Evolving role of artificial intelligence in gastrointestinal endoscopy

Parasher, Gulshan; Wong, Morgan; Rawat, Manmeet

doi:10.3748/wjg.v26.i46.7287

Cited by 23 publications

(15 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although the advantages of deep learning for the analysis of non-numerical data types is obvious, such as image data in endoscopy[ 39 - 41 ] and text or speech data in natural language processing[ 42 ], the utility of deep learning for the analysis of numerical data is less clear but remains promising. A recent study has demonstrated the utility of machine learning in predicting anti-TNF response in rheumatoid arthritis, but relied on genetic markers in addition to clinical data[ 43 ].…”

Section: Discussionmentioning

confidence: 99%

Deep learning vs conventional learning algorithms for clinical prediction in Crohn's disease: A proof-of-concept study

Con¹,

Langenberg²,

Vasudevan³

2021

WJG

View full text Add to dashboard Cite

BACKGROUND Traditional methods of developing predictive models in inflammatory bowel diseases (IBD) rely on using statistical regression approaches to deriving clinical scores such as the Crohn's disease (CD) activity index. However, traditional approaches are unable to take advantage of more complex data structures such as repeated measurements. Deep learning methods have the potential ability to automatically find and learn complex, hidden relationships between predictive markers and outcomes, but their application to clinical prediction in CD and IBD has not been explored previously. AIM To determine and compare the utility of deep learning with conventional algorithms in predicting response to anti-tumor necrosis factor (anti-TNF) therapy in CD. METHODS This was a retrospective single-center cohort study of all CD patients who commenced anti-TNF therapy (either adalimumab or infliximab) from January 1, 2010 to December 31, 2015. Remission was defined as a C-reactive protein (CRP) < 5 mg/L at 12 mo after anti-TNF commencement. Three supervised learning algorithms were compared: (1) A conventional statistical learning algorithm using multivariable logistic regression on baseline data only; (2) A deep learning algorithm using a feed-forward artificial neural network on baseline data only; and (3) A deep learning algorithm using a recurrent neural network on repeated data. Predictive performance was assessed using area under the receiver operator characteristic curve (AUC) after 10× repeated 5-fold cross-validation. RESULTS A total of 146 patients were included (median age 36 years, 48% male). Concomitant therapy at anti-TNF commencement included thiopurines (68%), methotrexate (18%), corticosteroids (44%) and aminosalicylates (33%). After 12 mo, 64% had CRP < 5 mg/L. The conventional learning algorithm selected the following baseline variables for the predictive model: Complex disease behavior, albumin, monocytes, lymphocytes, mean corpuscular hemoglobin concentration and gamma-glutamyl transferase, and had a cross-validated AUC of 0.659, 95% confidence interval (CI): 0.562-0.756. A feed-forward artificial neural network using only baseline data demonstrated an AUC of 0.710 (95%CI: 0.622-0.799; P = 0.25 vs conventional). A recurrent neural network using repeated biomarker measurements demonstrated significantly higher AUC compared to the conventional algorithm (0.754, 95%CI: 0.674-0.834; P = 0.036). CONCLUSION Deep learning methods are feasible and have the potential for stronger predictive performance compared to conventional model building methods when applied to predicting remission after anti-TNF therapy in CD.

show abstract

Section: Discussionmentioning

confidence: 99%

Deep learning vs conventional learning algorithms for clinical prediction in Crohn's disease: A proof-of-concept study

Con¹,

Langenberg²,

Vasudevan³

2021

WJG

View full text Add to dashboard Cite

show abstract

“…Deep learning algorithms can often be regarded as a complete "end-to-end" system, so this kind of recognition method directly uses the convolution layer at the front end of the deep network to extract features and uses the output layer at the back end to classify them, so as to realize the recognition of lesion images. For example, Parasher et al [20] used GoogLeNet to extract features and classify gastrointestinal images at the same time. Veitch et al [21] used classical CNN to realize feature extraction and classification of narrow-band imaging enteroscopy images to identify colon polyp images.…”

Section: Discussionmentioning

confidence: 99%

Utilization of Ultrasonic Image Characteristics Combined with Endoscopic Detection on the Basis of Artificial Intelligence Algorithm in Diagnosis of Early Upper Gastrointestinal Cancer

Wang

Song

Wang

et al. 2021

Journal of Healthcare Engineering

View full text Add to dashboard Cite

The aim of this study was to evaluate the diagnostic value of artificial intelligence algorithm combined with ultrasound endoscopy in early esophageal cancer and precancerous lesions by comparing the examination of conventional endoscopy and artificial intelligence algorithm combined with ultrasound endoscopy, and by comparing the real-time diagnosis of endoscopy and the ultrasonic image characteristics of artificial intelligence algorithm combined with endoscopic detection and pathological results. 120 cases were selected. According to the inclusion and exclusion criteria, 80 patients who met the criteria were selected and randomly divided into two groups: endoscopic examination combined with ultrasound imaging based on intelligent algorithm processing (cascade region-convolutional neural network (Cascade RCNN) model algorithm group) and simple use of endoscopy group (control group). This study shows that the ultrasonic image of artificial intelligence algorithm is effective, and the detection performance is better than that of endoscopic detection. The results are close to the gold standard of doctor recognition, and the detection time is greatly shortened, and the recognition time is shortened by 71 frames per second. Compared with the traditional convolutional neural network (CNN) algorithm, the accuracy and recall of image analysis and segmentation using feature pyramid network are increased. The detection rates of CNN model, Cascade RCNN model, and endoscopic detection alone in early esophageal cancer and precancerous lesions are 56.3% (45/80), 88.8% (71/80), and 44.1% (35/80), respectively. The detection rate of Cascade RCNN model and CNN model was higher than that of endoscopy alone, and the difference was statistically significant ( P < 0.05 ). The sensitivity, specificity, positive predictive value, and negative predictive value of Cascade RCNN model were higher than those of CNN model, which was close to the gold standard for physician identification. This provided a reference basis for endoscopic ultrasound identification of early upper gastrointestinal cancer or other gastrointestinal cancers.

show abstract

“…Our VC model, trained using videos, could predict the tumor invasion depth more reliably and accurately than an image-trained model (IC v 2), as observed in the unedited video ( Videos S1 and S2 ). The role of AI during the endoscopic procedure is decision support [ 18 , 19 ]. If the AI can give us information about the exact invasion depth during endoscopy, the endoscopist can make a final decision to predict the invasion depth easily and rapidly.…”

Section: Discussionmentioning

confidence: 99%

An Optimal Artificial Intelligence System for Real-Time Endoscopic Prediction of Invasion Depth in Early Gastric Cancer

Kim

Oh²,

Han

et al. 2022

Cancers

View full text Add to dashboard Cite

We previously constructed a VGG-16 based artificial intelligence (AI) model (image classifier [IC]) to predict the invasion depth in early gastric cancer (EGC) using endoscopic static images. However, images cannot capture the spatio-temporal information available during real-time endoscopy—the AI trained on static images could not estimate invasion depth accurately and reliably. Thus, we constructed a video classifier [VC] using videos for real-time depth prediction in EGC. We built a VC by attaching sequential layers to the last convolutional layer of IC v2, using video clips. We computed the standard deviation (SD) of output probabilities for a video clip and the sensitivities in the manner of frame units to observe consistency. The sensitivity, specificity, and accuracy of IC v2 for static images were 82.5%, 82.9%, and 82.7%, respectively. However, for video clips, the sensitivity, specificity, and accuracy of IC v2 were 33.6%, 85.5%, and 56.6%, respectively. The VC performed better analysis of the videos, with a sensitivity of 82.3%, a specificity of 85.8%, and an accuracy of 83.7%. Furthermore, the mean SD was lower for the VC than IC v2 (0.096 vs. 0.289). The AI model developed utilizing videos can predict invasion depth in EGC more precisely and consistently than image-trained models, and is more appropriate for real-world situations.

show abstract

Evolving role of artificial intelligence in gastrointestinal endoscopy

Cited by 23 publications

References 60 publications

Deep learning vs conventional learning algorithms for clinical prediction in Crohn's disease: A proof-of-concept study

Deep learning vs conventional learning algorithms for clinical prediction in Crohn's disease: A proof-of-concept study

Utilization of Ultrasonic Image Characteristics Combined with Endoscopic Detection on the Basis of Artificial Intelligence Algorithm in Diagnosis of Early Upper Gastrointestinal Cancer

An Optimal Artificial Intelligence System for Real-Time Endoscopic Prediction of Invasion Depth in Early Gastric Cancer

Contact Info

Product

Resources

About