AI and Better Assays Are Closing Acromegaly's 10-Year Diagnostic Gap

Acromegaly is a rare but serious endocrine disorder caused by sustained overproduction of growth hormone, typically from a pituitary adenoma, leading to chronically elevated IGF-1. Despite its distinct late-stage features, the average time from symptom onset to diagnosis stretches 6 to 10 years, during which patients accumulate significant cardiovascular, metabolic, and musculoskeletal damage.

This review, authored by endocrinologists from Cleveland Clinic and the Université Paris-Saclay, synthesizes current best practices and emerging innovations in acromegaly diagnosis. The authors emphasize that early symptoms — sleep apnea, carpal tunnel syndrome, arthralgia, and metabolic disturbances — are common in the general population and rarely trigger suspicion for acromegaly, making systematic screening strategies essential.

On the biochemical side, serum IGF-1 remains the preferred first-line screening test due to its stability and its reflection of integrated GH secretion over time. However, interpretation requires careful attention to age, sex, assay variability, and confounders including diabetes, obesity, liver or renal disease, pregnancy, and estrogen use. When clinical and biochemical findings are discordant, repeating IGF-1 and performing a GH suppression test during an oral glucose tolerance test (OGTT) is recommended. Advances in ultrasensitive GH assays have now lowered the nadir GH cutoff to approximately 0.4 µg/L.

Perhaps most forward-looking is the discussion of AI-assisted diagnosis. Comorbidity cluster analysis, AI-driven facial recognition software, electronic medical record mining, and radiomics applied to high-resolution MRI and PET/MRI are all identified as promising tools to flag at-risk patients earlier and improve diagnostic accuracy.

For clinicians, the practical takeaway is clear: acromegaly should be on the differential for patients presenting with clusters of seemingly unrelated conditions, and newer diagnostic tools are making earlier detection increasingly feasible.