Gülsüm Karaahmetli, Şefika Burçak Polat, Sevgül Fakı, Leyla Akdoğan, Feride Pınar Altay, Birgül Genç, Ayşegül Aksoy Altınboğa, Oya Topaloğlu, Reyhan Ersoy, B N Cakir
Objective: The 2022 WHO classification introduced differentiated high-grade thyroid carcinoma (DHGTC) as a distinct category characterized by high-grade features despite maintained differentiation. Therefore, this study aims to evaluate its clinical characteristics, treatment responses, and the prognostic impact of the Ki-67 proliferation index in this patient population. Methods: We retrospectively reviewed 3100 patients with differentiated thyroid carcinoma (DTC) between 2017 and 2024. From this baseline pool, a total of 56 patients (1.8%) were identified and re-classified as DHGTC based on mitotic count (≥5/2 mm2) and/or tumor necrosis. Additionally, 69 DTC patients (2.2%) with an elevated Ki-67 index (>5%) identified from the same baseline pool—representing an overlapping group with the DHGTC cohort—were analyzed to evaluate its clinical significance. Results: In the DHGTC group, tumor necrosis was present in 87.5% and high mitotic activity in 19.6% of cases. While all DHGTC patients were classified as high-risk under 2025 American Thyroid Association (ATA) guidelines, 42.8% showed biochemical or structural incomplete response at the last follow-up and 28.6% required additional salvage interventions. In the broader DTC cohort, Ki-67 ≥ 15% was significantly associated with older age, larger tumor size, extensive invasion, and poorer treatment response (p < 0.05). However, within the DHGTC subset, Ki-67 ≥ 15% was only significantly associated with increased lymphovascular invasion and more extensive surgery. Conclusions: DHGTC carries a significant burden of aggressive histopathological features and a high risk of structural disease persistence or recurrence. While an elevated Ki-67 index (≥15%) serves as an adverse marker in general DTC, its additional prognostic value within the high-grade DHGTC cohort remains inconclusive, potentially obscured by limited statistical power due to small subgroup sizes.