Colon Dataset Research Articles

Decoding the Methylation Patterns of ABC Transporters in Colorectal Cancer

Colorectal cancer (CRC) is a significant global health concern, posing a major threat to morbidity and mortality rates. The molecular mechanisms that drive CRC, particularly the DNA methylation patterns in ATP-binding cassette (ABC) genes, have attracted considerable attention because of their potential roles in CRC drug resistance, initiation, and progression. This study aimed to investigate the methylation patterns of ABC genes in CRC using a high-throughput microarray technology. Microarray methylation data from CRC and adjacent normal tissues were subjected to preprocessing, differential methylation analysis, and correlation analysis using the MethSurv tool for a detailed study of methylation patterns. The study revealed global hypomethylation of 43 ABC transporters and two pseudogenes in CRC compared to normal tissues. Receiver operating characteristic curve analysis identified 369 CpG sites as potential biomarkers for CRC diagnosis, with area under the curve values ranging from acceptable to outstanding. Survival analysis demonstrated the correlation between DNA methylation of individual CpG sites and patient survival probability in colon and rectal adenocarcinoma datasets from The Cancer Genome Atlas. The study provides insights into the epigenetic landscape of ABC genes in CRC, highlighting their potential roles in drug resistance, disease initiation, and progression. The findings offer opportunities for developing innovative therapeutic approaches and improving patient outcomes in the fight against CRC. Further research is needed to validate these results and investigate the functional implications of ABC gene methylation in CRC pathogenesis and treatment response.

WISE: Efficient WSI Selection for Active Learning in Histopathology

Deep neural network (DNN) models have been applied to a wide variety of medical image analysis tasks, often with the successful performance outcomes that match those of medical doctors. However, given that even minor errors in a model can impact patients’ life, it is critical that these models are continuously improved. Hence, active learning (AL) has garnered attention as an effective and sustainable strategy for enhancing DNN models for the medical domain. Extant AL research in histopathology has primarily focused on patch datasets derived from whole-slide images (WSIs), a standard form of cancer diagnostic images obtained from a high-resolution scanner. However, this approach has failed to address the selection of WSIs, which can impede the performance improvement of deep learning models and increase the number of WSIs needed to achieve the target performance.This study introduces a WSI-level AL method, termed WSI-informative selection (WISE). WISE is designed to select informative WSIs using a newly formulated WSI-level class distance metric. This method aims to identify diverse and uncertain cases of WSIs, thereby contributing to model performance enhancement. WISE demonstrates state-of-the-art performance across the Colon and Stomach datasets, collected in the real world, as well as the public DigestPath dataset, significantly reducing the required number of WSIs by more than threefold compared to the one-pool dataset setting, which has been dominantly used in the field.

Using multistate models with clinical trial data for a deeper understanding of complex disease processes.

A clinical trial represents a large commitment from all individuals involved and a huge financial obligation given its high cost; therefore, it is wise to make the most of all collected data by learning as much as possible. A multistate model is a generalized framework to describe longitudinal events; multistate hazards models can treat multiple intermediate/final clinical endpoints as outcomes and estimate the impact of covariates simultaneously. Proportional hazards models are fitted (one per transition), which can be used to calculate the absolute risks, that is, the probability of being in a state at a given time, the expected number of visits to a state, and the expected amount of time spent in a state. Three publicly available clinical trial datasets, colon, myeloid, and rhDNase, in the survival package in R were used to showcase the utility of multistate hazards models. In the colon dataset, a very well-known and well-used dataset, we found that the levamisole+fluorouracil treatment extended time in the recurrence-free state more than it extended overall survival, which resulted in less time in the recurrence state, an example of the classic "compression of morbidity." In the myeloid dataset, we found that complete response (CR) is durable, patients who received treatment B have longer sojourn time in CR than patients who received treatment A, while the mutation status does not impact the transition rate to CR but is highly influential on the sojourn time in CR. We also found that more patients in treatment A received transplants without CR, and more patients in treatment B received transplants after CR. In addition, the mutation status is highly influential on the CR to transplant transition rate. The observations that we made on these three datasets would not be possible without multistate models. We want to encourage readers to spend more time to look deeper into clinical trial data. It has a lot more to offer than a simple yes/no answer if only we, the statisticians, are willing to look for it.

Lesion-Decoupling-Based Segmentation With Large-Scale Colon and Esophageal Datasets for Early Cancer Diagnosis.

Lesions of early cancers often show flat, small, and isochromatic characteristics in medical endoscopy images, which are difficult to be captured. By analyzing the differences between the internal and external features of the lesion area, we propose a lesion-decoupling-based segmentation (LDS) network for assisting early cancer diagnosis. We introduce a plug-and-play module called self-sampling similar feature disentangling module (FDM) to obtain accurate lesion boundaries. Then, we propose a feature separation loss (FSL) function to separate pathological features from normal ones. Moreover, since physicians make diagnoses with multimodal data, we propose a multimodal cooperative segmentation network with two different modal images as input: white-light images (WLIs) and narrowband images (NBIs). Our FDM and FSL show a good performance for both single-modal and multimodal segmentations. Extensive experiments on five backbones prove that our FDM and FSL can be easily applied to different backbones for a significant lesion segmentation accuracy improvement, and the maximum increase of mean Intersection over Union (mIoU) is 4.58. For colonoscopy, we can achieve up to mIoU of 91.49 on our Dataset A and 84.41 on the three public datasets. For esophagoscopy, mIoU of 64.32 is best achieved on the WLI dataset and 66.31 on the NBI dataset.

Enhanced Medical Image Segmentation using Transfer Learning with Res101_UNet: Experimental Insights and Comparative Performance Analysis

Throughout the past few decades, artificial intelligence and machine learning have seen a lot of active research in areas such as computer vision, natural language processing, and speech processing. As a result, deep learning models became state-of-the-art for computer vision tasks such as object detection, classification, segmentation, and other allied tasks. Of course, the fruits of this research are extended to the design of robust and reliable digital health systems as well as other applications in the healthcare sector. Many clinical applications require the automatic segmentation of medical images. Recent deep learning-based approaches have demonstrated state-of-the-art performance in medical image segmentation tasks. In addition to their ability to automatically extract features and generalize over large amounts of data, transfer learning based deep learning models have proven to be handy for data scared areas like medical domains. In this research, we investigate and demonstrate the efficacy of a DCNN-based transfer learning model -Res101_Unet, which has been trained and/or fine-tuned to execute tumor tissue segmentation tasks in MRI, CT, PET, and X-RAY pictures of medical organ scans with little data. For our experimental study, we employed two image datasets: 'Liver Tumor' and 'Gland Colon Cancer', both obtained from the Kaggle portal. This experimental setup includes an Open-Source segmentation model API. Our findings indicate that domain similarity-based transfer learning can be used to data-scarce sectors. We achieved 98.47% accuracy and a IoU score of 0.9891 on Liver Tumor data and 0.6956 accuracy and a IoU score of 0.7043 on gland colon dataset.

Comparison of Matrix Decomposition in Null Space-Based LDA Method

Problems with small sample sizes and high dimensionality are common in pattern recognition. Almost all machine learning algorithms degrade in high-dimensional data, so that singularities in the scatter matrices, the main problem of the Linear Discriminant Analysis (LDA) technique, might result. A null space-based LDA (NLDA) has been conceived to address the singularity issue. NLDA aims to maximize the distance between classes in the null space of the within-class scatter matrix. In the first research, the NLDA method was performed by computing the eigenvalue decomposition and singular value decomposition (SVD). This research led to several new implementations of the NLDA method that use other matrix decompositions. The new implementations include NLDA using Cholesky decomposition and NLDA using QR decomposition. This paper compares the performance of three NLDA methods that use different matrix decompositions, namely, SVD, Cholesky decomposition, and QR decomposition. Two sets of data were used in the experiments that used three different NLDA algorithms. To determine the performance of the NLDA methods, the classification accuracy of the three methods was measured using the confusion matrix. The results show that the NLDA method using SVD has the best performance when compared to the other two methods, achieving a precision of 77.8% accuracy for the Colon dataset and a precision of 98.8% accuracy for the TKI-resistance dataset.

GammaGateR: semi-automated marker gating for single-cell multiplexed imaging.

Multiplexed immunofluorescence (mIF) is an emerging assay for multichannel protein imaging that can decipher cell-level spatial features in tissues. However, existing automated cell phenotyping methods, such as clustering, face challenges in achieving consistency across experiments and often require subjective evaluation. As a result, mIF analyses often revert to marker gating based on manual thresholding of raw imaging data. To address the need for an evaluable semi-automated algorithm, we developed GammaGateR, an R package for interactive marker gating designed specifically for segmented cell-level data from mIF images. Based on a novel closed-form gamma mixture model, GammaGateR provides estimates of marker-positive cell proportions and soft clustering of marker-positive cells. The model incorporates user-specified constraints that provide a consistent but slide-specific model fit. We compared GammaGateR against the newest unsupervised approach for annotating mIF data, employing two colon datasets and one ovarian cancer dataset for the evaluation. We showed that GammaGateR produces highly similar results to a silver standard established through manual annotation. Furthermore, we demonstrated its effectiveness in identifying biological signals, achieved by mapping known spatial interactions between CD68 and MUC5AC cells in the colon and by accurately predicting survival in ovarian cancer patients using the phenotype probabilities as input for machine learning methods. GammaGateR is a highly efficient tool that can improve the replicability of marker gating results, while reducing the time of manual segmentation. The R package is available at https://github.com/JiangmeiRubyXiong/GammaGateR.

UViT-Seg: An Efficient ViT and U-Net-Based Framework for Accurate Colorectal Polyp Segmentation in Colonoscopy and WCE Images.

Colorectal cancer (CRC) stands out as one of the most prevalent global cancers. The accurate localization of colorectal polyps in endoscopy images is pivotal for timely detection and removal, contributing significantly to CRC prevention. The manual analysis of images generated by gastrointestinal screening technologies poses a tedious task for doctors. Therefore, computer vision-assisted cancer detection could serve as an efficient tool for polyp segmentation. Numerous efforts have been dedicated to automating polyp localization, with the majority of studies relying on convolutional neural networks (CNNs) to learn features from polyp images. Despite their success in polyp segmentation tasks, CNNs exhibit significant limitations in precisely determining polyp location and shape due to their sole reliance on learning local features from images. While gastrointestinal images manifest significant variation in their features, encompassing both high- and low-level ones, a framework that combines the ability to learn both features of polyps is desired. This paper introduces UViT-Seg, a framework designed for polyp segmentation in gastrointestinal images. Operating on an encoder-decoder architecture, UViT-Seg employs two distinct feature extraction methods. A vision transformer in the encoder section captures long-range semantic information, while a CNN module, integrating squeeze-excitation and dual attention mechanisms, captures low-level features, focusing on critical image regions. Experimental evaluations conducted on five public datasets, including CVC clinic, ColonDB, Kvasir-SEG, ETIS LaribDB, and Kvasir Capsule-SEG, demonstrate UViT-Seg's effectiveness in polyp localization. To confirm its generalization performance, the model is tested on datasets not used in training. Benchmarking against common segmentation methods and state-of-the-art polyp segmentation approaches, the proposed model yields promising results. For instance, it achieves a mean Dice coefficient of 0.915 and a mean intersection over union of 0.902 on the CVC Colon dataset. Furthermore, UViT-Seg has the advantage of being efficient, requiring fewer computational resources for both training and testing. This feature positions it as an optimal choice for real-world deployment scenarios.

MMUNet: Morphological feature enhancement network for colon cancer segmentation in pathological images

Colon image analysis is an important step in diagnosing colon cancer, and achieving automated and accurate segmentation remains a challenging problem because of the diversity of cell shapes and boundaries in pathological sections. In this paper, we propose a U-shaped colon cancer segmentation network, which combines depth-separable convolution and morphological methods to reduce the number of model parameters and effectively improve segmentation accuracy. We improve the global and local feature capabilities by taking advantage of serial convolution and external focus as the underlying architecture for the model. We designed the skip connection to fuse the features from the encoder in a morphological way to enhance the morphological features. We introduced an edge enhancement module by extracting contour information using morphological methods to enhance edge features. We evaluated the proposed method on three colon cancer datasets, and the experimental results showed that our method with a small number of parameters has a Dice coefficient of 92.76% ± 5.86% on the Glas dataset, 86.11% ± 7.11% on the CoCaHis dataset, and 91.61% ± 11.25% on the Colon dataset. The code will be openly available at https://github.com/Yuanhaojun513/MMUNet.

Subspace learning using structure learning and non-convex regularization: Hybrid technique with mushroom reproduction optimization in gene selection

Gene selection as a problem with high dimensions has drawn considerable attention in machine learning and computational biology over the past decade. In the field of gene selection in cancer datasets, different types of feature selection techniques in terms of strategy (filter, wrapper and embedded) and label information (supervised, unsupervised, and semi-supervised) have been developed. However, using hybrid feature selection can still improve the performance. In this paper, we propose a hybrid feature selection based on filter and wrapper strategies. In the filter-phase, we develop an unsupervised features selection based on non-convex regularized non-negative matrix factorization and structure learning, which we deem NCNMFSL. In the wrapper-phase, for the first time, mushroom reproduction optimization (MRO) is leveraged to obtain the most informative features subset. In this hybrid feature selection method, irrelevant features are filtered-out through NCNMFSL, and most discriminative features are selected by MRO. To show the effectiveness and proficiency of the proposed method, numerical experiments are conducted on Breast, Heart, Colon, Leukemia, Prostate, Tox-171 and GLI-85 benchmark datasets. SVM and decision tree classifiers are leveraged to analyze proposed technique and top accuracy are 0.97, 0.84, 0.98, 0.95, 0.98, 0.87 and 0.85 for Breast, Heart, Colon, Leukemia, Prostate, Tox-171 and GLI-85, respectively. The computational results show the effectiveness of the proposed method in comparison with state-of-art feature selection techniques.

Sparse-to-dense coarse-to-fine depth estimation for colonoscopy

Colonoscopy, as the golden standard for screening colon cancer and diseases, offers considerable benefits to patients. However, it also imposes challenges on diagnosis and potential surgery due to the narrow observation perspective and limited perception dimension. Dense depth estimation can overcome the above limitations and offer doctors straightforward 3D visual feedback. To this end, we propose a novel sparse-to-dense coarse-to-fine depth estimation solution for colonoscopic scenes based on the direct SLAM algorithm. The highlight of our solution is that we utilize the scattered 3D points obtained from SLAM to generate accurate and dense depth in full resolution. This is done by a deep learning (DL)-based depth completion network and a reconstruction system. The depth completion network effectively extracts texture, geometry, and structure features from sparse depth along with RGB data to recover the dense depth map. The reconstruction system further updates the dense depth map using a photometric error-based optimization and a mesh modeling approach to reconstruct a more accurate 3D model of colons with detailed surface texture. We show the effectiveness and accuracy of our depth estimation method on near photo-realistic challenging colon datasets. Experiments demonstrate that the strategy of sparse-to-dense coarse-to-fine can significantly improve the performance of depth estimation and smoothly fuse direct SLAM and DL-based depth estimation into a complete dense reconstruction system.

A gravity inspired clustering algorithm for gene selection from high-dimensional microarray data

ABSTRACT The problem of selecting a subset of genes relevant to a particular disease classification task from high-dimensional gene microarray data has been an area of active research in the scientific community. A wide variety of methods ranging from assigning scores to genes based on their interactions with the output to the Gravity inspired clustering algorithm with the hybrid approach have been proposed. The feature sets selected by this approach have been evaluated by using the Support Vector Machine (SVM) with Leave One Out Cross Validation (LOOCV). The method has been evaluated on twelve different publicly available microarray cancer datasets and the proposed method provided good accuracy of 90.16% to the colon dataset, 98.59% to the ALL-AML dataset, 97.03% to the lung cancer dataset, 93.07% to the prostate cancer dataset, 98.41% to the meningioma dataset, 83.72% to the brain tumour dataset and has been shown to compete with existing methods.

An Atrous Convolved Hybrid Seg-Net Model with residual and attention mechanism for gland detection and segmentation in histopathological images

PurposeA clinically compatible computerized segmentation model is presented here that aspires to supply clinical gland informative details by seizing every small and intricate variation in medical images, integrate second opinions, and reduce human errors. ApproachIt comprises of enhanced learning capability that extracts denser multi-scale gland-specific features, recover semantic gap during concatenation, and effectively handle resolution-degradation and vanishing gradient problems. It is having three proposed modules namely Atrous Convolved Residual Learning Module in the encoder as well as decoder, Residual Attention Module in the skip connection paths, and Atrous Convolved Transitional Module as the transitional and output layer. Also, pre-processing techniques like patch-sampling, stain-normalization, augmentation, etc. are employed to develop its generalization capability. To verify its robustness and invigorate network invariance against digital variability, extensive experiments are carried out employing three different public datasets i.e., GlaS (Gland Segmentation Challenge), CRAG (Colorectal Adenocarcinoma Gland) and LC-25000 (Lung Colon-25000) dataset and a private HosC (Hospital Colon) dataset. ResultsThe presented model accomplished combative gland detection outcomes having F1-score (GlaS(Test A(0.957), Test B(0.926)), CRAG(0.935), LC 25000(0.922), HosC(0.963)); and gland segmentation results having Object-Dice Index (GlaS(Test A(0.961), Test B(0.933)), CRAG(0.961), LC-25000(0.940), HosC(0.929)), and Object-Hausdorff Distance (GlaS(Test A(21.77) and Test B(69.74)), CRAG(87.63), LC-25000(95.85), HosC(83.29)). In addition, validation score (GlaS (Test A(0.945), Test B(0.937)), CRAG(0.934), LC-25000(0.911), HosC(0.928)) supplied by the proficient pathologists is integrated for the end segmentation results to corroborate the applicability and appropriateness for assistance at the clinical level applications. ConclusionThe proposed system will assist pathologists in devising precise diagnoses by offering a referential perspective during morphology assessment of colon histopathology images.

Histopathological carcinoma classification using parallel, cross‐concatenated and grouped convolutions deep neural network

AbstractCancer is more alarming in modern days due to its identification at later stages. Among cancers, lung, liver and colon cancers are the leading cause of untimely death. Manual cancer identification from histopathological images is time‐consuming and labour‐intensive. Thereby, computer‐aided decision support systems are desired. A deep learning model is proposed in this paper to accurately identify cancer. Convolutional neural networks have shown great ability to identify the significant patterns for cancer classification. The proposed Parallel, Cross Concatenated and Grouped Convolutions Deep Neural Network (PC2GCDN2) has been developed to obtain accurate patterns for classification. To prove the robustness of the model, it is evaluated on the KMC and TCGA‐LIHC liver dataset, LC25000 dataset for lung and colon cancer classification. The proposed PC2GCDN2 model outperforms states‐of‐the‐art methods. The model provides 5.5% improved accuracy compared to the LiverNet proposed by Aatresh et. al on the KMC dataset. On the LC25000 dataset, 2% improvement is observed compared to existing models. Performance evaluation metrics like Sensitivity, Specificity, Recall, F1‐Score and Intersection‐Over‐Union are used to evaluate the performance. To the best of our knowledge, PC2GCDN2 can be considered as gold standard for multiple histopathology image classification. PC2GCDN is able to classify the KMC and TCGA‐LIHC liver dataset with 96.4% and 98.6% accuracy, respectively, which are the best results obtained till now. The performance has been superior on LC25000 dataset with 99.5% and 100% classification accuracy on lung and colon dataset, by utilizing less than 0.5 million parameters.

Analysis of Proposed and Traditional Boosting Algorithm with Standalone Classification Methods for Classifying Gene Expresssion Microarray Data Using a Reject Option

ABSTRACT In medical field, accurate decisions are very important as they risk human lives. decision support system (DSS) plays important role in making accurate decisions and used for classification/prediction. In gene expression analysis, genes are not only inflated by the external environmental conditions but also the expression values of certain genes are affected (like cancer, obesity etc). in this study, various traditional (Support Vector Machine, Decision Trees, and Linear Discriminant Analysis, naïve Bayes, logistic regression, and multilayer perceptron) and proposed methods (combination of traditional with ensemble and probabilistic classifiers) are used in order to perform the classification and prediction analysis. In this study we used the publicly available datasets comprised of Lymphoid, Leukemia and Colon Cancer. The classification performance on Colon dataset with traditional methods was obtained with accuracy (56%) and proposed probabilistic ensemble methods with accuracy (88%). For dataset, Leukemia, the accuracy was obtained using traditional methods (78%) and proposed methods (92%). Similarly, on Lymphoid dataset, the traditional methods yielded accuracy (75%) and proposed methods (87%). The results revealed that proposed methods yielded the improved detection performance. The proposed methods can be used as a better predictor for early diagnosis and improved diagnosis to improve the healthcare systems.

Dysregulated expression of suppressor loop of circadian rhythm genes in colorectal cancer pathogenesis.

Colorectal cancer (CRC) is a heterogeneous disease and activation of WNT and TGFβ mediated oncogenic pathways is frequently observed in this pathology. However, to date, limited reports have been published addressing the association of circadian clock with CRC pathogenesis and stratification. The current study aims at assessing the expression of important circadian markers, PER2, PER3 and NR1D1, in independent CRC cohorts and their associations with CRC-related pathways. Gene expression analysis was performed using available GEO (GSE39582) and TCGA datasets. Quantitative real time polymerase chain reaction was used to quantify the expression levels of PER2, PER3 and NRID1 in FFPE (formalin fixed paraffin embedded) CRC tissue samples. Furthermore, enrichment of circadian markers in WNT and TGFβ pathways-activated tumors was assessed. Statistically significant downregulation of PER3 was found in tumor versus control samples in GEO (P<0.0001) and TCGA colon and rectal adenocarcinoma datasets (P<0.05). Analysis of GEO dataset revealed a statistically significant upregulation of PER2 (P<0.01), and NR1D1 in colon adenocarcinoma, which was confirmed by qRT-PCR in CRC tumor samples versus controls in FFPE validation cohort. Higher expression of NR1D1 was associated with poor prognosis in colon adenocarcinoma. Contrastingly, PER3 was significantly downregulated in tumors (P<0.001) compared to controls and was associated with high-grade CRC tumors versus low-grade tumors. Tumors with WNT pathway activation had significantly low PER3 and slightly upregulated PER2 (<0.0001) expression. Interestingly, differential expression of PER3 and NR1D1 was significantly correlated with TGFβ1-expressing tumors (P<0.0001). Moreover, MYC- amplified tumors exhibited decreased PER3 levels. Thus, low PER3 expression in CRC and poor survival of patients with NR1D1-high tumors reveal that genes in the suppressor loop of circadian rhythm are dysregulated in CRC, hence pointing out to the importance of dissecting the circadian pathway in cancer.

Abstract 4081: Leveraging single-cell RNA sequencing data to design multi-targeting SeekRRNA therapeutics

Abstract SeekR™ chimeric molecules consist of dual flanking RNA aptamers designed to bind specific proteins (BINDERS), joined by a double stranded RNA bridge that encodes multiple siRNAs designed to silence select disease target genes (SILENCERS). Aptamers are folded nucleic acid chains that act much like chemical antibodies in that they can bind proteins in a sequence-specific manner. By identifying RNA aptamers capable of seeking and binding to specific cell surface protein receptors, SeekR™ RNA therapeutics can be engineered to self-deliver to a particular receptor defined tissue type. For example, a SeekR™ can be directed to prostate tumors that selectively express PSMA. The combination of two RNA aptamer binders with two siRNA silencers thus enables the selective delivery of the siRNAs to tumor cells expressing the binder target, wherein it directs sequence-specific silencing of cancer target genes. To identify and prioritize the right combination of SeekR targets, we have taken advantage of single-cell sequencing (scRNA-seq) data sets. By analyzing scRNA-seq data, we can identify specific cell populations within tumors to co-target with an aptamer and siRNA. We leveraged several publicly available colon and non-small cell lung cancer scRNA data sets for this analysis. A custom analysis pipeline was applied to these data to identify specific cell types within tumor samples. We then profiled specific cancer targets and determined percentages of target co-expression within a single tumor compared to normal cells. This provides information for cell-type-specific targeting to reduce potential toxicity within non-tumor tissue. In addition to targeting aptamers, we performed ligand-receptor analyses to prioritize specific siRNA targets in particular cell types within the tumor environment. This analysis revealed novel combinations of binders and silencers that will more precisely deliver the right multi-targeted gene silencing to the correct cell types. In summary, scRNA-seq data provide a rich resource of data to direct SeekR™ development and optimization. Citation Format: Susan Massey, Richard Hsiao, Shweta Yadav, David Azorsa, Spyro Mousses, Jeffrey A. Kiefer. Leveraging single-cell RNA sequencing data to design multi-targeting SeekRRNA therapeutics [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 4081.

Abstract 3759: Recurrent methylation heterogeneity and reduced expression of SPG20 is a common characteristic of different tumors

Abstract Background: The Spastic Paraplegia-20 (SPG20, also known as SPART), is a gene that encodes spartin protein. When SPG20 is mutated, spartin expression is downregulated and leads to a rare autosomal recessive disorder called Troyer Syndrome. We recently demonstrated that this gene is epigenetically silenced and downregulated in Diffuse Large B Cell Lymphoma (DLBCL) cell lines. In recent years, other authors demonstrated independently the epigenetic regulation of SPG20 in hepatocellular, gastric and colorectal cancers. In order to assess whether the differential methylation of SPG20 is a common characteristic of tumors of different histological origin, we analysed the methylation of 9 probes upstream of the transcription starting site (TSS) and the gene expression of SPG20 in tumor samples affecting 6 different anatomical sites. Methods: Illumina Infinium Human Methylation 450 BeadChip and RNAseq data extrapolated from The Cancer Genome Atlas (TCGA) database were used to evaluate the methylation levels and expression of SPG20. Bladder, Colon, Kidney, Liver, Lung and Prostate datasets were selected. At least 30 independent tumor and at least 5 normal tissue samples for each tumor site were analysed. The tumor and normal tissues were paired in each specific cohort. Transcriptome profiling was analysed only in datasets where both tumor and normal counterparts were available. Statistical analysis was performed by two-way ANOVA with a Sidak’s post hoc test for methylation analysis and multiple t-test with Holm-Sidak’s post hoc test for expression analysis. Results: SPG20 β-values ranging from 0.19 to 0.96 among the different tumor samples were observed in the promoter region covered by the 9 probes that we previously identified (Illumina 450k platform). In the normal group, the β-values were higher than 0.63. Statistically significant differences between the tumor and normal counterpart were observed in colon, liver and prostate and mainly in the probes closest the TSS. The heterogeneity in the methylation status of SPG20 was a common characteristic among the different tumor groups. The variability range of β-values between tumor and normal groups was significantly different with a p value &amp;lt;0.05. Differential expression was observed between tumor and normal counterpart in almost all sites. SPG20 was more expressed in normal than in tumor tissues and the differences were significant in bladder, lung and colon. Conclusion: Our analysis unveils that the methylation heterogeneity of SPG20 is a common hallmark in all the tumor samples, while the normal counterpart results homogenously distributed. In tumor samples there is an inverse correlation between SPG20 promoter methylation heterogeneity and gene expression. SPG20 is down-regulated in almost all the tumors of different histological origins. SPG20 could have a relavant but still undefined role in cancer as a tumor-suppressor gene. Citation Format: Vincenza Ylenia Cusenza, Raffaele Frazzi. Recurrent methylation heterogeneity and reduced expression of SPG20 is a common characteristic of different tumors [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 3759.

Machine learning-based lung and colon cancer detection using deep feature extraction and ensemble learning

Cancer is a fatal disease caused by a combination of genetic diseases and a variety of biochemical abnormalities. Lung and colon cancer have emerged as two of the leading causes of death and disability in humans. The histopathological detection of such malignancies is usually the most important component in determining the best course of action. Early detection of the ailment on either front considerably decreases the likelihood of mortality. Machine learning and deep learning techniques can be utilized to speed up such cancer detection, allowing researchers to study a large number of patients in a much shorter amount of time and at a lower cost. In this research work, we introduced a hybrid ensemble feature extraction model to efficiently identify lung and colon cancer. It integrates deep feature extraction and ensemble learning with high-performance filtering for cancer image datasets. The model is evaluated on histopathological (LC25000) lung and colon datasets. According to the study findings, our hybrid model can detect lung, colon, and (lung and colon) cancer with accuracy rates of 99.05%, 100%, and 99.30%, respectively. The study’s findings show that our proposed strategy outperforms existing models significantly. Thus, these models could be applicable in clinics to support the doctor in the diagnosis of cancers.

Real-time semantic segmentation of gastric intestinal metaplasia using a deep learning approach.

Background/AimsPrevious artificial intelligence (AI) models attempting to segment gastric intestinal metaplasia (GIM) areas have failed to be deployed in real-time endoscopy due to their slow inference speeds. Here, we propose a new GIM segmentation AI model with inference speeds faster than 25 frames per second that maintains a high level of accuracy. MethodsInvestigators from Chulalongkorn University obtained 802 histological-proven GIM images for AI model training. Four strategies were proposed to improve the model accuracy. First, transfer learning was employed to the public colon datasets. Second, an image preprocessing technique contrast-limited adaptive histogram equalization was employed to produce clearer GIM areas. Third, data augmentation was applied for a more robust model. Lastly, the bilateral segmentation network model was applied to segment GIM areas in real time. The results were analyzed using different validity values. ResultsFrom the internal test, our AI model achieved an inference speed of 31.53 frames per second. GIM detection showed sensitivity, specificity, positive predictive, negative predictive, accuracy, and mean intersection over union in GIM segmentation values of 93%, 80%, 82%, 92%, 87%, and 57%, respectively. ConclusionsThe bilateral segmentation network combined with transfer learning, contrast-limited adaptive histogram equalization, and data augmentation can provide high sensitivity and good accuracy for GIM detection and segmentation.