Faculty Bibliography

BACKGROUND AND OBJECTIVES/OBJECTIVE:Metastasectomy for melanoma provides durable disease control in carefully selected patients. Similarly, BRAF-targeted and immune checkpoint inhibition has improved median overall survival (OS) in metastatic patients. We hypothesized that there is an increasing role for metastasectomy in melanoma patients responding to these therapies. METHODS:Retrospective analysis of a prospectively maintained database identified 128 patients with stage IV melanoma who received targeted molecular and/or checkpoint inhibitors at an academic institution from 2006 to 2017. Records were reviewed to characterize clinicopathologic characteristics, response to treatment, and intent of surgery for those who underwent metastasectomy. OS was analyzed by the Kaplan-Meier method. RESULTS:Median OS from stage IV diagnosis was 31.3 months. A total of 81 patients received checkpoint inhibitors, 11 received targeted inhibitors, and 36 received both. A total of 73 patients underwent metastasectomy. Indications for surgery included the intent to render disease-free (54%), palliation (34%), and diagnostic confirmation (11%). Responders to systemic therapy who underwent metastasectomy had improved OS compared to responders who did not (84.3 vs. 42.9 months, P = .018). CONCLUSIONS:Metastasectomy for melanoma is associated with improved OS in patients that respond to targeted molecular or immunotherapy. Resection should be strongly considered in this cohort as multimodality treatment results in excellent OS.

Dermal invasion is a hallmark of malignant melanoma. Thought the molecular alterations driving the progression of primary melanoma to metastatic disease have been studied extensively, the early progression of non-invasive primary melanoma to an invasive state is poorly understood. To elucidate the mechanisms underlying the transition from radial to vertical growth, the first step in melanoma invasion, we developed a zebrafish melanoma model in which constitutive activation of ribosomal protein S6 kinase A1 (Rps6ka1; RSK1) drives tumor invasion. Transcriptomic analysis of ribosomal protein S6 kinase A1-activated tumors identified metabolic changes, including up-regulation of genes associated with oxidative phosphorylation. Vertical growth phase human melanoma cells show higher oxygen consumption and preferential utilization of glutamine compared to radial growth phase melanoma cells. Peroxisome proliferator activated receptor gamma (PPARG) coactivator 1 alpha (PPARGC1A), also known as PPARG coactivator-1α (PGC1α), has been proposed as a master regulator of tumor oxidative phosphorylation. In human primary melanoma specimens, PGC1α protein expression was found to be positively associated with increased tumor thickness and expression of the proliferative marker Ki-67 and the reactive oxygen species scavenger scavenger receptor class A member 3 . PGC1α depletion modulated cellular processes associated with primary melanoma growth and invasion, including oxidative stress. These results support a role for PGC1α in mediating glutamine-driven oxidative phosphorylation to facilitate the invasive growth of primary melanoma.

Mutational heterogeneity can contribute to therapeutic resistance in solid cancers. In melanoma, the frequency of inter- and intra-tumoral heterogeneity is controversial. We examined mutational heterogeneity within individual melanoma patients using multi-platform analysis of commonly mutated driver and non-passenger genes. We analyzed paired primary and metastatic tumors from 60 patients, and multiple metastatic tumors from 39 patients whose primary tumors were unavailable (n=271 tumors). We used a combination of multiplex SNaPshot assays, Sanger Sequencing, Mutation-specific PCR, or droplet digital PCR to determine the presence of BRAF^V600, NRAS^Q61, and TERT^-124C>T and TERT^-146C>T mutations. Mutations were detected in BRAF (39%), NRAS (21%) and/or TERT (78%). Thirteen patients had TERT^mutant discordant tumors; seven of these had a single tumor with both TERT^-124C>T and TERT^-146C>T mutations present at different allele frequencies. Two patients had both BRAF and NRAS mutations; one in different tumors and the other had a single tumor with both mutations. One patient with a BRAF^mutant primary lacked mutant BRAF in least one of their metastases. Overall, we identified mutational heterogeneity in 18/99 (18%) patients. These results suggest that some primary melanomas may be comprised of subclones with differing mutational profiles. Such heterogeneity may be relevant to treatment responses and survival outcomes.

BACKGROUND:The American Joint Committee on Cancer (AJCC) maintains that the eighth edition of its Staging Manual (AJCC8) has improved accuracy compared to the seventh (AJCC7). However, there are concerns that implementation may disrupt analysis of active clinical trials for stage III patients. We used an independent cohort of melanoma patients to test the extent to which AJCC8 has improved prognostic accuracy compared to AJCC7. METHODS:We analyzed a cohort of 1,315 prospectively enrolled patients. We assessed primary tumor and nodal classification of stage I-III patients using AJCC7 and AJCC8 to assign disease stages at diagnosis. We compared recurrence-free (RFS) and overall survival (OS) using Kaplan-Meier curves and log-rank tests. We then compared concordance indices of discriminatory prognostic ability and area under the curve (AUC) of 5-year survival to predict RFS/OS. All statistical tests were two-sided. RESULTS:Stage IIC continued to have worse outcomes than those for stage IIIA patients, with 5-year RFS of 26.5% (95%CI=12.8-55.1%) vs. 56.0% (95%CI=37.0-84.7%) by AJCC8 (P = 0.002). For stage I, removing mitotic index as T classification factor decreased its prognostic value, although not statistically significantly (RFS C-index=0.63 [95%CI=0.56-0.69] to 0.56 [95%CI=0.49-0.63], P = 0.07; OS C-index=0.48 [95%CI=0.38-0.58] to 0.48 [95%CI=0.41-0.56], P = 0.90). For stage II, prognostication remained constant (RFS C-index=0.65 [95%CI=0.57-0.72]; OS C-index=0.61 [95%CI=0.50-0.72]), and. For stage III, AJCC8 yielded statistically significantly enhanced prognostication for RFS (C-index=0.65 [95%CI=0.60-0.70] to 0.70 [95%CI=0.66-0.75], P = 0.01). CONCLUSIONS:Compared with AJCC7, we demonstrate that AJCC8 enables more accurate prognosis for patients with stage III melanoma. Restaging a large cohort of patients can enhance the analysis of active clinical trials.

Metastasis is the primary cause of death of cancer patients. Dissecting mechanisms governing metastatic spread may uncover important tumor biology and/or yield promising therapeutic insights. Here, we investigated the role of circular RNAs (circRNA) in metastasis, using melanoma as a model aggressive tumor. We identified silencing of cerebellar degeneration-related 1 antisense (CDR1as), a regulator of miR-7, as a hallmark of melanoma progression. CDR1as depletion results from epigenetic silencing of LINC00632, its originating long non-coding RNA (lncRNA) and promotes invasion in vitro and metastasis in vivo through a miR-7-independent, IGF2BP3-mediated mechanism. Moreover, CDR1as levels reflect cellular states associated with distinct therapeutic responses. Our study reveals functional, prognostic, and predictive roles for CDR1as and expose circRNAs as key players in metastasis.

PURPOSE/OBJECTIVE:Biomarkers for disease specific survival (DSS) in early stage melanoma are needed to select patients for adjuvant immunotherapy and accelerate clinical trial design. We present a pathology-based computational method using a deep neural network architecture for DSS prediction. EXPERIMENTAL DESIGN/METHODS:The model was trained on 108 patients from four institutions and tested on 104 patients from Yale School of Medicine (YSM). A receiver operating characteristic (ROC) curve was generated based on vote aggregation of individual image sequences, an optimized cutoff was selected, and the computational model was tested on a third independent population of 51 patients from Geisinger Health Systems (GHS). RESULTS:Area under the curve (AUC) in the YSM patients was 0.905 (p<0.0001). AUC in the GHS patients was 0.880 (p<0.0001). Using the cutoff selected in the YSM cohort, the computational model predicted DSS in the GHS cohort based on Kaplan-Meier (KM) analysis (p<0.0001). CONCLUSIONS:The novel method presented is applicable to digital images, obviating the need for sample shipment and manipulation and representing a practical advance over current genetic and IHC-based methods.

Next-generation sequencing has enabled genetic and genomic characterization of melanoma to an unprecedent depth. However, the high mutational background plus the limited deep-coverage whole-genome sequencing performed on cutaneous melanoma samples, make difficult the identification of novel driver mutations. We sought to explore the somatic mutation portfolio in exonic and gene regulatory regions in human melanoma samples, for which we performed targeted sequencing of tumors and matched germline DNA samples from 89 melanoma patients, identifying known and novel recurrent mutations. Two recurrent mutations found in the RPS27 promoter associated with decreased RPS27 mRNA levels in vitro. Data mining and IHC analyses revealed a bimodal pattern of RPS27 expression in melanoma, with RPS27-low patients displaying worse prognosis. In vitro characterization of RPS27-high and -low melanoma cell lines, as well as loss-of-function experiments, demonstrated that high RPS27 status provides increased proliferative and invasive capacities, while low RPS27 confers survival advantage in low-attachment and resistance to therapy. Additionally, we demonstrate that 10 other cancer types harbor bimodal RPS27 expression and in those, similarly to melanoma, RPS27-low expression associates with worse clinical outcomes. RPS27 promoter mutation could thus represent a mechanism of gene expression modulation in melanoma patients, which may have prognostic and predictive implications.

Background: Tumor-specific indicators, such as tumor mutation burden (TMB) have been shown to affect overall survival (OS) in melanoma. Recently, pan-cancer analyses from The Cancer Genome Atlas (TCGA) discovered specific tumor immune signatures predictive of overall survival (OS), yet it is unclear how these interact with other prognostic markers, independently of immunotherapy (IT). We aimed to combine the immune landscape signatures with TMB and other prognostic markers to improve melanoma OS prediction in patients, independent of IT.
Method(s): We examined the whole-exome data in conjunction with the molecular, clinical and immune features from 278 metastatic melanomas from TCGA, not treated by IT, to develop an improved prognostic model of melanoma OS. Using the discovery (N=139) and validation (N=139) design we performed multivariate Cox proportional hazards (Cox HR) models, adjusted for age and tumor stage at primary diagnosis, to identify interaction between TMB and melanoma immune features (n=59), refining the prediction of melanoma OS.
Result(s): We identified 4 immune features that were significantly associated with OS in both the discovery and validation cohorts. The multivariate Cox HR models revealed that IFN-c response (IFN-c R) and macrophage regulation (MR) signatures in combination with TMB were the most significantly associated with OS (p=8.80E-14). After further refinement, we observed that patients with high TMB, high IFN-c R and high MRhad significantly better OS compared to high TMB, low IFN-y R and low MR (HR=2.8, p=3.55E-08). This association was not observed in low TMB patients.
Conclusion(s): We show, for the first time, that TMB and tumor immune features are significantly associated with improved OS, independent of IT. Further analysis of patients revealed that high TMB associates with improved OS in patients with high IFN- c R and MR but not in low IFN- c R and MR. Hence, this data provides first evidence that patients with high TMB have distinct OS outcome depending on other tumor immune features. Beside biological link between TMB and IFN-y and MR, our data suggest that these associations may significantly improve the current melanoma prognostic models

Background Autoantibody landscapes are very specific to the individual, can remain stable for many years, and contain unique features reported in association with cancer, autoimmunity, infection, neurologic conditions, CD8+ T cell behavior, and checkpoint blockade adverse events [1-11]. The goal of this work was to determine whether pre-existing antigenspecific features in melanoma patient autoantibody landscapes would associate with clinical outcomes following checkpoint blockade. Methods Pre-treatment serum samples were collected from 117 melanoma patients prior to checkpoint blockade with anti-CTLA4 (N=60), anti-PD1 (N=38), or both in combination (N=16). All data was collected with approval of the NYU Institutional Review Board at the NYU Perlmutter Cancer Center with informed consent [11]. Serum samples were run on HuProt Human Proteome Microarrays containing >19,000 human proteins by CDI Laboratories. Raw serum IgG signal intensities were processed across staining cohorts via interquartile range normalization. Pre-existing antibody responses were defined as patient-specific IgG signals >3.5 median absolute deviations above cohort median IgG background (modified Z-score). Group statistics were computed (GraphPad Prism), and gene ontology enrichment analysis was performed (Enrichr) [12]. Results Several pre-existing antigen-specific IgG autoantibody targets were observed to have associations with good outcomes (SD/PR) or objective clinical responses (PR/CR) versus patients with progressive disease (POD). While final determination of the most predictive subsets is ongoing, many targets represent genes in an axis surrounding immune signaling pathways, hereditary neurodegenerative disease, and the ubiquitin proteasome pathway (ie, UBQLN1, UBQLN2). An exemplary example was observed in the autoantibody responses shared by >10% of all patients regardless of clinical outcome. Gene ontology enrichment analysis of these shared melanoma-patient autoantibodies versus KEGG 2019 [12] demonstrates this set of proteins is strongly enriched for neurotrophin signaling-associated proteins after multi-sample correction (P=0.004) (Table 1). Several other associations were observed cohort-wide for ontologies with tissuespecific enrichment in the brain, neurons, and neuronal processes. Conclusions In this pilot study, we found strong associations across the cohort for autoantibodies against nerve-growth-inducing neurotrophins and genes like UBQLN1 and UBQLN2 which have strong associations with amyotrophic lateral sclerosis, frontotemporal dementia, Parkinson's, and Alzheimer's - neurodegenerative diseases that are known to have incidences which correlate with melanoma [14-16]; this hints at a potential immunologic connection between the conditions, perhaps related to an antitumor / autoimmune axis involving the targets reported here. (Table Presented)