Melanoma Dataset Research Articles

Impact of senescence cell signature in patients with non-small cell carcinoma and melanoma receiving PD-L1/PD-1 inhibitors

Tumor cell senescence plays a crucial role in tumor immunity. We investigated whether the senescent cell signature (SCS) could predict prognosis in non-small cell carcinoma (NSCLC) and melanoma datasets treated with PD-L1/PD-1 inhibitors. Patients with high SCS expression exhibited elevated levels of interferon-gamma and T cell-inflamed signatures in three lung adenocarcinomas (LUAD), two squamous cell carcinoma (LUSC) and three melanoma datasets. The high SCS group was associated with PD-L1-related pathways such as IL6/JAK/STAT3 and TNF-alpha signaling via NF-kB in LUAD, LUSC, and melanoma datasets. A positive correlation was observed between several immune checkpoint markers and the SCS, indicating an immunosuppressive state in LUAD, LUSC and melanoma datasets. In patients treated with PD-1/PD-L1 inhibitors, a higher SCS was associated with a better prognosis, and a positive correlation between SCS and PD-L1 was observed in six independent NSCLC and three independent melanoma datasets. We used the LASSO Cox regression model to build a risk model focusing on the SCS genes that particularly predict prognosis. We confirmed that the model accurately predicts prognosis. However, the senescent immunohistochemical markers p16 and p21 could predict the response to PD-1/PD-L1 inhibitors in patients with LUSC and melanoma but not in patients with LUAD. SCS could serve as a valuable biomarker to complement PD-L1 expression in patients receiving PD-L1/PD-1 inhibitors.

SIMVI reveals intrinsic and spatial-induced states in spatial omics data.

Spatial omics technologies enable the analysis of gene expression and interaction dynamics in relation to tissue structure and function. However, existing computational methods may not properly distinguish cellular intrinsic variability and intercellular interactions, and may thus fail to capture spatial regulations for further biological discoveries. Here, we present Spatial Interaction Modeling using Variational Inference (SIMVI), an annotation-free framework that disentangles cell intrinsic and spatial-induced latent variables for modeling gene expression in spatial omics data. We derive theoretical support for SIMVI in disentangling intrinsic and spatial-induced variations. By this disentanglement, SIMVI enables estimation of spatial effects (SE) at a single-cell resolution, and opens up various opportunities for novel downstream analyses. To demonstrate the potential of SIMVI, we applied SIMVI to spatial omics data from diverse platforms and tissues (MERFISH human cortex, Slide-seqv2 mouse hippocampus, Slide-tags human tonsil, spatial multiome human melanoma, cohort-level CosMx melanoma). In all tested datasets, SIMVI effectively disentangles variations and infers accurate spatial effects compared with alternative methods. Moreover, on these datasets, SIMVI uniquely uncovers complex spatial regulations and dynamics of biological significance. In the human tonsil data, SIMVI illuminates the cyclical spatial dynamics of germinal center B cells during maturation. Applying SIMVI to both RNA and ATAC modalities of the multiome melanoma data reveals potential tumor epigenetic reprogramming states. Application of SIMVI on our newly-collected cohort-level CosMx melanoma dataset uncovers space-and-outcome-dependent macrophage states and the underlying cellular communication machinery in the tumor microenvironments.

Integrated multiomics revealed adenosine signaling predict immunotherapy response and regulate tumor ecosystem of melanoma

Extracellular adenosine is extensively involved in regulating the tumor microenvironment. Given the disappointing results of adenosine-targeted therapy trials, personalized treatment might be necessary, tailored to the microenvironment status of individual patients. Here, we introduce the adenosine signaling score (ADO-score) model using non-negative matrix fraction identified patient subtypes using publicly available melanoma dataset, which aimed to profile adenosine signaling-related genes and construct a model to predict prognosis. We analyzed 580 malignant melanoma samples and demonstrated its robust value for prognosis. Further investigation in immune checkpoint inhibitor dataset suggests its potential as a stratified factor of immune checkpoint inhibitor efficacy. We validated the power of the ADO-score at the protein level immunofluorescence in a melanoma cohort from Xiangya Hospital. More importantly, single-cell and spatial transcriptomic data highlighted the cell-specific expression patterns of adenosine signaling-related genes and the existence of adenosine signaling-mediated crosstalk between tumor cells and immune cells in melanoma. Our study reveals a robust connection between adenosine signaling and clinical benefits in melanoma patients and proposes a universally applicable adenosine signaling model, the ADO-score, in gene expression profiles and histological sections. This model enables us to more precisely and conveniently select patients who are likely to benefit from immunotherapy.

Leveraging single-cell sequencing analysis and bulk-RNA sequencing analysis to forecast necroptosis in cutaneous melanoma prognosis.

Cutaneous melanoma, a malignancy of melanocytes, presents a significant challenge due to its aggressive nature and rising global incidence. Despite advancements in treatment, the variability in patient responses underscores the need for further research into novel therapeutic targets, including the role of programmed cell death pathways such as necroptosis. The melanoma datasets used for analysis, GSE215120, GSE19234, GSE22153 and GSE65904, were downloaded from the GEO database. The melanoma data from TCGA were downloaded from the UCSC website. Using single-cell sequencing, we assess the heterogeneity of necroptosis in cutaneous melanoma, identifying distinct cell clusters and necroptosis-related gene expression patterns. A combination of 101 machine learning algorithms was employed to construct a necroptosis-related signature (NRS) based on key genes associated with necroptosis. The prognostic value of NRS was evaluated in four cohorts (one TCGA and three GEO cohorts), and the tumour microenvironment (TME) was analysed to understand the relationship between necroptosis, tumour mutation burden (TMB) and immune infiltration. Finally, we focused on the role of key target TSPAN10 in the prognosis, pathogenesis, immunotherapy relevance and drug sensitivity of cutaneous melanoma. Our study revealed significant heterogeneity in necroptosis among melanoma cells, with a higher prevalence in epithelial cells, myeloid cells and fibroblasts. The NRS, developed through rigorous machine learning techniques, demonstrated robust prognostic capabilities, distinguishing high-risk patients with poorer outcomes in all cohorts. Analysis of the TME showed that high NRS scores correlated with lower TMB and reduced immune cell infiltration, indicating a potential mechanism through which necroptosis influences melanoma progression. Finally, TSPAN10 has been identified as a key target for cutaneous melanoma and is highly associated with poor prognosis. The findings highlight the complex role of necroptosis in cutaneous melanoma and introduce the NRS as a novel prognostic tool with potential to guide therapeutic decisions.

Integrative deep learning with prior assisted feature selection.

Integrative analysis has emerged as a prominent tool in biomedical research, offering a solution to the "small and large " challenge. Leveraging the powerful capabilities of deep learning in extracting complex relationship between genes and diseases, our objective in this study is to incorporate deep learning into the framework of integrative analysis. Recognizing the redundancy within candidate features, we introduce a dedicated feature selection layer in the proposed integrative deep learning method. To further improve the performance of feature selection, the rich previous researches are utilized by an ensemble learning method to identify "prior information". This leads to the proposed prior assisted integrative deep learning (PANDA) method. We demonstrate the superiority of the PANDA method through a series of simulation studies, showing its clear advantages over competing approaches in both feature selection and outcome prediction. Finally, a skin cutaneous melanoma (SKCM) dataset is extensively analyzed by the PANDA method to show its practical application.

Comprehensive single-cell RNA-seq analysis using deep interpretable generative modeling guided by biological hierarchy knowledge.

Recent advances in microfluidics and sequencing technologies allow researchers to explore cellular heterogeneity at single-cell resolution. In recent years, deep learning frameworks, such as generative models, have brought great changes to the analysis of transcriptomic data. Nevertheless, relying on the potential space of these generative models alone is insufficient to generate biological explanations. In addition, most of the previous work based on generative models is limited to shallow neural networks with one to three layers of latent variables, which may limit the capabilities of the models. Here, we propose a deep interpretable generative model called d-scIGM for single-cell data analysis. d-scIGM combines sawtooth connectivity techniques and residual networks, thereby constructing a deep generative framework. In addition, d-scIGM incorporates hierarchical prior knowledge of biological domains to enhance the interpretability of the model. We show that d-scIGM achieves excellent performance in a variety of fundamental tasks, including clustering, visualization, and pseudo-temporal inference. Through topic pathway studies, we found that d-scIGM-learned topics are better enriched for biologically meaningful pathways compared to the baseline models. Furthermore, the analysis of drug response data shows that d-scIGM can capture drug response patterns in large-scale experiments, which provides a promising way to elucidate the underlying biological mechanisms. Lastly, in the melanoma dataset, d-scIGM accurately identified different cell types and revealed multiple melanin-related driver genes and key pathways, which are critical for understanding disease mechanisms and drug development.

Residual Analysis for Poisson-Exponentiated Weibull Regression Models with Cure Fraction

The use of cure-rate survival models has grown in recent years. Even so, proposals to perform the goodness of fit of these models have not been so frequent. However, residual analysis can be used to check the adequacy of a fitted regression model. In this context, we provide Cox–Snell residuals for Poisson-exponentiated Weibull regression with cure fraction. We developed several simulations under different scenarios for studying the distributions of these residuals. They were applied to a melanoma dataset for illustrative purposes.

Non-proportional hazards model with a PVF frailty term: application with a melanoma dataset

Survival data analysis often uses the Cox proportional hazards (PH) model. This model is widely applied due to its straightforward interpretation of the hazard ratio under the assumption that the hazard rates for two subjects remain constant over time. However, in several randomized clinical trials with long-term survival data comparing two new treatments, it is frequently observed that Kaplan-Meier plots exhibit crossing survival curves. This violation of the PH assumption of the Cox PH model can not be applied to evaluate the treatment's effect on survival. This paper introduces a novel long-term survival model with non-PH that incorporates a frailty term into the hazard function. This model allows us to examine the effect of prognostic factors on survival and quantify the degree of unobservable heterogeneity. The model parameters are estimated using the maximum likelihood estimation procedure, and we evaluate the performance of the proposed models through simulation studies. Additionally, we demonstrate the applicability of our approach by fitting the models to a real skin cancer dataset.

Strategies for improving the performance of prediction models for response to immune checkpoint blockade therapy in cancer

Immune checkpoint blockade (ICB) therapy holds promise for bringing long-lasting clinical gains for the treatment of cancer. However, studies show that only a fraction of patients respond to the treatment. In this regard, it is valuable to develop gene expression signatures based on RNA sequencing (RNAseq) data and machine learning methods to predict a patient’s response to the ICB therapy, which contributes to more personalized treatment strategy and better management of cancer patients. However, due to the limited sample size of ICB trials with RNAseq data available and the vast number of candidate gene expression features, it is challenging to develop well-performed gene expression signatures. In this study, we used several published melanoma datasets and investigated approaches that can improve the construction of gene expression-based prediction models. We found that merging datasets from multiple studies and incorporating prior biological knowledge yielded prediction models with higher predictive accuracies. Our finding suggests that these two strategies are of high value to identify ICB response biomarkers in future studies.

Abstract 1143: Dissection of melanoma and cutaneous lymphoma spatial molecular architecture by multimodal single-cell and spatial transcriptomics with generative AI

Abstract Advances in single-cell multiomic technologies have revolutionized our understanding of how heterogeneous cell types and cell states shape the tumor microenvironment (TME). However, dissociated single cell profiling lacks spatial context. To resolve the organization, cell-cell communication, and granular structures in the TME, new single-cell spatial transcriptomic (ST) characterizations are needed. Here, we dissected the spatial molecular architecture of the melanoma (30 tumors, 30 patients) and transformed cutaneous T-cell lymphoma (tCTCL, 12 tumors, 6 patients) TMEs by Multiplexed Error-Robust FISH (305 gene MERFISH), with benchmarking scRNAseq data from matched tissues (6 melanoma scFFPEseq, 6 tCTCL scVDJ/RNAseq) for cross-platform validation. We profiled 565,122 cells in melanoma and 92584 cells in tCTCL by MERFISH. We interrogated both datasets using a novel computational framework that includes cell typing by Leiden clustering and marker genes, spatial neighborhood and receptor-ligand (R-L) analyses, and spatial/scRNA imputation via deep learning. In the melanoma dataset, we identified tumor cells, major TME cell types, and rare immune cell types (TCF7+ stem-like T-cells and CD3+TCRα+PAX5+CD79a+ B/T dual-expressor lymphocytes enriched in tertiary lymphoid structures). All cell types were validated in scFFPEseq. As immune R-L interactions are critical therapeutic targets, we developed a novel computation method to quantify spatial R-L interactions by accounting for cell-cell distance, identifying spatially informed R-L pairs that differ from dissociation-based inferences. CNV inference from melanoma scFFPEseq revealed intratumoral heterogeneity (ITH). We therefore adapted the conditional variational autoencoder ENVI to simultaneously incorporate scRNA and MERFISH spatial data into a unified latent embedding. ENVI-ITH successfully learned ITH subgroup information, enabling imputation of phylogenetic lineages on spatial MERFISH. CTCL MERFISH data revealed major TME cell types, yet separation of malignant vs benign T-cells by clustering and marker approaches is hindered by their overlapping gene expression. To overcome this barrier, we deployed ENVI to impute patient-specific malignant and benign TCR clonotypes by training on the matched scVDJ/RNA data, yielding an ENVI-TCR pipeline that can delineate malignant vs benign T-cells in situ. We validated the spatial pattern of malignant T-cells by patient-specific in situ TCR probes. In sum, we have interrogated the melanoma and CTCL TME by MERFISH and presented a novel computational framework for robustly co-embedding scRNA data with imaging-based ST to spatially resolve melanoma ITH and malignant/benign T-cell populations. We believe this approach will be highly impactful for melanoma and CTCL and can be broadly applied to other solid tumors and lymphomas. Citation Format: Zichao Liu, Xiaofei Song, Jiang He, Justin He, Jodi Balasi, Jeffrey H. Chuang, Pei-Ling Chen. Dissection of melanoma and cutaneous lymphoma spatial molecular architecture by multimodal single-cell and spatial transcriptomics with generative AI [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1143.

Multi-omics analysis for ferroptosis-related genes as prognostic factors in cutaneous melanoma.

Melanoma is highly malignant and heterogeneous. It is essential to develop a specific prognostic model for improving the patients' survival and treatment strategies. Recent studies have shown that ferroptosis results from the overproduction of lipid peroxidation and is an iron-dependent form of programmed cell death. Despite this, ferroptosis-related genes (FRGs) and their clinical significances remain unknown in malignant melanoma. This study aims to assess the role of FRGs in melanoma, with the goal of developing a novel prognostic model that provides new insights into personalized treatment and improvement of therapeutic outcomes for melanoma. We systematically characterized the genetic alterations and mRNA expression of 73 FRGs in The Cancer Genome Atlas (TCGA)-skin cutaneous melanoma (SKCM) dataset in this study. The results were validated with real-time RT-PCR and Western blotting. Subsequently, a multi-gene feature model was constructed using the TCGA-SKCM cohort. Melanoma patients were classified into a high-risk group and a low-risk group based on the feature model. As a final step, correlations between ferroptosis-related signatures and immune features, immunotherapy efficacy, or drug response were analyzed. By analyzing melanoma samples from TCGA-SKCM dataset, FRGs exhibited a high frequency of genetic mutations and copy number variations (CNVs), significantly impacting gene expression. Additionally, compared with normal skin tissue, 30 genes with significantly differential expression were identified in melanoma tissues. A prognostic model related to FRGs, constructed using the LASSO Cox regression method, identified 13 FRGs associated with overall survival prognosis in patients and was validated with external datasets. Finally, functional enrichment and immune response analysis further indicated significant differences in immune cell infiltration, mutation burden, and hypoxia status between the high-risk group and the low-risk group, and the model was effective in predicting responses to immunotherapy and drug sensitivity. This study develops a strong ferroptosis-related prognostic signature model which could put forward new insights into target therapy and immunotherapy for patients with melanoma.

Machine Learning Methods for Gene Selection in Uveal Melanoma.

Uveal melanoma (UM) is the most common primary intraocular malignancy with a limited five-year survival for metastatic patients. Limited therapeutic treatments are currently available for metastatic disease, even if the genomics of this tumor has been deeply studied using next-generation sequencing (NGS) and functional experiments. The profound knowledge of the molecular features that characterize this tumor has not led to the development of efficacious therapies, and the survival of metastatic patients has not changed for decades. Several bioinformatics methods have been applied to mine NGS tumor data in order to unveil tumor biology and detect possible molecular targets for new therapies. Each application can be single domain based while others are more focused on data integration from multiple genomics domains (as gene expression and methylation data). Examples of single domain approaches include differentially expressed gene (DEG) analysis on gene expression data with statistical methods such as SAM (significance analysis of microarray) or gene prioritization with complex algorithms such as deep learning. Data fusion or integration methods merge multiple domains of information to define new clusters of patients or to detect relevant genes, according to multiple NGS data. In this work, we compare different strategies to detect relevant genes for metastatic disease prediction in the TCGA uveal melanoma (UVM) dataset. Detected targets are validated with multi-gene score analysis on a larger UM microarray dataset.

Defining Melanoma Immune Biomarkers-Desert, Excluded, and Inflamed Subtypes-Using a Gene Expression Classifier Reflecting Intratumoral Immune Response and Stromal Patterns.

The spatial distribution of tumor infiltrating lymphocytes (TILs) defines several histologically and clinically distinct immune subtypes-desert (no TILs), excluded (TILs in stroma), and inflamed (TILs in tumor parenchyma). To date, robust classification of immune subtypes still requires deeper experimental evidence across various cancer types. Here, we aimed to investigate, define, and validate the immune subtypes in melanoma by coupling transcriptional and histological assessments of the lymphocyte distribution in tumor parenchyma and stroma. We used the transcriptomic data from The Cancer Genome Atlas melanoma dataset to screen for the desert, excluded, and inflamed immune subtypes. We defined subtype-specific genes and used them to construct a subtype assignment algorithm. We validated the two-step algorithm in the qPCR data of real-world melanoma tumors with histologically defined immune subtypes. The accuracy of a classifier encompassing expression data of seven genes (immune response-related: CD2, CD53, IRF1, and CD8B; and stroma-related: COL5A2, TNFAIP6, and INHBA) in a validation cohort reached 79%. Our findings suggest that melanoma tumors can be classified into transcriptionally and histologically distinct desert, excluded, and inflamed subtypes. Gene expression-based algorithms can assist physicians and pathologists as biomarkers in the rapid assessment of a tumor immune microenvironment while serving as a tool for clinical decision making.

DEEP LEARNING-BASED CLASSIFICATION FOR MELANOMA DETECTION USING XCEPTIONNET AND DENSENET121

Melanoma, the deadliest form of skin cancer, poses a significant public health challenge worldwide. Early detection is crucial for improving patient outcomes, yet traditional methods often rely on subjective visual inspection by dermatologists, leading to variability and delay in diagnosis. This study presents a deep learningbased classification approach for melanoma detection, utilizing the XceptionNet and DenseNet121 convolutional neural network (CNN) architectures. Leveraging a large dataset of dermoscopic images, the proposed model is trained to automatically identify malignant melanomas from benign lesions with high accuracy. XceptionNet and DenseNet121 are employed as feature extractors, capturing intricate patterns and features from skin lesion images. Transfer learning techniques are utilized to fine-tune the pretrained models on the melanoma dataset, enhancing classification performance. Extensive experimentation and evaluation on benchmark datasets demonstrate the superior performance of the proposed approach compared to traditional methods and standalone CNN architectures. The deep learning-based classification model holds promise for aiding dermatologists in early melanoma detection, potentially reducing diagnostic variability and improving patient outcomes.

The Geometric Generalized Birnbaum–Saunders model with long-Term Survivors

Introduction: A cure rate survival model was developed based on the assumption that the number of competing reasons for the event of interest has the Geometric distribution and the time allocated to the event of interest follows the Generalized Birnbaum-Saunders distribution.
 Methods: The Geometric Generalized Birnbaum–Saunders distribution was defined and two useful representations were represented for its density function which contributes to the creation of some mathematical properties. Furthermore, the parameters of the model with cure rate were estimated by using the maximum likelihood method.
 Results: Several simulations were performed and a real data set was analyzed from the medical area for different sample sizes and censoring percentages.In the melanoma data set and regarding the AIC and SBC selection criteria, the Geometric Generalized Birnbaum–Saunders distribution model was preferred and was selected as the appropriate model in the present study.
 Conclusion: Geometric Generalized Birnbaum–Saunders distribution is a highly flexible lifetime model which allows for different degrees of Kurtosis and asymmetry.by considering the advantages of the Geometric Generalized Birnbaum–Saunders distribution model, the model can be implemented as an appropriate alternative to explain or predict the survival time for long-term individuals.

Generation of a Melanoma and Nevus Data Set From Unstandardized Clinical Photographs on the Internet

Artificial intelligence (AI) training for diagnosing dermatologic images requires large amounts of clean data. Dermatologic images have different compositions, and many are inaccessible due to privacy concerns, which hinder the development of AI. To build a training data set for discriminative and generative AI from unstandardized internet images of melanoma and nevus. In this diagnostic study, a total of 5619 (CAN5600 data set) and 2006 (CAN2000 data set; a manually revised subset of CAN5600) cropped lesion images of either melanoma or nevus were semiautomatically annotated from approximately 500 000 photographs on the internet using convolutional neural networks (CNNs), region-based CNNs, and large mask inpainting. For unsupervised pretraining, 132 673 possible lesions (LESION130k data set) were also created with diversity by collecting images from 18 482 websites in approximately 80 countries. A total of 5000 synthetic images (GAN5000 data set) were generated using the generative adversarial network (StyleGAN2-ADA; training, CAN2000 data set; pretraining, LESION130k data set). The area under the receiver operating characteristic curve (AUROC) for determining malignant neoplasms was analyzed. In each test, 1 of the 7 preexisting public data sets (total of 2312 images; including Edinburgh, an SNU subset, Asan test, Waterloo, 7-point criteria evaluation, PAD-UFES-20, and MED-NODE) was used as the test data set. Subsequently, a comparative study was conducted between the performance of the EfficientNet Lite0 CNN on the proposed data set and that trained on the remaining 6 preexisting data sets. The EfficientNet Lite0 CNN trained on the annotated or synthetic images achieved higher or equivalent mean (SD) AUROCs to the EfficientNet Lite0 trained using the pathologically confirmed public data sets, including CAN5600 (0.874 [0.042]; P = .02), CAN2000 (0.848 [0.027]; P = .08), and GAN5000 (0.838 [0.040]; P = .31 [Wilcoxon signed rank test]) and the preexisting data sets combined (0.809 [0.063]) by the benefits of increased size of the training data set. The synthetic data set in this diagnostic study was created using various AI technologies from internet images. A neural network trained on the created data set (CAN5600) performed better than the same network trained on preexisting data sets combined. Both the annotated (CAN5600 and LESION130k) and synthetic (GAN5000) data sets could be shared for AI training and consensus between physicians.

Learning how to detect: A deep reinforcement learning method for whole-slide melanoma histopathology images.

Cutaneous melanoma represents one of the most life-threatening malignancies. Histopathological image analysis serves as a vital tool for early melanoma detection. Deep neural network (DNN) models are frequently employed to aid pathologists in enhancing the efficiency and accuracy of diagnoses. However, due to the paucity of well-annotated, high-resolution, whole-slide histopathology image (WSI) datasets, WSIs are typically fragmented into numerous patches during the model training and testing stages. This process disregards the inherent interconnectedness among patches, potentially impeding the models' performance. Additionally, the presence of excess, non-contributing patches extends processing times and introduces substantial computational burdens. To mitigate these issues, we draw inspiration from the clinical decision-making processes of dermatopathologists to propose an innovative, weakly supervised deep reinforcement learning framework, titled Fast medical decision-making in melanoma histopathology images (FastMDP-RL). This framework expedites model inference by reducing the number of irrelevant patches identified within WSIs. FastMDP-RL integrates two DNN-based agents: the search agent (SeAgent) and the decision agent (DeAgent). The SeAgent initiates actions, steered by the image features observed in the current viewing field at various magnifications. Simultaneously, the DeAgent provides labeling probabilities for each patch. We utilize multi-instance learning (MIL) to construct a teacher-guided model (MILTG), serving a dual purpose: rewarding the SeAgent and guiding the DeAgent. Our evaluations were conducted using two melanoma datasets: the publicly accessible TCIA-CM dataset and the proprietary MELSC dataset. Our experimental findings affirm FastMDP-RL's ability to expedite inference and accurately predict WSIs, even in the absence of pixel-level annotations. Moreover, our research investigates the WSI-based interactive environment, encompassing the design of agents, state and reward functions, and feature extractors suitable for melanoma tissue images. This investigation offers valuable insights and references for researchers engaged in related studies. The code is available at: https://github.com/titizheng/FastMDP-RL.

ENHANCING MELANOMA CLASSIFICATION WITH GRAPH ATTENTION LAYERS AND GROUP METHOD OF DATA HANDLING - BASED FEATURE EXTRACTION

Melanoma, a deadly form of skin cancer, demands accurate and early diagnosis for effective treatment. In this study, we propose a novel approach to improve melanoma classification by integrating Graph Attention Layers (GALs) into the Group Method of Data Handling (GMDH) framework. Our method leverages the power of GMDH to automatically generate and select informative features from complex melanoma-related data. Simultaneously, GALs are employed to capture intricate relationships and dependencies within the data, often overlooked by traditional classification models. We construct a graph representation where nodes represent data elements (patients or genetic markers) and edges signify relationships between them. GALs are applied to the graph, allowing the model to attend to relevant nodes and connections, enhancing its ability to discern subtle patterns indicative of melanoma. We then train a classification model on this enriched feature set, aiming for superior accuracy in melanoma diagnosis. Experimental results on a diverse melanoma dataset demonstrate the effectiveness of our approach. The model consistently outperforms traditional methods in terms of accuracy, precision, and recall. This study highlights the potential of combining GMDH-based feature extraction with GALs in melanoma classification. This approach not only advances diagnostic accuracy but also provides valuable insights into the underlying factors driving melanoma risk. As early detection remains the key to melanoma treatment success, our proposed method holds promise for improving patient outcomes.

Dermokine mutations contribute to epithelial-mesenchymal transition and advanced melanoma through ERK/MAPK pathways.

To discover vulnerabilities associated with dermokine (DMKN) as a new trigger of the epithelial-mesenchymal transition (EMT) -driven melanoma, we undertook a genome-wide genetic screening using transgenic. Here, we showed that DMKN expression could be constitutively increased in human malignant melanoma (MM) and that this correlates with poor overall survival in melanoma patients, especially in BRAF-mutated MM samples. Furthermore, in vitro, knockdown of DMKN inhibited the cell proliferation, migration, invasion, and apoptosis of MM cancer cells by the activation of ERK/MAPK signaling pathways and regulator of STAT3 in downstream molecular. By interrogating the in vitro melanoma dataset and characterization of advanced melanoma samples, we found that DMKN downregulated the EMT-like transcriptional program by disrupting EMT cortical actin, increasing the expression of epithelial markers, and decreasing the expression of mesenchymal markers. In addition, whole exome sequencing was presented with p.E69D and p.V91A DMKN mutations as a novel somatic loss of function mutations in those patients. Moreover, our purposeful proof-of-principle modeled the interaction of ERK with p.E69D and p.V91A DMKN mutations in the ERK-MAPK kinas signaling that may be naturally associated with triggering the EMT during melanomagenesis. Altogether, these findings provide preclinical evidence for the role of DMKN in shaping the EMT-like melanoma phenotype and introduced DMKN as a new exceptional responder for personalized MM therapy.

TimiGP: Inferring cell-cell interactions and prognostic associations in the tumor immune microenvironment through gene pairs

Determining the prognostic association of different immune cell types in the tumor microenvironment is critical for understanding cancer biology and developing new therapeutic strategies. However, this is challenging in certain cancer types, where the abundance of different immune subsets is highly correlated. In this study, we develop a computational method named TimiGP to overcome this challenge. Based on bulk gene expression and survival data, TimiGP infers cell-cell interactions that reveal the association between immune cell relative abundance and prognosis. As demonstrated in metastatic melanoma, TimiGP prioritizes immune cells critical in prognosis based on the identified cell-cell interactions. Highly consistent results are obtained by TimiGP when applied to seven independent melanoma datasets and when different cell-type marker sets are used as inputs. Additionally, TimiGP can leverage single-cell RNA sequencing data to delineate the tumor immune microenvironment at high resolutions across a wide range of cancer types.