Papillary thyroid carcinoma in graves' disease: prevalence, clinicopathological features, and preoperative predictors

The coexistence of Graves' disease (GD) and papillary thyroid carcinoma (PTC) remains a subject of clinical debate. While PTC is frequently detected incidentally after thyroidectomy for GD, its prevalence, clinicopathological behavior, and preoperative predictors remain insufficiently defined.

This single-center retrospective cohort study included 602 patients who underwent thyroidectomy between 2020 and 2025. Patients were categorized as: GD + PTC (n = 51); GD-only (n = 109); and PTC-only (n = 442). Demographic, biochemical, radiological, surgical, and pathological data were analyzed. Univariate and multivariate logistic regression models were used to identify factors associated with PTC development in GD.

The prevalence of PTC among GD patients was 31.9%. Compared with sporadic PTC, GD-associated tumors were smaller (median 5 mm vs. 12 mm, p < 0.001) and demonstrated fewer aggressive features including lymphatic invasion, capsular invasion, multifocality, bilaterality, and nodal metastasis (all p < 0.01). GD + PTC patients were younger and showed a lower female predominance than those with sporadic PTC. When compared with GD-only patients, the GD + PTC group had significantly lower thyroid-stimulating Immunoglobulin (TSI) titers (median 3.8 vs. 7.65 IU/L, p = 0.007) and a higher prevalence of ultrasound-detected thyroid nodules (64.7% vs. 27.5%, p < 0.001). In multivariate analysis, only US-detected nodules (OR 3.56, p = 0.003) and lower TSI levels (OR 0.95, p = 0.03) independently predicted PTC in GD.

PTC is relatively common among surgically treated GD patients, yet presents predominantly as microcarcinoma with less aggressive histopathological features. The presence of ultrasound-detected thyroid nodules was the strongest preoperative predictor of malignancy. These findings support careful and systematic ultrasonographic assessment in patients with GD, with FNAB guided by established ultrasound risk patterns, nodule size thresholds, and high-risk clinical features, rather than indiscriminate lowering of biopsy thresholds.