Figure 1. Melanoma incidence. A, Incidence of regressing and all other malignant melanoma of the skin as reported in the Surveillance, Epidemiology, and End Results (SEER) database over time. B, Incidence of regressing melanoma of skin reported in SEER over time.
Figure 2. The use of the International Classification of Diseases for Oncology, 3rd Revision code for “melanoma, regressing” was highest in the New Mexico and Hawaii Surveillance, Epidemiology, and End Results (SEER) registries and lowest in Detroit and Iowa. Metro indicates metropolitan area; SF, San Francisco.
Martires KJ, Barnholtz-Sloan JS, Bordeaux JS. A Call for Consistent Reporting of Regression in Melanoma. Arch Dermatol. 2012;148(9):1081-1083. doi:10.1001/archdermatol.2012.1404
Author Affiliations: Case Comprehensive Cancer Center (Dr Barnholtz-Sloan), Case Western Reserve University School of Medicine (Ms Martires and Drs Barnholtz-Sloan and Bordeaux), Cleveland, Ohio; Department of Dermatology, University Hospitals Case Medical Center, Case Western Reserve University, Cleveland (Dr Bordeaux).
Spontaneous regression of melanoma is a commonly recognized but poorly characterized phenomenon.
Cases of primary malignant melanoma diagnosed between 1987 and 2007 from the 9 standard registries were analyzed using the National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) program.1 Because SEER data do not identify patients, institutional review board approval for this study was waived. Cases were classified by International Classification of Diseases for Oncology, 3rd Revision (ICD-O-3) as regressingsuperficial spreading, nodular, lentigo maligna, or acral lentiginous (ICD-O-3 codes 8723, 8721, 8742, 8743, and 8744, respectively). Excluded were cases that were not microscopically confirmed and/or those reported only by autopsy or death certificate. Age, race, sex, Breslow depth, ulceration, lymph node invasion, metastases, and overall survival were examined.
Cases of regressing melanomas were compared with malignant melanomas using t test and χ2 analysis. Overall survival analysis was performed using the Kaplan-Meier log-rank test and Cox proportional hazards modeling.
A total of 41 007 cases were analyzed, including only 374 cases of regressing melanoma. Use of the code for regressing melanomas did not occur until 1986, and its use has steadily increased since, in disproportion to the rising incidence of melanoma (Figure 1). Use of the code was highest in New Mexico and Hawaii and lowest in Detroit and Iowa (Figure 2). Compared with other cases of malignant melanoma, regressing melanomas occurred more often on the trunk and among men and blacks. These tumors had smaller Breslow depths and were less often ulcerated than other malignant melanomas; however, they were more likely to invade the lymph nodes and metastasize (Table 1).
Survival did not differ significantly between the regressing melanomas and other malignant melanomas (P = .72). By univariate analysis, older age (P < .001), greater Breslow depth (P < .001), presence of metastases (P < .001), and spread to regional or distant lymph nodes (P < .001, P = .03, respectively) were found to predict poor survival. Sex (P = .22), presence of ulceration (P > .99), and presence of tumor on the head or neck, upper extremities, or lower extremities vs the trunk (P = .96, P = .59, and P = .19, respectively) did not influence survival. Older age, greater Breslow depth, and lymph node invasion were significant prognostic factors for regressing melanomas in multivariable analysis (Table 2).
Our findings emphasize the need for consistent reporting of melanoma regression among clinicians and dermatopathologists. Spontaneous regression in melanoma is considered partial or complete resolution of a tumor in the absence of any treatment or therapy that is adequate to alter the course of malignancy. Between 13.8% and 50% of primary tumors are estimated to demonstrate spontaneous regression.2 In our population-based study, the diagnostic code for regressing melanomas was used only 374 times over a 20-year period, a 0.912% incidence rate. Despite the steady increase in its use over the 20-year period, this figure grossly underestimates the true incidence of primary spontaneous regression of melanomas. It reflects the lack of consistency in reporting this finding, leading to lack of abstraction of this feature for major incidence reporting at the national level. Coding of this morphologic subtype in ICD-O-3 codes usually occurs only when regression is indicated in the final diagnosis in place of a histologic subtype. Listing of regression is a site-specific factor in the database but is not required to be reported and is therefore of little utility. The inconsistency of recording of regression has been reported in Australia and Europe as well.3,4
The clinical significance of regression in melanomas is difficult to determine owing to the inability to assess true Breslow depth. It is therefore imperative to study this factor at the population-based level. Regressing melanomas present with a mixed prognostic picture—they are thinner, occur more often on the trunk, and are less likely to be ulcerated, but they occur more often among men and blacks and are more likely to invade. In our limited snapshot of regression reported on the national level, overall survival was no different than that for other melanomas. Studies have shown mixed data with regard to the prognostic significance of regression in melanomas.5 Both clinical and histopathologic characterization of regressing melanomas at the national level would be aided by consistency of reporting of this phenomenon.
Correspondence: Dr Bordeaux, Department of Dermatology, University Hospitals, Case Medical Center, Case Western Reserve University, 11100 Euclid Ave, 3500 Lakeside, Cleveland, OH 44106 (Jeremy.Bordeaux@uhhospitals.org).
Accepted for Publication: April 3, 2012.
Author Contributions: Ms Martires and Dr Bordeaux had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Martires, Barnholtz-Sloan, and Bordeaux. Acquisition of data: Martires and Barnholtz-Sloan. Analysis and interpretation of data: Martires, Barnholtz-Sloan, and Bordeaux. Drafting of the manuscript: Martires and Barnholtz-Sloan. Critical revision of the manuscript for important intellectual content: Barnholtz-Sloan and Bordeaux. Statistical analysis: Martires and Barnholtz-Sloan. Study supervision: Barnholtz-Sloan and Bordeaux.
Financial Disclosure: None reported.