文献速递垂体与肾上腺疾病

基于手部图像的深度学习自动识别肢端肥大症:多中心观察研究

Automatic acromegaly detection using deep learning on hand images: a multicenter observational study.
PubMed 文献 · J Clin Endocrinol Metab · 2026-01-01

模型在独立测试中心达到灵敏度0.89、特异度0.91、AUC 0.96,表现优于内分泌专科医生(F1-score范围0.43-0.63)。

要点

  • 背侧手与握拳征是肢端肥大症的有效诊断线索,可避免采集掌纹等隐私区域。
  • 模型灵敏度0.89、特异度0.91、AUC 0.96,优于内分泌专科医生。
  • 该隐私友好模型有望用于健康体检等公共场所的初筛。

核心结论

  • 基于手背与握拳征图像的深度学习模型能较准确识别肢端肥大症,性能优于内分泌专科医生。

研究设计与样本

  • 纳入标准为年龄≥18岁且在参与机构接受诊疗;12家中心数据用于训练/验证,3家中心用于测试。
  • 采用基于ResNet-50的PyTorch模型,配合数据增强与5折交叉验证,每位患者取4张图像预测均值。

主要发现

  • 模型灵敏度0.89、特异度0.91、阳性预测值0.88、阴性预测值0.93、F1-score 0.89。
  • 受试者工作特征曲线下面积(AUC)为0.96,优于专科医生(F1-score范围0.43-0.63)。

关键数值与推荐

  • 模型性能指标:灵敏度0.89、特异度0.91、阳性预测值0.88、阴性预测值0.93、F1-score 0.89、AUC 0.96。
  • 对照的专科医生F1-score范围为0.43-0.63。

临床怎么用

  • 可将手背与握拳征图像作为肢端肥大症筛查线索,模型隐私友好,适合在健康体检等公共场景部署初筛。

局限性

  • 需在包含健康人群及多种疾病的更大数据集上进一步验证。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

CONTEXT: Acromegaly poses clinical challenges in terms of early diagnosis and intervention. Therefore, the development of novel diagnostic tools is essential. Although artificial intelligence (AI) models based on external appearance have been proposed, privacy concerns have limited their use. OBJECTIVE: To develop a privacy-conscious deep learning model for detecting acromegaly using hand images. METHODS: This nationwide multicenter study enrolled 716 patients (317 with acromegaly and 399 controls) and 11 480 images from 15 Japanese pituitary centers. The inclusion criteria were age ≥18 years and care received at the participating facilities. Hand images focusing on the dorsal and fist sign, excluding the palm/fingerprint regions, were used to develop the model. The data were split into training/validation (12 centers) and test (3 centers) datasets. A ResNet-50-based model was trained using PyTorch with data augmentation and 5-fold cross-validation. For each patient, the predictions were averaged over 4 images. The performance of the model was compared with that of endocrinologists. RESULTS: The model achieved a sensitivity of 0.89, specificity of 0.91, positive predictive value of 0.88, negative predictive value of 0.93, F1-score of 0.89, and an area under the receiver operating characteristic curve of 0.96, outperforming specialists (F1-score range: 0.43-0.63). CONCLUSION: This study highlights the utility of dorsal hand and fist sign as diagnostic clues for acromegaly, which the AI model captured more accurately than endocrinologists. Using this privacy-conscious feature, this model can be deployed in public settings like health checkups. Further validation using larger datasets, including healthy individuals and diverse diseases, is necessary.

来源:J Clin Endocrinol Metab,查看原文。版权归原作者所有。

引用

  1. Ohmachi Y, Nishio M, Abe I, et al. Automatic acromegaly detection using deep learning on hand images: a multicenter observational study. J Clin Endocrinol Metab. 2026. PMID: 41757900. doi:10.1210/clinem/dgag027

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。