Artificial Intelligence Applications for Early Disease Detection in U.S. Healthcare Systems
Benny UHORANISHEMA *
Carnegie Mellon University, Pittsburgh, Pennsylvania, USA.
Phebian Odufuwa
Department of Biological Sciences, Boise State University, USA.
Kelvin Ebo Rabbles
College of Professional Studies, Northeastern University, USA.
Fidele Nsanzumukunzi
Carnegie Mellon University, USA.
Kalu Tasie Rhoda
Department of Industrial Engineering, University of Arkansas, Fayetteville, Arkansas, USA.
Bukola Elizabeth Shasere
Central Clinical core/Speciality laboratory, Mayo Clinic, DLMP, Rochester, MN, USA.
Musa Shamak Ibrahim
Cell Kinetics Lab, Hematopathology, DLMP, Mayo Clinic, Rochester, MN, USA.
*Author to whom correspondence should be addressed.
Abstract
High-burden diseases in the United States are frequently diagnosed at later stages, resulting in poorer outcomes and increased costs. Clinically validated artificial intelligence (AI) can enhance early diagnosis by identifying subtle patterns in routine clinical data. This review seeks to map AI technologies that facilitate early detection of cancers, cardiovascular, metabolic, and infectious diseases within U.S. healthcare systems and to summarise their performance, impact on care pathways, and regulatory readiness. A PCC-guided scoping review was conducted following PRISMA-ScR guidelines, with searches in MEDLINE/PubMed and IEEE Xplore spanning the years 2015 to 2025. Included in the review were peer-reviewed clinical validation studies applicable to the U.S. care context. Data was organised according to condition, setting, model type and inputs, validation design, reference standards, diagnostic metrics, workflow impacts, and FDA status, which were then synthesised into three main themes. A total of nineteen studies met the inclusion criteria. AI applications in imaging improved early screening for cancer and metabolic conditions across mammography, chest radiography, pathology, retinal imaging, and liver ultrasound, effectively reducing false recalls and low-yield reads. Additionally, routine-signal, electronic health record (EHR), and wearable AI technologies supported either opportunistic or continuous screening for ventricular dysfunction, valvular disease, atrial fibrillation, sepsis, and C. difficile, providing a lead time of hours to days compared to standard scores. When properly validated, aligned with guidelines, and monitored in real-world settings, AI has the potential to enhance early-diagnosis pathways in the U.S. healthcare system.
Keywords: Artificial intelligence, early disease detection, clinical validation, U.S. healthcare systems, diagnostic accuracy