Predicting the Development of Brain Metastases in Patients with Local/Regional Melanoma TIMOTHY L. FRANKEL, MD, 1 ZUBIN M. BAMBOAT, MD, 1 CHARLOTTE ARIYAN, MD, 1 DANIEL COIT, MD, 1 MICHAEL S. SABEL, MD, 2 AND MARY S. BRADY, MD 1 * 1 Department of Surgery, Gastric and Mixed Tumor Service, Memorial Sloan-Kettering Cancer Center, New York, New York 2 Department of Surgery, University of Michigan, Ann Arbor, Michigan Background: The brain is a common site of recurrence in melanoma patients. Brain recurrence may present as a seizure, hemorrhage, or death. We sought to determine predictors of brain metastases in patients with primary and regional melanoma in order to facilitate targeted screening. Methods: Prospectively maintained databases were used to identify patients treated for local or regional melanoma who developed stage IV melanoma with and without brain metastasis at initial recurrence. One hundred twenty patients were identified with brain relapse and compared to 487 patients without brain recurrence. Results: On univariate analysis, patients with brain metastases were younger (55 vs. 59yrs, P ¼ 0.04) but did not differ in primary site (head and neck 23% vs. 21%, P ¼ 0.20). Brain metastasis patients had thinner primaries (mean 3.4 vs. 4.5 mm, P ¼ 0.01). There were no other pathologic differences including ulceration (55% vs. 53%, P ¼ 0.75), mitoses (7 vs.7.5, P ¼ 0.61) or histologic subtype. Younger age and decreased Breslow thickness were independently associated with brain metastases at stage IV recurrence (OR ¼ 1.10 P ¼ 0.01 and OR ¼ 1.02 P ¼ 0.02, respectively). Conclusions: Our analysis, the largest to date, demonstrates that thinner Breslow depth and younger age were associated with brain recurrence at first presentation with Stage IV disease. J. Surg. Oncol. 2014;109:770–774 ß 2014 Wiley Periodicals, Inc. KEY WORDS: melanoma; metastases; brain INTRODUCTION The role of cross‐sectional imaging following treatment of clinically localized melanoma, as defined by the National Comprehensive Cancer Network (NCCN) is limited, with routine history and physical exam considered adequate in the absence of symptoms of recurrent disease [1]. Patients with stage III disease are more commonly subjected to routine imaging, particularly during the first 3 years following treatment. While signs and symptoms of visceral recurrence may be subtle, such as vague discomfort, fatigue, or nausea, recurrence in the brain may result in seizure, mental status changes or paralysis [2]. Current surveillance guidelines do not recommend routine brain imaging for asymptomatic patients with stage I–III melanoma [1,3]. If a high‐risk population could be identified, however, targeted screening could be offered, potentially allowing for diagnosis and treatment prior to development of symptoms. While no data exists regarding the impact of early detection of brain lesions on survival, prompt treatment with radiotherapy or surgical resection may decrease the likelihood of neurologic catastrophe due to edema, hemorrhage, or herniation [4,5]. We identified patients treated for primary and/or regional melanoma who developed brain metastases as part of their first distant recurrence. Baseline demographic characteristics and histopathologic data were used to compare these patients to those who developed distant non‐CNS metastases in order to identify predictors of brain metastasis as a component of systemic relapse. PATIENTS AND METHODS Approval for this study was obtained by the Institutional Review Boards (IRB) of Memorial Sloan‐Kettering Cancer Center (MSKCC) and the University of Michigan Medical Center (UMMC). Study Patients Prospectively maintained melanoma databases were queried to identify patients who underwent resection of primary or regional cutaneous melanoma at UMMC or MSKCC, and subsequently developed distant metastatic disease during follow‐up. Patients under the age of 18, no evidence of distant metastatic disease, or lost to follow‐ up were excluded from analysis. Those who presented with Stage IV disease at the time of their initial diagnosis and those with uveal or mucosal melanoma were also excluded. Patient and Tumor Characteristics Patient characteristics were collected from hospital records and included: gender, age, location of the primary melanoma (head/neck, trunk, lower extremity, and upper extremity) and stage at diagnosis. Histologic features of the primary melanoma were reported by pathologists with extensive experience in dermatopathology and included: Breslow thickness (in millimeters), Clark level, mitotic rate per square millimeter, histologic subtype, ulceration, perineural and lymphovascular invasion, regression, immune infiltrate and satellitosis. Sentinel lymph node (SLN) biopsy was performed, when appropriate, as Grant sponsor: There were no funding sources for the conduct of this study. The authors report no conflicts of interest in the conduct of this study. *Correspondence to: Mary S. Brady, MD, FACS, 1275 York Ave, New York, NY 10065. Fax: 212‐794‐5847. E‐mail: bradym@mskcc.org Received 13 November 2013; Accepted 14 January 2014 DOI 10.1002/jso.23574 Published online 20 February 2014 in Wiley Online Library (wileyonlinelibrary.com). Journal of Surgical Oncology 2014;109:770–774 ß 2014 Wiley Periodicals, Inc.