ARTIFICIAL INTELLIGENCE-POWERED MEDICAL IMAGING FOR EARLY CANCER DIAGNOSIS AND EPIDEMIOLOGICAL ANALYSIS: ADVANCING INNOVATION IN U.S. NATIONAL HEALTHCARE INFRASTRUCTURE

Abstract

This study addresses the persistent gap between the technical promise of artificial intelligence (AI) in cancer imaging and its uneven real-world impact on early diagnosis and epidemiological analysis within U.S. national healthcare infrastructure. The purpose is to quantify how technological readiness, organizational support, and data governance maturity relate to AI-powered medical imaging adoption and to perceived improvements in early cancer detection and population-level analytics. Using a quantitative cross-sectional, case-based survey design, data were collected from 320 professionals across four U.S. enterprise healthcare institutions that serve as AI-enabled cancer imaging and data analytics cases (response rate 64.0 percent). Key variables included technological readiness, organizational support, data governance, AI imaging adoption, big data analytics capability, early diagnosis effectiveness, and epidemiological analysis capability, all measured with 5-point Likert scales that showed strong reliability (Cronbach’s alpha 0.81–0.91). The analysis plan combined descriptive statistics, correlation analysis, and multiple regression, including an interaction term for governance. Mean scores indicated moderately strong readiness (3.62) and governance (3.89) but only intermediate AI imaging adoption (3.28). Regression models showed that readiness and support together explained 53.2 percent of variance in AI adoption, while adoption and analytics capability jointly explained 52.6 percent of variance in perceived early diagnosis effectiveness and 58.4 percent in epidemiological analysis capability; governance significantly strengthened the adoption–epidemiology link. These findings imply that AI-powered medical imaging delivers the greatest early diagnostic and epidemiological value when embedded in well-governed, analytics-capable enterprise environments rather than deployed as isolated algorithms.

Ähnliche Arbeiten