Multimodal Large Language Models for Pediatric Bone-Age Assessment: A Comparative Accuracy Analysis.
兒科骨齡評估之多模態大型語言模型:準確性比較分析
Acad Radiol 2025-08-13
這項研究比較了市面上熱門的多模態大型語言模型(像是 ChatGPT-4.5、o4-mini-high 和 Google Gemini 2.5 Pro)在判讀小孩手部X光、預測骨齡的準確度,並和放射科醫師做對照。雖然 Gemini 2.5 Pro 表現最好,但所有模型的誤差都很大(平均差2到3年),結果也不穩定,現階段還不適合臨床使用,未來還需要大幅改進。
相關文章PubMedDOI推理
Detecting Artificial Intelligence-Generated Text in Personal Statements of Adult Reconstruction Fellowship Applicants.
成人重建專科醫師申請者個人陳述中人工智慧生成文本的偵測
J Arthroplasty 2025-08-13
Can artificial intelligence accurately detect and summarize anatomy education literature? A comparative analysis of ChatGPT and ScholarGPT.
人工智慧能否準確偵測並摘要解剖學教育文獻?ChatGPT 與 ScholarGPT 的比較分析
Morphologie 2025-08-13
Assessing the Role of Large Language Models Between ChatGPT and DeepSeek in Asthma Education for Bilingual Individuals: Comparative Study.
在氣喘教育中評估大型語言模型(ChatGPT 與 DeepSeek)對雙語個體的角色:比較性研究
JMIR Med Inform 2025-08-13
這項研究比較 ChatGPT-4o 和 DeepSeek-v3 在中英文氣喘衛教的表現。兩者都能給出高品質、臨床相關的答案。DeepSeek-v3 在資訊完整性和更新度較好,特別適合需要詳細臨床資訊時使用;ChatGPT-4o 則說明更清楚、易懂,適合一般衛教。兩者都能協助病人自我管理,但不能取代醫師。未來 AI 工具應加強臨床推理和資訊更新,並需人工監督以確保安全。
相關文章PubMedDOI推理
Assessing ChatGPT's Educational Potential in Lung Cancer Radiotherapy From Clinician and Patient Perspectives: Content Quality and Readability Analysis.
從臨床醫師與病患觀點評估 ChatGPT 在肺癌放射治療教育潛力:內容品質與可讀性分析
JMIR Cancer 2025-08-13
Using Artificial Intelligence ChatGPT to Access Medical Information About Chemical Eye Injuries: Comparative Study.
使用人工智慧 ChatGPT 取得化學性眼損傷醫療資訊之比較研究
JMIR Form Res 2025-08-13
Evaluating large language models as clinical laboratory test recommenders in primary and emergency care: a crucial step in clinical decision making.
在初級與急診醫療中評估大型語言模型作為臨床實驗室檢驗建議工具:臨床決策中的關鍵一步
Clin Chem Lab Med 2025-08-13
Evaluating Hospital Course Summarization by an Electronic Health Record-Based Large Language Model.
以電子健康紀錄為基礎的大型語言模型對住院過程摘要的評估
JAMA Netw Open 2025-08-13