Purpose: To retrospectively test 2 nomograms recently validated as prognostic tools for patients with oropharyngeal cancer treated with curative radiation-based therapy, in 184 consecutive subjects treated from April 2004 to January 2016. Patient and Methods: Complete information regarding disease, survival status, and nomogram covariates was retrieved for 174 patients, with a median follow-up of 51 months. Calibration of the nomograms was obtained by evaluating the ability to accurately predict the absolute risk level; discriminative ability estimated by the Harrell C statistic was tested by assessing the ability of the nomograms to rank the patients from low risk to high risk. Results: The outcome predicted by the nomograms was worse than observed, especially for progression-free survival. This may be due to several factors: geographic and genetic background, smoking, alcohol use, the different national health system organization, which provides universal and free-of-charge coverage to all individuals, and the setting of treatment in clinical practice, outside of clinical trials. Conclusions: Our results show that the usefulness of nomograms to predict outcome of oropharyngeal cancer may not be generalizable to patients of different countries. Further studies are needed to clarify whether sociodemographic, epidemiologic, and cultural scenarios may seriously affect the prognostic ability of newly developed predictive tools.
Failure of Further Validation for Survival Nomograms in Oropharyngeal Cancer: Issues and Challenges
Bossi P.;
2018-01-01
Abstract
Purpose: To retrospectively test 2 nomograms recently validated as prognostic tools for patients with oropharyngeal cancer treated with curative radiation-based therapy, in 184 consecutive subjects treated from April 2004 to January 2016. Patient and Methods: Complete information regarding disease, survival status, and nomogram covariates was retrieved for 174 patients, with a median follow-up of 51 months. Calibration of the nomograms was obtained by evaluating the ability to accurately predict the absolute risk level; discriminative ability estimated by the Harrell C statistic was tested by assessing the ability of the nomograms to rank the patients from low risk to high risk. Results: The outcome predicted by the nomograms was worse than observed, especially for progression-free survival. This may be due to several factors: geographic and genetic background, smoking, alcohol use, the different national health system organization, which provides universal and free-of-charge coverage to all individuals, and the setting of treatment in clinical practice, outside of clinical trials. Conclusions: Our results show that the usefulness of nomograms to predict outcome of oropharyngeal cancer may not be generalizable to patients of different countries. Further studies are needed to clarify whether sociodemographic, epidemiologic, and cultural scenarios may seriously affect the prognostic ability of newly developed predictive tools.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
