scholarly journals Deep learning quantified mucus-tumor ratio predicting survival of patients with colorectal cancer using whole-slide images

2021 ◽  
Author(s):  
Ke Zhao ◽  
Lin Wu ◽  
Yanqi Huang ◽  
Su Yao ◽  
Zeyan Xu ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Deep Learning ◽  
Prognostic Value ◽  
Validation Cohort ◽  
Survival Rates ◽  
Training Cohort ◽  
Mucinous Component ◽  
Hematoxylin And Eosin ◽  
High Consistency ◽  
Whole Slide Images

Abstract Background In colorectal cancer (CRC), mucinous adenocarcinoma differs from other adenocarcinomas in gene-phenotype, morphology, and prognosis. However, mucinous components are present in a large amount of adenocarcinoma, and the prognostic value of mucus proportion has not been investigated. Artificial intelligence provides a way to quantify mucus proportion on whole-slide images (WSIs) accurately. We aimed to quantify mucus proportion by deep learning and further investigate its prognostic value in two CRC patients cohorts. Methods Deep learning was used to segment WSIs stained with hematoxylin and eosin. Mucus-tumor ratio (MTR) was defined as the proportion of mucinous component in the tumor area. A training cohort (N = 419) and a validation cohort (N = 315) were used to evaluate the prognostic value of MTR. Survival analysis was performed by the Cox proportional hazard model. Result Patients were stratified to mucus-low and mucus-high groups by 24.1% as the threshold. In the training cohort, patients with mucus-high had unfavorable outcomes (hazard ratio for high vs. low 1.88, 95% confidence interval 1.18-2.99, P = 0.008), with 5-year overall survival rates of 54.8% and 73.7% in mucus-high and mucus-low groups, respectively. The results were confirmed in the validation cohort (2.09, 1.21-3.60, 0.008; 79.8% vs. 62.8%). The prognostic value of MTR was maintained in multivariate analysis for both cohorts. Conclusion The deep learning quantified MTR was an independent prognostic factor in CRC. With the advantages of advanced efficiency and high consistency, our method is suitable for clinical application and promotes precision medicine development.

Identifying an immune-related gene-pair for prognosis prediction of metastatic colorectal cancer.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. e15565-e15565
Author(s):  
Qiqi Zhu ◽  
Du Cai ◽  
Wei Wang ◽  
Min-Er Zhong ◽  
Dejun Fan ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
High Risk ◽  
Prognostic Value ◽  
Validation Cohort ◽  
Risk Group ◽  
High Risk Group ◽  
Related Gene ◽  
Training Cohort ◽  
Gene Pairs ◽  
Immune Related Gene

e15565 Background: Few robust predictive biomarkers have been applied in clinical practice due to the heterogeneity of metastatic colorectal cancer (mCRC) . Using the gene pair method, the absolute expression value of genes can be converted into the relative order of genes, which can minimize the influence of the sequencing platform difference and batch effects, and improve the robustness of the model. The main objective of this study was to establish an immune-related gene pairs signature (IRGPs) and evaluate the impact of the IRGPs in predicting the prognosis in mCRC. Methods: A total of 205 mCRC patients containing overall survival (OS) information from the training cohort ( n = 119) and validation cohort ( n = 86) were enrolled in this study. LASSO algorithm was used to select prognosis related gene pairs. Univariate and multivariate analyses were used to validate the prognostic value of the IRGPs. Gene sets enrichment analysis (GSEA) and immune infiltration analysis were used to explore the underlying biological mechanism. Results: An IRGPs signature containing 22 gene pairs was constructed, which could significantly separate patients of the training cohort ( n = 119) and validation cohort ( n = 86) into the low-risk and high-risk group with different outcomes. Multivariate analysis with clinical factors confirmed the independent prognostic value of IRGPs that higher IRGPs was associated with worse prognosis (training cohort: hazard ratio (HR) = 10.54[4.99-22.32], P < 0.001; validation cohort: HR = 3.53[1.24-10.08], P = 0.012). GSEA showed that several metastasis and immune-related pathway including angiogenesis, TGF-β-signaling, epithelial-mesenchymal transition and inflammatory response were enriched in the high-risk group. Through further analysis of the immune factors, we found that the proportions of CD4+ memory T cell, regulatory T cell, and Myeloid dendritic cell were significantly higher in the low-risk group, while the infiltrations of the Macrophage (M0) and Neutrophil were significantly higher in the high-risk group. Conclusions: The IRGPs signature could predict the prognosis of mCRC patients. Further prospective validations are needed to confirm the clinical utility of IRGPs in the treatment decision.

scholarly journals The Prognostic Significance of Tumor Deposit Count for Colorectal Cancer Patients after Radical Surgery

2020 ◽  
Vol 2020 ◽  
pp. 1-8 ◽  
Author(s):  
Kuo Zheng ◽  
Nanxin Zheng ◽  
Cheng Xin ◽  
Leqi Zhou ◽  
Ge Sun ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Prognostic Value ◽  
Validation Cohort ◽  
Cutoff Point ◽  
Disease Specific Survival ◽  
Optimal Cutoff ◽  
Training Cohort ◽  
Tumor Deposit ◽  
Optimal Cutoff Point ◽  
Disease Specific

Background. The prognostic value of tumor deposit (TD) count in colorectal cancer (CRC) patients has been rarely evaluated. This study is aimed at exploring the prognostic value of TD count and finding out the optimal cutoff point of TD count to differentiate the prognoses of TD-positive CRC patients. Method. Patients diagnosed with CRC from Surveillance, Epidemiology, and End Results (SEER) database from January 1, 2010, to December 31, 2012, were analyzed. X-tile program was used to identify the optimal cutoff point of TD count in training cohort, and a validation cohort was used to test this cutoff point after propensity score matching (PSM). Univariate and multivariate Cox proportional hazard models were used to assess the risk factors of survival. Results. X-tile plots identified 3 (P<0.001) as the optimal cutoff point of TD count to divide the patients of training cohort into high and low risk subsets in terms of disease-specific survival (DSS). This cutoff point was validated in validation cohort before and after PSM (P<0.001, P=0.002). More TD count, which was defined as more than 3, was validated as an independent risk prognostic factor in univariate and multivariate analysis (P<0.001). Conclusion. More TD count (TD count≥4) was significantly associated with poor disease-specific survival in CRC patients.

scholarly journals Long non-coding RNA profile study identifies a metabolism-related signature for colorectal cancer

2021 ◽  
Vol 27 (1) ◽  
Author(s):  
Yongqu Lu ◽  
Wendong Wang ◽  
Zhenzhen Liu ◽  
Junren Ma ◽  
Xin Zhou ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Risk Score ◽  
Roc Curve ◽  
Prognostic Value ◽  
Clinical Decision Making ◽  
Validation Cohort ◽  
Gene Set Enrichment Analysis ◽  
Training Cohort ◽  
Non Coding Rna ◽  
Long Non Coding Rna

Abstract Background Heterogeneity in colorectal cancer (CRC) patients provides novel strategies in clinical decision-making. Identifying distinctive subgroups in patients can improve the screening of CRC and reduce the cost of tests. Metabolism-related long non-coding RNA (lncRNA) can help detection of tumorigenesis and development for CRC patients. Methods RNA sequencing and clinical data of CRC patients which extracted and integrated from public databases including The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) were set as training cohort and validation cohort. Metabolism-related genes were acquired from Kyoto Encyclopedia of Genes and Genomes (KEGG) and the metabolism-related lncRNAs were filtered using correlation analysis. The risk score was calculated based on lncRNAs with prognostic value and verified through survival curve, receiver operating characteristic (ROC) curve and risk curve. Prognostic factors of CRC patients were also analyzed. Nomogram was constructed based on the results of cox regression analyses. The different immune status was observed in the single sample Gene Set Enrichment Analysis (ssGSEA). Results The training cohort and the validation cohort enrolled 432 and 547 CRC patients respectively. A total of 23 metabolism-related lncRNAs with prognostic value were screened out and 10 of which were significantly differentially expressed between tumour and normal tissues. Finally, 8 lncRNAs were used to establish a risk score (DICER1-AS1, PCAT6, GAS5, PRR7-AS1, MCM3AP-AS1, GAS6-AS1, LINC01082 and ADIRF-AS1). Patients were divided into high-risk and low-risk groups according to the median of risk scores in training cohort and the survival curves indicated that the survival prognosis was significantly different. The area under curve (AUC) of the ROC curve in two cohorts were both greater than 0.6. The age, tumour stage and risk score were selected as independent factors and used to construct a nomogram to predict CRC patients' survival rate with the c-index of 0.806. The ssGSEA indicated that the risk score was associated with immune cells and functions. Conclusions Our systematic study established a metabolism-related lncRNA signature to predict outcomes of CRC patients which may contribute to individual prevention and treatment.

scholarly journals The Prognostic Value of Preoperative Serum Lactate Dehydrogenase (LDH) Levels in Patients Underwent Curative-intent Hepatectomy For Colorectal Liver Metastases: A Two-Center Cohort Study

2021 ◽  
Author(s):  
Long Bai ◽  
Ze-Yu Lin ◽  
Yun-Xin Lu ◽  
Qin Chen ◽  
Han Zhou ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Lactate Dehydrogenase ◽  
Cancer Patients ◽  
Prognostic Value ◽  
Colorectal Liver Metastases ◽  
Validation Cohort ◽  
Curative Intent ◽  
Training Cohort ◽  
Preoperative Serum ◽  
Colorectal Cancer Patients

Abstract Background: The prognostic value of lactate dehydrogenase (LDH) in colorectal cancer patients has remained inconsistent between non-metastatic and metastatic settings. So far very few studies have included LDH in prognostic analysis for patients with colorectal liver metastases (CRLM) who underwent curative-intent hepatectomy. Patients and Methods: Consecutive metastatic colorectal cancer patients who underwent curative-intent resection for CRLM from two Chinese medical centers treated in 2000-2019 were enrolled in the training cohort (434 patients) and the validation cohort (146 patients). Overall survival (OS) was the primary endpoint. Cox regression model was performed to identify the prognostic values of LDH and other clinicopathology variables. A modification of the established Fong scoring system comprising LDH was developed within this Chinese population.Results: In the training cohort, preoperative LDH > upper limit of normal (ULN) was the strongest independent prognostic factor both for RFS (HR 2.11, 95% confidence intervals [CIs], 1.54-2.89; P < .001) and OS (HR 2.41, 95% CI, 1.72-3.39; P < .001) in multivariate analysis. 5-year survival rates were 23.7% and 52.9% in the LDH > ULN group and LDH < ULN group, respectively. These data were also confirmed in the validation cohort and then in pooled cohort. Replacing carcinoembryonic antigen (CEA) with LDH in the Fong score contributed to an improvement in the predictive value.Conclusions: Preoperative serum LDH is a reliable and independent predictor for curative-intent CRLM resection. Composite of LDH and Fong score is a potential stratification tool for CRLM resection.

scholarly journals Noninvasive KRAS mutation estimation in colorectal cancer using a deep learning method based on routine CT imaging

2020 ◽  
Author(s):  
kan He ◽  
Xiaoming Liu ◽  
Mingyang Li ◽  
Xueyan Li ◽  
Hualin Yang ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Deep Learning ◽  
Kras Mutation ◽  
Validation Cohort ◽  
Ct Images ◽  
Ct Imaging ◽  
Kras Mutations ◽  
Training Cohort ◽  
Contrast Enhanced Ct ◽  
Pre Treatment

Abstract Background: The detection of KRAS gene mutations in colorectal cancer (CRC) is key to the optimal design of individualized therapeutic strategies. The noninvasive prediction of the KRAS status in CRC is challenging. Deep learning (DL) in medical imaging has shown its high performance in diagnosis, classification, and prediction in recent years. In this paper, we investigated predictive performance by using a DL method with a residual neural network ( ResNet ) to estimate the KRAS mutation status in CRC patients based on routine pre-treatment contrast-enhanced CT imaging. Methods: We have collected a dataset consisting of 157 patients with pathology-confirmed CRC who were randomly divided into a training cohort (n = 117) and a validation cohort (n = 40). We developed an ResNet model that used portal venous phase CT images to estimate KRAS mutations in the axial, coronal, and sagittal directions of the training cohort and validated the model in the validation cohort. Several groups of expended ROI patches were generated for the ResNet model, to explore whether tissues around the tumor can contribute to cancer assessment. We also explored a radiomics model with the random forest classifier (RFC) to predict KRAS mutations and compared it with the DL model. Results: The ResNet model in the axial direction achieved the higher area under the curve (AUC) value (0.90) in the validation cohort and peaked at 0.93 with an input of “ROI and 20-pixel” surrounding area. In the training cohort, the AUC was 0.945 (sensitivity: 0.75; specificity: 0.94), and in the validation cohort, the AUC was0.818 (sensitivity: 0.70; specificity: 0.85). In comparison, the ResNet model showed better predictive ability . Conclusions: Our experiments reveal that the computerized assessment of the pre-treatment CT images of CRC patients using a DL model has the potential to precisely predict KRAS mutations. This new model has the potential to assist in noninvasive KRAS mutation estimation. Keywords: Colorectal Neoplasm, Mutation, Deep Learning

scholarly journals Two Ensemble-CNN Approaches for Colorectal Cancer Tissue Type Classification

2021 ◽  
Vol 7 (3) ◽  
pp. 51
Author(s):  
Emanuela Paladini ◽  
Edoardo Vantaggiato ◽  
Fares Bougourzi ◽  
Cosimo Distante ◽  
Abdenour Hadid ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Deep Learning ◽  
Digital Pathology ◽  
Texture Features ◽  
Tissue Type ◽  
Cancer Tissue ◽  
Learning Methods ◽  
Feature Based ◽  
Type Classification ◽  
Whole Slide Images

In recent years, automatic tissue phenotyping has attracted increasing interest in the Digital Pathology (DP) field. For Colorectal Cancer (CRC), tissue phenotyping can diagnose the cancer and differentiate between different cancer grades. The development of Whole Slide Images (WSIs) has provided the required data for creating automatic tissue phenotyping systems. In this paper, we study different hand-crafted feature-based and deep learning methods using two popular multi-classes CRC-tissue-type databases: Kather-CRC-2016 and CRC-TP. For the hand-crafted features, we use two texture descriptors (LPQ and BSIF) and their combination. In addition, two classifiers are used (SVM and NN) to classify the texture features into distinct CRC tissue types. For the deep learning methods, we evaluate four Convolutional Neural Network (CNN) architectures (ResNet-101, ResNeXt-50, Inception-v3, and DenseNet-161). Moreover, we propose two Ensemble CNN approaches: Mean-Ensemble-CNN and NN-Ensemble-CNN. The experimental results show that the proposed approaches outperformed the hand-crafted feature-based methods, CNN architectures and the state-of-the-art methods in both databases.

scholarly journals Accurate recognition of colorectal cancer with semi-supervised deep learning on pathological images

2020 ◽  
Author(s):  
Gang Yu ◽  
Ting Xie ◽  
Chao Xu ◽  
Xing-Hua Shi ◽  
Chong Wu ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Deep Learning ◽  
Supervised Learning ◽  
Medical Sciences ◽  
Patient Level ◽  
General Utility ◽  
Significant Difference ◽  
Multi Center Study ◽  
The Cost ◽  
Whole Slide Images

Abstract Background: The machine-assisted recognition of colorectal cancer has been mainly focused on supervised deep learning that suffer from a significant bottleneck of requiring massive labeled data. We hypothesize that semi-supervised deep learning leveraging a small number of labeled data can provide a powerful alternative strategy.Method: We proposed a semi-supervised model based on mean teacher that provide pathological predictions at both patch-level and patient-level. We demonstrated the general utility of the model utilizing 13,111 whole slide images from 8,803 subjects gathered from 13 centers. We compared our proposed method with the prevailing supervised learning and six pathologists.Results: with a small amount of labeled training patches (~3,150 labeled, ~40,950 unlabeled or ~6,300 labeled,~37,800 unlabeled), the semi-supervised model performed significantly better than the supervised model (AUC: 0.90 ± 0.06 vs. 0.84 ± 0.07, P value = 0.02 or AUC: 0.98 ± 0.01 vs 0.92 ± 0.04, P value = 0.0004). Moreover, we found no significant difference between the supervised model using massive ~44,100 labeled patches and the semi-supervised model (~6,300 labeled, ~37,800 unlabeled) at patch-level diagnoses (AUC: 0.98 ± 0.01 vs 0.987 ± 0.01, P value = 0.134) and patient-level diagnoses (average AUC: 97.40% vs. 97.96%, P value = 0.117) . Our model was close to human pathologists (average AUC: 97.17% vs. 96.91%).Conclusions: We reported that semi-supervised learning can achieve excellent performance through a multi-center study. We thus suggested that semi-supervised learning has great potentials to build artificial intelligence (AI) platforms, which will dramatically reduce the cost of labeled data and greatly facilitate the development and application of AI in medical sciences.

scholarly journals Breast Tumor Cellularity Assessment using Deep Neural Networks

2019 ◽  
Author(s):  
Alexander Rakhlin ◽  
Aleksei Tiulpin ◽  
Alexey A. Shvets ◽  
Alexandr A. Kalinin ◽  
Vladimir I. Iglovikov ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Deep Learning ◽  
Automatic Assessment ◽  
Appropriate Treatment ◽  
Image Patches ◽  
Expert Pathologist ◽  
Hematoxylin And Eosin ◽  
Treated Breast ◽  
Intra Class Correlation Coefficient ◽  
Whole Slide Images

AbstractBreast cancer is one of the main causes of death world-wide. Histopathological cellularity assessment of residual tumors in post-surgical tissues is used to analyze a tumor’s response to a therapy. Correct cellularity assessment increases the chances of getting an appropriate treatment and facilitates the patient’s survival. In current clinical practice, tumor cellularity is manually estimated by pathologists; this process is tedious and prone to errors or low agreement rates between assessors. In this work, we evaluated three strong novel Deep Learning-based approaches for automatic assessment of tumor cellularity from post-treated breast surgical specimens stained with hematoxylin and eosin. We validated the proposed methods on the BreastPathQ SPIE challenge dataset that consisted of 2395 image patches selected from whole slide images acquired from 64 patients. Compared to expert pathologist scoring, our best performing method yielded the Cohen’s kappa coefficient of 0.69 (vs. 0.42 previously known in literature) and the intra-class correlation coefficient of 0.89 (vs. 0.83). Our results suggest that Deep Learning-based methods have a significant potential to alleviate the burden on pathologists, enhance the diagnostic workflow, and, thereby, facilitate better clinical outcomes in breast cancer treatment.

scholarly journals A deep learning quantified stroma-immune score to predict survival of patients with stage II–III colorectal cancer

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Zeyan Xu ◽  
Yong Li ◽  
Yingyi Wang ◽  
Shenyan Zhang ◽  
Yanqi Huang ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
T Cells ◽  
Deep Learning ◽  
External Validation ◽  
Survival Rates ◽  
Tnm Stage ◽  
Stage Ii ◽  
Disease Stage ◽  
Immune Score ◽  
Development Group

Abstract Background Profound heterogeneity in prognosis has been observed in colorectal cancer (CRC) patients with intermediate levels of disease (stage II–III), advocating the identification of valuable biomarkers that could improve the prognostic stratification. This study aims to develop a deep learning-based pipeline for fully automatic quantification of immune infiltration within the stroma region on immunohistochemical (IHC) whole-slide images (WSIs) and further analyze its prognostic value in CRC. Methods Patients from two independent cohorts were divided into three groups: the development group (N = 200), the internal (N = 134), and the external validation group (N = 90). We trained a convolutional neural network for tissue classification of CD3 and CD8 stained WSIs. A scoring system, named stroma-immune score, was established by quantifying the density of CD3+ and CD8+ T-cells infiltration in the stroma region. Results Patients with higher stroma-immune scores had much longer survival. In the development group, 5-year survival rates of the low and high scores were 55.7% and 80.8% (hazard ratio [HR] for high vs. low 0.39, 95% confidence interval [CI] 0.24–0.63, P < 0.001). These results were confirmed in the internal and external validation groups with 5-year survival rates of low and high scores were 57.1% and 78.8%, 63.9% and 88.9%, respectively (internal: HR for high vs. low 0.49, 95% CI 0.28–0.88, P = 0.017; external: HR for high vs. low 0.35, 95% CI 0.15–0.83, P = 0.018). The combination of stroma-immune score and tumor-node-metastasis (TNM) stage showed better discrimination ability for survival prediction than using the TNM stage alone. Conclusions We proposed a stroma-immune score via a deep learning-based pipeline to quantify CD3+ and CD8+ T-cells densities within the stroma region on WSIs of CRC and further predict survival.

scholarly journals A Correlation Study of Prognostic Risk Prediction for Colorectal Cancer Based on Autophagy Signature Genes

2021 ◽  
Vol 11 ◽  
Author(s):  
Haibi Zhao ◽  
Chengzhi Huang ◽  
Yuwen Luo ◽  
Xiaoya Yao ◽  
Yong Hu ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Risk Prediction ◽  
Roc Curve ◽  
Prognostic Value ◽  
Cox Regression ◽  
Survival Rates ◽  
Risk Groups ◽  
Correlation Study ◽  
Colorectal Tumor ◽  
Risk Parameters

Autophagy plays a complex role in tumors, sometimes promoting cancer cell survival and sometimes inducing apoptosis, and its role in the colorectal tumor microenvironment is controversial. The purpose of this study was to investigate the prognostic value of autophagy-related genes (ARGs) in colorectal cancer. We identified 37 differentially expressed autophagy-related genes by collecting TCGA colorectal tumor transcriptome data. A single-factor COX regression equation was used to identify 11 key prognostic genes, and a prognostic risk prediction model was constructed based on multifactor COX analysis. We classified patients into high and low risk groups according to prognostic risk parameters (p &lt;0.001) and determined the prognostic value they possessed by survival analysis and the receiver operating characteristic (ROC) curve in the training and test sets of internal tests. In a multifactorial independent prognostic analysis, this risk value could be used as an independent prognostic indicator (HR=1.167, 95% CI=1.078-1.264, P&lt;0.001) and was a robust predictor without any staging interference. To make it more applicable to clinical procedures, we constructed nomogram based on risk parameters and parameters of key clinical characteristics. The area under ROC curve for 3-year and 5-year survival rates were 0.735 and 0.718, respectively. These will better enable us to monitor patient prognosis, thus improve patient outcomes.

Sign in / Sign up

Export Citation Format

Share Document