreve, MD Amin Moein, MD Cameron B Hilton All Patients!!! K12 Nathan Radakovich Rachel Shirly, PhD Yazan Rouphail Stephen Kemura

Cleveland Clinic Every Life Deserves World Class Care.