New nomograms to predict overall and cancer‐specific survival of angiosarcoma

Abstract

Objective This study was designed to establish and validate promising and reliable nomograms for predicting the survival of angiosarcoma (AS) patients. Methods The Surveillance, Epidemiology, and End Results database was queried to collect the clinical information of 785 AS patients between 2004 and 2015. Data were split into a training cohort (n = 549) and a validation cohort (n = 236) without any preference. Univariate Cox and multivariate Cox regression analyses were performed to analyze the clinical parameters. Independent prognostic factors were then identified. Two nomograms were constructed to predict overall survival (OS) and cancer‐specific survival (CSS) at 3 and 5 years. Finally, the models were evaluated using concordance indices (C‐indices), calibration plots, and decision curve analysis (DCA). Results Based on the inclusion and exclusion criteria, 785 individuals were included in this analysis. Univariate and multivariate Cox regression analyses revealed that age, tumor size, and stage were prognostic factors independently associated with the OS of AS. Tumor site, tumor size, and stage were associated with the CSS of AS. Based on the statistical results and clinical significance of variables, nomograms were built. The nomograms for OS and CSS had C‐indices of 0.666 and 0.654, respectively. The calibration curves showed good agreement between the predictive values and the actual values. DCA also indicated that the nomograms were clinically useful. Conclusion We established nomograms with good predictive ability that could provide clinicians with better predictions about the clinical outcomes of AS patients.

Angiosarcomas of the head and neck: Impact of large data analysis on clinical management

“…Using this approach, Liu et al 44 identified and validated a nomogram driven by tumor site, tumor size, and stage as predictive of DSS in angiosarcoma. Several other studies have developed nomograms for this disease which decrease predictive uncertainty further 45,46 .…”

“…Validation of these nomograms has been achieved as well, by dividing patients into a training set and a validation set, showing good overal agreement.Using this approach, Liu et al44 identified and validated a nomogram driven by tumor site, tumor size, and stage as predictive of DSS in angiosarcoma. Several other studies have developed nomograms for this disease which decrease predictive uncertainty further.…”