Creating Prognostic Systems for Well-Differentiated Thyroid Cancer Using Machine Learning

Abstract

Updates to staging models are needed to reflect a greater understanding of tumor behavior and clinical outcomes for well-differentiated thyroid carcinomas. We used a machine learning algorithm and disease-specific survival data of differentiated thyroid carcinoma from the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute to integrate clinical factors to improve prognostic accuracy. The concordance statistic (C-index) was used to cut dendrograms resulting from the learning process to generate prognostic groups. We created one computational prognostic model (7 prognostic groups with C-index = 0.8583) based on tumor size (T), regional lymph nodes (N), status of distant metastasis (M), and age to mirror the contemporary American Joint Committee on Cancer (AJCC) staging system (C-index = 0.8387). We showed that adding histologic type (papillary and follicular) improved the survival prediction of the model. We also showed that 55 is the best cutoff of age in the model, consistent with the changes from the most recent 8th edition staging manual from AJCC. The demonstrated approach has the potential to create prognostic systems permitting data driven and real time analysis that can aid decision-making in patient management and prognostication.

Machine Learning for Predicting Complications in Head and Neck Microvascular Free Tissue Transfer

“…Prudent and appropriate implementation of ML algorithms should be no exception. Several recent publications highlight the early achievements of investigators who have employed ML algorithms for addressing scenarios that range from prognosticating outcomes in well-differentiated thyroid cancer based on imaging characteristics 9 to identifying factors that predict delays in adjuvant radiation treatment in patients with oral cavity squamous cell carcinoma. 10 The overall accuracy for correctly predicting a complication, as reflected in F scores of the GBDT models, ranged from 65% to 75% (Table III).…”

“…Other investigators have applied ML modeling to the Surveillance, Epidemiology, and End Results database with a goal of improving prognostic predictions in patients with well-differentiated thyroid cancer. 9 Shew et al similarly used data from the NCDB and a decision tree ML model to predict delays in adjuvant radiation in those undergoing surgery for head and neck cancer. 10 Still others have applied ML algorithms to imaging data to create models that predict treatment outcomes in patients with sinonasal squamous cell carcinoma.…”

Metastatic site discriminates survival benefit of primary tumor surgery for differentiated thyroid cancer with distant metastases

“…We placed Hurthle cell carcinoma (ICD-O-3 = 8290) into the category of follicular carcinomas, as used in Lim et al study. [ 14 , 15 ] To investigate the benefit of primary tumor surgery on the basis of metastasis sites, the variable was categorized into single organ and multiple organ metastases. The single organ metastasis was further classified into bone-only, liver-only, lung-only and brain-only metastasis, and multiple organ metastases were classified into multiply organ metastases including brain or excluding brain.…”

Clinical characteristics and disease-specific prognostic nomogram for primary gliosarcoma: a SEER population-based analysis

“…Another systematic review of predictive models for resectable pancreatic cancer reported that within the 16 developed models, 11 used the Cox regression method 28 . There are also reports of the application of machine learning in the development of a clinical prognostic model 29,30 . However, the Cox proportional hazards regression method is still the most widely used method when establishing prognostic models.…”