Clinical Scorecard: Hand Photos May Flag Rare Hormone Disorder

At a Glance

Key Highlights

• Deep learning model analyzes hand images for acromegaly detection.
• Sensitivity of 0.89 and specificity of 0.91 achieved in independent test dataset.
• Diagnostic delays exceeding 10 years are common in acromegaly patients.
• Model outperformed board-certified endocrinologists in image evaluation.
• Tool intended to support clinical evaluation, not replace it.

Guideline-Based Recommendations

Diagnosis

• Utilize hand image analysis as a potential screening tool for acromegaly.

Management

• Early referral for endocrine evaluation in patients with subtle clinical features.

Monitoring & Follow-up

• Regular assessment of growth hormone levels and physical changes.

Risks

• Potential for diagnostic delays and misidentification in diverse populations.

Patient & Prescribing Data

716 adults, including 317 with acromegaly.

AI model provides a non-invasive method for early detection.

Clinical Best Practices

• Incorporate AI-assisted tools in routine evaluations for acromegaly.
• Ensure diverse population representation in diagnostic studies.

Related Resources & Content

• The Journal of Clinical Endocrinology & Metabolism