A study by Inventorus, covering data up to July 31, 2026, reveals that AI is utilized across nearly every stage of medical research, from hypothesis generation to clinical trial recruitment. Despite 4,670 primary studies published, only 2.4% have undergone clinical validation. The report highlights that 88.2% of these studies remain in the preclinical stage, with only 113 reaching randomized controlled trials.
Oncology leads the field, accounting for 32.5% of all primary research, followed by musculoskeletal system diseases at 14.1%. Radiology and pathology are identified as the most mature areas for AI integration due to the ease of standardizing images. Meanwhile, generative models saw 4,610 publications between January 2022 and September 2025, yet only 19 of these works involved full randomized trials.
Inventorus CEO Evgeny Elfimov emphasizes that the industry currently lacks external validation and prospective studies. The report predicts a shift from single-task models to multimodal systems and digital twins. Future industry maturity will be measured by the speed at which the gap between the volume of academic publications and proven clinical evidence is reduced, rather than by model size alone.
