Prognostic scoring models in parotid gland carcinoma
Marie Westergaard‐Nielsen, Sören Möller, Christian Godballe, Jesper Grau Eriksen, Stine Rosenkilde Larsen, Katalin Kiss, Tina Klitmøller Agander, Benedicte Parm Ulhøi, Birgitte Charabi, Tejs Ehlers Klug, Henrik Jacobsen, Jørgen Johansen, Claus Andrup Kristensen, Elo Andersen, Maria Andersen, Kristine Bjørndal
Abstract
BACKGROUND: The aim was to identify prognostic factors and test three prognostic scoring models that predicted the risk of recurrence in patients with parotid gland carcinoma. METHODS: All Danish patients with parotid gland carcinoma, treated with curative intent, from 1990 to 2015 (n = 726) were included. Potential prognostic factors were evaluated using Cox regression and competing risk analyses. The concordance of each prognostic model was estimated using Harrel's C index. RESULTS: The study population consisted of 344 men and 382 women, with a median age of 63 years. Age above 60 years, high grade histology, T3/T4 tumor, regional lymph node metastases, and involved surgical margins were all associated with a significant reduction in recurrence-free survival. The prognostic model that agreed best with actual outcomes had a C-index of 0.76. CONCLUSION: Prognostic scoring models may improve individualized follow-up strategies after curatively intended treatment for patients with parotid gland carcinoma.