Artificial intelligence enabled automated diagnosis and grading of ulcerative colitis endoscopy images

Sutton, Reed T.; Ane, Osmar R Zai; Goebel, Randolph; Baumgart, Daniel C.

doi:10.1038/s41598-022-06726-2

Cited by 54 publications

(36 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We note that our present work has a number of simplifications particularly with respect to the protein coat. The binding and unbinding of these anisotropic curvature inducing proteins is curvature dependent ( 8, 19, 58 ). The accumulation of these protein dimers and the interaction between their curvature sensing and curvature generation abilities also depends on their adsorption, diffusion, and aggregation in the plane of the membrane ( 59 ).…”

Section: Discussionmentioning

confidence: 99%

Formation of protein-mediated bilayer tubes is governed by a snapthrough transition

Mahapatra

Rangamani

2022

Preprint

View full text Add to dashboard Cite

Plasma membrane tubes are ubiquitous in cellular membranes and in the membranes of intracellular organelles. They play crucial roles in trafficking, ion transport, and cellular motility. The formation of plasma membrane tubes can be due to localized forces acting on the membrane or by curvature-induced by membrane-bound proteins. Here, we present a mathematical framework to model cylindrical tubular protrusions formed by proteins that induce anisotropic spontaneous curvature. Our analysis revealed that the tube radius depends on an effective tension that includes contributions from the bare membrane tension and the protein-induced curvature. We also found that the length of the tube undergoes an abrupt transition from a short, dome-shaped membrane to a long cylinder and this transition is characteristic of a snapthrough instability. Finally, we show that the snapthrough instability depends on the different parameters including coat area, bending modulus, and extent of protein-induced curvature. Our findings have implications for tube formation due to BAR-domain proteins in processes such as endocytosis, t-tubule formation in myocytes, and cristae formation in mitochondria.1Significance StatementCylindrical tubes are ubiquitous on cellular membranes and many studies have focused on how forces can give rise to tubes. How do curvature-inducing proteins, in the absence of forces, form cylindrical tubes on lipid bilayers? The answer to this question relies on minimizing the bending energy of the membrane to accommodate two different principal curvatures such that a cylindrical shape is an energy minimizer. Here, we build on previous works to develop such a model and show that the formation of an elongated cylindrical tube is accompanied by a snapthrough transition in the bending energy. Furthermore, we identify how this transition depends on different membrane properties, providing a landscape for comparing with different experimental observations. These findings have implications for our understanding how tubes forms in different cellular processes including trafficking, mitochondrial inner membrane cristae formation, and T-tubule formation in myocytes.

show abstract

Section: Discussionmentioning

confidence: 99%

Formation of protein-mediated bilayer tubes is governed by a snapthrough transition

Mahapatra

Rangamani

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In [ 43 ] researchers employed CNN and capsule network for stomach disease classification and deformation analysis using a Kvasir V2 dataset having 8000 images with an accuracy of 94.73%. In [ 44 ] the authors classified ulcerative colitis from challenging datasets Kvasir, Kvasir V2, and hyper-Kvasir for binary classification using deep learning and achieved an accuracy of 87.50%. Our proposed deep feature extraction and optimization technique for stomach cancer classification was applied on 17,500 images of different stomach disease classes for stomach diseases recognition from Kvasir and Kvasir v2 and two classes of healthy and bleeding obtained from [ 9 ] and achieved the highest accuracy of 99.8% on the.…”

Section: Discussionmentioning

confidence: 99%

Deep Feature Fusion and Optimization-Based Approach for Stomach Disease Classification

Mohammad

Al-Razgan

2022

Sensors

View full text Add to dashboard Cite

Cancer is the deadliest disease among all the diseases and the main cause of human mortality. Several types of cancer sicken the human body and affect organs. Among all the types of cancer, stomach cancer is the most dangerous disease that spreads rapidly and needs to be diagnosed at an early stage. The early diagnosis of stomach cancer is essential to reduce the mortality rate. The manual diagnosis process is time-consuming, requires many tests, and the availability of an expert doctor. Therefore, automated techniques are required to diagnose stomach infections from endoscopic images. Many computerized techniques have been introduced in the literature but due to a few challenges (i.e., high similarity among the healthy and infected regions, irrelevant features extraction, and so on), there is much room to improve the accuracy and reduce the computational time. In this paper, a deep-learning-based stomach disease classification method employing deep feature extraction, fusion, and optimization using WCE images is proposed. The proposed method comprises several phases: data augmentation performed to increase the dataset images, deep transfer learning adopted for deep features extraction, feature fusion performed on deep extracted features, fused feature matrix optimized with a modified dragonfly optimization method, and final classification of the stomach disease was performed. The features extraction phase employed two pre-trained deep CNN models (Inception v3 and DenseNet-201) performing activation on feature derivation layers. Later, the parallel concatenation was performed on deep-derived features and optimized using the meta-heuristic method named the dragonfly algorithm. The optimized feature matrix was classified by employing machine-learning algorithms and achieved an accuracy of 99.8% on the combined stomach disease dataset. A comparison has been conducted with state-of-the-art techniques and shows improved accuracy.

show abstract

“…We envision AI-assisted histological analysis as an integral component of how big data will contribute to precision medicine for patients with IBD [30], analogous to its established role in advancing oncology research, often demonstrating the ability to predict clinically relevant molecular features, refine prognostic information, guide therapeutic selection, or predict gene expression signatures from H&E images [31], and allow rich data extraction from H&E images [32]. In the immediate future, explainable AI can function as an assist device with the potential to reduce inter-observer variability, facilitate decentralized trials, or could serve as an automated predictor equivalent to a centralized reader, similar to the demonstrated utility in evaluation of endoscopic images [33][34][35]. This has particular relevance in mild disease, where inter-observer variability is greatest.…”

Section: Discussionmentioning

confidence: 99%

Artificial Intelligence Enables Quantitative Assessment of Ulcerative Colitis Histology

Najdawi

Sucipto

Mistry

et al. 2022

Preprint

View full text Add to dashboard Cite

Ulcerative colitis (UC) is a chronic inflammatory bowel disease that is characterized by a relapsing and remitting course. Appropriate assessment of disease activity is critical for adequate treatment decisions. In addition to endoscopic mucosal healing, histologic remission is emerging as a treatment target and a key factor in the evaluation of disease activity and therapeutic efficacy. However, there is no standardized definition of histologic remission, limiting the utility of histologic scoring, and manual pathologist evaluation is subject to intra-and inter-observer variability. Machine learning approaches are increasingly being developed to aid pathologists in accurate and reproducible scoring of histology, and can enable sensitive assessment of clinically relevant features. Here we report a proof-of-concept study using the PathAI platform to develop ML models for identification and quantification of UC histological features directly from hematoxylin and eosin (H&E)-stained whole slide images. Model-predicted histological features were used to quantify tissue area proportions and cell count proportions and densities, which correlated with disease severity and pathologist-assigned Nancy Histological Index (NHI) scores. Moreover, using multivariate analysis based on selected model-predicted histological features, we were able to accurately predict NHI scores, with a weighted kappa (k=0.93) and Spearman correlation (ρ=0.93, p<0.001) when compared to manual pathologist consensus NHI scores. We were also able to predict histological remission, based on the resolution of active inflammation, with high accuracy of 0.94. These results demonstrate the accuracy of ML models in quantifying histologic features of UC and predicting NHI scores, and highlight the potential of this approach to enable standardized and robust assessment of histologic remission for improved evaluation of disease activity and prognosis.

show abstract

Artificial intelligence enabled automated diagnosis and grading of ulcerative colitis endoscopy images

Cited by 54 publications

References 38 publications

Formation of protein-mediated bilayer tubes is governed by a snapthrough transition

Formation of protein-mediated bilayer tubes is governed by a snapthrough transition

Deep Feature Fusion and Optimization-Based Approach for Stomach Disease Classification

Artificial Intelligence Enables Quantitative Assessment of Ulcerative Colitis Histology

Contact Info

Product

Resources

About