Grade 2 subclinical hypothyroidism (G2SCH), defined as a serum thyroid-stimulating hormone (TSH) level exceeding 10 mU/L, often requires treatment. A team from the Second Affiliated Hospital of Zhejiang University School of Medicine used machine learning algorithms on routine physical examination indicators and identified four factors most strongly associated with G2SCH: age, hemoglobin, red blood cell count, and systolic blood pressure. They further developed a prediction model for G2SCH with an area under the curve (AUC) of 0.870, a sensitivity of 86.8%, and a specificity of 70.8%. This model is expected to enable early detection of G2SCH using low-cost routine tests.
Inappropriate ultrasound screening is a major cause of overdiagnosis of thyroid diseases, especially thyroid nodules and cancer, and also leads to the waste of medical resources. Larios et al. reviewed data from the Mayo Clinic and found that approximately 1 in 13 thyroid ultrasound orders at this top medical institution were inappropriate. Patients receiving such inappropriate orders were more likely to be young individuals, patients with hyperthyroidism, and those without an in-person consultation, and the orders were more frequently issued by non-endocrinologists. Compared with appropriately indicated ultrasounds, inappropriate ultrasound orders had significantly lower rates of thyroid nodule detection (23% vs. 65%) and confirmed thyroid cancer (1.8% vs. 4.0%), indicating that they resulted in ineffective use of medical resources.