Cancer Subtype Classification Research Articles

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 Cascade Flexible Neural Forest Model for Cancer Subtypes Classification on Gene Expression Data.

The correct classification of cancer subtypes is of great significance for the in-depth study of cancer pathogenesis and the realization of accurate treatment for cancer patients. In recent years, the classification of cancer subtypes using deep neural networks and gene expression data has become a hot topic. However, most classifiers may face the challenges of overfitting and low classification accuracy when dealing with small sample size and high-dimensional biological data. In this paper, the Cascade Flexible Neural Forest (CFNForest) Model was proposed to accomplish cancer subtype classification. CFNForest extended the traditional flexible neural tree structure to FNT Group Forest exploiting a bagging ensemble strategy and could automatically generate the model's structure and parameters. In order to deepen the FNT Group Forest without introducing new hyperparameters, the multilayer cascade framework was exploited to design the FNT Group Forest model, which transformed features between levels and improved the performance of the model. The proposed CFNForest model also improved the operational efficiency and the robustness of the model by sample selection mechanism between layers and setting different weights for the output of each layer. To accomplish cancer subtype classification, FNT Group Forest with different feature sets was used to enrich the structural diversity of the model, which make it more suitable for processing small sample size datasets. The experiments on RNA-seq gene expression data showed that CFNForest effectively improves the accuracy of cancer subtype classification. The classification results have good robustness.

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.

Classification of breast cancer types, sub-types and grade from histopathological images using deep learning technique

Classification of breast cancer types, sub-types and grade from histopathological images using deep learning technique

Reducing variability of breast cancer subtype predictors by grounding deep learning models in prior knowledge

Deep learning neural networks have improved performance in many cancer informatics problems, including breast cancer subtype classification. However, many networks experience underspecificationwheremultiplecombinationsofparametersachievesimilarperformance, bothin training and validation. Additionally, certain parameter combinations may perform poorly when the test distribution differs from the training distribution. Embedding prior knowledge from the literature may address this issue by boosting predictive models that provide crucial, in-depth information about a given disease. Breast cancer research provides a wealth of such knowledge, particularly in the form of subtype biomarkers and genetic signatures. In this study, we draw on past research on breast cancer subtype biomarkers, label propagation, and neural graph machines to present a novel methodology for embedding knowledge into machine learning systems. We embed prior knowledge into the loss function in the form of inter-subject distances derived from a well-known published breast cancer signature. Our results show that this methodology reduces predictor variability on state-of-the-art deep learning architectures and increases predictor consistency leading to improved interpretation. We find that pathway enrichment analysis is more consistent after embedding knowledge. This novel method applies to a broad range of existing studies and predictive models. Our method moves the traditional synthesis of predictive models from an arbitrary assignment of weights to genes toward a more biologically meaningful approach of incorporating knowledge.

Raman spectroscopy and machine learning for the classification of breast cancers

Breast cancer is a major health threat for women. The drug responses associated with different breast cancer subtypes have obvious effects on therapeutic outcomes; therefore, the accurate classification of breast cancer subtypes is critical. Breast cancer subtype classification has recently been examined using various methods, and Raman spectroscopy has emerged as an effective technique that can be used for noninvasive breast cancer analysis. However, the accurate and rapid classification of breast cancer subtypes currently requires a great deal of effort and experience with the processing and analysis of Raman spectra data. Here, we adopted Raman spectroscopy and machine learning techniques to simplify and accelerate the process used to distinguish normal from breast cancer cells and classify breast cancer subtypes. Raman spectra were obtained from cultured breast cancer cell lines, and the data were analyzed by two machine learning algorithms: principal component analysis (PCA)–discriminant function analysis (DFA) and PCA–support vector machine (SVM). The accuracies with which these two algorithms were able to distinguish normal breast cells from breast cancer cells were both greater than 97%, and the accuracies of breast cancer subtype classification for both algorithms were both greater than 92%. Moreover, our results showed evidence to support the use of characteristic Raman spectral features as cancer cell biomarkers, such as the intensity of intrinsic Raman bands, which increased in cancer cells. Raman spectroscopy combined with machine learning techniques provides a rapid method for breast cancer analysis able to reveal differences in intracellular compositions and molecular structures among subtypes.

Improved heterogeneous data fusion and multi‐scale feature selection method for lung cancer subtype classification

SummaryThe diagnosis of the disease requires a variety of data and indicators. In order to make the computer perform intelligent computation and diagnosis from a doctor's perspective, complementary information between different modal data needs to be taken into account. Meanwhile, the redundancy features with key information should be selected in order to reduce the complexity of calculation. In this study, an adaptive dynamic loss function is proposed to weight different scales in the multi‐scale expansion network of pathological images according to the doctor's diagnosis process. And an ant colony algorithm based on maximum information coefficient correlation was designed for unsupervised feature selection of fusion features combined image feature and patient differential genes. Experimental results show that the addition of pathological image information and genetic information plays an important role in the classification of lung cancer subtypes. Compared with other feature selection methods, the proposed algorithm can quickly converge. Combining pathological image and gene expression matrix for cancer diagnosis can improve the diagnostic accuracy of specific patients, with an accuracy of 95.62 and AUC achieves 0.897. The proposed method has high effectiveness and superior performance in the classification of lung cancer.

A Cascade Deep Forest Model for Breast Cancer Subtype Classification Using Multi-Omics Data

Automated diagnosis systems aim to reduce the cost of diagnosis while maintaining the same efficiency. Many methods have been used for breast cancer subtype classification. Some use single data source, while others integrate many data sources, the case that results in reduced computational performance as opposed to accuracy. Breast cancer data, especially biological data, is known for its imbalance, with lack of extensive amounts of histopathological images as biological data. Recent studies have shown that cascade Deep Forest ensemble model achieves a competitive classification accuracy compared with other alternatives, such as the general ensemble learning methods and the conventional deep neural networks (DNNs), especially for imbalanced training sets, through learning hyper-representations through using cascade ensemble decision trees. In this work, a cascade Deep Forest is employed to classify breast cancer subtypes, IntClust and Pam50, using multi-omics datasets and different configurations. The results obtained recorded an accuracy of 83.45% for 5 subtypes and 77.55% for 10 subtypes. The significance of this work is that it is shown that using gene expression data alone with the cascade Deep Forest classifier achieves comparable accuracy to other techniques with higher computational performance, where the time recorded is about 5 s for 10 subtypes, and 7 s for 5 subtypes.

Efficient Multi-Organ Multi-Center Cell Nuclei Segmentation Method Based on Deep Learnable Aggregation Network

Automated cell nuclei delineation in whole-slide imaging (WSI) is a fundamental step for many tasks like cancer cell recognition, cancer grading, and cancer subtype classification. Although numerous computational methods have been proposed for segmenting nuclei in WSI images based on image processing and deep learning, existing approaches face major challenges such as color variation due to the use of different stains, the various structures of cell nuclei, and the overlapping and clumped cell nuclei. To circumvent these challenges in this article, we propose an efficient and accurate cell nuclei segmentation method based on deep learning, in which a set of accurate individual cell nuclei segmentation models are developed to predict rough segmentation masks, and then a learnable aggregation network (LANet) is used to predict the final nuclei masks. Besides, we develop cell nuclei segmentation software (with a graphical user interface—GUI) that includes the proposed method and other deep-learning-based cell nuclei segmentation methods. A challenging WSI dataset collected from different centers and organs is used to demonstrate the efficiency of our method. The experimental results reveal that our method obtains a competitive performance compared to the existing approaches in terms of the aggregated Jaccard index (AJI=89.25%) and F1-score (F1=73.02%). The developed nuclei segmentation software can be downloaded from https://github.com/loaysh2010/Cell-Nuclei-Segmentation-GUI-Application.

Molecular genetic classification of colorectal cancer subtypes: current state of the problem

Today, colorectal cancer (CRC) is the third most common cancer and therefore an urgent problem of oncology. Despite all modern diagnostic capabilities, the rates of advanced cases are growing steadily. CRC was proven to be a result of a phased dysplastic change in the colon mucosa, molecular genetic changes that determine the molecular biology of the tumor, its properties, morphology, disease prognosis and response to therapy. The following mechanisms of CRC tumor progression are distinguished: chromosomal instability, microsatellite instability, "methylator" phenotype, and serrated pathway of adenocarcinoma development. Application of molecular and diagnostic methods has become a promising direction in recent years. This led to the development of a molecular genetic classifi cation with 4 CRC subtypes differing not only in their molecular genetic characteristics, but also in clinical course and response to therapy.

ARTIFICIAL INTELLIGENCE AND PATHOLOGY

The microscopic diagnosis of cancer remains challenging. Recent data from our group within the French (nationwide) LymphoPath network shows that 20% of diagnoses are inaccurate, with direct impact on patient care. Indeed, molecular techniques, that are not affordable for all pathology departments, have become critical for the final diagnosis. Currently, automated solutions that could help pathologists with diagnostic decisions or histological grading are lacking, and it is not reasonable to expect pathologists to be experts on rare tumours if they only see a few cases per year. Digital microscopy offers unique features which are not available in conventional optical microscopy. Assisted by dedicated software tools, digital microscopy enables dynamic and prompt access to any detail of the stained slides at several microscopic magnifications. Furthermore, the calibrated qualities and the number of discrete pixels in a digital slide allow automated image analysis and quantification by using computer vision and, in particular, deep learning approaches. Image feature extraction methods based on pixel detection or less often on object segmentation has brought much hope in improving the accuracy of the human eye. Automatic analysis of cancer whole-slide images has recently been performed to allow experts for predicting tumour classification, gene mutations and survival outcomes. However, the positioning of such technologies in routine practice is limited by the trained networks which frequently do not meet industrial constraints required for a general application such as certification, qualification and explainability (black box effect) of algorithms. This presentation will describe different approaches of machine learning aiming at classifying cancer subtypes through the analysis of histopathologic features. Advantages and drawbacks of these techniques will be discussed with a special emphasis on the risk of biased performance assessment of deep learning systems. The critical role of data sets quality will be discussed and different strategies will be envisaged to broaden the use deep neural networks. Providing that these limitations are taken into account and circumvented, it seems that artificial intelligence solutions dedicated to cancer histopathology should pave the road to precision medicine through data integration into a holistic patient dashboard for oncology care teams. Keywords: Bioinformatics; Computational and Systems Biology, Genomics, Epigenomics, and Other -Omics, Tumor Biology and Heterogeneity Conflicts of interests pertinent to the abstract P. Brousset Consultant or advisory role: Roche, MSB, Janssen Cilag laboratory Research funding: Roche, Pierre Fabre

Building patient-specific models for receptor tyrosine kinase signaling networks.

Cancer progresses due to changes in the dynamic interactions of multidimensional factors associated with gene mutations. Cancer research has actively adopted computational methods, including data-driven and mathematical model-driven approaches, to identify causative factors and regulatory rules that can explain the complexity and diversity of cancers. A data-driven, statistics-based approach revealed correlations between gene alterations and clinical outcomes in many types of cancers. A model-driven mathematical approach has elucidated the dynamic features of cancer networks and identified the mechanisms of drug efficacy and resistance. More recently, machine learning methods have emerged that can be used for mining omics data and classifying patient. However, as the strengths and weaknesses of each method becoming apparent, new analytical tools are emerging to combine and improve the methodologies and maximize their predictive power for classifying cancer subtypes and prognosis. Here, we introduce recent advances in cancer systems biology aimed at personalized medicine, with focus on the receptor tyrosine kinase signaling network.

A self-organizing deep neuro-fuzzy system approach for classification of kidney cancer subtypes using miRNA genomics data

Kidney cancer is a dangerous disease affecting many patients all over the world. Early-stage diagnosis and correct identification of kidney cancer subtypes play an essential role in the patient's survival; therefore, its subtypes diagnosis and classification are the main challenges in kidney cancer treatment. Medical studies have proved that miRNA dysregulation can increase the risk of cancer. Thus, in this paper, we propose a new machine learning approach for significant miRNAs identification and kidney cancer subtype classification to design an automatic diagnostic tool. The proposed method contains two main steps: feature selection and classification. First, we apply the feature selection algorithm to choose the candidate miRNAs for each subtype. The feature selection algorithm utilizes the AMGM measure to select significant miRNAs with high discriminant power. Next, the candidate miRNAs are fed to a classifier to evaluate the candidate features. In the classification step, the proposed self-organizing deep neuro-fuzzy system is employed to classify kidney cancer subgroups. The new deep neuro-fuzzy system consists of a deep structure in the rule layer and novel architecture in the fuzzifier layer. The proposed self-organizing deep neuro-fuzzy system can help us to overcome the main obstacles in the field of neuro-fuzzy system applications, such as the curse of dimensionality. The goal of this paper is to illustrate that the neuro-fuzzy system can very useful in high dimensional data, such as genomics data, using the proposed deep neuro-fuzzy system. The obtained results illustrated that our proposed method has succeeded in classifying kidney cancer subtypes with high accuracy based on the selected miRNAs.

Clinical cancer genomic profiling.

Technological innovation and rapid reduction in sequencing costs have enabled the genomic profiling of hundreds of cancer-associated genes as a component of routine cancer care. Tumour genomic profiling can refine cancer subtype classification, identify which patients are most likely to benefit from systemic therapies and screen for germline variants that influence heritable cancer risk. Here, we discuss ongoing efforts to enhance the clinical utility of tumour genomic profiling by integrating tumour and germline analyses, characterizing allelic context and identifying mutational signatures that influence therapy response. We also discuss the potential clinical utility of more comprehensive whole-genome and whole-transcriptome sequencing and ultra-sensitive cell-free DNA profiling platforms, which allow for minimally invasive, serial analyses of tumour-derived DNA in blood.

Diagnostic segregation of human breast tumours using Fourier-transform infrared spectroscopy coupled with multivariate analysis: Classifying cancer subtypes

The present study aimed to investigate whether attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy coupled with multivariate analysis could be applied to discriminate and classify among breast tumour molecular subtypes based on the unique spectral “fingerprints” of their biochemical composition. The different breast cancer tissues and normal breast tissues were collected and identified by pathology and ATR-FTIR spectroscopy respectively. The study indicates that the levels of the lipid-to-protein, nucleic acid-to-lipid, phosphate-to-carbohydrate and their secondary structure ratio, including RNA-to-DNA, Amide I-to-Amide II, and RNA-to-lipid ratios were significantly altered among the molecular subtype of breast tumour compared with normal breast tissues, which helps explain the changes in the biochemical structure of different molecular phenotypes of breast cancer. Tentatively-assigned characteristic peak ratios of infrared (IR) spectra reflect the changes of the macromolecule structure in different issues to a great extent and can be used as a potential biomarker to predict the molecular subtype of breast tumour. The present study acts as the first case study to show the successful application of IR spectroscopy in classifying subtypes of breast cancer with biochemical alterations. Therefore, the present study is likely to help to provide a new diagnostic approach for the accurate diagnosis of breast tumours and differential molecular subtypes and has the potential to be used for further intraoperative management.

Identification of potential serum biomarkers for simultaneously classifying lung adenocarcinoma, squamous cell carcinoma and small cell carcinoma.

Histological subtypes of lung cancer are crucial for making treatment decisions. However, multi-subtype classifications including adenocarcinoma (AC), squamous cell carcinoma (SqCC) and small cell carcinoma (SCLC) were rare in the previous studies. This study aimed at identifying and screening potential serum biomarkers for the simultaneous classification of AC, SqCC and SCLC. A total of 143 serum samples of AC, SqCC and SCLC were analyzed by 1HNMR and UPLC-MS/MS. The stepwise discriminant analysis (DA) and multilayer perceptron (MLP) were employed to screen the most efficient combinations of markers for classification. The results of non-targeted metabolomics analysis showed that the changes of metabolites of choline, lipid or amino acid might contribute to the classification of lung cancer subtypes. 17 metabolites in those pathways were further quantified by UPLC-MS/MS. DA screened out that serum xanthine, S-adenosyl methionine (SAM), carcinoembryonic antigen (CEA), neuron-specific enolase (NSE) and squamous cell carcinoma antigen (SCC) contributed significantly to the classification of AC, SqCC and SCLC. The average accuracy of 92.3% and the area under the receiver operating characteristic curve of 0.97 would be achieved by MLP model when a combination of those five variables as input parameters. Our findings suggested that metabolomics was helpful in screening potential serum markers for lung cancer classification. The MLP model established can be used for the simultaneous diagnosis of AC, SqCC and SCLC with high accuracy, which is worthy of further study.

An Accurate Tool for Uncovering Cancer Subtypes by Fast Kernel Learning Method to Integrate Multiple Profile Data.

In recent years, cancer has become a severe threat to human health. If we can accurately identify the subtypes of cancer, it will be of great significance to the research of anti-cancer drugs, the development of personalized treatment methods, and finally conquer cancer. In this paper, we obtain three feature representation datasets (gene expression profile, isoform expression and DNA methylation data) on lung cancer and renal cancer from the Broad GDAC, which collects the standardized data extracted from The Cancer Genome Atlas (TCGA). Since the feature dimension is too large, Principal Component Analysis (PCA) is used to reduce the feature vector, thus eliminating the redundant features and speeding up the operation speed of the classification model. By multiple kernel learning (MKL), we use Kernel target alignment (KTA), fast kernel learning (FKL), Hilbert-Schmidt Independence Criterion (HSIC), Mean to calculate the weight of kernel fusion. Finally, we put the combined kernel function into the support vector machine (SVM) and get excellent results. Among them, in the classification of renal cell carcinoma subtypes, the maximum accuracy can reach 0.978 by using the method of MKL (HSIC calculation weight), while in the classification of lung cancer subtypes, the accuracy can even reach 0.990 with the same method (FKL calculation weight).

SSIG: Single‐Sample Information Gain Model for Integrating Multi‐Omics Data to Identify Cancer Subtypes

Different living environments of cancer samples lead to different molecular mechanisms of cancer development, which in turn leads to different cancer subtypes. How to identify cancer subtypes is a key issue for the realization of precision medicine. With the development of high-throughput technologies, multi-omics data which can better understand different causes of cancer have emerged. However, the current methods of analyzing cancer subtypes using multi-omics data is mostly derived from population cancer sample data and ignores the differences between different cancer samples. Therefore, the joint analysis of multi-omics based on a single sample may reveal more information about the differences between individual cancers. A strategy for identifying cancer subtypes is proposed based on Single-sample information gain (SSIG) which construct sample feature matrix by considering the heterogeneity of sample. Applying this strategy to current popular subtype identification methods, cancer subtypes can be identified more accurately and the mechanism of cancer can be found from the perspective of a single sample. By comparing different methods in different clustering measure, and using survival analysis, it is shown that SSIG is more suitable for cancer subtype identification than the original multi-omics data, and it is easier to mine the cancer subtype classification mechanism hidden behind the data.