Artificial Intelligence in Diabetic Retinopathy

Grzybowski, Andrzej; Brona, Piotr

doi:10.1007/978-3-030-78601-4_11

Cited by 5 publications

(5 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Various studies have shown that deep learning is ideally suited to segment and classify OCT data at a level at least similar to that of human examiners [28, 29]. Despite its enormous scientific success with more than 20 000 medical publications, deep learning has hardly been used in routine clinical practice to date [30, 31]. For physicians and practitioners to trust deep learning applications, they need to understand how deep learning makes predictions.…”

Section: Discussionmentioning

confidence: 99%

Human selection bias drives the linear nature of the more ground truth effect in explainable deep learning optical coherence tomography image segmentation

Maloca,

Pfau,

Janeschitz‐Kriegl

et al. 2023

Journal of Biophotonics

View full text Add to dashboard Cite

Supervised Deep Learning (DL) algorithms are highly dependent on training data for which human graders are assigned, e.g. for optical coherence tomography (OCT) image annotation. Despite the tremendous success of DL, due to human judgement, these ground truth labels can be inaccurate and/or ambiguous and cause a human selection bias. We therefore investigated the impact of the size of the ground truth and variable numbers of graders on the predictive performance of the same DL architecture and repeated each experiment 3 times. The largest training dataset delivered a prediction performance close to that of human experts. All DL systems utilized were highly consistent. Nevertheless, the DL under‐performers could not achieve any further autonomous improvement even after repeated training. Furthermore, a quantifiable linear relationship between ground truth ambiguity and the beneficial effect of having a larger amount of ground truth data was detected and marked as the more‐ground‐truth effect.This article is protected by copyright. All rights reserved.

show abstract

Section: Discussionmentioning

confidence: 99%

Human selection bias drives the linear nature of the more ground truth effect in explainable deep learning optical coherence tomography image segmentation

Maloca,

Pfau,

Janeschitz‐Kriegl

et al. 2023

Journal of Biophotonics

View full text Add to dashboard Cite

show abstract

“…Stále více se prosazuje vyhodnocení výsledků vyšetření pomocí neuronových sítí (umělé inteligence) a jejich přenos cestou telemedicíny. Své uplatnění tyto technologie nalezly zatím ve screeningu, ale lze očekávat jejich rozšíření i do specializované oftalmologické péče [37][38][39].…”

Section: Diagnostika Diabetické Retinopateunclassified

Diabetic Retinopathy – Diagnostics and Treatment Guidelines

Studnička,

Němčanský,

Vysloužilová

et al. 2023

CSO

View full text Add to dashboard Cite

Diabetic retinopathy is one of the most common complications of diabetes mellitus and represents a serious health, social and economic problem. With the expected increase in the number of patients with diabetes, it is becoming the leading cause of severe vision loss in the working-age population. The presented guidelines summarize the current knowledge about this disease in order to standardize and update the procedures for the diagnosis, classification and treatment of diabetic retinopathy.

show abstract

“…However, it is essential to control the plasma concentration of antidiabetic drugs through a directed drug delivery system thereby regulating the dosage of the drug to suit the patient's tolerance level and harshness of the disease [8] . Agreeing to statistics, the number of people with diabetes rose from 108 million in 1980 to 422 million in 2014 [9] . Consequentially, situation changes and modernity have during the past five decades resulted in remarkable changes in our respective societies, political systems, the environment, and human exploitation [10] .…”

Section: Introductionmentioning

confidence: 99%

“…[8] Agreeing to statistics, the number of people with diabetes rose from 108 million in 1980 to 422 million in 2014. [9] Consequentially, situation changes and modernity have during the past five decades resulted in remarkable changes in our respective societies, political systems, the environment, and human exploitation. [10] As a result, a good number of people with diabetes has increased substantially in both developed and developing countries with the greatest burden tending to fall on socially underprivileged groups and Indigenous people.…”

Section: Introductionmentioning

confidence: 99%

Modelling of Tungsten (C₅₉W), Osmium (C₅₉O_s), and Platinum (C₅₉Pt) Doped Fullerenes for Drug Delivery of Biguanides (BNG) and Metformin (MET): DFT Perspective

et al. 2023

View full text Add to dashboard Cite

Owing to the vast range of exciting ideas and chances for the diagnosis and treatment of diabetes, the number of people with the disease has increased significantly in both developed and developing nations, with the greatest burden typically falling on indigenous people and socially disadvantaged groups. Nanostructures, however, have drawn a lot of interest because of their capacity to specifically target diabetes cells without damaging healthy cells. Herein, the potential of fullerene (C60) surface doped with C59Os, C59W, and C59Pt is explored as drug delivery systems for biguanides and metformin using density functional theory (DFT) computation at the M06‐2X/Gen/LanL2DZ/6‐311++G(d,p) level of theory. Interestingly, results obtained in this investigation specify that metformin interacts with both the doped and updoped surfaces more strongly than biguanides. Fascinatingly, the energy gap obtained from frontier molecular orbital (FMO) accounted for the significant reactivity and stability strength in the following order: MET@C59W>MET@C59Pt>MET@C59Os>MET@C60, whereas BGN@C59 W>BGN@C59Pt>BGN@C59Os>BGN@C60. Substantially, hydrogen bonding was found to be crucial in the interactions between the drug molecules and all doped and undoped surfaces through non‐covalent interaction (NCI) and quantum theory of atoms in molecules (QTAIM). Also worth mentioning is the complexes′ propensity for forward electron transfer, which is aided by their severe electronegative values, thus explicating that the doped surfaces are a preferable carrier for both metformin and biguanides. Although, theoretically, it follows that doped surfaces are promising drug delivery targets most notably for metformin.

show abstract

Artificial Intelligence in Diabetic Retinopathy

Cited by 5 publications

References 42 publications

Human selection bias drives the linear nature of the more ground truth effect in explainable deep learning optical coherence tomography image segmentation

Human selection bias drives the linear nature of the more ground truth effect in explainable deep learning optical coherence tomography image segmentation

Diabetic Retinopathy – Diagnostics and Treatment Guidelines

Modelling of Tungsten (C₅₉W), Osmium (C₅₉O_s), and Platinum (C₅₉Pt) Doped Fullerenes for Drug Delivery of Biguanides (BNG) and Metformin (MET): DFT Perspective

Contact Info

Product

Resources

About

Artificial Intelligence in Diabetic Retinopathy

Cited by 5 publications

References 42 publications

Human selection bias drives the linear nature of the more ground truth effect in explainable deep learning optical coherence tomography image segmentation

Human selection bias drives the linear nature of the more ground truth effect in explainable deep learning optical coherence tomography image segmentation

Diabetic Retinopathy – Diagnostics and Treatment Guidelines

Modelling of Tungsten (C59W), Osmium (C59Os), and Platinum (C59Pt) Doped Fullerenes for Drug Delivery of Biguanides (BNG) and Metformin (MET): DFT Perspective

Contact Info

Product

Resources

About

Modelling of Tungsten (C₅₉W), Osmium (C₅₉O_s), and Platinum (C₅₉Pt) Doped Fullerenes for Drug Delivery of Biguanides (BNG) and Metformin (MET): DFT Perspective