Abstract
Graves' disease (GD) is an autoimmune disorder that can be difficult to distinguish from other diseases due to symptom similarity. The exacerbation of GD owing to delayed diagnosis is a serious issue, and a novel accessible health screening system is needed. Therefore, this study investigated the association between GD and thyroid hormone levels in women's hair and evaluated the prediction accuracy of this non-invasive type of sample. By optimizing pretreatment and analysis techniques using liquid chromatography-mass spectrometry (LC-MS), free triiodothyronine (FT3) and thyroxine (FT4) could be detected in only 2 mg of hair with high sensitivity. Compared with healthy controls, the thyroid hormone levels in the hair of GD patients were significantly higher in correlation with blood levels. The predictive ability of hair thyroid hormones was analyzed using a receiver operating characteristic (ROC) curve, and the optimal cut-off value was determined via the Youden index. As a result, the area under the curve (AUC) was 0.974 (95% confidence interval (CI): 0.935-1.000) for FT3 and 0.900 (95% CI: 0.807-0.993) for FT4. The cut-off value was 0.133 pg/mg (sensitivity: 91.2%; specificity: 100%; positive predictive value (PPV): 100%; negative predictive value (NPV): 76.9%) for FT3 and 0.067 pg/mg (sensitivity: 70.6%; specificity: 100%; PPV: 100%; NPV: 50.0%) for FT4. Collectively, our new approach offers the possibility of accurately and non-invasively detecting GD using hair samples. Since hair can be stored and transported at room temperature, this system facilitates large-scale screening at locations including hair salons and homes, potentially enabling the early determination of GD outside of medical facilities.