Nomogram-Based Prediction of the Risk of Diabetic Retinopathy: A Retrospective Study

Prediction of diabetic retinopathy among type 2 diabetic patients in University of Gondar Comprehensive Specialized Hospital, 2006-2021: A prognostic model

BACKGROUND

There has been a paucity of evidence for the development of a prediction model for diabetic retinopathy (DR) in Ethiopia. Predicting the risk of developing DR based on the patient's demographic, clinical, and behavioral data is helpful in resource-limited areas where regular screening for DR is not available and to guide practitioners estimate the future risk of their patients.

METHODS

A retrospective follow-up study was conducted at the University of Gondar (UoG) Comprehensive Specialized Hospital from January 2006 to May 2021 among 856 patients with type 2 diabetes (T2DM). Variables were selected using the Least Absolute Shrinkage and Selection Operator (LASSO) regression. The data were validated by 10-fold cross-validation. Four ML techniques (naïve Bayes, K-nearest neighbor, decision tree, and logistic regression) were employed. The performance of each algorithm was measured, and logistic regression was a well-performing algorithm. After multivariable logistic regression and model reduction, a nomogram was developed to predict the individual risk of DR.

RESULTS

Logistic regression was the best algorithm for predicting DR with an area under the curve of 92%, sensitivity of 87%, specificity of 83%, precision of 84%, F1-score of 85%, and accuracy of 85%. The logistic regression model selected seven predictors: total cholesterol, duration of diabetes, glycemic control, adherence to anti-diabetic medications, other microvascular complications of diabetes, sex, and hypertension. A nomogram was developed and deployed as a web-based application. A decision curve analysis showed that the model was useful in clinical practice and was better than treating all or none of the patients.

CONCLUSIONS

The model has excellent performance and a better net benefit to be utilized in clinical practice to show the future probability of having DR. Identifying those with a higher risk of DR helps in the early identification and intervention of DR.

Association between blood urea nitrogen levels and diabetic retinopathy in diabetic adults in the United States (NHANES 2005-2018)

Objective

To investigate the association between blood urea nitrogen (BUN) levels and diabetic retinopathy (DR) in adults with diabetes mellitus (DM).

Methods

Seven cycles of cross-sectional population information acquired from NHANES(national health and nutrition examination surveys) 2005-2018 were collected, from which a sample of diabetic adults was screened and separated into two groups based on whether or not they had DR, followed by weighted multivariate regression analysis. This study collected a complete set of demographic, biological, and sociological risk factor indicators for DR. Demographic risk factors comprised age, gender, and ethnicity, while biological risk factors included blood count, blood pressure, BMI, waist circumference, and glycated hemoglobin. Sociological risk factors included education level, deprivation index, smoking status, and alcohol consumption.

Results

The multiple regression model revealed a significant connection between BUN levels and DR [odds ratio =1.04, 95% confidence interval (1.03-1.05), p-value <0.0001],accounting for numerous variables. After equating BUN levels into four groups, multiple regression modeling showed the highest quartile (BUN>20 mg/dl) was 2.22 times more likely to develop DR than the lowest quartile [odds ratio =2.22, 95% confidence interval (1.69-2.93), p- value <0.0001]. Subgroup analyses revealed that gender, race, diabetes subtype, and duration of diabetes had a regulating effect on the relationship between BUN and DR.

Conclusion

BUN levels were related with an increased prevalence of DR, particularly in individuals with BUN >20 mg/dl. These findings highlight the significance of BUN level in assessing the risk of DR.

Plaque enhancement predicts recurrence in acute ischemic stroke patients with large artery intracranial atherosclerosis

BACKGROUND

The association between the degree of plaque enhancement and ischemic brain stroke recurrence remains unclear. We aimed to establish models to predict plaque enhancement and stroke recurrence.

METHODS

Seventy-eight participants with acute ischemic brain stroke due to intracranial arterial stenosis were recruited and divided into high enhancement (HE) and non-HE groups. The relationship between imaging characteristics (degree of stenosis, minimal lumen area, intraplaque hemorrhage, and plaque burden) and the degree of plaque contrast enhancement was analyzed. Inflammatory cytokine expression was examined by flow cytometry. Independent predictors of stroke recurrence were investigated via multivariate Cox proportional hazards regression analysis. Nomogram was used to construct a prediction model. Harrell's concordance indices (c-indices) and calibration curves were used to assess the discrimination of the nomogram. A risk prediction nomogram for prognosis was constructed.

RESULTS

Thirty-three participants were assigned to the HE group and 45 to the non-HE group. The degree of stenosis and plaque burden in the HE group was higher than that in the non-HE group (P<0.05). Multiple linear regression analysis showed the degree of stenosis was associated with HE (β=0.513; P=0.000). After adjusting for confounding factors, age (HR=1.115; 95%CI=1.034-1.203, P=0.005) and HE plaques (HR=10.457; 95%CI=1.176-93.018; P=0.035) were independent risk factors of stroke recurrence, whereas cytokine levels were not statistically significant between two group.

CONCLUSIONS

HE of intracranial atherosclerosis plaques is an independent factor for ischemic brain stroke recurrence.

Longitudinal Changes of Parafoveal Vessel Density in Diabetic Patients without Clinical Retinopathy Using Optical Coherence Tomography Angiography

PURPOSE

The purpose of this study was to identify the rate of parafoveal vessel density (VD) changes associated with the progression from non-diabetic retinopathy (NDR) to early stages of DR over a year.

METHODS

This longitudinal cohort study enrolled diabetic patients from the Guangzhou community in China. The patients with NDR at baseline were included and underwent comprehensive examinations at baseline and after 1 year. A commercial OCTA device (Triton Plus, Topcon, Tokyo, Japan) was employed to quantify the parafoveal VD in the superficial and deep capillary plexuses. The rates of change in parafoveal VD over time in the incident DR and NDR groups were compared after a year.

RESULTS

A total of 448 NDR patients were included in the study. Among them, 382 (83.2%) were stable and 66 (14.4%) developed incident DR during the 1-year follow-up. The average parafoveal VD in the superficial capillary plexus (SCP) reduced significantly more quickly in the incident DR group than in the NDR group (-1.95 ± 0.45%/year vs. -0.45 ± 0.19/year, p = 0.002). The VD reduction rate for the deep capillary plexus (DCP) was not significantly different for the groups (p = 0.156).

CONCLUSIONS

The incident DR group experienced a significantly faster reduction in parafoveal VD in the SCP compared with the stable group. Our findings further provide supporting evidence that parafoveal VD in the SCP may be used as an early indicator of the pre-clinical stages of DR.

Risk prediction model for distinguishing Gram-positive from Gram-negative bacteremia based on age and cytokine levels: A retrospective study

BACKGROUND

Severe infection often results in bacteremia, which significantly increases mortality rate. Different therapeutic strategies are employed depending on whether the blood-borne infection is Gram-negative (G^-) or Gram-positive (G⁺). However, there is no risk prediction model for assessing whether bacteremia patients are infected with G^- or G⁺ pathogens.

AIM

To establish a clinical prediction model to distinguish G^- from G⁺ infection.

METHODS

A total of 130 patients with positive blood culture admitted to a single intensive care unit were recruited, and Th1 and Th2 cytokine concentrations, routine blood test results, procalcitonin and C-reactive protein concentrations, liver and kidney function test results and coagulation function were compared between G⁺ and G^- groups. Least absolute shrinkage and selection operator (LASSO) regression analysis was employed to optimize the selection of predictive variables by running cyclic coordinate descent and K-fold cross-validation (K = 10). The predictive variables selected by LASSO regression analysis were then included in multivariate logistic regression analysis to establish a prediction model. A nomogram was also constructed based on the prediction model. Calibration chart, receiver operating characteristic curve and decision curve analysis were adopted for validating the prediction model.

RESULTS

Age, plasma interleukin 6 (IL-6) concentration and plasma aspartate aminotransferase concentration were identified from 57 measured variables as potential factors distinguishing G⁺ from G^- infection by LASSO regression analysis. Inclusion of these three variables in a multivariate logistic regression model identified age and IL-6 as significant predictors. In receiver operating characteristic curve analysis, age and IL-6 yielded an area under the curve of 0.761 and distinguished G⁺ from G^- infection with specificity of 0.756 and sensitivity of 0.692. Serum IL-6 and IL-10 levels were upregulated by more than 10-fold from baseline in the G^- bacteremia group but by less than ten-fold in the G⁺ bacteremia group. The calibration curve of the model and Hosmer-Lemeshow test indicated good model fit (P > 0.05). When the decision curve analysis curve indicated a risk threshold probability between 0% and 68%, a nomogram could be applied in clinical settings.

CONCLUSION

A simple prediction model distinguishing G^- from G⁺ bacteremia can be constructed based on reciprocal association with age and IL-6 level.

Risk prediction model for distinguishing Gram-positive from Gram-negative bacteremia based on age and cytokine levels: A retrospective study

BACKGROUND

Severe infection often results in bacteremia, which significantly increases mortality rate. Different therapeutic strategies are employed depending on whether the blood-borne infection is Gram-negative (G-) or Gram-positive (G+). However, there is no risk prediction model for assessing whether bacteremia patients are infected with G- or G+ pathogens.

AIM

To establish a clinical prediction model to distinguish G- from G+ infection.

METHODS

A total of 130 patients with positive blood culture admitted to a single intensive care unit were recruited, and Th1 and Th2 cytokine concentrations, routine blood test results, procalcitonin and C-reactive protein concentrations, liver and kidney function test results and coagulation function were compared between G+ and G- groups. Least absolute shrinkage and selection operator (LASSO) regression analysis was employed to optimize the selection of predictive variables by running cyclic coordinate descent and K-fold cross-validation (K = 10). The predictive variables selected by LASSO regression analysis were then included in multivariate logistic regression analysis to establish a prediction model. A nomogram was also constructed based on the prediction model. Calibration chart, receiver operating characteristic curve and decision curve analysis were adopted for validating the prediction model.

RESULTS

Age, plasma interleukin 6 (IL-6) concentration and plasma aspartate aminotransferase concentration were identified from 57 measured variables as potential factors distinguishing G+ from G- infection by LASSO regression analysis. Inclusion of these three variables in a multivariate logistic regression model identified age and IL-6 as significant predictors. In receiver operating characteristic curve analysis, age and IL-6 yielded an area under the curve of 0.761 and distinguished G+ from G- infection with specificity of 0.756 and sensitivity of 0.692. Serum IL-6 and IL-10 levels were upregulated by more than 10-fold from baseline in the G- bacteremia group but by less than ten-fold in the G+ bacteremia group. The calibration curve of the model and Hosmer-Lemeshow test indicated good model fit (P > 0.05). When the decision curve analysis curve indicated a risk threshold probability between 0% and 68%, a nomogram could be applied in clinical settings.

CONCLUSION

A simple prediction model distinguishing G- from G+ bacteremia can be constructed based on reciprocal association with age and IL-6 level.

Development of predictive nomograms clinical use to quantify the risk of diabetic foot in patients with type 2 diabetes mellitus

Objective

The aim of the study was to explore the risk factors for diabetic foot disease in patients with type 2 diabetes mellitus and to establish and verify the nomogram model of DF risk in patients with T2DM.

Methods

The clinical data of 705 patients with type 2 diabetes who were hospitalized in our hospital from January 2015 to December 2022 were analyzed retrospectively. According to random sampling, the patients were divided into two groups: the training set (DF = 84; simple T2DM = 410) and the verification set (DF = 41; simple T2DM = 170). Univariate and multivariate logistic regression analysis was used to screen the independent risk factors for DF in patients with T2DM in the training set. According to the independent risk factors, the nomogram risk prediction model is established and verified.

Results

Logistic regression analysis showed age (OR = 1.093, 95% CI 1.062-1.124, P <0.001), smoking history (OR = 3.309, 95% CI 1.849-5.924, P <0.001), glycosylated hemoglobin (OR = 1.328, 95% CI 1.173-1.502, P <0.001), leukocyte (OR = 1.203, 95% CI 1.076-1.345, and LDL-C (OR = 2.002, 95% CI 1.463-2.740), P <0.001) was independent risk factors for T2DM complicated with DF. The area of the nomogram model based on the above indexes under the ROC curve of the training set and the verification set is 0.827 and 0.808, respectively; the correction curve shows that the model has good accuracy; and the DCA results show that when the risk threshold is between 0.10-0.85 (training set) and 0.10-0.75 (verification set), the clinical practical value of the model is higher.

Conclusion

The nomogram model constructed in this study is of high value in predicting the risk of DF in patients with T2DM and is of reference value for clinicians to identify people at high risk of DF and provide them with early diagnosis and individual prevention.

Serum cholinesterase is associated with incident diabetic retinopathy: the Shanghai Nicheng cohort study

BACKGROUND

Serum cholinesterase (ChE) is positively associated with incident diabetes and dyslipidemia. We aimed to investigate the relationship between ChE and the incidence of diabetic retinopathy (DR).

METHODS

Based on a community-based cohort study followed for 4.6 years, 1133 participants aged 55-70 years with diabetes were analyzed. Fundus photographs were taken for each eye at both baseline and follow-up investigations. The presence and severity of DR were categorized into no DR, mild non-proliferative DR (NPDR), and referable DR (moderate NPDR or worse). Binary and multinomial logistic regression models were used to estimate the risk ratio (RR) and 95% confidence interval (CI) between ChE and DR.

RESULTS

Among the 1133 participants, 72 (6.4%) cases of DR occurred. The multivariable binary logistic regression showed that the highest tertile of ChE (≥ 422 U/L) was associated with a 2.01-fold higher risk of incident DR (RR 2.01, 95%CI 1.01-4.00; P for trend < 0.05) than the lowest tertile (< 354 U/L). The multivariable binary and multinomial logistic regression showed that the risk of DR increased by 41% (RR 1.41, 95%CI 1.05-1.90), and the risk of incident referable DR was almost 2-fold higher than no DR (RR 1.99, 95%CI 1.24-3.18) with per 1-SD increase of log_e-transformed ChE. Furthermore, multiplicative interactions were found between ChE and elderly participants (aged 60 and older; P for interaction = 0.003) and men (P for interaction = 0.044) on the risk of DR.

CONCLUSIONS

In this study, ChE was associated with the incidence of DR, especially referable DR. ChE was a potential biomarker for predicting the incident DR.

Development and validation of a diabetic retinopathy risk prediction model for middle-aged patients with type 2 diabetes mellitus

Objectives

The study aims to establish a predictive nomogram of diabetic retinopathy(DR) for the middle-aged population with type 2 diabetes mellitus (T2DM).

Methods

This retrospective study screened 931 patients with T2DM between 30 and 59 years of age from the 2011-2018 National Health and Nutrition Examination Survey database. The development group comprised 704 participants from the 2011-2016 survey, and the validation group included 227 participants from the 2017-2018 survey. The least absolute shrinkage and selection operator regression model was used to determine the best predictive variables. The logistic regression analysis built three models: the full model, the multiple fractional polynomial (MFP) model, and the stepwise (stepAIC) selected model. Then we decided optimal model based on the receiver operating characteristic curve (ROC). ROC, calibration curve, Hosmer-Lemeshow test, and decision curve analysis (DCA) were used to validate and assess the model. An online dynamic nomogram prediction tool was also constructed.

Results

The MFP model was selected to be the final model, including gender, the use of insulin, duration of diabetes, urinary albumin-to-creatinine ratio, and serum phosphorus. The AUC was 0.709 in the development set and 0.704 in the validation set. According to the ROC, calibration curves, and Hosmer-Lemeshow test, the nomogram demonstrated good coherence. The nomogram was clinically helpful, according to DCA.

Conclusion

This study established and validated a predictive model for DR in the mid-life T2DM population, which can assist clinicians quickly determining who is prone to develop DR.

Network pharmacology study of the mechanism underlying the therapeutic effect of Zhujing pill and its main component oleanolic acid against diabetic retinopathy

Diabetic retinopathy (DR) is the leading cause of blindness in the working population worldwide, with few effective drugs available for its treatment in the early stages. The Zhujing pill (ZJP) is well-established to enhance the early symptoms of DR, but the mechanism underlying its therapeutic effect remains unclear. In the present study, we used systems biology and multidirectional pharmacology to screen the main active ingredients of ZJP and retrieved DrugBank and Genecards databases to obtain 'drug-disease' common targets. Using bioinformatics analysis, we obtained the core targets, and potential mechanisms of action of ZJP and its main components for the treatment of DR. Molecular docking was used to predict the binding sites and the binding affinity of the main active ingredients to the core targets. The predicted mechanism was verified in animal experiments. We found that the main active ingredient of ZJP was oleanolic acid, and 63 common 'drug-disease' targets were identified. Topological analysis and cluster analysis based on the protein-protein interaction network of the Metascape database screened the core targets as PRKCA, etc. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showed that these core targets were significantly enriched in the pro-angiogenic pathway of the VEGF signaling pathway. Molecular docking and surface plasmon resonance revealed that ZJP and its main active component, oleanolic acid had the highest binding affinity with PKC-α, the core target of the VEGF signaling pathway. Animal experiments validated that ZJP and oleanolic acid could improve DR.

A risk prediction model for type 2 diabetes mellitus complicated with retinopathy based on machine learning and its application in health management

Objective

This study aimed to establish a risk prediction model for diabetic retinopathy (DR) in the Chinese type 2 diabetes mellitus (T2DM) population using few inspection indicators and to propose suggestions for chronic disease management.

Methods

This multi-centered retrospective cross-sectional study was conducted among 2,385 patients with T2DM. The predictors of the training set were, respectively, screened by extreme gradient boosting (XGBoost), a random forest recursive feature elimination (RF-RFE) algorithm, a backpropagation neural network (BPNN), and a least absolute shrinkage selection operator (LASSO) model. Model I, a prediction model, was established through multivariable logistic regression analysis based on the predictors repeated ≥3 times in the four screening methods. Logistic regression Model II built on the predictive factors in the previously released DR risk study was introduced into our current study to evaluate the model's effectiveness. Nine evaluation indicators were used to compare the performance of the two prediction models, including the area under the receiver operating characteristic curve (AUROC), accuracy, precision, recall, F1 score, balanced accuracy, calibration curve, Hosmer-Lemeshow test, and Net Reclassification Index (NRI).

Results

When including predictors, such as glycosylated hemoglobin A1c, disease course, postprandial blood glucose, age, systolic blood pressure, and albumin/urine creatinine ratio, multivariable logistic regression Model I demonstrated a better prediction ability than Model II. Model I revealed the highest AUROC (0.703), accuracy (0.796), precision (0.571), recall (0.035), F1 score (0.066), Hosmer-Lemeshow test (0.887), NRI (0.004), and balanced accuracy (0.514).

Conclusion

We have built an accurate DR risk prediction model with fewer indicators for patients with T2DM. It can be used to predict the individualized risk of DR in China effectively. In addition, the model can provide powerful auxiliary technical support for the clinical and health management of patients with diabetes comorbidities.

Nomogram-Based Prediction of the Futile Recanalization Risk Among Acute Ischemic Stroke Patients Before and After Endovascular Therapy: A Retrospective Study

Background and Purpose

Futile recanalization (FRC) is common among large artery occlusion (LAO) patients after endovascular therapy (EVT). We developed nomogram models to identify LAO patients at a high risk of FRC pre- and post-EVT to help neurologists select the optimal candidates for EVT.

Methods

From April 2020 to July 2022, EVT and mTICI score ≥2b LAO patients were recruited. Nomogram models was developed by two-step approach for predicting the outcomes of LAO patients. First, the least absolute shrinkage and selection operator (LASSO) regression analysis was to optimize variable selection. Then, a multivariable analysis was to construct an estimation model with significant indicators from the LASSO. The accuracy of the model was verified using receiver operating characteristic (ROC), calibration curve, and decision curve analyses (DCA), along with validation cohort (VC).

Results

Using LASSO, age, sex, hypertension history, baseline NIHSS, ASPECTS and baseline SBP upon admission were identified from the pre-EVT variables. Model 1 (pre-EVT) showed good predictive performance, with an area under the ROC curve (AUC) of 0.815 in the training cohort (TrC) and 0.904 in VC. Under the DCA, the generated nomogram was clinically applicable where risk cut-off was between 15%-85% in the TrC and 5%-100% in the VC. Moreover, age, ASPECTS upon admission, onset duration, puncture-to-recanalization (PTR) duration, and lymphocyte-to-monocyte ratio (LMR) were screened by LASSO. Model 2 (post-EVT) also demonstrated good predictive performance with AUCs of 0.888 and 0.814 for TrC and VC, respectively. Under the DCA, the generated nomogram was clinically applicable if the risk cut-off was between 13-100% in the TrC and 22-85% of VC.

Conclusion

In this study, two nomogram models were generated that showed good discriminative performance, improved calibration, and clinical benefits. These nomograms can potentially accurately predict the risk of FRC in LAO patients pre- and post-EVT and help to select appropriate candidates for EVT.

Nomogram model predicts the risk of visual impairment in diabetic retinopathy: a retrospective study

BACKGROUND

To develop a model for predicting the risk of visual impairment in diabetic retinopathy (DR) by a nomogram.

METHODS

Patients with DR who underwent both optical coherence tomography angiography (OCTA) and fundus fluorescein angiography (FFA) were retrospectively enrolled. FFA was conducted for DR staging, swept-source optical coherence tomography (SS-OCT) of the macula and 3*3-mm blood flow imaging by OCTA to observe retinal structure and blood flow parameters. We defined a logarithm of the minimum angle of resolution visual acuity (LogMAR VA) ≥0.5 as visual impairment, and the characteristics correlated with VA were screened using binary logistic regression. The selected factors were then entered into a multivariate binary stepwise regression, and a nomogram was developed to predict visual impairment risk. Finally, the model was validated using the area under the receiver operating characteristic (ROC) curve (AUC), calibration plots, decision curve analysis (DCA), and clinical impact curve (CIC).

RESULTS

A total of 29 parameters were included in the analysis, and 13 characteristics were used to develop a nomogram model. Finally, diabetic macular ischaemia (DMI) grading, disorganization of the retinal inner layers (DRIL), outer layer disruption, and the vessel density of choriocapillaris layer inferior (SubVD) were found to be statistically significant (P < 0.05). The model was found to have good accuracy based on the ROC (AUC = 0.931) and calibration curves (C-index = 0.930). The DCA showed that risk threshold probabilities in the (3-91%) interval models can be used to guide clinical practice, and the proportion of people at risk at each threshold probability is illustrated by the CIC.

CONCLUSION

The nomogram model for predicting visual impairment in DR patients demonstrated good accuracy and utility, and it can be used to guide clinical practice.

TRIAL REGISTRATION

Chinese Clinical Trial Registry, ChiCTR2200059835. Registered 12 May 2022, https://www.chictr.org.cn/edit.aspx?pid=169290&htm=4.

A Nomogram Model for Prediction of Tracheostomy in Patients With Traumatic Cervical Spinal Cord Injury

OBJECTIVE

To develop a nomogram for the prediction of tracheostomy in patients with traumatic cervical spinal cord injury (TCSCI).

METHODS

A total of 689 TCSCI patients were included in our study. First, the variable selection was performed using between-group comparisons and LASSO regression analysis. Second, a multivariate logistic regression analysis (MLRA) with a step-by-step method was performed. A nomogram model was developed based on the MLRA. Finally, the model was validated on the training set and validation set.

RESULTS

The nomogram prediction model incorporated 5 predictors, including smoking history, dislocation, thoracic injury, American Spinal Injury Association (ASIA) grade, and neurological level of injury (NLI). The area under curve in the training group and in the validation group were 0.883 and 0.909, respectively. The Hosmer-Lemeshow test result was p = 0.153. From the decision curve analysis curve, the model performed well and was feasible to make beneficial clinical decisions.

CONCLUSION

The nomogram combining dislocation, thoracic injury, ASIA grade A, NLI, and smoking history was validated as a reliable model for the prediction of tracheostomy.

Nomogram-based prediction of clinically significant macular edema in diabetes mellitus patients

AIMS

The aim of the study was to construct and validate a risk nomogram for clinically significant macular edema (CSME) prediction in diabetes mellitus (DM) patients using systemic variables.

METHODS

In this retrospective study, DM inpatients who underwent routine diabetic retinopathy screening were recruited and divided into training and validation sets according to their admission date. Ninety-three demographic and systemic variables were collected. The least absolute shrinkage and selection operator was used to select the predictive variables from the training set. The selected variables were used to construct the CSME prediction nomogram. Internal and external validations were performed. The C-index, calibration curve and decision curve analysis (DCA) were reported.

RESULTS

A total of 349 patients were divided into the training set (240, 68.77%) and the validation set (109, 31.23%). The presence of diabetic peripheral neuropathy (DPN) symptoms, uric acid, use of insulin only or not for treatment, insulin dosage, urinary protein grade and disease duration were chosen for the nomogram. The C-index of the prediction nomogram was 0.896, 0.878 and 0.837 in the training set, internal validation and external validation, respectively. The calibration curves of the nomogram showed good agreement between the predicted and actual outcomes. DCA demonstrated that the nomogram was clinically useful.

CONCLUSIONS

A nomogram with good performance for predicting CSME using systemic variables was developed. It suggested that DPN symptoms and renal function may be crucial risk factors for CSME. Moreover, this nomogram may be a convenient tool for non-ophthalmic specialists to rapidly recognize CSME in patients and to transfer them to ophthalmologists for early diagnosis and treatment.

Development and validation of a nomogram for predicting mild cognitive impairment in middle-aged and elderly people

BACKGROUND

Mild cognitive impairment (MCI) is a clinical cognitive impairment state between dementia and normal aging. Early identification of MCI is beneficial, and it can delay the development of dementia. We aimed to develop and validate a prediction model to predict MCI of middle-aged and elderly people (aged 45 years and over).

METHODS

According to 478 middle-aged and elderly people (48-85 years old) from a cross-sectional study, we developed and validated a predictive nomogram. The least absolute shrinkage and selection operator (LASSO) regression model and multivariate logistic regression analysis were used to select variables and develop a prediction model. The performance of the nomogram was evaluated in terms of its discriminative power, calibration, and decision curve analysis (DCA).

RESULTS

The predictive nomogram was composed of the following: age, gender, education level, residence, and reading. The model showed good discrimination power (area under receiver-operating characteristic (ROC) curve was 0.8704) and good calibration. Similar results were seen in 10-fold cross-validation. The nomogram showed clinically useful in DCA analysis.

CONCLUSION

This predictive nomogram provides researchers with a practical tool for predicting MCI. The variables included in this nomogram were readily available. The population used for this nomogram was middle-aged and elderly people.

Prediction of Intracranial Infection in Patients under External Ventricular Drainage and Neurological Intensive Care: A Multicenter Retrospective Cohort Study

Objective: To generate an optimal prediction model along with identifying major contributors to intracranial infection among patients under external ventricular drainage and neurological intensive care. Methods: A retrospective cohort study was conducted among patients admitted into neurointensive care units between 1 January 2015 and 31 December 2020 who underwent external ventricular drainage due to traumatic brain injury, hydrocephalus, and nonaneurysmal spontaneous intracranial hemorrhage. Multivariate logistic regression in combination with the least absolute shrinkage and selection operator regression was applied to derive prediction models and optimize variable selections. Other machine-learning algorithms, including the support vector machine and K-nearest neighbor, were also applied to derive alternative prediction models. Five-fold cross-validation was used to train and validate each model. Model performance was assessed by calibration plots, receiver operating characteristic curves, and decision curves. A nomogram analysis was developed to explicate the weights of selected features for the optimal model. Results: Multivariate logistic regression showed the best performance among the three tested models with an area under curve of 0.846 ± 0.006. Six variables, including hemoglobin, albumin, length of operation time, American Society of Anesthesiologists grades, presence of traumatic subarachnoid hemorrhage, and a history of diabetes, were selected from 37 variable candidates as the top-weighted prediction features. The decision curve analysis showed that the nomogram could be applied clinically when the risk threshold is between 20% and 100%. Conclusions: The occurrence of external ventricular-drainage-associated intracranial infections could be predicted using optimal models and feature-selection approaches, which would be helpful for the prevention and treatment of this complication in neurointensive care units.

Prediction Model for the Risk of HIV Infection among MSM in China: Validation and Stability

The impact of psychosocial factors on increasing the risk of HIV infection among men who have sex with men (MSM) has attracted increasing attention. We aimed to develop and validate an integrated prediction model, especially incorporating emerging psychosocial variables, for predicting the risk of HIV infection among MSM. We surveyed and collected sociodemographic, psychosocial, and behavioral information from 547 MSM in China. The participants were split into a training set and a testing set in a 3:1 theoretical ratio. The prediction model was constructed by introducing the important variables selected with the least absolute shrinkage and selection operator (LASSO) regression, applying multivariate logistic regression, and visually assessing the risk of HIV infection through the nomogram. Receiver operating characteristic curves (ROC), Kolmogorov–Smirnov test, calibration plots, Hosmer–Lemeshow test and population stability index (PSI) were performed to test validity and stability of the model. Four of the 15 selected variables—unprotected anal intercourse, multiple sexual partners, involuntary subordination and drug use before sex—were included in the prediction model. The results indicated that the comprehensive prediction model we developed had relatively good predictive performance and stability in identifying MSM at high-risk for HIV infection, thus providing targeted interventions for high-risk MSM.

Diabetic retinopathy risk prediction in patients with type 2 diabetes mellitus using a nomogram model

BACKGROUND

This study aims to develop a diabetic retinopathy (DR) hazard nomogram for a Chinese population of patients with type 2 diabetes mellitus (T2DM).

METHODS

We constructed a nomogram model by including data from 213 patients with T2DM between January 2019 and May 2021 in the Affiliated Hospital of Zunyi Medical University. We used basic statistics and biochemical indicator tests to assess the risk of DR in patients with T2DM. The patient data were used to evaluate the DR risk using R software and a least absolute shrinkage and selection operator (LASSO) predictive model. Using multivariable Cox regression, we examined the risk factors of DR to reduce the LASSO penalty. The validation model, decision curve analysis, and C-index were tested on the calibration plot. The bootstrapping methodology was used to internally validate the accuracy of the nomogram.

RESULTS

The LASSO algorithm identified the following eight predictive variables from the 16 independent variables: disease duration, body mass index (BMI), fasting blood glucose (FPG), glycated hemoglobin (HbA1c), homeostatic model assessment-insulin resistance (HOMA-IR), triglyceride (TG), total cholesterol (TC), and vitamin D (VitD)-T3. The C-index was 0.848 (95% CI: 0.798-0.898), indicating the accuracy of the model. In the interval validation, high scores (0.816) are possible from an analysis of a DR nomogram's decision curve to predict DR.

CONCLUSION

We developed a non-parametric technique to predict the risk of DR based on disease duration, BMI, FPG, HbA1c, HOMA-IR, TG, TC, and VitD.

Using Machine Learning Techniques to Develop Risk Prediction Models for the Risk of Incident Diabetic Retinopathy Among Patients With Type 2 Diabetes Mellitus: A Cohort Study

OBJECTIVE

To construct and validate prediction models for the risk of diabetic retinopathy (DR) in patients with type 2 diabetes mellitus.

METHODS

Patients with type 2 diabetes mellitus hospitalized over the period between January 2010 and September 2018 were retrospectively collected. Eighteen baseline demographic and clinical characteristics were used as predictors to train five machine-learning models. The model that showed favorable predictive efficacy was evaluated at annual follow-ups. Multi-point data of the patients in the test set were utilized to further evaluate the model's performance. We also assessed the relative prognostic importance of the selected risk factors for DR outcomes.

RESULTS

Of 7943 collected patients, 1692 (21.30%) developed DR during follow-up. Among the five models, the XGBoost model achieved the highest predictive performance with an AUC, accuracy, sensitivity, and specificity of 0.803, 88.9%, 74.0%, and 81.1%, respectively. The XGBoost model's AUCs in the different follow-up periods were 0.834 to 0.966. In addition to the classical risk factors of DR, serum uric acid (SUA), low-density lipoprotein cholesterol (LDL-C), total cholesterol (TC), estimated glomerular filtration rate (eGFR), and triglyceride (TG) were also identified to be important and strong predictors for the disease. Compared with the clinical diagnosis method of DR, the XGBoost model achieved an average of 2.895 years prior to the first diagnosis.

CONCLUSION

The proposed model achieved high performance in predicting the risk of DR among patients with type 2 diabetes mellitus at each time point. This study established the potential of the XGBoost model to facilitate clinicians in identifying high-risk patients and making type 2 diabetes management-related decisions.

Prognostic models of diabetic microvascular complications: a systematic review and meta-analysis

BACKGROUND

Many prognostic models of diabetic microvascular complications have been developed, but their performances still varies. Therefore, we conducted a systematic review and meta-analysis to summarise the performances of the existing models.

METHODS

Prognostic models of diabetic microvascular complications were retrieved from PubMed and Scopus up to 31 December 2020. Studies were selected, if they developed or internally/externally validated models of any microvascular complication in type 2 diabetes (T2D).

RESULTS

In total, 71 studies were eligible, of which 32, 30 and 18 studies initially developed prognostic model for diabetic retinopathy (DR), chronic kidney disease (CKD) and end stage renal disease (ESRD) with the number of derived equations of 84, 96 and 51, respectively. Most models were derived-phases, some were internal and external validations. Common predictors were age, sex, HbA1c, diabetic duration, SBP and BMI. Traditional statistical models (i.e. Cox and logit regression) were mostly applied, otherwise machine learning. In cohorts, the discriminative performance in derived-logit was pooled with C statistics of 0.82 (0.73‑0.92) for DR and 0.78 (0.74‑0.83) for CKD. Pooled Cox regression yielded 0.75 (0.74‑0.77), 0.78 (0.74‑0.82) and 0.87 (0.84‑0.89) for DR, CKD and ESRD, respectively. External validation performances were sufficiently pooled with 0.81 (0.78‑0.83), 0.75 (0.67‑0.84) and 0.87 (0.85‑0.88) for DR, CKD and ESRD, respectively.

CONCLUSIONS

Several prognostic models were developed, but less were externally validated. A few studies derived the models by using appropriate methods and were satisfactory reported. More external validations and impact analyses are required before applying these models in clinical practice.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42018105287.

A predictive nomogram: a cross-sectional study on a simple-to-use model for screening 12-year-old children for severe caries in middle schools

BACKGROUND

A nomogram is a tool that transforms complex regression equations into simple and visual graphs and enables clinicians and patients to conveniently compute output probabilities without needing medical knowledge and complex formulas. The aim of this study was to develop and validate a predictive nomogram to screen for severe caries among 12-year-old children based on risk factors in Sichuan Province, China.

METHODS

A cross-sectional study of 4573 12-year-olds was conducted up to May 2016 in middle schools from three districts and three counties in Sichuan Province, China. All the children underwent oral examinations and completed questionnaires to assess general information, oral impacts on daily performance, dietary habits, subjective health conditions, history of dental trauma, frequency of toothache, dental visits, and knowledge, attitudes, and behaviours toward oral hygiene. Univariate analysis and multivariate logistic regression analysis were used to determine which variables were significantly associated with severe caries (operationalized as DMFT ≥ 3). A nomogram was developed and validated by using the 'rms' package and two cross-validation methods.

RESULTS

Severe caries was found in 537 of the 4573 children (11.74%). Multivariate logistic regression analysis revealed that the following variables predicted a higher risk of severe caries: 'female' [odds ratio (OR) = 1.985, 95% confidence interval (95% CI): 1.63-2.411], 'urban' (OR = 2.389, 95% CI: 1.96-2.91), 'non-only child' (OR = 1.317, 95% CI: 1.07-1.625), 'very poor self-assessment of oral health status' (OR = 2.157, 95% CI: 1.34-3.467) and 'visited a dentist less than 6 months' (OR = 1.861, 95% CI: 1.38-2.505). Multivariate logistic regression analysis also indicated that the following variables predicted a lower risk of severe caries: 'middle level of urbanization' (OR = 0.395, 95% CI: 0.32-0.495) and 'high level of urbanization' (OR = 0.466, 95% CI: 0.37-0.596). Both the fivefold and leave-one-out cross-validation methods indicated that the nomogram model built by these 6 variables displayed good disease recognition ability.

CONCLUSIONS

The nomogram was a simple-to-use model to screen children for severe caries. This model was found to facilitate non-dental professionals in assessing risk values without oral examinations and making referrals to dental professionals.

Nomogram Prediction Model for Diabetic Retinopathy Development in Type 2 Diabetes Mellitus Patients: A Retrospective Cohort Study

BACKGROUND

This study is aimed at investigating the systemic risk factors of diabetic retinopathy and further establishing a risk prediction model for DR development in T2DM patients.

METHODS

This is a retrospective cohort study including 330 type 2 diabetes mellitus (T2DM) patients who were followed up from December 2012 to November 2020. Multivariable cox regression analysis identifying factors associated with the hazard of developing diabetic retinopathy (DR) was used to construct the DR risk prediction model in the form of nomogram.

RESULTS

50.6% of participants (mean age: 58.60 ± 10.55) were female, and mean duration of diabetes was 7.09 ± 5.36 years. After multivariate cox regression, the risk factors for developing DR were age (HR 1.068, 95%Cl 1.021-1.118, P = 0.005), diabetes duration (HR 1.094, 95%Cl 1.018-1.177, P = 0.015), HbA1c (HR 1.411, 95%Cl 1.113-1.788, P = 0.004), albuminuria (HR 6.908, 95%Cl 1.794-26.599, P = 0.005), and triglyceride (HR 1.554, 95%Cl 1.037-2.330, P = 0.033). The AUC values of the nomogram for predicting developing DR at 3-, 4-, and 5-year were 0.854, 0.845, and 0.798.

CONCLUSION

Combining age, diabetes duration, HbA1c, albuminuria, and triglyceride, the nomogram model is effective for early recognition and intervention of individuals at high risk of DR development.

Nomogram Model for Screening the Risk of Type II Diabetes in Western Xinjiang, China

OBJECTIVE

A simple type 2 diabetes mellitus (T2DM) screening model was established preciously based on easily available variables for identifying high-risk individuals in western Xinjiang, China.

METHODS

A total of 458,153 cases participating in the national health examination were recruited. Logistic regression and the least absolute shrinkage and selection operator (LASSO) models were used for univariate analysis, factors selection, and the establishment of prediction model. Receiver operating characteristic (ROC) curve, Hosmer-Lemeshow test and clinical decision curve (CDA) were applied for evaluating the discrimination, calibration and clinical validity, respectively. The optimal threshold for predicting risk factors for T2DM has been estimated as well.

RESULTS

The nomogram depicted the risk of T2DM based on different genders, the factors mainly consisted of age, family history of T2DM (FHOT), waist circumference (WC), total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDLc), body mass index (BMI), high-density lipoprotein cholesterol (HDLc), etc. The area under ROC of men and women was 0.864 and 0.816 in the development group, similarly in the validation group, which was 0.865 and 0.815, respectively. The calibration curve showed that the nomogram was accurate for predicting the risk of T2DM, and the CDA proved great clinical application value of the nomogram. Threshold values of the age, WC, TC, TG, HDLc, BMI in different genders were 52.5 years old (men) and 48.5 years old (women), 85.50 cm (men) and 89.9 cm (women), 4.94 mmol/L (men) and 4.94mmol/L (women), 1.26mmol/L (men) and 1.67mmol/L (women), 1.40mmol/L (men) and 1.40mmol/L (women), 24.70kg/m² (men) and 24.95kg/m² (women), respectively.

CONCLUSION

Our results give a clue that the nomogram may be useful for identifying adults who have high risk for diabetes, which is simple, affordable, with high credibility and can be widely implemented. Further studies are needed to evaluate the utility and feasibility of this model in various settings.

A Nomogram Model that Predicts the Risk of Diabetic Nephropathy in Type 2 Diabetes Mellitus Patients: A Retrospective Study

OBJECTIVE

To construct a novel nomogram model that predicts the risk of diabetic nephropathy (DN) incidence in Chinese patients with type 2 diabetes mellitus (T2DM).

METHODS

Questionnaire surveys, physical examinations, routine blood tests, and biochemical index evaluations were conducted on 1095 patients with T2DM from Guilin. A least absolute contraction selection operator (LASSO) regression and multivariable logistic regression analysis were used to screen out DN risk factors. A logistic regression analysis incorporating the screened risk factors was used to establish a predictive nomogram model. The performance of the nomogram model was evaluated using the C-index, an area under the receiver operating characteristic curve (AUC), calibration plots, and a decision curve analysis. Bootstrapping was applied for internal validation.

RESULTS

Independent predictors for DN incidence risk included gender, age, hypertension, medicine use, duration of diabetes, body mass index, blood urea nitrogen level, serum creatinine level, neutrophil to lymphocyte ratio, and red blood cell distribution width. The nomogram model exhibited moderate prediction ability with a C-index of 0.819 (95% confidence interval (CI): 0.783-0.853) and an AUC of 0.813 (95%CI: 0.778-0.848). The C-index from internal validation reached 0.796 (95%CI: 0.763-0.829). The decision curve analysis displayed that the DN risk nomogram was clinically applicable when the risk threshold was between 1 and 83%.

CONCLUSION

Our novel and simple nomogram containing 10 factors may be useful in predicting DN incidence risk in T2DM patients.