Department of Ultrasound, The Second Affiliated Hospital of Zhejiang Chinese Medical University, Hangzhou 310000, Zhejiang Province, China.
Department of Ultrasound, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai 200233, China.
Int J Med Sci. 2024 Jun 24;21(9):1701-1709. doi: 10.7150/ijms.95022. eCollection 2024.
We aimed to explore the predictive value of an ultrasound-based radiomics model for the central lymph node metastasis of papillary thyroid carcinoma. A total of 126 patients with papillary thyroid carcinoma treated between February 2021 and February 2023 were retrospectively enrolled and assigned into metastasis group (n=59, with cervical central lymph node metastasis) or non-metastasis group (n=67, without metastasis) based on surgical and pathological findings. Intergroup comparisons were conducted on the results of contrast-enhanced ultrasonography, preoperative conventional ultrasonography, as well as real-time shear wave elastography. The maximum lesion diameter, echo, margin, capsule invasion, calcification, average elasticity modulus (Eavg), rising time (RT), and peak intensity (PI) had diagnostic value for papillary thyroid carcinoma, and their combination exhibited higher diagnostic value (area under the curve: 0.817). The logistic regression model was built, and the maximum lesion diameter, hypoechoic/extremely hypoechoic, lobulated or irregular margin (95% confidence interval: 1.451-6.755), capsule invasion, microcalcification/macrocalcification or peripheral calcification, high-level Eavg, low-level RT and high-level PI served as risk elements affecting papillary thyroid carcinoma from the aspect of central lymph node metastasis (odds ratio>1, P<0.05). According to the logistic regression model, the model was reliable and stable (area under the curve: 0.889, P<0.05). The established ultrasound-based radiomics model can be utilized for early identifying the central lymph node metastasis of papillary thyroid carcinoma.
我们旨在探讨基于超声的放射组学模型对甲状腺乳头状癌中央淋巴结转移的预测价值。回顾性纳入 2021 年 2 月至 2023 年 2 月期间治疗的 126 例甲状腺乳头状癌患者，根据手术和病理结果将其分为转移组（n=59，有颈中央淋巴结转移）和非转移组（n=67，无转移）。对增强超声、术前常规超声和实时剪切波弹性成像的结果进行组间比较。最大病变直径、回声、边界、包膜侵犯、钙化、平均弹性模量（Eavg）、上升时间（RT）和峰值强度（PI）对甲状腺乳头状癌具有诊断价值，它们的组合具有更高的诊断价值（曲线下面积：0.817）。建立逻辑回归模型，最大病变直径、低回声/极低回声、分叶状或不规则边界（95%置信区间：1.451-6.755）、包膜侵犯、微钙化/大钙化或周围钙化、高水平 Eavg、低水平 RT 和高水平 PI 作为影响甲状腺乳头状癌中央淋巴结转移的风险因素（优势比>1，P<0.05）。根据逻辑回归模型，该模型具有可靠性和稳定性（曲线下面积：0.889，P<0.05）。基于超声的放射组学模型可用于早期识别甲状腺乳头状癌的中央淋巴结转移。
