The dynamics of death in melanoma

Authors


Robin T. Vollmer, MD, MS,

Laboratory Medicine 113, VA Medical Center,

508 Fulton Street, Durham, NC 27705, USA

Tel: +1 919 286 0411

Fax: +1 919 286 6818

e-mail: Robin.Vollmer@med.va.gov

Abstract

Background

Melanoma has been recently characterized as an over-diagnosed tumor, and some have suggested that the ‘epidemic’ in melanoma is spurious. Nevertheless, a fraction of melanoma patients continue to die of this tumor. For any tumor, the hazard function provides information about the timing and intensity of fatalities, and to examine the details of fatality in melanoma, herein the hazard functions for melanoma are derived and examined.

Methods

Data for this study came from the SEER data base, from AJCC data and from previously published studies of melanoma, and the hazard function was derived using the sum of two gamma functions and a nonlinear least-squares fitting algorithm.

Results

The derived hazard functions for melanoma peak at 1–3 years, a result that implies a form of rapidly evolving and fatal melanoma that is not consistently identified by routine prognostic factors. Yet in recent years the hazard function for all melanomas has declined suggesting that much of the epidemic in melanoma is due to non-fatal tumors.

Conclusion

Analyses of the hazard functions in a large number of melanoma patients has uncovered details about the dynamics of death that otherwise are not apparent.

Ancillary