Novel prognostic nomogram for predicting recurrence‐free survival in medullary thyroid carcinoma
Yagiz A Aksoy,Bin Xu,Kartik Viswanathan,Mahsa S Ahadi,Abir Al Ghuzlan,Bayan Alzumaili,Mohamed‐Amine Bani,Justine A Barletta,Nicole Chau,Angela Chou,Adele Clarkson,Roderick J Clifton‐Bligh,Antonio De Leo,Snjezana Dogan,Ian Ganly,Ronald Ghossein,Matti L Gild,Anthony R Glover,Julien Hadoux,Livia Lamartina,Daniel J Lubin,Kelly Magliocca,Fedaa Najdawi,Arad Nigam,Alex Papachristos,Andrea Repaci,Bruce G Robinson,Amy Sheen,Qiuying Shi,Stan B Sidhu,Loretta Sioson,Erica Solaroli,Mark S Sywak,Giovanni Tallini,Venessa Tsang,John Turchini,Brian R Untch,Anthony J Gill,Talia L Fuchs
DOI: https://doi.org/10.1111/his.15141
2024-01-24
Histopathology
Abstract:Building upon the International Medullary Thyroid Carcinoma (MTC) Grading System (IMTCGS) developed in 2022, this study presents a prognostic nomogram to predict recurrence‐free survival in MTC. The model was created using data from 300 MTC patients and was externally validated in a separate cohort. The nomogram is available at https://nomograms.shinyapps.io/MTC_ML_DFS/. Aims Recently, there have been attempts to improve prognostication and therefore better guide treatment for patients with medullary thyroid carcinoma (MTC). In 2022, the International MTC Grading System (IMTCGS) was developed and validated using a multi‐institutional cohort of 327 patients. The aim of the current study was to build upon the findings of the IMTCGS to develop and validate a prognostic nomogram to predict recurrence‐free survival (RFS) in MTC. Methods and Results Data from 300 patients with MTC from five centres across the USA, Europe, and Australia were used to develop a prognostic nomogram that included the following variables: age, sex, AJCC stage, tumour size, mitotic count, necrosis, Ki67 index, lymphovascular invasion, microscopic extrathyroidal extension, and margin status. A process of 10‐fold cross‐validation was used to optimize the model's performance. To assess discrimination and calibration, the area‐under‐the‐curve (AUC) of a receiver operating characteristic (ROC) curve, concordance‐index (C‐index), and dissimilarity index (D‐index) were calculated. Finally, the model was externally validated using a separate cohort of 87 MTC patients. The model demonstrated very strong performance, with an AUC of 0.94, a C‐index of 0.876, and a D‐index of 19.06. When applied to the external validation cohort, the model had an AUC of 0.9. Conclusions Using well‐established clinicopathological prognostic variables, we developed and externally validated a robust multivariate prediction model for RFS in patients with resected MTC. The model demonstrates excellent predictive capability and may help guide decisions on patient management. The nomogram is freely available online at https://nomograms.shinyapps.io/MTC_ML_DFS/.
pathology,cell biology