AI Model Can Diagnose Acromegaly Using Dorsal Hand, Fist Images

THURSDAY, March 5, 2026 (HealthDay News) — An artificial intelligence model using the dorsal hand and fist sign can diagnose acromegaly, according to a study published online Feb. 27 in the Journal of Clinical Endocrinology & Metabolism.

Yuka Ohmachi, from the Kobe University Graduate School of Medicine in Japan, and colleagues developed a privacy-conscious deep learning model for detecting acromegaly using hand images in a nationwide study involving 716 patients and 11,480 images from 15 Japanese pituitary centers. The model was developed using hand images focusing on the dorsal and fist sign, excluding the palm/fingerprint regions. Data were split into training/validation and test datasets (from 12 and three centers, respectively). The predictions for each patient were averaged over four images. Model performance was compared to that of endocrinologists.

The researchers found that the model achieved a sensitivity and specificity of 0.89 and 0.91, respectively, and a positive and negative predictive value of 0.88 and 0.93, respectively. The F1-score was 0.89 and area under the receiver operating characteristic curve was 0.96, outperforming specialists (F1 score range, 0.43 to 0.63).

“I was surprised that the diagnostic accuracy reached such a high level using only photographs of the back of the hand and the clenched fist,” Ohmachi said in a statement. “What struck me as particularly significant was achieving this level of performance without facial features, which makes this approach a great deal more practical for disease screening.”

Abstract/Full Text