Single-Centre Retrospective Training Cohort Using Artificial Intelligence for Prognostic Prediction of Encephalopathy, Mortality, and Liver Dysfunction after Early TIPS Creation

2021 ◽  
Author(s):  
Bin-Yan Zhong ◽  
Wan-Sheng Wang ◽  
Jian Shen ◽  
Hang Du ◽  
Shuai Zhang ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Liver Dysfunction ◽  
Single Centre ◽  
Training Cohort ◽  
Prognostic Prediction ◽  
Early Tips

scholarly journals A radiomic nomogram based on arterial phase of CT for differential diagnosis of ovarian cancer

2021 ◽  
Author(s):  
Yumin Hu ◽  
Qiaoyou Weng ◽  
Haihong Xia ◽  
Tao Chen ◽  
Chunli Kong ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Ovarian Cancer ◽  
Clinical Data ◽  
Arterial Phase ◽  
Multivariable Logistic Regression Analysis ◽  
Clinical Factors ◽  
Good Discrimination ◽  
Training Cohort ◽  
Ovarian Cancers ◽  
Radiomics Signature

Abstract Purpose To develop and validate a radiomic nomogram based on arterial phase of CT to discriminate the primary ovarian cancers (POCs) and secondary ovarian cancers (SOCs). Methods A total of 110 ovarian cancer patients in our hospital were reviewed from January 2010 to December 2018. Radiomic features based on the arterial phase of CT were extracted by Artificial Intelligence Kit software (A.K. software). The least absolute shrinkage and selection operation regression (LASSO) was employed to select features and construct the radiomics score (Rad-score) for further radiomics signature calculation. Multivariable logistic regression analysis was used to develop the predicting model. The predictive nomogram model was composed of rad-score and clinical data. Nomogram discrimination and calibration were evaluated. Results Two radiomic features were selected to build the radiomics signature. The radiomics nomogram that incorporated 2 radiomics signature and 2 clinical factors (CA125 and CEA) showed good discrimination in training cohort (AUC 0.854), yielding the sensitivity of 78.8% and specificity of 90.7%, which outperformed the prediction model based on radiomics signature or clinical data alone. A visualized differential nomogram based on the radiomic score, CEA, and CA125 level was established. The calibration curve demonstrated the clinical usefulness of the proposed nomogram. Conclusion The presented nomogram, which incorporated radiomic features of arterial phase of CT with clinical features, could be useful for differentiating the primary and secondary ovarian cancers.

Construction and Validation of a Prognostic Prediction Model for Gastric Cancer Using a Series of Genes Related to Lactate Metabolism

2021 ◽  
Author(s):  
Siyu Wang ◽  
Yuxin Wang ◽  
Xin Wang ◽  
Shuqiang Yuan
Keyword(s):  
Gastric Cancer ◽  
Prediction Model ◽  
Function Analysis ◽  
Differential Expression Analysis ◽  
External Validation ◽  
High Morbidity ◽  
Lactate Metabolism ◽  
Training Cohort ◽  
Cox Regression Analysis ◽  
Prognostic Prediction

Abstract Background: Gastric cancer is one of the most common clinical malignant tumors worldwide, with high morbidity and mortality. The commonly used TNM staging and some common biomarkers have a certain value in predicting the prognosis of GC patients, but they gradually failed to meet the clinical demands. Therefore, we aim to construct a prognostic prediction model for GC patients.Methods: A total of 350 cases were included as TCGA-STAD entire cohort, including TCGA-STAD training cohort (n=176) and TCGA-STAD testing cohort (n=174). GSE15459 (n=191), GSE62254 (n=300) and some cases in our center (n=12) were for external validation.Results: Through differential expression analysis and univariate Cox regression analysis in TCGA-STAD training cohort, we screened out 5 genes among 600 genes related to lactate metabolism for the construction of our prognostic prediction model. The internal and external validations showed the same result, that is, patients with higher risk score were associated with poor prognosis (all P<0.05), and our model works well without regard of patients' age, gender, tumor grade, clinical stage or TNM stage, which supports the availability, validity and stability of our model. Gene function analysis, tumor-infiltrating immune cells analysis, tumor microenvironment analysis and clinical treatment exploration were performed to improve the practicability of the model, and hope to provide a new basis for more in-depth study of the molecular mechanism for GC and for clinicians to formulate more reasonable and individualized treatment plans.Conclusions: We screened out and used 5 genes related to lactate metabolism to develop a prognostic prediction model for GC patients based on them. The prediction performance of the model is confirmed by a series of bioinformatics and statistical analysis.

Artificial intelligence for streamlined immunofluorescence-based biomarker discovery in prostate cancer.

2020 ◽  
Vol 38 (6_suppl) ◽  
pp. 279-279
Author(s):  
Claire Marie de la Calle ◽  
Hao Gia Nguyen ◽  
Ehsan Hosseini-Asl ◽  
Clarence So ◽  
Richard Socher ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Normal Tissue ◽  
Prediction Models ◽  
Biomarker Discovery ◽  
Mean Shift ◽  
Visual Assessment ◽  
Algorithm Analysis ◽  
Ki 67 ◽  
Training Cohort ◽  
Deep Model

279 Background: Immunofluorescence (IF) performed on tissue microarrays (TMA) is used for biomarker discovery but is limited by the arduous and subjective human visual assessment with an IF microscope. We aim to implement deep learning-based artificial intelligence (AI) models to automate and speed up the analysis of numerous biomarkers and generate prediction models of recurrence and metastasis after surgery. Methods: A TMA was constructed consisting of 648 samples (424 tumors, 224 normal tissue) generated from prostatectomy specimens. IF staining was performed on the TMA using anti Ki-67, ERG antibodies and analyzed for differential expression using “gold standard” manual microscopy and using an AI algorithm. Analysis was done blinded to any clinicopathological data. For manual microscopy, relative mean fluorescence intensity of cancerous versus normal tissue was determined. The AI algorithm was generated using a training cohort of digitized images. To do so the Otsu method thresholding algorithm combined with mean shift clustering was employed to find cell centers, followed by a level-set algorithm, to compute cell boundaries.These predictions were then combined with pixel predictions of a fully convolutional deep model to refine the regions of overlapping epithelium, stroma, and artifact. The algorithm was then validated using a separate cohort. Results from the algorithm were then compared to the data from manual microscopy. Results: Ki-67 and ERG expression levels generated by the algorithm showed only a 5% variance compared to the manually generated results. The algorithm was able to pick out which tumor were positive for ERG with 100% accuracy in spite of variance from artifacts. The algorithm also had the ability to improve its accuracy after each iteration of modifications and feedback through the training cohort. Conclusions: The AI algorithm produced similar outcomes than manual quantification with high accuracy but with more efficiency, cost effectiveness and objectivity. We are now developing more complex algorithms that will include the differential pattern of expression of PTEN, MYC and others with the objectives of streamlining biomarker discovery.

scholarly journals Reimagining T Staging Through Artificial Intelligence and Machine Learning Image Processing Approaches in Digital Pathology

2020 ◽  
pp. 1039-1050
Author(s):  
Kaustav Bera ◽  
Ian Katz ◽  
Anant Madabhushi
Keyword(s):  
Artificial Intelligence ◽  
Domain Knowledge ◽  
Imaging Techniques ◽  
Prognostic Significance ◽  
Digital Pathology ◽  
Staging System ◽  
Tumor Stage ◽  
Tnm Staging ◽  
Molecular Alterations ◽  
Prognostic Prediction

Tumor stage and grade, visually assessed by pathologists from evaluation of pathology images in conjunction with radiographic imaging techniques, have been linked to outcome, progression, and survival for a number of cancers. The gold standard of staging in oncology has been the TNM (tumor-node-metastasis) staging system. Though histopathological grading has shown prognostic significance, it is subjective and limited by interobserver variability even among experienced surgical pathologists. Recently, artificial intelligence (AI) approaches have been applied to pathology images toward diagnostic-, prognostic-, and treatment prediction–related tasks in cancer. AI approaches have the potential to overcome the limitations of conventional TNM staging and tumor grading approaches, providing a direct prognostic prediction of disease outcome independent of tumor stage and grade. Broadly speaking, these AI approaches involve extracting patterns from images that are then compared against previously defined disease signatures. These patterns are typically categorized as either (1) handcrafted, which involve domain-inspired attributes, such as nuclear shape, or (2) deep learning (DL)–based representations, which tend to be more abstract. DL approaches have particularly gained considerable popularity because of the minimal domain knowledge needed for training, mostly only requiring annotated examples corresponding to the categories of interest. In this article, we discuss AI approaches for digital pathology, especially as they relate to disease prognosis, prediction of genomic and molecular alterations in the tumor, and prediction of treatment response in oncology. We also discuss some of the potential challenges with validation, interpretability, and reimbursement that must be addressed before widespread clinical deployment. The article concludes with a brief discussion of potential future opportunities in the field of AI for digital pathology and oncology.

scholarly journals From the Immune Profile to the Immunoscore: Signatures for Improving Postsurgical Prognostic Prediction of Pancreatic Neuroendocrine Tumors

2021 ◽  
Vol 12 ◽  
Author(s):  
Miaoyan Wei ◽  
Jin Xu ◽  
Jie Hua ◽  
Qingcai Meng ◽  
Chen Liang ◽  
...  
Keyword(s):  
Neuroendocrine Tumors ◽  
Low Risk ◽  
Individual Risk ◽  
Recurrence Free Survival ◽  
Pancreatic Neuroendocrine Tumors ◽  
Free Survival ◽  
Training Cohort ◽  
Immune Infiltration ◽  
Risk Patients ◽  
Prognostic Prediction

ObjectiveImmune infiltration plays an important role in tumor development and progression and shows promising prognostic value in numerous tumors. In this study, we aimed to identify the role of immune infiltration in pancreatic neuroendocrine tumors (Pan-NETs) and to establish an Immunoscore system to improve the prediction of postsurgical recurrence-free survival.MethodsTo derive transcriptional signatures and deconvolute specific immune populations, two GEO datasets containing 158 Pan-NET patients were reanalyzed to summarize the immune infiltration landscape and identify immune-related signatures. Using real-time reverse transcription-polymerase chain reaction, immunofluorescence and immunochemistry methods, candidate signatures were further detected. The least absolute shrinkage and selection operator (LASSO) logistic regression model used statistically significant survival predicators in the training cohort (n=125) to build an Immunoscore system. The prognostic and predictive accuracy was validated in an external independent cohort of 77 patients.ResultsThe immune infiltration profile in Pan-NETs showed significant heterogeneity, among which accumulated immune cells, T lymphocytes and macrophages were predominant. Fourteen statistically significant immune-related signatures were further identified in the screening cohort. The Immunoscore system for Pan-NETs (ISpnet) consisting of six immune features (CCL19, IL-16, CD163, IRF4, CD8PT and CD8IT) was constructed to classify patients as high and low risk in the training cohort (cutoff value = 2.14). Low-risk patients demonstrated longer 5-year recurrence-free survival (HR, 0.061; 95% CI, 0.026 to 0.14; p &lt; 0.0001), with fewer recurrences and better prognoses. To predict the individual risk of recurrence, a nomogram incorporating both immune signatures and clinicopathological characteristics was developed.ConclusionOur model, ISpnet, captures immune feature-associated prognostic indicators in Pan-NETs and represents the first immune feature-based score for the postsurgical prognostic prediction. The nomogram based on the ISpnet and independent clinical risk factors might facilitate decision-making regarding early recurrence risk monitoring, identify high-risk patients in need of adjuvant therapy, and provide auxiliary guidance for patients with Pan-NETs that may benefit from immunotherapy in clinical trials.

Sign in / Sign up

Export Citation Format

Share Document