Medical reports often contain complex terminology that can confuse patients, hindering their understanding of critical health information. A recent study by a team from the Technical University of Munich (TUM) explored how artificial intelligence (AI) can simplify **CT** reports, making them more accessible to patients. The findings show a significant decrease in reading time and an increase in patient satisfaction with AI-generated simplified reports.
To achieve this, the researchers utilized an open-source large language model that operates under strict data protection regulations at the TUM University Hospital. The AI transformed intricate medical jargon into clearer language. For instance, a technical statement like, “The cardiomediastinal silhouette is midline. The cardiac chambers are normally opacified. […] A small pericardial effusion is noted,” was reframed as: “Heart: The report notes a small amount of fluid around your heart. This is a common finding, and your doctor will determine if it needs any attention.” This translation underscores the potential of AI to bridge the gap between medical professionals and patients.
Importance of Understandable Medical Language
The significance of making medical terminology accessible extends beyond mere convenience. As Felix Busch, assistant physician at the Institute for Diagnostic and Interventional Radiology and co-author of the study, emphasized: “Ensuring that patients understand their reports, examinations, and treatments is a central pillar of modern medicine. This is the only way to guarantee informed consent and strengthen health literacy.” This perspective highlights that patient comprehension is foundational to effective healthcare.
Previous research had indicated that AI could enhance the readability of medical texts, but this study specifically assessed its impact on actual patient experiences. The team involved 200 patients who underwent CT imaging due to a cancer diagnosis. They divided the participants into two groups: one receiving the original report and the other the AI-simplified version.
Results: Enhanced Readability and Patient Satisfaction
The outcomes were compelling. The average reading time for the original reports was seven minutes, dropping to just two minutes for the simplified versions. Notably, patients who read the AI-generated summaries reported a significantly easier reading experience: 81% found them easier to read compared to just 17% for the originals. Additionally, 80% indicated the simplified reports were easier to understand, contrasting sharply with only 9% for the original texts. Patients also rated the simplified versions as more helpful (82% vs. 29%) and informative (82% vs. 27%).
Various objective measurements also confirmed the improved readability of the simplified reports.
Felix Busch, assistant physician, Institute for Diagnostic and Interventional Radiology
While these findings are promising, the research team acknowledges the need for further studies to establish whether improved understanding correlates with tangible improvements in patient health outcomes. Nonetheless, they propose that providing automatically simplified reports alongside specialist reports could be a feasible service in clinical settings, provided optimized and secure AI systems are available.
Importance of Professional Oversight
Despite the benefits of AI in simplifying medical language, the study’s authors caution against relying solely on AI tools like chatbots for medical advice. Dr. Philipp Prucker, the first author of the study, warned of potential inaccuracies, stating that 6% of AI-generated reports contained factual errors, 7% omitted critical information, and 3% included misleading details. Before patients received the simplified reports, they underwent verification to correct any inaccuracies. “Language models are useful tools, but they are no substitute for medical staff. Without trained specialists verifying the findings, patients may, in the worst case, receive incorrect information about their illness,” concluded Prucker.
In summary, the TUM study reveals a promising application of AI in healthcare, demonstrating its potential to enhance patient understanding of complex medical reports. As AI technology continues to evolve, it may play an integral role in improving patient communications, provided it is used responsibly and supplemented by professional oversight.
Source:
Technical University of Munich (TUM)
Journal reference:
Prucker et al. “A Prospective Controlled Trial of Large Language Model–based Simplification of Oncologic CT Reports for Patients with Cancer”. Radiology (2025). DOI: 10.1148/radiol.251844.
Google DeepMind Hires Boston Dynamics CTO Aaron Saunders to Advance Robotics Efforts
Anthropic Reveals AI Misalignment Risks Linked to Reward Hacking in New Study
ServiceNow Partners with Microsoft to Elevate AI Governance, Targeting $20.3B Revenue by 2028
Nvidia’s Jensen Huang Donates $5M to San Francisco Opera, Boosts Arts with ‘The Monkey King’ Premiere
Google Extends Free Year of AI Pro for US College Students Until 2027 with Gemini 3 Pro
