RONALD M. RAZMI, MD is a cardiologist and the co-founder and Managing Director of Zoi Capital, a venture capital firm that invests in the applications of AI in healthcare. Dr. Razmi completed his medical training at the Mayo Clinic and holds an MBA from Northwestern University's Kellogg School of Management. He was a McKinsey consultant before launching a population health management software company at the dawn of digital health. He saw firsthand the confluence of clinical, technical, and business factors that need to come together for new technologies to gain a foothold in healthcare. He is a co-author of the Handbook of Cardiovascular Magnetic Resonance Imaging.

About the Author xi

Foreword xiii

Preface xix

Acknowledgments xxiii

Part I Roadmap of AI in Healthcare 1

1 History of AI and Its Promise in Healthcare 3

1.1 What is AI? 5

1.2 A Classification System for Underlying AI/ML Algorithms 14

1.3 AI and Deep Learning in Medicine 17

1.4 The Emergence of Multimodal and Multipurpose Models in Healthcare 20

References 23

2 Building Robust Medical Algorithms 27

2.1 Obtaining Datasets That are Big Enough and Detailed Enough for Training 30

2.2 Data Access Laws and Regulatory Issues 33

2.3 Data Standardization and Its Integration into Clinical Workflows 34

2.4 Federated AI as a Possible Solution 36

2.5 Synthetic Data 40

2.6 Data Labeling and Transparency 43

2.7 Model Explainability 45

2.8 Model Performance in the Real World 50

2.9 Training on Local Data 52

2.10 Bias in Algorithms 53

2.11 Responsible AI 60

References 62

3 Barriers to AI Adoption in Healthcare 67

3.1 Evidence Generation 71

3.2 Regulatory Issues 74

3.3 Reimbursement 76

3.4 Workflow Issues with Providers and Payers 78

3.5 Medical- Legal Barriers 81

3.6 Governance 83

3.7 Cost and Scale of Implementation 85

3.8 Shortage of Talent 86

References 86

4 Drivers of AI Adoption in Healthcare 91

4.1 Availability of Data 92

4.2 Powerful Computers, Cloud Computing, and Open Source Infrastructure 93

4.3 Increase in Investments 94

4.4 Improvements in Methodology 95

4.5 Policy and Regulatory 95

4.5.1 Fda 95

4.5.2 Other Bodies 100

4.6 Reimbursement 102

4.7 Shortage of Healthcare Resources 105

4.8 Issues with Mistakes, Inefficient Care Pathways, and Non- personalized Care 106

References 110

Part II Applications of AI in Healthcare 113

5 Diagnostics 115

5.1 Radiology 115

5.2 Pathology 122

5.3 Dermatology 124

5.4 Ophthalmology 125

5.5 Cardiology 127

5.6 Neurology 132

5.7 Musculoskeletal 133

5.8 Oncology 134

5.8.1 Diagnosis and Treatment of Cancer 136

5.8.2 Histopathological Cancer Diagnosis 136

5.8.3 Tracking Tumor Development 136

5.8.4 Prognosis Detection 137

5.9 Gi 139

5.10 Covid- 19 139

5.11 Genomics 140

5.12 Mental Health 141

5.13 Diagnostic Bots 142

5.14 At Home Diagnostics/Remote Monitoring 144

5.15 Sound AI 148

5.16 AI in Democratizing Care 149

References 150

6 Therapeutics 157

6.1 Robotics 158

6.2 Mental Health 159

6.3 Precision Medicine 161

6.4 Chronic Disease Management 164

6.5 Medication Supply and Adherence 167

6.6 Vr 168

References 169

7 Clinical Decision Support 171

7.1 AI in Decision Support 176

7.2 Initial Use Cases 180

7.3 Primary Care 182

7.4 Specialty Care 185

7.4.1 Cancer Care 185

7.4.2 Neurology 185

7.4.3 Cardiology 186

7.4.4 Infectious Diseases 187

7.4.5 Covid- 19 187

7.5 Devices 188

7.6 End- of- Life AI 189

7.7 Patient Decision Support 190

References 191

8 Population Health and Wellness 195

8.1 Nutrition 196

8.2 Fitness 200

8.3 Stress and Sleep 201

8.4 Population Health and Management 204

8.5 Risk Assessment 206

8.6 Use of Real World Data 208

8.7 Medication Adherence 208

8.8 Remote Engagement and Automation 209

8.9 Sdoh 211

8.10 Aging in Place 212

References 214

9 Clinical Workflows 217

9.1 Documentation Assistants 218

9.2 Quality Measurement 225

9.3 Nursing and Clinical Assistants 225

9.4 Virtual Assistants 227

References 230

10 Administration and Operations 233

10.1 Providers 234

10.1.1 Documentation, Coding, and Billing 234

10.1.2 Practice Management and Operations 238

10.1.3 Hospital Operations 240

10.2 Payers 243

10.2.1 Payer Administrative Functions 244

10.2.2 Fraud 246

10.2.3 Personalized Communications 247

References 248

11 AI Applications in Life Sciences 251

11.1 Drug Discovery 252

11.2 Clinical Trials 261

11.2.1 Information Engines 264

11.2.2 Patient Stratification 267

11.2.3 Clinical Trial Operations 268

11.3 Medical Affairs and Commercial 271

References 272

Part III the Business Case for Ai in Healthcare 275

12 Which Health AI Applications Are Ready for Their Moment? 277

12.1 Methodology 278

12.2 Clinical Care 281

12.3 Administrative and Operations 289

12.4 Life Sciences 291

References 293

13 The Business Model for Buyers of Health AI Solutions 295

13.1 Clinical Care 298

13.2 Administrative and Operations 305

13.3 Life Sciences 309

13.4 Guide for Buyer Assessment of Health AI Solutions 312

References 313

14 How to Build and Invest in the Best Health AI Companies 315

14.1 Barriers to Entry and Intellectual Property (IP) 316

14.1.1 Creating Defensible Products 318

14.2 Startups Versus Large Companies 319

14.3 Sales and Marketing 321

14.4 Initial Customers 324

14.5 Direct- to- Consumer (D2C) 325

14.6 Planning Your Entrepreneurial Health AI Journey 327

14.7 Assessment of Companies by Investors 329

14.7.1 Key Areas to Explore for a Health AI Company for Investment 329

References 330

Index 333