scholarly journals How can artificial intelligence models assist PD-L1 expression scoring in breast cancer: results of multi-institutional ring studies

2021 ◽  
Vol 7 (1) ◽  
Author(s):  
Xinran Wang ◽  
Liang Wang ◽  
Hong Bu ◽  
Ningning Zhang ◽  
Meng Yue ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Artificial Intelligence ◽  
Clinical Practice ◽  
Subjective Assessment ◽  
Current Guideline ◽  
Scoring Method ◽  
Programmed Death ◽  
Intelligence Models ◽  
Artificial Intelligence Models ◽  
Scoring Accuracy

AbstractProgrammed death ligand-1 (PD-L1) expression is a key biomarker to screen patients for PD-1/PD-L1-targeted immunotherapy. However, a subjective assessment guide on PD-L1 expression of tumor-infiltrating immune cells (IC) scoring is currently adopted in clinical practice with low concordance. Therefore, a repeatable and quantifiable PD-L1 IC scoring method of breast cancer is desirable. In this study, we propose a deep learning-based artificial intelligence-assisted (AI-assisted) model for PD-L1 IC scoring. Three rounds of ring studies (RSs) involving 31 pathologists from 10 hospitals were carried out, using the current guideline in the first two rounds (RS1, RS2) and our AI scoring model in the last round (RS3). A total of 109 PD-L1 (Ventana SP142) immunohistochemistry (IHC) stained images were assessed and the role of the AI-assisted model was evaluated. With the assistance of AI, the scoring concordance across pathologists was boosted to excellent in RS3 (0.950, 95% confidence interval (CI): 0.936–0.962) from moderate in RS1 (0.674, 95% CI: 0.614–0.735) and RS2 (0.736, 95% CI: 0.683–0.789). The 2- and 4-category scoring accuracy were improved by 4.2% (0.959, 95% CI: 0.953–0.964) and 13% (0.815, 95% CI: 0.803–0.827) (p < 0.001). The AI results were generally accepted by pathologists with 61% “fully accepted” and 91% “almost accepted”. The proposed AI-assisted method can help pathologists at all levels to improve the PD-L1 assay (SP-142) IC assessment in breast cancer in terms of both accuracy and concordance. The AI tool provides a scheme to standardize the PD-L1 IC scoring in clinical practice.

scholarly journals Detección automática de la Fibrilación Auricular a través de un Smartwatch

2019 ◽  
Vol 11 (4) ◽  
pp. 3
Author(s):  
Anna Abad Torrent ◽  
Helena Benito Naverac
Keyword(s):  
Artificial Intelligence ◽  
Atrial Fibrillation ◽  
Clinical Practice ◽  
General Population ◽  
Cardiac Arrhythmia ◽  
Automatic Detection ◽  
Insuficiencia Cardiaca ◽  
Intelligence Models ◽  
Artificial Intelligence Models ◽  
Electrocardiogram Ecg

La fibrilación auricular es la arritmia cardiaca más frecuente en la práctica clínica. La prevalencia se sitúa en torno al 0,4 - 1 % de la población general. Aumenta con la edad, llegando hasta el 8% a partir de los 80 años. Esta arritmia es la principal causa a nivel mundial de accidente cerebrovascular (20-30% de los casos son debidos a la fibrilación auricular), insuficiencia cardíaca o muerte súbita. Muchas veces , es clínicamente silente o se manifiesta con síntomas vagos como las palpitaciones, que pueden atribuirse erróneamente a ansiedad y retrasar el diagnóstico. La instauración temprana de anticoagulación (en determinados casos) reduce, de forma significativa la incidencia de fenómenos tromboembólicos. ABSTRACT Automatic detection of atrial fibrillation using a Smartwatch Atrial fibrillation is the most common cardiac arrhythmia in clinical practice. The prevalence is around 0.4 — 1% of the general population. It increases with age, reaching up to 8% from 80 years. In cardiology, the standard for the diagnosis of a cardiac arrhythmia is based on the performance of an electrocardiogram (ECG). From the monitoring of KardiaBand ™ and SmartRhythm ™, AliveCor launches the first platform for Apple Watch series 4, which combines an electrocardiography device approved by the FDA and certain analysis algorithms with artificial intelligence models, which help to detect the atrial fibrillation.  

scholarly journals BRCA Variations Risk Assessment in Breast Cancers Using Different Artificial Intelligence Models

Genes ◽  
2021 ◽  
Vol 12 (11) ◽  
pp. 1774
Author(s):  
Niyazi Senturk ◽  
Gulten Tuncel ◽  
Berkcan Dogan ◽  
Lamiya Aliyeva ◽  
Mehmet Sait Dundar ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Neural Network ◽  
Artificial Intelligence ◽  
Risk Factors ◽  
Risk Assessment ◽  
Cancer Patients ◽  
Breast Cancers ◽  
Breast Cancer Patients ◽  
Intelligence Models ◽  
Artificial Intelligence Models

Artificial intelligence provides modelling on machines by simulating the human brain using learning and decision-making abilities. Early diagnosis is highly effective in reducing mortality in cancer. This study aimed to combine cancer-associated risk factors including genetic variations and design an artificial intelligence system for risk assessment. Data from a total of 268 breast cancer patients have been analysed for 16 different risk factors including genetic variant classifications. In total, 61 BRCA1, 128 BRCA2 and 11 both BRCA1 and BRCA2 genes associated breast cancer patients’ data were used to train the system using Mamdani’s Fuzzy Inference Method and Feed-Forward Neural Network Method as the model softwares on MATLAB. Sixteen different tests were performed on twelve different subjects who had not been introduced to the system before. The rates for neural network were 99.9% for training success, 99.6% for validation success and 99.7% for test success. Despite neural network’s overall success was slightly higher than fuzzy logic accuracy, the results from developed systems were similar (99.9% and 95.5%, respectively). The developed models make predictions from a wider perspective using more risk factors including genetic variation data compared with similar studies in the literature. Overall, this artificial intelligence models present promising results for BRCA variations’ risk assessment in breast cancers as well as a unique tool for personalized medicine software.

scholarly journals Stress testing reveals gaps in clinic readiness of image-based diagnostic artificial intelligence models

2021 ◽  
Vol 4 (1) ◽  
Author(s):  
Albert T. Young ◽  
Kristen Fernandez ◽  
Jacob Pfau ◽  
Rasika Reddy ◽  
Nhat Anh Cao ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Real World ◽  
Stress Testing ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Skin Lesions ◽  
Stress Tests ◽  
Performance Metric ◽  
Intelligence Models ◽  
Artificial Intelligence Models

AbstractArtificial intelligence models match or exceed dermatologists in melanoma image classification. Less is known about their robustness against real-world variations, and clinicians may incorrectly assume that a model with an acceptable area under the receiver operating characteristic curve or related performance metric is ready for clinical use. Here, we systematically assessed the performance of dermatologist-level convolutional neural networks (CNNs) on real-world non-curated images by applying computational “stress tests”. Our goal was to create a proxy environment in which to comprehensively test the generalizability of off-the-shelf CNNs developed without training or evaluation protocols specific to individual clinics. We found inconsistent predictions on images captured repeatedly in the same setting or subjected to simple transformations (e.g., rotation). Such transformations resulted in false positive or negative predictions for 6.5–22% of skin lesions across test datasets. Our findings indicate that models meeting conventionally reported metrics need further validation with computational stress tests to assess clinic readiness.

Sign in / Sign up

Export Citation Format

Share Document