PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Pigment Cell Melanoma Res. Author manuscript; available in PMC 2011 December 1.
Published in final edited form as:
PMCID: PMC3107973
NIHMSID: NIHMS296009

Genetic and morphologic features for melanoma classification

Summary

Melanoma is comprised of biologically distinct subtypes. The defining clinical, histomorphologic and molecular features are not fully established. This study sought to validate the association between genetic and histomorphologic features previously described, determine their reproducibility, and association with important clinical variables.

Detailed clinical and histomorphologic features of 365 primary cutaneous melanomas were assessed by 11 pathologists and correlated with mutation status of BRAF and NRAS. There was substantial agreement in the quantitative assessment of histomorphologic features showing similar or better interobserver reproducibility than the established WHO classification scheme. We confirmed that melanomas with BRAF mutations showed characteristic morphologic features (p<0.0001) and metastasized more frequently to regional lymph nodes (p=0.046). Importantly, melanomas without mutations were a heterogeneous group, with a subset having very similar features clinical and morphological features than those with BRAF mutation raising the possibility that they are biologically related.

Our study confirms an association between histomorphologic features, mutation status and pattern of metastasis, providing criteria for a refined melanoma classification aimed at defining biologically homogeneous disease subgroups.

Keywords: melanoma, BRAF, mutation, classification, histomorphology

Introduction

Melanoma is an aggressive cancer with an increasing incidence (John F Thompson et al. 2009). Clinical, histopathologic, epidemiologic, and molecular data indicate that melanoma is comprised of biologically distinct subtypes (Curtin et al. 2005). The current WHO classification of skin tumors distinguishes four main subtypes of melanomas: superficial spreading melanoma (SSM), lentigo maligna melanoma (LMM), nodular melanoma (NM), and acral lentiginous melanoma (ALM) (Clark et al. 1969), (Reed 1976). Another classification scheme distinguishes two types of melanoma on non-glabrous1 skin: melanomas occurring in non-chronically sun-damaged skin (non-CSD) with frequent BRAF mutations and no KIT mutations and melanomas on CSD skin with infrequent BRAF mutations and KIT mutations in approximately 20% ((B C Bastian et al. 2000), (Curtin et al. 2005)(Curtin et al. 2006)). Melanomas originating on glabrous skin or the nail apparatus were distinguished as acral melanomas based on genetic features such as gene amplifications and mutations or amplification of KIT in 40% (B C Bastian et al. 2000), (Curtin et al. 2005), (Curtin et al. 2006). Mucosal melanomas (i.e. primary melanomas arising from mucosal membranes) while also expressing frequent amplifications and mutations or amplification of KIT (40% of cases), were proposed as separate based on the patterns of chromosomal aberrations (Curtin et al. 2006), (Curtin et al. 2005). Molecular studies also set apart uveal melanomas and intradermal melanocytic proliferations of the skin that morphologically fall under the spectrum of blue nevi: both showed frequent mutations in the Gq family of G-proteins GNAQ (Van Raamsdonk et al. 2009) and GNA11 (Van Raamsdonk et al., submitted for publication).

The similarities between classification approaches based purely on histopathologic and clinical characteristics and those based on molecular features is remarkable and provides strong genetic support for the existence of different melanoma subtypes. Specifically, BRAF mutations are most common in SSMs and less frequent in LMMs and ALMs, while the categories enriched for KIT mutations overlap with ALM and LMM. However, important differences between the two schemes exist, suggesting that the incorporation of molecular features can help refine classification. For example, molecular studies did not find any characteristic genetic or clinical features associated with the WHO melanoma subtype of NM. This indicates that NM may not describe a subtype of the same taxonomic rank as the other categories, but instead represent a subcategory under the various higher levels that undergoes an accelerated progression. In taxonomic terms SSM, ALM, LMM thus may represent genera of the melanoma family, while NM represents a type that may occur in each or several of these genera.

The precise delineation of biologically distinct subtypes is likely to impact on clinical management and prevention strategies. Several of the oncogenic alterations identified in melanoma are being successfully exploited as therapeutic targets bringing new relevance for improved disease classification to identify patients who may benefit from targeted therapy. While it seems straightforward to base therapy stratification entirely on the molecular genotype, this approach may be too narrow. A considerable portion of melanomas have no known oncogenic alterations, and it is conceivable that some of them may have mutations in functionally related genes and thus also may respond to agents targeting the same signaling pathway. Assuming a causal relationship between the underlying oncogenic alterations and at least some of the phenotypic characteristics used for classification, it is plausible that these potential responders will be found within the same genus of melanoma. While the spectrum of oncogenic mutations in melanoma is being filled in, it is important to continuously integrate newly generated genetic information with clinical and histomorphologic correlates. This effort will help refine criteria for classification but at the same time generate testable hypotheses that may help to identify new oncogenic alterations in specific subsets and assist in the design of clinical trials with targeted agents.

In a first attempt to integrate morphologic aspects with underlying genetic alterations, we previously defined a series of refined morphometric features and found that melanomas with BRAF mutations had characteristic histopathological and clinical features. Using a combination of morphometric features BRAF mutation status could be predicted with 82% accuracy (Viros et al. 2008).

In the present study we establish the practicability of the proposed morphologic criteria by showing that they can be reproducibly applied among pathologists and validate their association with BRAF mutation status. Finally, we demonstrate that melanomas for which the altered oncogenes are currently not known are a heterogeneous group, in which subsets have strong phenotypical similarities with subsets with known oncogenic alterations. These findings have implications for disease classification, and, possibly, for the design of clinical trials and efforts to identify additional genetic alterations in melanoma.

Results

Table 1 shows the breakdown of gender and age, WHO melanoma subtype, mutation frequencies of BRAF and NRAS, and anatomic site of the primary tumor of the cohort analyzed in this study and the previous study (Viros et al. 2008) for comparison. With the exception of a greater median tumor thickness (2.0 mm vs. 1.3 mm) and a greater number of NMs (23.6% vs. 5.3%) in the present cohort, the compositions were very similar. Figure 1 shows the median weighted kappa scores as a measure of interobserver agreement ranging from 0.35 to 0.69, corresponding to fair (scores between 0.2 and 0.4) and to substantial (scores between 0.6 and 0.8) agreement. The best agreement was obtained with the criteria pigmentation, scatter of intraepidermal melanocytes, and solar elastosis, which showed better inter-observer agreement than the established WHO melanoma subtype.

Figure 1
Interobserver agreement on morphologic features and WHO melanoma subtype
Table 1
Clinical characteristics of patients and melanomas by mutation status of BRAF and NRAS in both cohorts.

As shown in Figure 2 and Table S2, BRAF mutant primaries had increased upward scatter and nest formation of intraepidermal melanocytes, consisted of larger and rounder, more heavily pigmented melanocytes, displayed a thickened epidermis of their radial growth phase areas, were better circumscribed, and had less solar elastosis of their surrounding skin than primaries that did not have BRAF mutations. The features showed a linear association with BRAF mutation status, i.e. the strength of the association increased monotonously with higher (or lower) scores. Patients with BRAF-mutant primaries were also significantly younger than patients whose melanomas did not have BRAF mutations. BRAF mutations were further significantly more common in melanomas on the trunk compared to those in acral location. The associations between the morphologic and clinical features are similar in direction as in the previous study by Viros et al. (Viros et al. 2008) but the strength of the association, as expressed by the odds ratios, varies. In particular, the associations for scatter and nesting are weaker in the present study.

Figure 2
Association of histomorphologic features,anatomic site, and clinical characteristics and BRAF mutation status

When applying the binary decision tree determined in the previous study in order to predict BRAF mutation status, prediction could be made with 60.3% accuracy (52.0% sensitivity, 72.2 % specificity, p=0.0004, calculated by Fisher's Exact test), compared to a probability of 49.7% to predict BRAF mutation by chance (Figure 3a). The original decision tree determined in the prior study uses scatter, nesting, and pigmentation as decision nodes for predicting BRAF status. Because the current study established that the feature “nesting” used in this tree had comparatively lower kappa scores, we tested alternative trees composed of features with higher reproducibility but comparable associations with BRAF status. The resulting binary tree that best predicted BRAF mutation status had cell shape at the entry node followed by nodes using pigmentation and solar elastosis (Figure S1). This classification tree created by single binary classification tree classifier was too complex to be practicable. Therefore, smaller alternative trees were created taking into account the information obtained in the binary classification tree classifier as well as inter-rater agreement and odds ratios of association of characteristics and BRAF mutation status (Figure 3 b–d).

Figure 3
Prediction algorithms of BRAF mutation status using only morphologic variables

As scoring of solar elastosis can be affected by variation in the hematoxylin eosin staining protocol, we excluded those 90 slides that were not processed at UCSF. This increased the prediction accuracy from 61.9% (sensitivity 79.0%, specificity 44.9%, p<0.0001) to 65.5% (sensitivity 80.2%, specificity 51.2%, p<0.0001), using the prediction algorithm that included solar elastosis as the entry node (Figure 3d, Table S1).

Independent of BRAF mutation status, NRAS mutations were found to be significantly associated with low or absent scatter and better circumscription (p=0.026 and p<0.03 respectively, Table S2). However, because the frequency of NRAS mutations was lower and NRAS status could not be established in 32 cases that were wild type for BRAF, our power to identify features associated with NRAS mutations was limited.

In the prior study, we took advantage of the association between BRAF and age and used age as a proxy for BRAF status to analyze possible associations with survival and pattern of metastasis in a large melanoma series from a registry, for which no mutation status was available (Viros et al. 2008). In that study, an age cutoff of 55 years was determined in the study cohort to best predict BRAF mutation status. Patients in the registry whose melanomas were predicted to be BRAF mutant because they were younger than 55 years had a significantly better survival and were more prone to relapse in the regional lymph node basin than patients aged 55 years or older. The same association was seen in the current cohort (Table S3 and Figure S2a) (Mehta & N. R. Patel 1986). The available follow-up information in the current study allowed us to test more directly, whether these associations reflect any effect of BRAF mutation status on survival and pattern of metastasis. Supporting the prior studies, we found that BRAF-mutant primaries were more likely to first metastasize to regional lymph nodes while melanomas without BRAF mutations were more likely to show non-nodal metastases (logistic regression, p=0.046; Fisher's Exact test, p=0.069) (Table 2). By contrast, when we tested the role of BRAF mutations on overall and disease-free survival directly, no significant associations were found (p-values p=0.58 and p=0.39, respectively) (Figure S2b+c).

Table 2
Primaries with BRAF mutations are more likely to metastasize to the regional node.

Discussion

The primary purpose of this study was to determine the practicability and reproducibility of the previously defined morphometric features and validate their association with BRAF mutation status of the primary. Most of the criteria, in particular the criteria that best predict BRAF mutation status, showed better inter-observer reproducibility than the established WHO-classification scheme, indicating that they can be reliably used by pathologists.

As in the previous study, melanomas carrying BRAF mutations showed increased upward migration and nest formation of intraepidermal melanocytes in the RGP portion of the tumors, thickening of the involved epidermis, sharper demarcation to the surrounding skin, larger and rounder, more pigmented tumor cells and tend to occur in skin without marked solar elastosis in anatomic sites such as the trunk. As in previous studies, BRAF mutations were found more commonly in younger patients (Maldonado et al. 2003), (Viros et al. 2008), (Thomas et al. 2010). While BRAF status did not have a significant association with overall or disease-free survival it was associated with lymph node metastasis (Table S4). Taken together, these findings strongly support the notion that BRAF-mutant melanomas are part of a biologically distinct melanoma type that not only differs in its clinical and histopathologic aspects, but also in its pattern of metastasis.

27.1% of cases in our cohort had no mutations in BRAF or NRAS. While a small minority of these cases, in particular acral and CSD melanomas, are likely to have KIT mutations instead (Curtin et al. 2006), the majority are expected to harbor mutations in yet to be discovered oncogenes. It is therefore possible, if not likely, that some of these mutations could be functionally related to BRAF, e.g. affect genes acting in the same pathway, and thus place those melanomas in the same melanoma type as those with BRAF mutations. As we only analyzed exon 15 of BRAF, mutations elsewhere in BRAF may even be present in some of these cases. Alternative candidate genes in the pathway include MEK1 and 2, acting immediately downstream of BRAF (Emery et al. 2009). A similar argument can be made for NRAS mutant melanomas, which show other histomorphologic characteristics and thus are likely to be part of a different melanoma subtype. In an attempt to test this hypothesis using the data from the current study, we compared the morphologic and clinical features of two groups of melanomas. One was comprised of melanomas that were either BRAF mutant or were BRAF and NRAS wild type but, based on the decision tree shown in Figure 3a, predicted to be BRAF mutant. The other group were melanomas that either harbored NRAS mutations or that were NRAS and BRAF wild-type but, based on the same decision tree, were predicted to be BRAF wild type. As shown in Figure 4 and Table S4, the association of the majority of microscopic and clinical features was strengthened by this maneuver, compared to an assessment where BRAF mutation alone was used in the analysis (features of the decision tree which were used to predict BRAF status had to be eliminated to avoid circularity). In other words, a combination of genetic information – BRAF and NRAS mutation status - combined with scatter, nesting, and pigmentation resulted in melanoma groups that were more homogeneous for other clinical and pathologic features than using genetic information alone. This strongly suggests that BRAF-mutant melanomas are a part of a larger group of melanomas that form a biologically distinct subtype, which is currently imperfectly characterized by BRAF mutation status alone. If confirmed, this concept would have important implications. It would mean that, at least until the full complement of oncogenic mutation in melanoma is discovered and the functional relationship of these alterations is characterized, morphologic and clinical features will remain of critical importance for disease classification. If BRAF mutations are actively causing some of the associated phenotypic features, it would be conceivable that melanomas, which morphologically are predicted to have BRAF mutations but do not carry BRAF mutations, may still respond to a therapeutic intervention that is effective in BRAF-mutant melanomas. This will of course depend on whether the yet to be discovered oncogenes(s) mutated in these cases resides downstream of the target of therapy. Additional studies are required to determine the possible clinical relevance for therapeutic trials aimed at specific melanoma subsets defined by a combination of genetic and clinicopathologic features.

Figure 4
Melanomas without mutations in BRAF or NRAS are a heterogeneous group

Materials and Methods

Study design

A cohort of 365 formalin-fixed, paraffin-embedded, invasive primary cutaneous melanomas for which a minimum of three year follow-up was available was retrieved from four different sources: 90 cases from the EORTC Melanoma group, 50 cases from the German Cancer Research Center (DKFZ), 180 cases from the University of Vienna, Austria, and 64 cases from the University of California, San Francisco (UCSF). With the exception of the EORTC cases, new sections were cut and stained at UCSF for all cases. The median follow-up was 5.1 years. Eleven pathologists with expertise in melanoma were divided into three teams, with three or four members. Each team examined subsets of 115 specimens, and an additional subset of 20 specimens was assessed by all pathologists. All histopathologic assessments were performed blinded to the clinical and genetic information.

The histomorphologic criteria were intraepidermal scatter of melanocytes, nest formation, lateral circumscription, epidermal thickness, tumor cell pigmentation, solar elastosis (CSD), cell size and shape defined as previously published (Viros et al. 2008). In addition, WHO melanoma subtype and presence of an associated nevus were recorded according to standard criteria. Detailed descriptions and illustrations for all criteria were provided to all experts.

Mutation analysis

The EORTC cases had been previously characterized for multi-gene expression (Winnepenninckx et al. 2006) and BRAF mutations (Kannengiesser et al. 2008). For the other cases, tumor bearing areas were manually microdissected and DNA extracted for sequencing as described previously (Curtin et al. 2005). BRAF exon 15 was sequenced in all cases and NRAS was sequenced only in cases in which no BRAF exon 15 mutations were found, according to previous studies indicating that these mutations are mutually exclusive in the vast majority of cases (Omholt et al. 2003). Due to lack of amplifiable DNA or insufficient tissue, BRAF or NRAS status could not be obtained in 45 and 69 cases, respectively.

Interpretation of observer scores

Establishment of consensus scores

Consensus scores were only determined if at least two observers scored a given feature. For features with categorical scores (WHO melanoma subtype and associated nevus), the majority vote was used for analysis. In cases where two outcomes were tied, one outcome was chosen randomly. Categorical scores with complete discordance of observers were excluded.

CSD-melanomas were defined as primaries with a non-acral location with a score for solar elastosis of seven or above, and primaries on non-acral sites with lower solar elastosis scores were defined as non-CSD melanomas, as described previously (Balch et al. 2000). Acral location was defined as palms, soles, and nail beds and determined by the anatomic site provided and by histopathology.

Statistical analysis

Light's Kappa (Conger 1980) scores were computed for inter-rater agreements for ordinal and categorical variables, using squared weighted Kappa for ordinal variables. Confidence intervals were obtained from adjusted bootstrap percentile (BCa) intervals from 1000 ordinary bootstrap replicates (Davison & Hinkley 1997).

For correlation of mutation status and clinical and morphologic features, only cases in which the mutation status (wild type and mutant) as well as information on clinical characteristics could be determined were included. Association of binary mutation status outcomes with phenotypic variables were tested by fitting univariate logistic regression models. Higher order polynomials were fitted for ordinal variables not linearly associated with mutation status. For continuous variable thickness, one-tenth of the smallest non-zero value was added to the values and then log10 transformed to render its distribution more normal. The unordered categorical variables (anatomic site, WHO melanoma subtype, Acral/CSD/non-CSD grouping, gender, sentinel lymph node status, recurrence and ulceration) were treated as factor variables.

Prediction of BRAF mutation status

Prediction of mutation status was made using single binary classification tree classifier (Breiman et al. 1984). Binary classification trees were generated using the rpart package of the statistical software R. Kappa inter-rater agreement scores of the morphologic variables were used as weights in the model. P-value and accuracy were analyzed by Fisher's Exact Test using mode/median values (consensus vote) of the variables on all features.

Clinical data in both cohorts were compared (gender, age at diagnosis, clinical outcome) and correlations with age at diagnosis, time to relapse, time to death, tumor thickness, and mutation status were done by calculating median values and mode. Survival analysis was performed using Kaplan-Meier log rank tests.

Supplementary Material

Supplement

Acknowledgments

Funding: Supported by grants from the Melanoma Research Foundation and the National Cancer Institute R01 CA131524-01A1 to BCB. IO acknowledges support by Intendis Austria and the “Österreichische Nationalbank” grant number 10848 to IO, JvdO and AS from the Melanoma group of the European Organization for Research and Treatment on Cancer, and RAS from the Cancer Institute New South Wales and National Health and Medical Research Council, respectively. The funding agencies did not have a role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Abbreviations

ALM
acral lentiginous melanoma
CSD
chronic sun-induced damage
LMM
lentigo maligna melanoma
NC
not classifiable
NM
nodular melanoma
OR
odds ratio
RGP
radial growth phase
SSM
superficial spreading melanoma
VGP
vertical growth phase
WHO
World Health Organization

Footnotes

Competing Interests: The authors have declared that no competing interests exist.

1Glabrous skin is skin that is hairless as found on volar surfaces of fingers and toes, palms, soles, lips, labia minora, penis, lower anal canal and external anus.

References

1. Balch CM, Buzaid AC, Atkins MB, Cascinelli N, Coit DG, Fleming ID, Houghton A, Jr., Kirkwood JM, Mihm MF, Morton DL, et al. A new American Joint Committee on Cancer staging system for cutaneous melanoma. Cancer. 2000;88(6):1484–1491. [PubMed]
2. Bastian BC, Kashani-Sabet M, Hamm H, Godfrey T, Moore DH, 2nd, Bröcker EB, LeBoit PE, Pinkel D. Gene amplifications characterize acral melanoma and permit the detection of occult tumor cells in the surrounding skin. Cancer Res. 2000;60(7):1968–1973. [PubMed]
3. Breiman L, Friedman J, Stone CJ, Olshen RA. Classification and Regression Trees. 1st ed. Chapman and Hall/CRC; 1984.
4. Clark WH, From L, Bernardino EA, Mihm MC. The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res. 1969;29(3):705–727. [PubMed]
5. Conger A. Integration and generalisation of Kappas on multiple raters. 1980;88(2):322–328.
6. Curtin JA, Busam K, Pinkel D, Bastian BC. Somatic activation of KIT in distinct subtypes of melanoma. J. Clin. Oncol. 2006;24(26):4340–4346. [PubMed]
7. Curtin JA, Fridlyand J, Kageshita T, Patel HN, Busam KJ, Kutzner H, Cho KH, Aiba S, Broecker EB, LeBoit EB, et al. Distinct sets of genetic alterations in melanoma. N. Engl. J. Med. 2005;353(20):2135–2147. [PubMed]
8. Davison AC, Hinkley DV. Bootstrap Methods and Their Application. 1st ed. Cambridge University Press; 1997.
9. Emery CM, Vijayendran KG, Zipser MC, Sawyer AM, Niu L, Kim JJ, Hatton C, Chopra R, Oberholzer PA, Karpova MB, et al. MEK1 mutations confer resistance to MEK and B-RAF inhibition. Proc. Natl. Acad. Sci. USA. 2009;106(48):20411–20416. [PubMed]
10. Kannengiesser C, Spatz A, Michiels S, Eychene A, Dessen P, Lazar V, Winnepenninckx V, Lesueur F, Druillennec S, Robert C, et al. Gene expression signature associated with BRAF mutations in human primary cutaneous melanomas. Mol. Oncol. 2008;1(4):425–430. [PubMed]
11. Maldonado JL, Fridlyand J, Patel H, Jain AN, Busam K, Kageshita T, Ono T, Albertson DG, Pinkel D, Bastian BC. Determinants of BRAF mutations in primary melanomas. J. Natl. Cancer Inst. 2003;95(24):1878–1890. [PubMed]
12. Mehta CR, Patel NR. ALGORITHM 643: FEXACT: a FORTRAN subroutine for Fisher's exact test on unordered r×c contingency tables. ACM Trans. Math. Softw. 1986;12(2):154–161.
13. Omholt K, Platz A, Kanter L, Ringborg U, Hansson J. NRAS and BRAF mutations arise early during melanoma pathogenesis and are preserved throughout tumor progression. Clin. Cancer Res. 2003;9(17):6483–6488. [PubMed]
14. Reed R. New Concepts in Surgical Pathology of the Skin. John Wiley & Sons, Inc.; New York: 1976. Acral lentiginous melanoma; pp. 89–90.
15. Thomas NE, Kanetsky PA, Begg CB, Conway K, Berwick M. Melanoma molecular subtypes: unifying and paradoxical results. J. Invest. Dermatol. 2010;130(1):12–14. [PMC free article] [PubMed]
16. Thompson JF, Scolyer RA, Kefford RF. Cutaneous melanoma in the era of molecular profiling. Lancet. 2009;374(9687):362–365. [PubMed]
17. Van Raamsdonk CD, Bezrookove V, Green G, Bauer J, Gaugler L, O'Brien JM, Simpson EM, Barsh GS, Bastian BC. Frequent somatic mutations of GNAQ in uveal melanoma and blue naevi. Nature. 2009;457(7229):599–602. [PMC free article] [PubMed]
18. Viros A, Fridlyand J, Bauer J, Lasithiotakis K, Garbe C, Pinkel D, Bastian BC. Improving melanoma classification by integrating genetic and morphologic features. PLoS Med. 2008;5(6):e120. [PubMed]
19. Winnepenninckx V, Lazar V, Michiels S, Dessen P, Stas M, Alonso SR, Avril MF, Ortiz Romero PL, Robert T, Balacescu O, et al. Gene expression profiling of primary cutaneous melanoma and clinical outcome. J. Natl. Cancer Inst. 2006;98(7):472–482. [PubMed]