Clinical significance of BRAF mutations in metastatic melanoma
© Chang et al. 2004
Received: 07 December 2004
Accepted: 21 December 2004
Published: 21 December 2004
Skip to main content
© Chang et al. 2004
Received: 07 December 2004
Accepted: 21 December 2004
Published: 21 December 2004
Forty to eighty percent of melanoma tumors have activating mutations in BRAF although the clinical importance of these mutations is not clear. We previously reported an analysis of BRAF mutations in metastatic melanoma samples from 68 patients. In this study, we correlated patient baseline characteristics, prognostic factors, and/or clinical outcomes with the presence of BRAF mutations. No significant differences were observed in age, gender, location of primary melanoma, stage at the diagnosis, and depth of primary tumor between patients with and without BRAF mutations. Melanomas harboring BRAF mutations were more likely to metastasize to liver (P = 0.02) and to metastasize to multiple organs (P = 0.048). Neither time to progression to stage IV nor overall survival were associated with BRAF mutations. In conclusion, we observed no significant differences in clinical characteristics or outcomes between melanomas with or without BRAF mutations. Although there was an increased frequency of liver metastasis and tendency to metastasize to multiple organs in tumors with BRAF mutations, there was no detectable effect on survival. Future prospective studies should include analysis of whether BRAF mutations in melanoma tumors correlate with an increased tendency to metastasize to liver or to multiple organs.
The mitogen-activated protein kinase (MAPK) pathway mediates cellular responses to growth signals and activation of this pathway has been shown to be critical in tumor formation, particularly in melanoma [1–3]. Recently, activating BRAF mutations were found with high frequency in malignant melanomas, including primary tumors and cell lines [4, 5]. Suppression of activating BRAF mutations in cultured human melanoma cells inhibited the MAPK cascade causing growth arrest and promoting apoptosis , further suggesting the potential critical role of activating BRAF mutations in malignant transformation in melanoma.
We have reported the analysis of BRAF mutations in a cohort of metastatic melanoma patients  and noted a mutation proportion of 44%. As expected from previous reports, the most frequent mutation was BRAF V599E , which was found in 40% of samples. Since little is known about the clinical implications of activating BRAF mutations in melanoma tumors, we examined whether the melanoma tumors harboring BRAF mutations in this cohort showed different clinical or biological features compared to the melanoma tumors without mutations.
Cryopreserved metastatic melanoma samples from 68 patients were selected from the Memorial Sloan-Kettering Cancer Center Tumor Bank. Patient demographic data were collected on the 68 patients whose tumors we had previously analyzed for BRAF mutations . Data collected included: location of primary tumor, thickness, ulceration, stage of disease (according to American Joint Committee on Cancer Staging System), sites of metastasis, site of tumor biopsy, and history of and responsiveness to chemotherapy. This retrospective analysis was performed with IRB approval which determined that this was exempt research under 45 CFR 46.101.b(4).
BRAF (exons 11 and 15) was sequenced as previously reported . For 65/68 patients, a single metastatic site was sequenced for BRAF. In three patients, two to four metastatic sites were available for sequencing. For patients with multiple specimens, we considered only the first acquisition of tissue in assigning patients to mutant or wild type categories.
The patients were first seen at MSKCC between June 1993 and April 2000. Clinical follow up was available through April, 2003. Comparisons between mutated and wild type were made using either the χ2 test, t-test or Cochran-Armitage test to trend. Survival distributions were estimated using the Kaplan-Meier method and compared using the log-rank test. Stage IV patients were stratified into two categories: those with stage M1a or M1b (lymph nodes, soft tissues and/or lung metastasis) and those with stage M1c (all other sites).
BRAF mutations and clinical characteristics
Mutation N = 30 (44.1%)
Wild Type N = 38 (55.9%)
Stage at Diagnosis
Thickness (Number available)
(N = 18)
(N = 22)
1.75 (0.2, 20)
2.80 (0.4, 35)
Patients' age ranged from 29 to 97 years; there was no statistically significant difference in patients' age with regards to BRAF mutations (p = 0.12). Similarly, there was no difference in the distribution of primary sites and stages at diagnosis between patients with and without BRAF mutations. We noted that among the 7 melanomas arising from the head and neck region, only 1 harbored a BRAF mutation. Although there were too few of these patients for a meaningful statistical analysis, this observation is consistent with a recent report indicating that mucosal melanomas did not harbor BRAF mutations [8, 9]. The mean thickness of primary tumor was 2.98 mm (range: 0.2, 20 mm) for patients with BRAF mutations, and 4.83 mm (range: 0.4, 35 mm) for patients without (p = 0.29). The effect of BRAF mutation on other known prognostic features of primary tumor such as the presence or absence of ulceration, regression, tumor-infiltrating lymphocytes, lymph-vascular invasion, and mitotic index could not be assessed because this information was available for only a small proportion of patients.
Correlation between BRAF mutations and number of metastasis among patients with stage IV melanoma
Sites of Metastasis
Mutation N = 27 (%)
Wild Type N = 24 (%)
Soft Tissue/Lymph Nodes/Lung only
Non-soft tissue site
Association of BRAF mutations with the number of metastatic sites in patients with stage IV melanoma
Number of Sites Per Patients
Mutation N = 27 (%)
Wild Type N = 24 (%)
p = 0.048
We examined the response to systemic therapy (chemotherapy or biochemotherapy) for the 33 patients who received such treatments. For patients with BRAF mutations, 18 patients received systemic therapy of who two patients achieved complete remission (response rate 11.1%). Fifteen patients with wild-type BRAF received systemic therapy of whom three patients achieved complete remission and two achieved partial remission (response rate 33.3%) (p = 0.12).
There was no statistically significant difference between time to progression to stage IV disease either from the time of diagnosis or from stage III in patients with or without BRAF mutations (data not shown). As this is a retrospective study, we cannot rule out the possibility that differences in interval assessments affected our ability to detect a difference in time to progression. On the other hand, date of death is an endpoint not affected by interval assessment times. There was no statistically significant difference between patients with BRAF mutations and those without BRAF mutations.
High frequency of BRAF mutations has been reported in malignant melanoma [4, 5, 7], however, there has been little clinical correlation data elucidating the biological effects of these mutations in patients. We initiated this study in an attempt to address this question.
The observation that BRAF mutations are common in melanocytic nevi  has led to the assumption that mutations in BRAF occur early in melanocytic transformation and play an important role in the initiation of malignant transformation. Recently, an alternative view has been suggested by Dong et al who confirmed the high frequency of BRAF mutations present both in nevi and later stage melanomas but found few BRAF mutations in early stage radial growth phase melanomas . They interpret these findings to mean that BRAF mutations are not involved in the initiation of the majority of melanoma, but rather play a role later in progression.
Since little information was available on the biological effects of activating BRAF mutations in melanoma, we analyzed the clinical characteristics of 68 melanoma patients whose tumors we had previously analyzed for BRAF . We found that patients with tumors harboring a BRAF mutation were more likely to have metastasis to the liver and tended to have more organs involved with melanoma than patients without mutations. This is consistent with the idea that activating BRAF mutations affect the pattern of metastatic spread in melanoma, although we await confirmation of these findings in a prospective study.
In our cohort of subjects, there were 33 patients who received systemic therapy (18 patients with BRAF mutations, 15 patients without detectable mutations). There was a trend towards lower response rates among patients with mutations, although this trend was not statistically significant and is confounded by the small number of patients, the heterogeneity of treatments these patients received, and the retrospective nature of these analyses. This is a question that deserves to be revisited in a prospective manner.
Kumar and colleagues found that melanoma patients with BRAF mutations showed a statistically significant diminished duration of response to treatment compared to those without the mutations [12, 13]. Their retrospective analysis consisted of 38 patients with metastatic melanoma (stage III or IV) who had been treated with chemoimmunotherapy (dacarbazine, vincristine, bleomycin, lomustine, and human leukocyte interferon). This cohort of patients had a surprisingly high response rate of 55%. Although the likelihood of response did not correlate with the presence of a BRAF mutation, multivariate analysis revealed that among patients who had responded, patients with BRAF mutations had a shorter duration of response compared to patients without any BRAF mutations (median 3.4 versus 9.8 months). They did not analyze the effect of BRAF mutations on the site of metastatic spread or other biological characteristics of the tumor.
Houben et al. reported that the presence of BRAF mutation in a metastatic melanoma lesion was associated with a poor prognosis as measured by shortened survival . In our study, we did not detect any impact on either progression free or overall survival by the presence of BRAF mutation. The patient characteristics were not reported by Houben and colleagues but they indicate that most patients had soft-tissue metastases (M1a or M1b). In contrast, most of our patients had M1c melanoma and this could account for the different findings.
In three patients, multiple metastatic samples were available for analysis; in 2 of these patients, there was discordance in the presence of detectable BRAF mutations. In one patient in whom 2 lung metastasis collected over a period of one month were analyzed, one metastasis contained a BRAF V599E mutation; the other metastasis was wild-type for BRAF. In another patient, metastasis from lung, gastrointestinal (GI) tract, lymph node, and soft tissue were collected of a period of 34 months. All tumors harbored the BRAF V599E mutation except for the GI metastasis which was wild-type. It is possible that this discordance represents a problem with assay sensitivity, but we cannot rule out the possibility that there is true heterogenicity among metastasis with regard to BRAF mutations. Although this discordance among metastasis seems to contradict the observation that BRAF mutations are an early event in melanocytic nevi transformation, one possibility is that in melanomas arising from non-nevus melanocytes, BRAF mutation is a late event occurring in individual metastasis. Consistent with this, Shinozaki et al. recently reported that the incidence of BRAF mutation of primary melanoma did not correlate with Breslow thickness, and there was significantly higher frequency of BRAF mutation in metastasis than in primary melanoma, arguing that BRAF mutation maybe acquired during development of metastasis . Houben also reported that in 3/22 cases, the BRAF mutational status of the primary and metastasis did not correlate . This issue merits further investigation.
In summary, this analysis represents the largest study to date correlating BRAF mutations and clinical outcomes in metastatic melanoma. Although we observed a statistically significant higher frequency of liver metastasis and tendency to metastasize to multiple organs in patients with BRAF mutations, there was no significant effect on survival or response to systemic therapy detected by this study. Although this analysis is limited by its retrospective nature and the relatively small number of patients, it appears unlikely from these observations that there will be a major qualitative difference in the biological behavior between melanomas with and without BRAF mutations. Larger prospective studies are required to verify these observations and to clarify other biological consequences of BRAF mutations in melanoma.
David Z. Chang is supported by the AACR-Bristol-Myers Squibb Oncology Fellowship in Clinical Research, ASCO Young Investigator Award sponsored by Roche Laboratories, CALGB Clinical Research Fellowship sponsored by Aventis Oncology, and Ladies Auxiliary Veteran of Foreign Wars Cancer Research Grant.
Paul B. Chapman is supported by NCI grant K24 CA81293 and the Swim Across America Foundation.
The authors wish to thank Jennifer Guido, Susan Clinco and Ami Patel for their help with the database.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.