Determination of prognosis in metastatic melanoma through integration of clinico‐pathologic, mutation, mRNA, microRNA, and protein information

Abstract

In patients with metastatic melanoma, the identification and validation of accurate prognostic biomarkers will assist rational treatment planning. Studies based on "-omics" technologies have focussed on a single high-throughput data type such as gene or microRNA transcripts. Occasionally, these features have been evaluated in conjunction with limited clinico-pathologic data. With the increased availability of multiple data types, there is a pressing need to tease apart which of these sources contain the most valuable prognostic information. We evaluated and integrated several data types derived from the same tumor specimens in AJCC stage III melanoma patients-gene, protein, and microRNA expression as well as clinical, pathologic and mutation information-to determine their relative impact on prognosis. We used classification frameworks based on prevalidation and bootstrap multiple imputation to compare the prognostic power of each data source, both individually as well as integratively. We found that the prognostic utility of clinico-pathologic information was not out-performed by any of the various "-omics" platforms. Rather, a combination of clinico-pathologic variables and mRNA expression data performed best. Furthermore, a patient-based classification analysis revealed that the prognostic accuracy of various data types was not the same for different patients. This indicates that ongoing development in the individualized evaluation of melanoma patients must take account of the value of both traditional and novel "-omics" measurements.

microRNA-7-5p inhibits melanoma cell proliferation and metastasis by suppressing RelA/NF-κB

“…Data on survival and mRNA expression for 463 melanoma patients was analyzed by Cox regression, and after adjustment for age and sex, high expression of RelA (hazard ratio (HR) = 1.55, p -value = 0.007; Figure 4B), POLE4 (HR = 1.45, p -value = 0.023) and SP1 (HR = 1.40, p -value = 0.032) was predictive of poorer overall survival (Table 1). To validate this finding, we performed meta-analyses of these genes across three additional melanoma cohorts [35–37], in addition to the TCGA study, to increase the sample size and statistical power. Comparisons across these four datasets showed high expression of either RelA (consensus HR = 1.47), POLE4 (consensus HR = 1.39) or SP1 (consensus HR = 1.48) was consistently associated with an increased risk of death in melanoma (Figure 4A and 4B).…”

“…Comparisons across these four datasets showed high expression of either RelA (consensus HR = 1.47), POLE4 (consensus HR = 1.39) or SP1 (consensus HR = 1.48) was consistently associated with an increased risk of death in melanoma (Figure 4A and 4B). Furthermore, low expression of miR-7-1, which is the predominant isoform that gives rise to mature miR-7-5p, was associated with an increased risk of death in melanoma patients, with HR = 0.558, p = 0.05 (Figure 4C) [37]. Additionally, TCGA melanoma data was analyzed for an inverse correlation between expression of miR-7 isoforms and putative target gene expression, and revealed a negative association between the miR-7-1 isoform and RelA with R = −0.12, p = 0.013, but no significant negative correlations with the other genes that were also associated with survival (Figure 4D and Supplementary Table S6).…”

“…The Kabbarah cohort consisted of 81 melanoma patients, 54 men and 27 women, with 50 primary and 31 metastatic melanomas, and a mean age of 58 years (GEO accession number GSE46517 [36]). The Jayawardana cohort consisted of 74 melanoma patients, with 29 stage I, 29 stage II, and 20 stage III tumors, and 1 melanoma of unknown staging (GEO accession numbers GSE54467 and GSE59334 [37]).…”

A multi-step classifier addressing cohort heterogeneity improves performance of prognostic biomarkers in three cancer types

“…We previously observed that there were subsets of AJCC stage III metastatic melanoma patients whose prognosis, death within a year of resection or survival for more than four years after resection, was more easily predicted than others for a variety of biomarkers [4]. This insight is demonstrated in Figure 1 which shows the classification performance at a patient level for a biomarker constructed with either gene expression data or clinico-pathologic and mutation variables (“clinical” variables) in the melanoma cohort [4].…”

“…The inherent complexity of most human cancer cohorts and practical limits of identifying enough patients of a similar demographic and pathologic stage create a reality whereby partitioning patients is still not guaranteed to produce a sufficiently large homogeneous study cohort. Demonstrating this, we previously observed that there were subsets of AJCC stage III metastatic melanoma patients whose clinical outcome could be more easily explained by different clinical pathological or molecular biomarkers than others [4], an observation subsequently confirmed in breast cancer [5]. One interpretation of these results is that even after prospectively restricting study inclusion criteria to a specific pathologic stage – e.g., metastatic lymph node samples from patients with AJCC stage III disease in the melanoma study – there is a further subset of patients for whom the different clinical, pathological or molecular measurements point to alternative, and even competing, predicted outcomes for a given individual patient.…”