Association between halo nevi and melanoma in adults: A multicenter retrospective case series

Abstract

To the Editor: The halo nevus is thought to be of little concern in children. In adults, however, a new-onset halo nevus has been suggested to be a harbinger of melanoma either within the halo nevus or at distant cutaneous or noncutaneous sites, according to case reports and small case series.1Epstein W.L. Sagebeil R. Spitler L. Wybran J. Reed W.B. Blois M.S. Halo nevi and melanoma.JAMA. 1973; 225: 373-377Crossref PubMed Scopus (62) Google Scholar,2Pellegrini J.R. Wagner Jr., R.F. Nathanson L. Halo nevi and melanoma.Am Fam Physician. 1984; 30: 157-159PubMed Google Scholar Multiple widely used dermatologic reference texts3Bolognia J. Schaffer J.V. Cerroni L. Dermatology: 2-Volume Set. Elsevier, Philadelphia, PA2017Google Scholar,4James W.D. Berger T.G. Elston D.M. Andrews' Diseases of the Skin. Saunders/Elsevier, London, England2011Google Scholar and online references (eg, UpToDate.com, DermNetNZ.org) advocate extensive melanoma screening in adults with new-onset halo nevus, including full cutaneous, oral, ophthalmic, and vaginal examinations, despite limited evidence supporting an association between halo nevus and melanoma. We aimed to further investigate the association between new-onset halo nevus and melanoma in adults by evaluating the incidence of melanoma in the year after a new halo nevus diagnosis. A multicenter retrospective chart review of clinical and histopathologic records at 8 university hospitals identified 879 patients in whom 888 halo nevi were diagnosed at aged 18 years or older (Brigham and Women's Hospital [n = 80], Massachusetts General Hospital [n = 166], New York University [n = 7], Northwestern University [n = 36], Oregon Health & Science University [n = 103], University of Pennsylvania [n = 27], Huntsman Cancer Institute and the University of Utah [n = 364], and Yale University [n = 96]). Ethical approval was obtained from each university's institutional review board. Patients being treated with immunotherapy for melanoma were excluded. Mean age at halo nevus diagnosis was 36.3 years (standard deviation 13.2 years) (Table I). Clinical records review identified 95 occurrences of melanoma in these 879 patients. Only 9 halo nevi were diagnosed within 1 year before melanoma diagnosis, representing a melanoma incidence rate of 0.01 (95% confidence interval 0.004-0.017) per person per year. All 9 of these melanomas represented primary cutaneous melanoma; there were no occurrences of primary noncutaneous melanoma, metastatic melanoma, or melanoma within the incident halo nevus. None of these 9 patients presented with multiple halo nevi. The remaining 86 melanoma diagnoses occurred either before the halo nevus diagnosis (n = 78) or greater than 1 year after (n = 8) (mean 5.75 years [standard deviation 4.30 years]). Personal history of vitiligo was not significantly associated with odds of melanoma development within the year after an incident halo nevus (odds ratio 3.97; 95% confidence interval 0.65-24.2; P = .13).Table IPatient demographicsCharacteristicAll patients with HN, n = 879Patients with melanoma diagnosed in the year after a new HN, n = 9Patients without melanoma diagnoses, n = 784Mean age at HN diagnosis ± SD, y36.3 ± 13.239.1 ± 11.935.3 ± 12.7 18–39580 (66)5 (55.6)540 (68.9) 40–59212 (24.1)4 (44.4)174 (22.2) 60–7968 (7.7)051 (6.5) ≥801 (0.1)01 (0.1) Unknown18 (2.0)018 (2.3)Mean age at melanoma diagnosis ± SD, y39.4 ± 14.740.4 ± 11.8NASex Men348 (39.6)5 (55.6)307 (39.3) Women529 (60.2)4 (44.4)475 (60.6) Unknown2 (0.2)02 (0.3)Race White799 (90.9)9 (100)711 (90.7) Nonwhite18 (2.0)017 (2.2) Unknown62 (7.1)056 (7.2)Halo nevus biopsied Yes396 (45.1)6 (66.7)345 (44.0) No474 (53.9)3 (33.3)431 (55.0) Unknown9 (1.0)08 (1.0)Vitiligo Yes78 (8.9)2 (22.2)70 (8.9) No450 (51.2)3 (33.3)391 (49.9) Unknown351 (39.9)4 (44.4)323 (41.2)Values are provided as No. (%) unless otherwise indicated.HN, Halo nevus; NA, not applicable; SD, standard deviation. Open table in a new tab Values are provided as No. (%) unless otherwise indicated. HN, Halo nevus; NA, not applicable; SD, standard deviation. Limitations of the study include the retrospective design and heterogeneity by which each site identified halo nevus cases. Halo nevus and melanoma were identified by documentation and coding that may not accurately reflect the true date of onset or incidence. In conclusion, this study found that adult-onset halo nevi were associated with a 1% risk of primary cutaneous melanoma development in the year after halo nevus diagnosis, with no cases of primary noncutaneous or metastatic melanoma. This melanoma risk is comparable to that of individuals with a history of atypical nevi or personal or family history of melanoma.5Varedi A. Bishop M.D. Boucher K.M. Kim C.C. Grossman D. Powering a prospective melanoma chemoprevention trial in high-risk cohorts.Int J Dermatol. 2019; 58: e232-e234Crossref PubMed Scopus (3) Google Scholar Given these findings, we recommend annual total body skin examinations in adults with a new diagnosis of halo nevus and without additional risk factors, and we do not advocate reflexive screening for primary noncutaneous or metastatic melanoma.