Cancer Subtype Classification Research Articles

Deep learning approach for cancer subtype classification using high-dimensional gene expression data

MotivationStudies have shown that classifying cancer subtypes can provide valuable information for a range of cancer research, from aetiology and tumour biology to prognosis and personalized treatment. Current methods usually adopt gene expression data to perform cancer subtype classification. However, cancer samples are scarce, and the high-dimensional features of their gene expression data are too sparse to allow most methods to achieve desirable classification results.ResultsIn this paper, we propose a deep learning approach by combining a convolutional neural network (CNN) and bidirectional gated recurrent unit (BiGRU): our approach, DCGN, aims to achieve nonlinear dimensionality reduction and learn features to eliminate irrelevant factors in gene expression data. Specifically, DCGN first uses the synthetic minority oversampling technique algorithm to equalize data. The CNN can handle high-dimensional data without stress and extract important local features, and the BiGRU can analyse deep features and retain their important information; the DCGN captures key features by combining both neural networks to overcome the challenges of small sample sizes and sparse, high-dimensional features. In the experiments, we compared the DCGN to seven other cancer subtype classification methods using breast and bladder cancer gene expression datasets. The experimental results show that the DCGN performs better than the other seven methods and can provide more satisfactory classification results.

Pan-cancer gene expression analysis of tissue microarray using EdgeSeq oncology biomarker panel and a cross-comparison with HER2 and HER3 immunohistochemical analysis.

Molecular and protein biomarker profiling are key to oncology drug development. Antibody-drug conjugates (ADCs) directly deliver chemotherapeutic agents into tumor cells based on unique cancer cell biomarkers. A pan-cancer tissue microarray (TMA) data set and gene panel were validated and gene signature analyses were conducted on normal and cancer tissues to refine selection of ADC targets. Correlation of mRNA and protein levels, and human epidermal growth factor receptor (HER) expression patterns were assessed. An EdgeSeq biomarker panel (2862 genes) was used across 8531 samples (23 solid cancer types/subtypes; 16 normal tissues) with an established TMA data set, and immune cell and cell cycle gene signatures were analyzed. Discriminating gene expression signatures were defined based on pathological classification of cancer subtypes. Correlative analyses of HER2 and HER3 mRNA (EdgeSeq) and protein expression (immunohistochemistry [IHC]) were performed and compared with publicly available data (The Cancer Genome Atlas [TCGA]; Cancer Cell Line Encyclopedia [CCLE]). Gene expression patterns among cancer types in the TMA (EdgeSeq) and TCGA (RNA-seq) were similar. EdgeSeq gene signature analyses aligned with the majority of pathological cancer types/subtypes and identified cancer-specific gene expression patterns. TMA IHC H-scores for HER3 varied across cancer types/subtypes. In a few cancer types, HER3 mRNA and protein expression did not align, including lower liver hepatocellular carcinoma IHC H-score, compared with mRNA. Although all TNBC and ovarian cancer subtypes expressed mRNA, some had lower protein expression. This was seen in TMA and TCGA data sets, but not in CCLE. The EdgeSeq TMA data set can expand upon current biomarker data by including cancers not currently in TCGA. The primary analysis of EdgeSeq and IHC comparison suggested a unique protein-level regulation of HER3 in some tumor subtypes and highlights the importance of investigating protein levels of ADC targets in both tumor and normal tissues.

A Gene Selection Method Based on Outliers for Breast Cancer Subtype Classification

Breast cancer is the second most common cancer type and is the leading cause of cancer-related deaths worldwide. Since it is a heterogeneous disease, subtyping breast cancer plays an important role in performing a specific treatment. Gene expression data is a viable alternative to be employed on cancer subtype classification, as they represent the state of a cell at the molecular level, but generally has a relatively small number of samples compared to a large number of genes. Gene selection is a promising approach that addresses this uneven high-dimensional matrix of genes versus samples and plays an important role in the development of efficient cancer subtype classification. In this work, an innovative outlier-based gene selection (OGS) method is proposed to select relevant genes for efficiently and effectively classify breast cancer subtypes. Experiments show that our strategy presents an F1 score of 1.0 for basal and 0.86 for her 2, the two subtypes with the worst prognoses, respectively. Compared to other methods, our proposed method outperforms in the F1 score using 80% less genes. In general, our method selects only a few highly relevant genes, speeding up the classification, and significantly improving the classifier's performance.

Breast Cancer Subtypes Classification with Hybrid Machine Learning Model.

Breast cancer is the most prevailing heterogeneous disease among females characterized with distinct molecular subtypes and varied clinicopathological features. With the emergence of various artificial intelligence techniques especially machine learning, the breast cancer research has attained new heights in cancer detection and prognosis. Recent development in computer driven diagnostic system has enabled the clinicians to improve the accuracy in detecting various types of breast tumors. Our study is to develop a computer driven diagnostic system which will enable the clinicians to improve the accuracy in detecting various types of breast tumors. In this article, we proposed a breast cancer classification model based on the hybridization of machine learning approaches for classifying triple-negative breast cancer and non-triple negative breast cancer patients with clinicopathological features collected from multiple tertiary care hospitals/centers. The results of genetic algorithm and support vector machine (GA-SVM) hybrid model was compared with classics feature selection SVM hybrid models like support vector machine-recursive feature elimination (SVM-RFE), LASSO-SVM, Grid-SVM, and linear SVM. The classification results obtained from GA-SVM hybrid model outperformed the other compared models when applied on two distinct hospital-based datasets of patients investigated with breast cancer in North West of African subcontinent. To validate the predictive model accuracy, 10-fold cross-validation method was applied on all models with the same multicentered datasets. The model performance was evaluated with well-known metrics like mean squared error, logarithmic loss, F1-score, area under the ROC curve, and the precision-recall curve. The hybrid machine learning model can be employed for breast cancer subtypes classification that could help the medical practitioners in better treatment planning and disease outcome.

Inflammatory aging clock: A cancer clock to characterize the patients' subtypes and predict the overall survival in glioblastoma.

Background: Many biological clocks related to aging have been linked to the development of cancer. A recent study has identified that the inflammatory aging clock was an excellent indicator to track multiple diseases. However, the role of the inflammatory aging clock in glioblastoma (GBM) remains to be explored. This study aimed to investigate the expression patterns and the prognostic values of inflammatory aging (iAge) in GBM, and its relations with stem cells. Methods: Inflammation-related genes (IRG) and their relations with chronological age in normal samples from the Cancer Genome Atlas (TCGA) were identified by the Spearman correlation analysis. Then, we calculated the iAge and computed their correlations with chronological age in 168 patients with GBM. Next, iAge was applied to classify the patients into high- and low-iAge subtypes. Next, the survival analysis was performed. In addition, the correlations between iAge and stem cell indexes were evaluated. Finally, the results were validated in an external cohort. Results: Thirty-eight IRG were significantly associated with chronological age (|coefficient| > 0.5), and were used to calculate the iAge. Correlation analysis showed that iAge was positively correlated with chronological age. Enrichment analysis demonstrated that iAge was highly associated with immune cells and inflammatory activities. Survival analysis showed the patients in the low-iAge subtype had significantly better overall survival (OS) than those in the high-iAge subtype (p < 0.001). In addition, iAge outperformed the chronological age in revealing the correlations with stem cell stemness. External validation demonstrated that iAge was an excellent method to classify cancer subtypes and predict survival in patients with GBM. Conclusions: Inflammatory aging clock may be involved in the GBM via potential influences on immune-related activities. iAge could be used as biomarkers for predicting the OS and monitoring the stem cell.

Multimodal Dimension Reduction and Subtype Classification of Head and Neck Squamous Cell Tumors.

Traditional analysis of genomic data from bulk sequencing experiments seek to group and compare sample cohorts into biologically meaningful groups. To accomplish this task, large scale databases of patient-derived samples, like that of TCGA, have been established, giving the ability to interrogate multiple data modalities per tumor. We have developed a computational strategy employing multimodal integration paired with spectral clustering and modern dimension reduction techniques such as PHATE to provide a more robust method for cancer sub-type classification. Using this integrated approach, we have examined 514 Head and Neck Squamous Carcinoma (HNSC) tumor samples from TCGA across gene-expression, DNA-methylation, and microbiome data modalities. We show that these approaches, primarily developed for single-cell sequencing can be efficiently applied to bulk tumor sequencing data. Our multimodal analysis captures the dynamic heterogeneity, identifies new and refines subtypes of HNSC, and orders tumor samples along well-defined cellular trajectories. Collectively, these results showcase the inherent molecular complexity of tumors and offer insights into carcinogenesis and importance of targeted therapy. Computational techniques as highlighted in our study provide an organic and powerful approach to identify granular patterns in large and noisy datasets that may otherwise be overlooked.

Cancer Subtyping via Embedded Unsupervised Learning on Transcriptomics Data.

Cancer is one of the deadliest diseases worldwide. Accurate diagnosis and classification of cancer subtypes are indispensable for effective clinical treatment. Promising results on automatic cancer subtyping systems have been published recently with the emergence of various deep learning methods. However, such automatic systems often overfit the data due to the high dimensionality and scarcity. In this paper, we propose to investigate automatic subtyping from an unsupervised learning perspective by directly constructing the underlying data distribution itself, hence sufficient data can be generated to alleviate the issue of overfitting. Specifically, we bypass the strong Gaussianity assumption that typically exists but fails in the unsupervised learning subtyping literature due to small-sized samples by vector quantization. Our proposed method better captures the latent space features and models the cancer subtype manifestation on a molecular basis, as demonstrated by the extensive experimental results.

A Novel Hybrid Convolutional Neural Network Approach for the Stomach Intestinal Early Detection Cancer Subtype Classification.

There may be different types of cancer that cause fatal effects in the human body. In general, cancer is nothing but the unnatural growth of blood cells in different parts of the body and is named accordingly. It may be skin cancer, breast cancer, uterus cancer, intestinal cancer, stomach cancer, etc. However, every type of cancer consists of unwanted blood cells which cause issues in the body starting from the minor to death. Cancer cells have the common features in them, and these common features we have used in our work for the processing. Cancer has a significant death rate; however, it is frequently curable with simple surgery if detected in its early stages. A quick and correct diagnosis may be extremely beneficial to both doctors and patients. In several medical domains, the latest deep-learning-based model's performance is comparable to or even exceeds that of human specialists. We have proposed a novel methodology based on a convolutional neural network that may be used for almost all types of cancer detection. We have collected different datasets of different types of common cancer from different sources and used 90% of the sample data for the training purpose, then we reduced it by 10%, and an 80% image set was used for the validation of the model. After that for testing purposes, we fed a sample dataset and obtain the results. The final output clearly shows that the proposed model outperforms the previous model when we compared our methodology with the existing work.

Abstract 1221: Comprehensive cell-type classification of tumor and normal cells from single cell RNA sequencing in pan cancer settings

Abstract Single-cell RNA sequencing (scRNA-seq) allows for the study of the transcriptome at a cellular level, where populations of cells are annotated based on the expression of marker genes, providing a tool to gain cell-specific, subtle insights on cancer biology. However, precise annotation of cell type remains a challenge, hindering the efficiency of data interpretation. Several existing tools for cell-type annotation have been developed to improve resolution and reproducibility, yet their performance is reduced when the reference dataset contains many cell types, subclasses of similar cell types, or malignant cells. Interpatient malignant cell heterogeneity often leads to reduced accuracy when classifying cancer cells as most methods rely on correlations to a reference from a different source. Given the challenges in the annotation of scRNA-seq data of cancer and its high impact for elucidating mechanisms associated with tumor heterogeneity, pathogenesis, and treatment, we developed a comprehensive, hierarchically organized, multi-layered classifier spanning diverse malignant and normal cells of the tumor microenvironment. We found that performance improves when each layer focuses on a smaller number of classes and each cell sequentially moves down a series of classifiers with increased cell type resolution. When applied to an external validation dataset of over 300 primary solid tumor biopsies spanning diverse cancer types, the classifier accurately annotated the tissue of origin of malignant cells, and relevant subtypes of stromal and blood cells, with average F1 scores of 0.91, 0.95 and 0.99 respectively. Using confidence thresholds at each layer, the classifier abstains from classifying ambiguous cells. We applied 4 existing annotators provided with the same reference to the external test dataset and found that cancer cells are misclassified or unclassified, while the blood and stromal cells are accurately classified, highlighting our tool’s unique ability to classify cancer cells. Moreover, we applied our classifier’s to scRNA-seq data derived from breast cancer metastasis to the liver and were able to uncover the tissue of origin, demonstrating the potential use for determining the source of a metastatic tumour. Finally, given that our classifier is modular, we leveraged two recently published single cell breast cancer atlases to add a breast cancer subtype classification layer, that consistently identified the correct clinical subtype of single breast cancer cells in external data. This study provides a flexible model for the annotation of cells comprising the tumor microenvironment in pan cancer settings, while existing methods require tissue-specific references for every cancer type. Our classifier provides a powerful method for investigating intercellular communication pathways between tumor cells and non-malignant cells of the tumor microenvironment. Citation Format: Ido Nofech-Mozes, Philip Awadalla, Sagi Abelson. Comprehensive cell-type classification of tumor and normal cells from single cell RNA sequencing in pan cancer settings [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1221.

Dynamic Meta-data Network Sparse PCA for Cancer Subtype Biomarker Screening

Previous research shows that each type of cancer can be divided into multiple subtypes, which is one of the key reasons that make cancer difficult to cure. Under these circumstances, finding a new target gene of cancer subtypes has great significance on developing new anti-cancer drugs and personalized treatment. Due to the fact that gene expression data sets of cancer are usually high-dimensional and with high noise and have multiple potential subtypes’ information, many sparse principal component analysis (sparse PCA) methods have been used to identify cancer subtype biomarkers and subtype clusters. However, the existing sparse PCA methods have not used the known cancer subtype information as prior knowledge, and their results are greatly affected by the quality of the samples. Therefore, we propose the Dynamic Metadata Edge-group Sparse PCA (DM-ESPCA) model, which combines the idea of meta-learning to solve the problem of sample quality and uses the known cancer subtype information as prior knowledge to capture some gene modules with better biological interpretations. The experiment results on the three biological data sets showed that the DM-ESPCA model can find potential target gene probes with richer biological information to the cancer subtypes. Moreover, the results of clustering and machine learning classification models based on the target genes screened by the DM-ESPCA model can be improved by up to 22–23% of accuracies compared with the existing sparse PCA methods. We also proved that the result of the DM-ESPCA model is better than those of the four classic supervised machine learning models in the task of classification of cancer subtypes.

Leveraging Deep Learning Techniques and Integrated Omics Data for Tailored Treatment of Breast Cancer.

Multiomics data of cancer patients and cell lines, in synergy with deep learning techniques, have aided in unravelling predictive problems related to cancer research and treatment. However, there is still room for improvement in the performance of the existing models based on the aforementioned combination. In this work, we propose two models that complement the treatment of breast cancer patients. First, we discuss our deep learning-based model for breast cancer subtype classification. Second, we propose DCNN-DR, a deep convolute.ion neural network-drug response method for predicting the effectiveness of drugs on in vitro and in vivo breast cancer datasets. Finally, we applied DCNN-DR for predicting effective drugs for the basal-like breast cancer subtype and validated the results with the information available in the literature. The models proposed use late integration methods and have fairly better predictive performance compared to the existing methods. We use the Pearson correlation coefficient and accuracy as the performance measures for the regression and classification models, respectively.

Molecular Subtypes of High-Grade Neuroendocrine Carcinoma (HGNEC): What is YAP1-Positive HGNEC?

Small cell lung cancer (SCLC) subtype classification, based on high-level expression of key transcriptional regulators; ASCL1 (SCLC-A), NEUROD1 (SCLC-N), POU2F3 (SCLC-P), and YAP1 (SCLC-Y), has recently been proposed. YAP1 (and POU2F3) has attracted attention as an important factor for non-neuroendocrine (non-NE) phenotypic subtyping of SCLC. However, subsequent studies reported that YAP1 expression alone cannot define a single group in primary SCLC, which makes it difficult to understand what SCLC-Y is by focusing only on SCLC. In this review, we concluded that YAP1 is an essential anti-neuroendocrine factor in both SCLC and non-small cell lung cancer (NSCLC) based on previous studies, including our own analysis of the cell lines and primary tumors of SCLC and NSCLC. The classification of SCLC-Y is a concept mainly established from the analysis of cell lines, and SCLC-Y cell lines correspond to "variant type" SCLC cell lines. Primary SCLC and large cell neuroendocrine carcinoma (LCNEC) are typically heterogeneous tumors composed mostly of NE-type cells, but they contain a small number of non-NE-type cells. Importantly, individual cells with NE features exhibit YAP1 loss, whereas the non-NE-type cells exhibit YAP1 expression. Although rare in primary SCLC, some cases of primary LCNEC have many YAP1-positive cells, which is correlated with chemotherapy resistance. YAP1 staining may be useful in the exclusion diagnosis of SCLC or in the selection of treatment for LCNEC.

MoGCN: A Multi-Omics Integration Method Based on Graph Convolutional Network for Cancer Subtype Analysis.

In light of the rapid accumulation of large-scale omics datasets, numerous studies have attempted to characterize the molecular and clinical features of cancers from a multi-omics perspective. However, there are great challenges in integrating multi-omics using machine learning methods for cancer subtype classification. In this study, MoGCN, a multi-omics integration model based on graph convolutional network (GCN) was developed for cancer subtype classification and analysis. Genomics, transcriptomics and proteomics datasets for 511 breast invasive carcinoma (BRCA) samples were downloaded from the Cancer Genome Atlas (TCGA). The autoencoder (AE) and the similarity network fusion (SNF) methods were used to reduce dimensionality and construct the patient similarity network (PSN), respectively. Then the vector features and the PSN were input into the GCN for training and testing. Feature extraction and network visualization were used for further biological knowledge discovery and subtype classification. In the analysis of multi-dimensional omics data of the BRCA samples in TCGA, MoGCN achieved the highest accuracy in cancer subtype classification compared with several popular algorithms. Moreover, MoGCN can extract the most significant features of each omics layer and provide candidate functional molecules for further analysis of their biological effects. And network visualization showed that MoGCN could make clinically intuitive diagnosis. The generality of MoGCN was proven on the TCGA pan-kidney cancer datasets. MoGCN and datasets are public available at https://github.com/Lifoof/MoGCN. Our study shows that MoGCN performs well for heterogeneous data integration and the interpretability of classification results, which confers great potential for applications in biomarker identification and clinical diagnosis.

Optimized LSTM with Dimensionality Reduction Based Gene Expression Data Classification

The classification of cancer subtypes is substantial for the diagnosis and treatment of cancer. However, the gene expression data used for cancer subtype classification are high dimensional in nature and small in sample size. In this paper, an efficient dimensionality reduction with optimized long short term memory, algorithm (OLSTM) is used for gene expression data classification. The main three stages of the proposed method are explicitly pre-processing, dimensional reduction, and gene expression data classification. In the pre-processing method, the missing values and redundant values are removed for high-quality data. Following, the dimensional reduction is done by orthogonal locality preserving projections (OLPP). Finally, gene classification is done by an OLSTM classifier. Here the traditional long short term memory (LSTM) is modified using parameter optimization which uses the adaptive artificial flora optimization (AAFO) algorithm. Based on the migration and flora reproduction process, the AAFO algorithm is stimulated. Using the accuracy, sensitivity, specificity, precision, recall, and f-measure, the proposed performance is analyzed. The test outcomes illustrate the effectiveness of the gene expression data classification with a 94.19% of accuracy value. The proposed gene expression data classification is implemented in the MATLAB platform.

Identifying Cancer Subtypes Using a Residual Graph Convolution Model on a Sample Similarity Network.

Cancer subtype classification helps us to understand the pathogenesis of cancer and develop new cancer drugs, treatment from which patients would benefit most. Most previous studies detect cancer subtypes by extracting features from individual samples, ignoring their associations with others. We believe that the interactions of cancer samples can help identify cancer subtypes. This work proposes a cancer subtype classification method based on a residual graph convolutional network and a sample similarity network. First, we constructed a sample similarity network regarding cancer gene co-expression patterns. Then, the gene expression profiles of cancer samples as initial features and the sample similarity network were passed into a two-layer graph convolutional network (GCN) model. We introduced the initial features to the GCN model to avoid over-smoothing during the training process. Finally, the classification of cancer subtypes was obtained through a softmax activation function. Our model was applied to breast invasive carcinoma (BRCA), glioblastoma multiforme (GBM) and lung cancer (LUNG) datasets. The accuracy values of our model reached 82.58%, 85.13% and 79.18% for BRCA, GBM and LUNG, respectively, which outperformed the existing methods. The survival analysis of our results proves the significant clinical features of the cancer subtypes identified by our model. Moreover, we can leverage our model to detect the essential genes enriched in gene ontology (GO) terms and the biological pathways related to a cancer subtype.

Dual-elemental analysis of single particles using quadrupole-based inductively coupled plasma-mass spectrometry

Single-particle inductively coupled plasma-mass spectrometry (SP-ICP-MS) is used for elemental analysis of single particles and biological cells. Time-of-flight (TOF) mass analyzers are widely used for multiple element analysis of individual particles. Owing to the sequential nature of the mass analyzer, quadrupole-based ICP-MS generally gives poor analytical performance when more than one element are being monitored. In this study, we present the first accurate and precise dual-mass measurement of individual particles using quadrupole-based ICP-MS, with the assistance of oxygen collision gas. Simultaneous measurement of the intensity of 107Ag and 109Ag of Ag nanoparticles (AgNP) showed particle recovery of 100% and Pearson correlation coefficient of 0.97, indicating proper sampling of all particles in the ICP. This technique gives good measurement precision (RSD <8%) and high accuracy in size measurement (error <3%). This technique was further applied to determine the elemental content and isotope ratios of particles and to study cell viability after Cisplatin staining. The results are comparable to that of existing TOF and multi-collector ICP-MS, indicating that quadrupole-based ICP-MS can be a cost-effective alternative for simultaneous measurement of two isotopes. Acquisition with longer integration time and shorter settling time is also proposed to further improve the sensitivity and number of isotopes monitored. This study will potentially open up more possibilities of using quadrupole based ICP-MS in biomedical research as quantification of multi-elements in single cells is far more informative. Other possible applications include classification of cancer subtypes according to the abundance of several biomarkers, as well as elemental bio-imaging of transcripts and proteins in tissues by laser ablation.

Lung cancer subtype classification using histopathological images based on weakly supervised multi-instance learning

Objective. Subtype classification plays a guiding role in the clinical diagnosis and treatment of non-small-cell lung cancer (NSCLC). However, due to the gigapixel of whole slide images (WSIs) and the absence of definitive morphological features, most automatic subtype classification methods for NSCLC require manually delineating the regions of interest (ROIs) on WSIs. Approach. In this paper, a weakly supervised framework is proposed for accurate subtype classification while freeing pathologists from pixel-level annotation. With respect to the characteristics of histopathological images, we design a two-stage structure with ROI localization and subtype classification. We first develop a method called multi-resolution expectation-maximization convolutional neural network (MR-EM-CNN) to locate ROIs for subsequent subtype classification. The EM algorithm is introduced to select the discriminative image patches for training a patch-wise network, with only WSI-wise labels available. A multi-resolution mechanism is designed for fine localization, similar to the coarse-to-fine process of manual pathological analysis. In the second stage, we build a novel hierarchical attention multi-scale network (HMS) for subtype classification. HMS can capture multi-scale features flexibly driven by the attention module and implement hierarchical features interaction. Results. Experimental results on the 1002-patient Cancer Genome Atlas dataset achieved an AUC of 0.9602 in the ROI localization and an AUC of 0.9671 for subtype classification. Significance. The proposed method shows superiority compared with other algorithms in the subtype classification of NSCLC. The proposed framework can also be extended to other classification tasks with WSIs.

The Role of Ki67 in Evaluating Neoadjuvant Endocrine Therapy of Hormone Receptor-Positive Breast Cancer.

Ki67 is a proliferation marker. It has been proposed as a useful clinical marker for breast cancer subtype classification, prognosis, and prediction of therapeutic response. But the questionable analytical validity of Ki67 prevents its widespread adoption of these measures for treatment decisions in breast cancer. Currently, Ki67 has been tested as a predictive marker for chemotherapy using clinical and pathological response as endpoints in neoadjuvant endocrine therapy. Ki67 can be used as a predictor to evaluate the recurrence-free survival rate of patients, or its change can be used to predict the preoperative “window of opportunity” in neoadjuvant endocrine therapy. In this review, we will elaborate on the role of Ki67 in neoadjuvant endocrine therapy in breast cancer.

The impact of scan number and its preprocessing in micro-FTIR imaging when applying machine learning for breast cancer subtypes classification

The breast cancer molecular subtype is an important classification to outline the prognostic. Gold-standard assessing using immunohistochemistry adds subjectivity due to interlaboratory and interobserver variations. In order to increase the diagnosis confidence, other techniques need to be examined, where the FTIR spectroscopy imaging allied with machine learning techniques may provide additional and quantitative information regarding the molecular composition. However, the impact of co-added scans acquisition parameter into machine learning classifications still needs better evaluation. In this study, FTIR images of Luminal B and HER2 subtypes were acquired varying the scan number and preprocessing techniques. It was demonstrated a spectral quality improvement when the scan number was increased, decreasing the standard deviation and outliers. Six machine learning models were used to classify the subtypes: Linear Discriminant Analysis, Partial Least Squares Discriminant Analysis, K-Nearest Neighbors, Support Vector Machine, Random Forest and Extreme Gradient Boosting. Best mean accuracy of 0.995 was achieved by Extreme Gradient Boosting model. It was found that all models achieved similar high accuracies with groups b256_064 (256 background and 064 scans), b256_128 and b128_128. Besides assessing the performance of different models, the b256_064 was established as the optimal group due to the minimum acquisition time. Therefore, this work indicates b256_064 for breast cancer subtype classification and also as a basis for other studies using machine learning for cancer evaluation.

A laminar augmented cascading flexible neural forest model for classification of cancer subtypes based on gene expression data

BackgroundCorrectly classifying the subtypes of cancer is of great significance for the in-depth study of cancer pathogenesis and the realization of personalized treatment for cancer patients. In recent years, classification of cancer subtypes using deep neural networks and gene expression data has gradually become a research hotspot. However, most classifiers may face overfitting and low classification accuracy when dealing with small sample size and high-dimensional biology data.ResultsIn this paper, a laminar augmented cascading flexible neural forest (LACFNForest) model was proposed to complete the classification of cancer subtypes. This model is a cascading flexible neural forest using deep flexible neural forest (DFNForest) as the base classifier. A hierarchical broadening ensemble method was proposed, which ensures the robustness of classification results and avoids the waste of model structure and function as much as possible. We also introduced an output judgment mechanism to each layer of the forest to reduce the computational complexity of the model. The deep neural forest was extended to the densely connected deep neural forest to improve the prediction results. The experiments on RNA-seq gene expression data showed that LACFNForest has better performance in the classification of cancer subtypes compared to the conventional methods.ConclusionThe LACFNForest model effectively improves the accuracy of cancer subtype classification with good robustness. It provides a new approach for the ensemble learning of classifiers in terms of structural design.