Objective:

To evaluate the effectiveness of a deep learning model in detecting acromegaly through analysis of hand images.

Approach:

Key Findings:

• The model achieved a sensitivity of 0.89 and specificity of 0.91.
• Positive predictive value was 0.88 and negative predictive value was 0.93.
• The F1-score of the model was 0.89, with an area under the ROC curve of 0.96.
• Endocrinologists' F1-scores ranged from 0.43 to 0.63 when evaluating the same images.

Interpretation:

The findings indicate that AI can effectively identify physical signs of acromegaly, potentially facilitating earlier diagnosis and referral for endocrine evaluation.

Limitations:

• Higher proportion of acromegaly cases in the dataset compared to typical clinical populations.
• Study included only Japanese participants, affecting generalizability.
• The tool is intended to support clinical evaluation, not replace it.

Conclusion:

Image-based AI may assist in the early detection of acromegaly, helping to address diagnostic delays.

Sources:

• The Journal of Clinical Endocrinology & Metabolism
Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.