Ruifeng Liu1, Yu Xia1, Cheng Chen1, Tiantian Ye1, Xuepei Huang1, Li Ma1, Ya Hu2, Yuxin Jiang3. 1. Department of Ultrasound, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shuaifuyuan 1#, Dongcheng District, 100730, Beijing, China. 2. Department of General Surgery, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shuaifuyuan 1#, Dongcheng District, 100730, Beijing, China. huyapumch@126.com. 3. Department of Ultrasound, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shuaifuyuan 1#, Dongcheng District, 100730, Beijing, China. jiangyuxinxh@163.com.
Abstract
PURPOSES: Parathyroid cancer (PC) is rare, but fatal condition. Preoperative prediction of PC remains challenging but meaningful. The aim of this study was to determine an effective model to predict PC in patients with parathyroid lesions >1.5 cm. METHODS: In this retrospective case-control study, we enrolled 30 patients with PC matched to 60 patients with parathyroid adenoma or hyperplasia by admission year. All patients were diagnosed with primary hyperparathyroidism (pHPT) and had parathyroid lesions >1.5 cm. Ultrasonic features of the two patient groups, as well as demographic, clinical, and biochemical characteristics were retrospectively compared. Best subset selection and multivariate logistic regression analysis were conducted to identify the independent risk factors of PC. ROC curve and decision curve analysis were developed to evaluate the applicability of the new model. RESULTS: The best subset selection method and multiple logistic regression analysis showed that ultrasonic features of DR (two diameters' ratio of the lesion) and tumor infiltration in conjunction with intact parathyroid hormone (iPTH) level (collective model) were independent predictors of malignancy. Meanwhile, DR, shape, and tumor infiltration (ultrasound model) were found to be risk factors when only ultrasonic features were included in the multivariate analysis. The decision curve analysis showed that collective model outperforms ultrasound model with a better net benefit and a wider range of threshold probabilities. CONCLUSIONS: Ultrasonic features in combination with iPTH level may be an applicable model for predicting PC and has a better potential to facilitate decision-making preoperatively.
PURPOSES: Parathyroid cancer (PC) is rare, but fatal condition. Preoperative prediction of PC remains challenging but meaningful. The aim of this study was to determine an effective model to predict PC in patients with parathyroid lesions >1.5 cm. METHODS: In this retrospective case-control study, we enrolled 30 patients with PC matched to 60 patients with parathyroid adenoma or hyperplasia by admission year. All patients were diagnosed with primary hyperparathyroidism (pHPT) and had parathyroid lesions >1.5 cm. Ultrasonic features of the two patient groups, as well as demographic, clinical, and biochemical characteristics were retrospectively compared. Best subset selection and multivariate logistic regression analysis were conducted to identify the independent risk factors of PC. ROC curve and decision curve analysis were developed to evaluate the applicability of the new model. RESULTS: The best subset selection method and multiple logistic regression analysis showed that ultrasonic features of DR (two diameters' ratio of the lesion) and tumor infiltration in conjunction with intact parathyroid hormone (iPTH) level (collective model) were independent predictors of malignancy. Meanwhile, DR, shape, and tumor infiltration (ultrasound model) were found to be risk factors when only ultrasonic features were included in the multivariate analysis. The decision curve analysis showed that collective model outperforms ultrasound model with a better net benefit and a wider range of threshold probabilities. CONCLUSIONS: Ultrasonic features in combination with iPTH level may be an applicable model for predicting PC and has a better potential to facilitate decision-making preoperatively.
Authors: Julia A Krupinova; Alina R Elfimova; Olga Yu Rebrova; Iya A Voronkova; Anna K Eremkina; Elena V Kovaleva; Irina S Maganeva; Anna M Gorbacheva; Ekaterina E Bibik; Andrey A Deviatkin; Galina A Melnichenko; Natalia G Mokrysheva Journal: J Pathol Inform Date: 2022-08-27