Prognostic evaluation models for primary thyroid lymphoma, based on the SEER database and an external validation cohort

PURPOSE: Primary thyroid lymphoma (PTL) is a rare malignancy, and the literature is limited to small case series and case reports. This study aimed to assess the epidemiologic characteristics, survival, and prognostic factors of patients with PTL. METHODS: We analyzed 2215 PTL patients from the Surveillance, Epidemiology, and End Results database medical records, between 1983 and 2015, as the training cohort. We enrolled 105 patients from the Cancer Hospital, Chinese Academy of Medical Sciences, for the external validation cohort. The nomograms for predicting the 1-, 5-, and 10-year overall survival (OS) and lymphoma-specific survival (LSS) were constructed. RESULTS: PTL incidence steadily increased from 1977 to 1994, with an annual percentage change of 3.2% (95% confidence interval [CI]: 1.2–5.2, P < 0.05). The 1-, 5-, and 10-year OS and LSS rates were 84.66%, 71.61%, and 55.95%; and 90.5%, 85.7%, and 82.2%, respectively. Multivariate Cox regression analysis revealed that shorter OS association with age ≥ 60 years (hazard ratio [HR], 3.94; 95% CI 3.31–4.69; P < 0.001), unmarried status (HR, 1.55; 95% CI 1.37–1.75; P < 0.001), Ann Arbor stage III-IV (HR, 1.55; 95% CI 1.37–1.75; P = 0.020), diffuse large B-cell lymphoma (HR, 2.60; 95% CI 1.15–5.87; P = 0.022), and T cell non–Hodgkin lymphoma (HR, 3.53; 95% CI 1.12–11.10; P = 0.031). In the multivariate competing-risk analyzes, age, stages III-IV, year of diagnosis, surgery, radiation, chemotherapy, and histology were strongly predictive of PTL-specific risk of death. To estimate the 1-, 5-, and 10-year LSS and OS rates, respectively, nomograms were built. In the validation cohort, the results also confirmed the utility. CONCLUSIONS: This study presents the first prognostic model with an external validation that could help clinicians identify patients with high-risk PTL to improve their prognosis. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s40618-021-01712-3.