Risk-Profile and Feature Selection Comparison in Diabetic Retinopathy

Maeda-Gutiérrez, Valeria; Galván-Tejada, Carlos E.; Cruz, Miguel; Galván-Tejada, Jorge I.; Gamboa-Rosales, Hamurabi; García-Hernández, Alejandra; Luna-García, Huizilopoztli; González-Curiel, Irma; Martínez-Acuña, Mónica

doi:10.3390/jpm11121327

Cited by 6 publications

(5 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The 24 studies enrolled 5596 DR patients and 17907 non-DR patients to explore the association between DR and abdominal obesity in patients with diabetes. Seventeen studies were conducted in Asia, four in Europe [ 15 – 18 ] and one each in Australia [ 19 ], North America [ 20 ] and Africa [ 21 ]. All studies were conducted in populations over 18 years.…”

Section: Resultsmentioning

confidence: 99%

Association between abdominal obesity and diabetic retinopathy in patients with diabetes mellitus: A systematic review and meta-analysis

Zhang

et al. 2023

PLoS ONE

View full text Add to dashboard Cite

Objective Previous studies have reported different opinions regarding the association between abdominal obesity and diabetic retinopathy (DR) in patients with diabetes mellitus (DM). In this study, we aimed to investigate this problem through a systematic review and meta-analysis to provide a basis for clinical interventions. Methods A comprehensive search was conducted in the PubMed, Embase, and Web of Science databases up to May 1, 2022, for all eligible observational studies. Standardized mean differences (SMD) and 95% confidence intervals (CI) were evaluated using a random-effects model in the Stata software. We then conducted, publication bias assessment, heterogeneity, subgroup and sensitivity analyses. Results A total of 5596 DR patients and 17907 non-DR patients were included from 24 studies. The results of the meta-analysis of abdominal obesity parameters showed statistically significant differences between DR and non-DR patients in both type 1 and type 2 diabetes. Waist circumference (WC) was higher in patients with DR than in the non-DR patients. In the waist-hip ratio (WHR) subgroup, the level of WHR was higher in patients with DR than that in non-DR patients. The association between abdominal obesity and mild to moderate nonproliferative DR or vision-threatening DR groups did not show any statistical difference. Subgroup analysis according to ethnicity showed that Caucasians had higher levels of combined abdominal obesity parameters than Asians. Conclusion We found that abdominal obesity measured by WC and WHR is associated with DR in patients with type 1 and type 2 diabetes. This association is stronger in Caucasians than in Asians, where isolated abdominal obesity might be more related to DR. However, no correlation was found between abdominal obesity and varying degrees of diabetic retinopathy. Further prospective cohort studies with larger sample sizes are yet to be conducted to clarify our findings.

show abstract

Section: Resultsmentioning

confidence: 99%

Association between abdominal obesity and diabetic retinopathy in patients with diabetes mellitus: A systematic review and meta-analysis

Zhang

et al. 2023

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…The characteristics of the studies are summarized in Table 1. The search revealed that the articles that represented Latin America (LA) only focused on the Mexican population, from which studies were found focused on the analysis of the risk factors involved in the development of DR and clinical studies, finding a positive association with the levels of HDL, creatinine, glomerular filtration ratio, cholesterol [32], SBP, HbA1c, and albuminuria [33], proinflammatory cytokines such as TNF-a and IL-6 [34], mRNA such as TUBD1 [35], and others ruled out the association that could exist between DR and some integrin a2 polymorphisms [36]. In the studies in non-LA, the approach was differentiated between imaging trials analyzing values of venule caliber, arterial tortuosity, insulin dose, and duration of DM [37], while others associated it with clinical characteristics such as albuminuria [38,39], another focused on morphometric values such as body mass index and waist circumference [40], and another was related to conditions that may be important factors in DR such as psoriasis [41].…”

Section: Resultsmentioning

confidence: 99%

“…In contrast, the Latin American analysis did not show a significant difference between DBP pressures and the presence of DR (p = 0.54) (C). Maximum Likelihood Model (MLE) [32][33][34][35][36][37][38][39]41].…”

Section: ) (A) the Exclusion Ofmentioning

confidence: 99%

Risk Factors for Diabetic Retinopathy in Latin America (Mexico) and the World: A Systematic Review and Meta-Analysis

Vivanco-Rojas,

López-Letayf,

Londoño-Angarita

et al. 2023

JCM

View full text Add to dashboard Cite

Diabetic retinopathy (DR) is one of the main complications of diabetes, and the management of the main control parameters explains only an 11% reduction in the risk of progressing to DR, leaving 89% to be explained by other factors or correlations between the usual factors that are currently unknown. The objective of this systematic review and meta-analysis is to evaluate the similarities and differences between the possible risk factors for developing DR when comparing the world to Latin American populations. The search was performed first for Latin American (LA) populations and a second search for non-Latin American (Non-LA) populations. Using the PRISMA guidelines, five articles were found to be relevant for each of the groups. The patients who had elevated systolic blood pressure (SBP) developed DR more frequently than the patients without retinopathy (Z = 2.1, p = 0.03), an effect measured in the population at a global level (GL), behavior that becomes not significant when the LA and non-LA populations are grouped separately; relevant to this is that the diagnosis of hypertension (HBP) grouped globally and stratified does not present a risk factor for DR (Z = 0.79, p = 0.42). This indicates that SBP is a risk factor for the world population and that, by separating it into different regions, the omission could cause it not to be considered a possible risk factor. In conclusion, the relationship between the increase in DR associated with the risk factors present in different populations, the limited research conducted in Latin America, and the cultural, social, economic, and genetic differences makes for a complex condition, which reflects the necessity of researching in a more integrated way.

show abstract

“…By using Boruta wrapper feature selection algorithms with CART decision tree, [19] have improved the classification accuracy of medical datasets effectively. Moreover, Boruta has been used to identify the significant features of medical datasets [10], [22], [23], [25]. The researchers recommended Boruta to select relevant variables in high-dimensional datasets [37], [38].…”

Section: A Identifying the Important Variablesmentioning

confidence: 99%

Detecting Need-Attention Patients using Machine Learning

Law,

Ting,

Zakariah

2024

JOIV : Int. J. Inform. Visualization

View full text Add to dashboard Cite

In healthcare, detecting patients who need immediate attention is difficult. Identifying the critical variables is challenging in patient detection because human intervention in variable selection is required. Consequently, patients who need immediate attention often experience prolonged waiting times. Researchers have investigated various approaches to identify those who require attention. One of the techniques is leveraging Artificial Intelligence (AI). However, identifying the optimal feature set and predictive model is complex. Therefore, this study has attempted to (i) identify the critical features and (ii) develop and evaluate predictive models in detecting those who need attention. The dataset is collected from one of the healthcare companies. The dataset collected contains 67 variables and 51102 records. It consists of patient information and questionnaires answered by each participant registered in the Selangor Saring Program. Important features were identified in detecting those who need attention on treated data. Multiple classifiers were developed due to their simplicity. The models were evaluated before and after hyperparameter tuning based on accuracy, precision, recall, F1-score, Geometric Mean, and Area Under the Curve. The findings showed that the Stacking Classifier produced the highest accuracy (69.9%) when using the blood dataset. In contrast, Extreme Gradient Boosting achieved the highest accuracy (81.7%) when the urine dataset was used. This work can be extended to explore the incorporation of Points of Interest and geographical data near patients’ residences and study other ensemble models to enhance the performance of detecting those who need attention.

show abstract

Risk-Profile and Feature Selection Comparison in Diabetic Retinopathy

Cited by 6 publications

References 35 publications

Association between abdominal obesity and diabetic retinopathy in patients with diabetes mellitus: A systematic review and meta-analysis

Association between abdominal obesity and diabetic retinopathy in patients with diabetes mellitus: A systematic review and meta-analysis

Risk Factors for Diabetic Retinopathy in Latin America (Mexico) and the World: A Systematic Review and Meta-Analysis

Detecting Need-Attention Patients using Machine Learning

Contact Info

Product

Resources

About