小児腫瘍病理および教育におけるChatGPT-4からChatGPT-5への進化する応用
Mai He1, Jinyi Weng1
1Department of Pathology & Immunology, Washington University in St. Louis School of Medicine, St. Louis, MO, USA.
Introduction:
Large language models (LLMs) such as ChatGPT are emerging tools in pathology. This study evaluated their diagnostic utility in pediatric neoplastic pathology.
Materials And Methods:
Thirty-three pediatric tumor cases were retrospectively analyzed. Clinical data and histology images were input to ChatGPT-4o (03-06/2025) and GPT-5 (10/2025) for diagnostic suggestions, immunohistochemical (IHC) panels, and report generation. GPT outputs were graded for diagnostic concordance (2 = concordant, 1 = partial, 0 = discordant).
Results:
ChatGPT-4o achieved at least partial concordance in 27/35 (77.1%) cases (toal and mean score 42, 1.2), with highest accuracy in small round blue cell tumors (SRBCT, 100%) and IHC interpretation accuracy 74.4%. GPT-5 showed 19/33 (57.6%) concordance (total and mean 25, 0.76), highest in SRBCT (87.5%) and IHC accuracy 60.7%. (p < 0.05 between 4o and 5, Mann-Whitney U).
Conclusions:
ChatGPT demonstrates promise as a diagnostic and educational adjunct in pediatric pathology, though expert oversight and further validation remain essential.
関連する概念動画
02:07Tumor Progression
Colon cancer is one of the best-documented examples of tumor progression. Early mutation in the APC gene in colon cells causes a small growth on the colon wall called a polyp. With time, this polyp grows into a benign, pre-cancerous tumor. Further...
01:24Pharmacokinetics in Pediatric Patients: Drug Metabolism