ASOURCE Navi

The evolution of AI (artificial intelligence) in the medical field has attracted particular attention in the field of diagnostic support systems. AI can instantly process massive amounts of data and support diagnoses, making a significant contribution to the early detection of disease and improving diagnostic accuracy. In this fourth installment, we will summarize the current state of diagnostic support using AI, specific examples in the medical field, the importance of collaboration with doctors, and future challenges.

The most prominent area in which AI is being utilized in the medical field is radiological imaging diagnosis. Image diagnostic AI can accurately identify abnormal findings in images such as CT and MRI, and quickly indicate the presence or possibility of disease. This allows doctors to devote more time to diagnosing each patient and reduces the risk of overlooking a diagnosis. For example, there have been reported cases in which AI has detected minute abnormalities in the early detection of lung cancer and brain tumors and alerted doctors, raising expectations for improved diagnostic accuracy.
AI is also playing a revolutionary role in the field of pathological diagnosis. By combining the digitization of pathological specimens with machine learning technology, AI can accurately detect abnormalities in cells and tissues from microscopic images and support diagnosis. In particular, in cancer pathological diagnosis, AI can quickly analyze large volumes of pathological slides and identify subtle cellular changes and tumor characteristics. By adding objective analysis to a diagnostic process that previously relied on the eyes of experienced pathologists, AI is improving the accuracy and consistency of diagnoses. Furthermore, AI can also contribute to the detection of rare pathological findings and minute lesions, expanding the possibility of early diagnosis. It also helps to reduce the burden on pathologists, who are in short supply.

The introduction of AI diagnosis has also produced results in the field of ophthalmology. For example, in diagnosing diabetic retinopathy, AI analyzes retinal images and achieves accuracy on par with traditional human diagnoses. The use of AI is also rapidly advancing in endoscopic examinations. In gastrointestinal endoscopic examinations, AI is now able to detect minute lesions and early-stage cancers with high accuracy from mucosal images, making a significant contribution to the early detection of colon and stomach cancer in particular. AI is expected to improve patient prognosis because it can accurately identify small lesions less than 10 mm in size that were previously easily overlooked. This has enabled early detection and improved patient outcomes in an increasing number of cases. Furthermore, during the COVID-19 pandemic, AI was used to detect signs of infection from chest images, improving diagnostic speed and reducing the burden on medical facilities.

AI is not only making diagnoses more efficient, but also contributing to improved accuracy. The parts that previously relied on doctors' experience and knowledge are now complemented by AI's data analysis capabilities, reducing diagnostic variability. Furthermore, because AI can detect new patterns by comparing data with past diagnostic data, it is also useful for discovering rare diseases and pointing out lesions that are often overlooked. This improves the quality of medical care and reduces the burden on patients.
With the revision of medical fees for fiscal year 2022, an additional charge for diagnostic imaging assistance using artificial intelligence (AI) technology (plain and computed tomography) has been made applicable to insurance.

AI does not replace doctors, but rather acts as a support. The process in which doctors review the diagnostic results presented by AI and make the final decision is crucial. This collaboration enables more accurate and faster diagnoses, allowing doctors to devote more time to communicating with patients and explaining treatment plans. AI-based diagnosis is also useful for doctor education, and is expected to serve as a tool to support the development of skills among junior doctors.

However, challenges remain with AI diagnosis. Discussion is ongoing regarding the risk of misdiagnosis due to data bias, the skill required for doctors to properly interpret AI analysis results, and legal responsibility. Furthermore, because the implementation status of AI systems varies from medical institution to medical institution, standardization and uniform quality control are required. However, overcoming these challenges is expected to lead to further development of AI in the medical field.
Diagnostic support AI has great potential to support doctors' decision-making and improve the quality of medical care. Collaboration between doctors and AI will enable earlier and more accurate diagnoses, contributing to maintaining patients' health. AI will likely become an indispensable partner in the future of medicine.