Meeting News

Molecular testing requires additional validation before ready for prime time in melanoma

NEW YORK — Preliminary and validation studies have been conducted on molecular testing methods for melanoma prognostication, but these methods have not yet been prospectively validated or incorporated into clinical treatment guidelines, according to a presenter at HemOnc Today Melanoma and Cutaneous Malignancies.

Molecular testing methods for the prognostication of melanoma that have undergone the most study include whole-genome RNA microarray, whole-genome microRNA (miRNA) profiling and 31 gene expression profiling, Eddy C. Hsueh, MD, surgical oncologist at Saint Louis University, said during his presentation.

In terms of whole-genome RNA profiling, Jönsson and colleagues evaluated archival tumor tissues from 57 patients with stage IV melanoma. Researchers profiled 3,000 of the most variable genes to identify four types of gene expression based on a signature of 503 genes.

In the validation set, patients with pigmentation and proliferative signatures had poorer prognosis compared with patients with high-immune response or normal-like signature (P = .04).

These data were independently validated in 214 melanoma tumor tissues. Results of this multivariate analysis showed distant metastatic–free survival was poorer in patients with the pigmentation signature (HR = 1.8; 95% CI, 1-3.7). Patients with the proliferative signature and poorer distant metastatic–free survival (HR = 2.7; 95% CI, 1.37-5.36) and disease-specific survival (HR = 2.8; 95% CI, 1.19-6.65).

The gene signature classification has been further refined into low grade and high grade, which appeared prognostic for OS in an analysis of 223 melanoma tissues (P = 2.48 x 10-11) and in a secondary independent validation of 300 archived tissues (P = .0001). A prospective trial on this is ongoing, Hsueh said.

MicroRNA markers that have been studied include miRNA-21, miRNA-137, miRNA-203, miRNA-205 and miRNA-150.

High miRNA-21 expression correlated with Breslow thickness and advanced clinical stage, as well as with worse 5-year DFS and OS. Conversely, low miRNA-137 expression correlated with TNM stage, ulceration and primary site, as well as with poor OS. Low miRNA-203 also was associated with Breslow thickness and poor survival.

miRNA-205 showed an association with survival in discovery (P = .0289) and validation (P = .0386) cohorts.

“MicroRNA-150 is one of the most studied microRNA agents of prognosis for melanoma, and is usually included in a panel of different microRNA signatures,” Hsueh said.

Down-regulation of miRNA-150-5p and miRNA142-3p/142-5p duplex correlated with poor survival.

DecisionDx-Melanoma (Castle Biosciences) gene expression profiling assay is a 31-gene panel that contains 28 discriminating genes and three control genes. The read-out classifies patients as class I, or low risk, and class II, or high risk. Study results show these classes correlated with survival in a training set of 164 patients with stage 0 to IV disease and in a validation set of 104 patients with stage I to IV melanoma (P < .0001 for both).

Hsueh and colleagues also conducted an independent single-institution prospective validation of DecisionDx-Melanoma in 174 patients with cutaneous melanoma undergoing sentinel lymph node biopsy, which was positive in 20 patients (13%).

After a median follow-up of 18 months, patients with class I disease (n = 117) had better OS than patients with class II disease (n = 42; P < .0001). Researchers conducted a DFS update analysis after a median follow-up of 27 months and found gene expression profiling was significant on multivariable analysis, and the association with survival persisted (P < .0001).

“Is molecular testing for melanoma ready for prime time?” Hsueh said. “It depends on your definition of prime time. Whole-genome miRNA profiling, whole-genome RNA microarray and gene expression profiling all have preliminary and validation studies. Prospective validation is lacking or ongoing, and incorporation into clinical treatment guidelines is lacking as well.” – by Alexandra Todak

Reference:

Hsueh EC. Molecular testing in melanoma: Ready for prime time? Presented at: HemOnc Today Melanoma and Cutaneous Malignancies; March 24-25, 2017; New York.

Disclosure: Hsueh reports speakers bureau roles with Amgen and Castle Biosciences.

NEW YORK — Preliminary and validation studies have been conducted on molecular testing methods for melanoma prognostication, but these methods have not yet been prospectively validated or incorporated into clinical treatment guidelines, according to a presenter at HemOnc Today Melanoma and Cutaneous Malignancies.

Molecular testing methods for the prognostication of melanoma that have undergone the most study include whole-genome RNA microarray, whole-genome microRNA (miRNA) profiling and 31 gene expression profiling, Eddy C. Hsueh, MD, surgical oncologist at Saint Louis University, said during his presentation.

In terms of whole-genome RNA profiling, Jönsson and colleagues evaluated archival tumor tissues from 57 patients with stage IV melanoma. Researchers profiled 3,000 of the most variable genes to identify four types of gene expression based on a signature of 503 genes.

In the validation set, patients with pigmentation and proliferative signatures had poorer prognosis compared with patients with high-immune response or normal-like signature (P = .04).

These data were independently validated in 214 melanoma tumor tissues. Results of this multivariate analysis showed distant metastatic–free survival was poorer in patients with the pigmentation signature (HR = 1.8; 95% CI, 1-3.7). Patients with the proliferative signature and poorer distant metastatic–free survival (HR = 2.7; 95% CI, 1.37-5.36) and disease-specific survival (HR = 2.8; 95% CI, 1.19-6.65).

The gene signature classification has been further refined into low grade and high grade, which appeared prognostic for OS in an analysis of 223 melanoma tissues (P = 2.48 x 10-11) and in a secondary independent validation of 300 archived tissues (P = .0001). A prospective trial on this is ongoing, Hsueh said.

MicroRNA markers that have been studied include miRNA-21, miRNA-137, miRNA-203, miRNA-205 and miRNA-150.

High miRNA-21 expression correlated with Breslow thickness and advanced clinical stage, as well as with worse 5-year DFS and OS. Conversely, low miRNA-137 expression correlated with TNM stage, ulceration and primary site, as well as with poor OS. Low miRNA-203 also was associated with Breslow thickness and poor survival.

miRNA-205 showed an association with survival in discovery (P = .0289) and validation (P = .0386) cohorts.

“MicroRNA-150 is one of the most studied microRNA agents of prognosis for melanoma, and is usually included in a panel of different microRNA signatures,” Hsueh said.

Down-regulation of miRNA-150-5p and miRNA142-3p/142-5p duplex correlated with poor survival.

DecisionDx-Melanoma (Castle Biosciences) gene expression profiling assay is a 31-gene panel that contains 28 discriminating genes and three control genes. The read-out classifies patients as class I, or low risk, and class II, or high risk. Study results show these classes correlated with survival in a training set of 164 patients with stage 0 to IV disease and in a validation set of 104 patients with stage I to IV melanoma (P < .0001 for both).

Hsueh and colleagues also conducted an independent single-institution prospective validation of DecisionDx-Melanoma in 174 patients with cutaneous melanoma undergoing sentinel lymph node biopsy, which was positive in 20 patients (13%).

After a median follow-up of 18 months, patients with class I disease (n = 117) had better OS than patients with class II disease (n = 42; P < .0001). Researchers conducted a DFS update analysis after a median follow-up of 27 months and found gene expression profiling was significant on multivariable analysis, and the association with survival persisted (P < .0001).

“Is molecular testing for melanoma ready for prime time?” Hsueh said. “It depends on your definition of prime time. Whole-genome miRNA profiling, whole-genome RNA microarray and gene expression profiling all have preliminary and validation studies. Prospective validation is lacking or ongoing, and incorporation into clinical treatment guidelines is lacking as well.” – by Alexandra Todak

Reference:

Hsueh EC. Molecular testing in melanoma: Ready for prime time? Presented at: HemOnc Today Melanoma and Cutaneous Malignancies; March 24-25, 2017; New York.

Disclosure: Hsueh reports speakers bureau roles with Amgen and Castle Biosciences.

    See more from HemOnc Today New York