Calibration Performance Research Articles

Raising awareness of potential biases in medical machine learning: Experience from a Datathon.

To challenge clinicians and informaticians to learn about potential sources of bias in medical machine learning models through investigation of data and predictions from an open-source severity of illness score. Over a two-day period (total elapsed time approximately 28 hours), we conducted a datathon that challenged interdisciplinary teams to investigate potential sources of bias in the Global Open Source Severity of Illness Score. Teams were invited to develop hypotheses, to use tools of their choosing to identify potential sources of bias, and to provide a final report. Five teams participated, three of which included both informaticians and clinicians. Most (4/5) used Python for analyses, the remaining team used R. Common analysis themes included relationship of the GOSSIS-1 prediction score with demographics and care related variables; relationships between demographics and outcomes; calibration and factors related to the context of care; and the impact of missingness. Representativeness of the population, differences in calibration and model performance among groups, and differences in performance across hospital settings were identified as possible sources of bias. Datathons are a promising approach for challenging developers and users to explore questions relating to unrecognized biases in medical machine learning algorithms. Disadvantaged groups are at risk of being adversely impacted by biased medical machine learning models. To avoid these undesirable outcomes, developers and users must understand the challenges involved in identifying potential biases. We conducted a datathon aimed at challenging a diverse group of participants to explore an open-source patient severity model for potential biases. Five groups of clinicians and informaticians used tools of their choosing to evaluate possible sources of biases, applying a range of analytic techniques and exploring multiple features. By engaging diverse participants with hands-on data experience with meaningful data, datathons have the potential to raise awareness of potential biases and promote best practices in developing fair and equitable medical machine learning models.

To Determine the Risk-Based Screening Interval for Diabetic Retinopathy: Development and Validation of Risk Algorithm from a Retrospective Cohort Study.

The optimal screening interval for diabetic retinopathy (DR) remains controversial. This study aimed to develop a risk algorithm to predict the individual risk of referable sight-threatening diabetic retinopathy (STDR) in a mainly Chinese population and to provide evidence for risk-based screening intervals. The retrospective cohort data from 117,418 subjects who received systematic DR screening in Hong Kong between 2010 and 2016 were included to develop and validate the risk algorithm using a parametric survival model. The risk algorithm can be used to predict the individual risk of STDR within a specific time interval, or the time to reach a specific risk margin and thus to allocate a screening interval. The calibration performance was assessed by comparing the cumulative STDR events versus predicted risk over 2 years, and discrimination by using receiver operative characteristics (ROC) curve. Duration of diabetes, glycosylated hemoglobin, systolic blood pressure, presence of chronic kidney disease, diabetes medication, and age were included in the risk algorithm. The validation of prediction performance showed that there was no significant difference between predicted and observed STDR risks in males (5.6% vs. 5.1%, P=0.724) or females (4.8% vs. 4.6%, P=0.099). The area under the receiver operating characteristic curve was 0.80 (95% confidence interval [CI], 0.78 to 0.81) for males and 0.81 (95% CI, 0.79 to 0.83) for females. The risk algorithm has good prediction performance for referable STDR. Using a risk-based screening interval allows us to allocate screening visits disproportionally more to those at higher risk, while reducing the frequency of screening of lower risk people.

Target-specified reference-based deep learning network for joint image deblurring and resolution enhancement in surgical zoom lens camera calibration

Background and objectiveFor the augmented reality of surgical navigation, which overlays a 3D model of the surgical target on an image, accurate camera calibration is imperative. However, when the checkerboard images for calibration are captured using a surgical microscope having high magnification, blur owing to the narrow depth of focus and blocking artifacts caused by limited resolution around the fine edges occur. These artifacts strongly affect the localization of corner points of the checkerboard in these images, resulting in inaccurate calibration, which leads to a large displacement in augmented reality. To solve this problem, in this study, we proposed a novel target-specific deep learning network that simultaneously enhances both the blur and spatial resolution of an image for surgical zoom lens camera calibration. MethodsAs a scheme of an end-to-end convolutional deep neural network, the proposed network is specifically intended for the checkerboard image enhancement used in camera calibration. Through the symmetric architecture of the network, which consists of encoding and decoding layers, the distinctive spatial features of the encoding layers are transferred and merged with the output of the decoding layers. Additionally, by integrating a multi-frame framework including subpixel motion estimation and ideal reference image with the symmetric architecture, joint image deblurring and enhanced resolution were efficiently achieved. ResultsFrom experimental comparisons, we verified the capability of the proposed method to improve the subjective and objective performances of surgical microscope calibration. Furthermore, we confirmed that the augmented reality overlap ratio, which quantitatively indicates augmented reality accuracy, from calibration with the enhanced image of the proposed method is higher than that of the previous methods. ConclusionsThese findings suggest that the proposed network provides sharp high-resolution images from blurry low-resolution inputs. Furthermore, we demonstrate superior performance in camera calibration by using surgical microscopic images, thus showing its potential applications in the field of practical surgical navigation.

Validation of an Actuarial Instrument to Predict Repeat Family Violence for Frontline Police: Calibration, Discrimination and Classification Performance of the VP-SAFvR

Abstract Purpose Police are increasingly tasked with identifying, responding to, and preventing intimate partner and family violence (IPFV). Several police organisations have adopted validated risk assessment instruments to assist frontline police officers to assess and manage the risk of future IPFV. This paper reports a population-based evaluation of the predictive and discriminative validity of an actuarial risk assessment instrument, the Victoria Police Screening Assessment for Family Violence Risk (VP-SAFvR), used by frontline police officers in Victoria, Australia. Methods The entire population (n = 38,454) of family violence reports (FVRs) where police used the VP-SAFvR between 1 September 2019 and 31 March 2020 were tracked for 365 days to ascertain whether the VP-SAFvR was able to predict subsequent family violence and discriminate between cases with and without subsequent family violence. Calibration and survival analysis was performed on the instrument. Results Approximately 33% of the sample was involved in subsequent IPFV over the 12 month outcome period. In line with the development (McEwan et al., 2019) and validation (Spivak et al., 2020) studies of the same instrument, the results show the VP-SAFvR continues to provide adequate consistent discriminative and predictive performance across time. The AUC for repeat IPFV was 0.65, with the recommended threshold score of 4 (or above) identifying 75% of individuals involved in repeat family violence and correctly screening out approximately 50% of non-repeat cases. Conclusions The paper extends the previous research on VP-SAFvR by providing some indication that the tool is well-calibrated. Incidence curves show the largest increase in repeat family violence among those scoring above threshold occurs in the first 90 days post index. The implications and limitations of the study are discussed.

Nationwide study on development and validation of a risk prediction model for CIN3+ and cervical cancer in Estonia

Transitioning to an individualized risk-based approach can significantly enhance cervical cancer screening programs. We aimed to derive and internally validate a prediction model for assessing the risk of cervical intraepithelial neoplasia grade 3 or higher (CIN3+) and cancer in women eligible for screening. This retrospective study utilized data from the Estonian electronic health records, including 517,884 women from the health insurance database and linked health registries. We employed Cox proportional hazard regression, incorporating reproductive and medical history variables (14 covariates), and utilized the least absolute shrinkage and selection operator (LASSO) for variable selection. A 10-fold cross-validation for internal validation of the model was used. The main outcomes were the performance of discrimination and calibration. Over the 8-year follow-up, we identified 1326 women with cervical cancer and 5929 with CIN3+, with absolute risks of 0.3% and 1.1%, respectively. The prediction model for CIN3 + and cervical cancer had good discriminative power and was well calibrated Harrell’s C of 0.74 (0.73–0.74) (calibration slope 1.00 (0.97–1.02) and 0.67 (0.66–0.69) (calibration slope 0.92 (0.84–1.00) respectively. A developed model based on nationwide electronic health data showed potential utility for risk stratification to supplement screening efforts. This work was supported through grants number PRG2218 from the Estonian Research Council, and EMP416 from the EEA (European Economic Area) and Norway Grants.

Predicting vision-threatening diabetic retinopathy in patients with type 2 diabetes mellitus: Systematic review, meta-analysis, and prospective validation study.

Delayed diagnosis and treatment of vision-threatening diabetic retinopathy (VTDR) is a common cause of visual impairment in individuals with type 2 diabetes mellitus (T2DM). Identification of VTDR predictors is the key to early prevention and intervention, but the predictors from previous studies are inconsistent. This study aims to conduct a systematic review and meta-analysis of the existing evidence for VTDR predictors, then to develop a risk prediction model after quantitatively summarising the predictors across studies, and finally to validate the model with two Chinese cohorts. We systematically retrieved cohort studies that reported predictors of VTDR in T2DM patients from PubMed, Ovid, Embase, Scopus, Cochrane Library, Web of Science, and ProQuest from their inception to December 2023. We extracted predictors reported in two or more studies and combined their corresponding relative risk (RRs) using meta-analysis to obtain pooled RRs. We only selected predictors with statistically significant pooled RRs to develop the prediction model. We also prospectively collected two Chinese cohorts of T2DM patients as the validation set and assessed the discrimination and calibration performance of the prediction model by the time-dependent ROC curve and calibration curve. Twenty-one cohort studies involving 622 490 patients with T2DM and 57 107 patients with VTDR were included in the meta-analysis. Age of diabetes onset, duration of diabetes, glycosylated haemoglobin (HbA1c), estimated glomerular filtration rate (eGFR), hypertension, high albuminuria and diabetic treatment were used to construct the prediction model. We validated the model externally in a prospective multicentre cohort of 555 patients with a median follow-up of 52 months (interquartile range = 39-77). The area under the curve (AUC) of the prediction model was all above 0.8 for 3- to 10-year follow-up periods and different cut-off value of each year provided the optimal balance between sensitivity and specificity. The data points of the calibration curves for each year closely surround the corresponding dashed line. The risk prediction model of VTDR has high discrimination and calibration performance based on validation cohorts. Given its demonstrated effectiveness, there is significant potential to expand the utilisation of this model within clinical settings to enhance the detection and management of individuals at high risk of VTDR.

Delayed gastric emptying risk stratification in patients with pancreatic ductal adenocarcinoma after pancreatoduodenectomy: An international validation cohort study.

Currently, there is still a lack of an accurate predictive model for delayed gastric emptying (DGE) following pancreaticoduodenectomy (PD) in patients with pancreatic ductal adenocarcinoma (PDAC). The aim of this study was to develop a concise model that could effectively predict the risk of DGE. This retrospective cohort study included a training cohort of 1251 consecutive PDAC patients who underwent PD from the US multicenter ACS-NSQIP database. Additionally, a validation cohort of 934 consecutive PDAC patients who underwent PD was included from the National Cancer Center in China. A total of 46 perioperative indicators were incorporated in the analysis. The DGE risk stratification (DGERS) model was then developed and validated using Lasso-logistic regression. After screening using Lasso-logistic regression, we identified four independent predictors that were significantly correlated with DGE: days to pancreatic drain removal (HR, 1.05; 95% CI, 1.02-1.08; p<0.001), pancreatic fistula (HR, 2.61; 95% CI, 1.65-4.12; p<0.001), sepsis/septic shock (HR, 2.46; 95% CI, 1.52-3.91; p<0.001), and reoperation (HR, 4.16; 95% CI, 2.27-7.57; p<0.001). Based on these factors, we developed a nomogram to predict postoperative DGE. The model demonstrated excellent calibration and optimal performance in the validation cohorts (AUC, 0.73; 95% CI, 0.67-0.73). In the validation cohort, the DGERS exhibited significant risk stratification ability, with AUC values of 0.7, 0.61, and 0.74 for the low-, moderate-, and high-risk groups, respectively. This study identified four factors that independently increased the occurrence of DGE in patients with PDAC after PD, including days to pancreatic drain removal, pancreatic fistula, sepsis/septic shock, and reoperation. Based on these findings, we developed a personalized and straightforward DGERS that enables dynamic and precise prediction of DGE risk, allowing for effective stratification of individuals based on their risk profiles.

Scoring and validation of a simple model for predicting diabetic retinopathy in patients with type 2 diabetes based on a meta-analysis approach of 21 cohorts

Aim To develop and validate a model for predicting diabetic retinopathy (DR) in patients with type 2 diabetes. Methods All risk factors with statistical significance in the DR prediction model were scored by their weights. Model performance was evaluated by the area under the receiver operating characteristic (ROC) curve, Kaplan–Meier curve, calibration curve and decision curve analysis. The prediction model was externally validated using a validation cohort from a Chinese hospital. Results In this meta-analysis, 21 cohorts involving 184,737 patients with type 2 diabetes were examined. Sex, smoking, diabetes mellitus (DM) duration, albuminuria, glycated haemoglobin (HbA1c), systolic blood pressure (SBP) and TG were identified to be statistically significant. Thus, they were all included in the model and scored according to their weights (maximum score: 35.0). The model was validated using an external cohort with median follow-up time of 32 months. At a critical value of 16.0, the AUC value, sensitivity and specificity of the validation cohort are 0.772 ((95% confidence interval (95%CI): 0.740–0.803), p < .01), 0.715 and 0.775, respectively. The calibration curve lied close to the ideal diagonal line. Furthermore, the decision curve analysis demonstrated that the model had notably higher net benefits. The external validation results proved the reliability of the risk prediction model. Conclusions The simple DR prediction model developed has good overall calibration and discrimination performance. It can be used as a simple tool to detect patients at high risk of DR.

Comparison of deep-learning multimodality data fusion strategies in mandibular osteoradionecrosis NTCP modelling using clinical variables and radiation dose distribution volumes

Objective.Normal tissue complication probability (NTCP) modelling is rapidly embracing deep learning (DL) methods, acknowledging the importance of spatial dose information. Finding effective ways to combine information from radiation dose distribution maps (dosiomics) and clinical data involves technical challenges and requires domain knowledge. We propose different multi-modality data fusion strategies to facilitate future DL-based NTCP studies.Approach.Early, joint and late DL multi-modality fusion strategies were compared using clinical and mandibular radiation dose distribution volumes. These were contrasted with single-modality models: a random forest trained on non-image data (clinical, demographic and dose-volume metrics) and a 3D DenseNet-40 trained on image data (mandibular dose distribution maps). The study involved a matched cohort of 92 osteoradionecrosis cases and 92 controls from a single institution.Main results.The late fusion model exhibited superior discrimination and calibration performance, while the join fusion achieved a more balanced distribution of the predicted probabilities. Discrimination performance did not significantly differ between strategies. Late fusion, though less technically complex, lacks crucial inter-modality interactions for NTCP modelling. In contrast, joint fusion, despite its complexity, resulted in a single network training process which included intra- and inter-modality interactions in its model parameter optimisation.Significance.This study is a pioneering effort in comparing different strategies for including image data into DL-based NTCP models in combination with lower dimensional data such as clinical variables. The discrimination performance of such multi-modality NTCP models and the choice of fusion strategy will depend on the distribution and quality of both types of data. Multiple data fusion strategies should be compared and reported in multi-modality NTCP modelling using DL.

Digital background calibration algorithm for pipelined ADC based on time-delay neural network with genetic algorithm feature selection

This paper presents a novel background calibration method for pipelined analog-to-digital converters (ADCs) using a time-delay neural network (TDNN), which is optimized through genetic algorithm (GA) techniques. The proposed technique leverages TDNN to create enhanced feature sets, significantly improving the calibration of nonlinear errors exhibiting memory effects. It harnesses the GA's global optimization capabilities for feature selection, effectively reducing the feature dimension and consequently alleviating the NN's computational burden. A parallel pipeline architecture is devised for the calibration circuit, with its implementation realized on FPGA to facilitate forward inference processing. The inference circuit is synthesized using TSMC's 90 nm CMOS process, achieving a power consumption of 40.11 mW and an area of 0.45 mm2. Simulations based on MATLAB for a 14-bit Pipelined ADC demonstrate that the proposed calibration method significantly improves the SFDR from 59.77 dB to 165.52 dB, and ENOB from 8.79 bits to 19.23 bits, surpassing the target ADC's specifications. Moreover, the dimensionality of features is effectively reduced by up to 34 % without compromising the calibration performance.

Quantitative MRI biomarker for classification of clinically significant prostate cancer: Calibration for reproducibility across echo times.

The purpose of the present study is to develop a calibration method to account for differences in echo times (TE) and facilitate the use of restriction spectrum imaging restriction score (RSIrs) as a quantitative biomarker for the detection of clinically significant prostate cancer (csPCa). This study included 197 consecutive patients who underwent MRI and biopsy examination; 97 were diagnosed with csPCa (grade group ≥ 2). RSI data were acquired three times during the same session: twice at minimum TE ~75ms and once at TE=90ms (TEmin1, TEmin2, and TE90, respectively). A linear regression model was determined to match the C-maps of TE90 to the reference C-maps of TEmin1 within the interval ranging from 95th to 99th percentile of signal intensity within the prostate. RSIrs comparisons were made at the 98th percentile within each patient's prostate. We compared RSIrs from calibrated TE90 (RSIrsTE90corr) and uncorrected TE90 (RSIrsTE90) to RSIrs from reference TEmin1 (RSIrsTEmin1) and repeated TEmin2 (RSIrsTEmin2). Calibration performance was evaluated with sensitivity, specificity and area under the ROC curve (AUC). Scaling factors for C1, C2, C3, and C4 were estimated as 1.68, 1.33, 1.02, and 1.13, respectively. In non-csPCa cases, the 98th percentile of RSIrsTEmin2 and RSIrsTEmin1 differed by 0.27±0.86SI (mean±standard deviation), whereas RSIrsTE90 differed from RSIrsTEmin1 by 1.82±1.20SI. After calibration, this bias was reduced to -0.51±1.21SI, representing a 72% reduction in absolute error. For patients with csPCa, the difference was 0.54±1.98SI between RSIrsTEmin2 and RSIrsTEmin1 and 2.28±2.06SI between RSIrsTE90 and RSIrsTEmin1. After calibration, the mean difference decreased to -1.03SI, a 55% reduction in absolute error. At the Youden index for patient-level classification of csPCa (8.94SI), RSIrsTEmin1 has a sensitivity of 66% and a specificity of 72%. The proposed linear calibration method produces similar quantitative biomarker values for acquisitions with different TE, reducing TE-induced error by 72% and 55% for non-csPCa and csPCa, respectively.

Analysis and prediction of ablation zone absorption in papillary thyroid microcarcinoma undergoing microwave ablation.

This study aims to investigate the factors that influence the absorption of the ablation zone in patients with papillary thyroid microcarcinoma (PTMC) following microwave ablation (MWA) and construct a nomogram for predicting the absorption of the ablation zone. Data from 150 patients with 187 PTMCs who received MWA between April 2020 and April 2023 were analyzed. PTMCs were randomly divided into training and validation sets in a 7:3 ratio. Univariable and multivariable analyses of Cox regression were utilized to identify the independent variables associated with the absorption of the ablation zone in PTMC post-MWA, and a nomogram was established. The discrimination and calibration performance of the nomogram was assessed using the time-dependent receiver operating characteristic curves and calibration curves. At 12 months post-MWA, a 53% proportion of complete disappearance of the ablation zone was observed. Energy delivered per milliliter of volume measured in contrast-enhanced ultrasound (CEUS) mode immediately post-MWA (Edv) and the CEUS margin at 1-month post-WMA were identified as independently correlated with the ablation zone absorption post-MWA (P = 0.001, P < 0.001 respectively). A nomogram incorporating these two factors was constructed. The areas under the receiver operating characteristic curve were all above 0.78 in the training and validation sets. Edv and the CEUS margin at 1-month post-MWA were found to be significantly associated with complete absorption of the ablation zone in PTMC patients following MWA. The established nomogram can assist practitioners in formulating more appropriate ablation strategies and provide a clinical basis for explaining the recovery status to patients.

Fast and simultaneous prediction of inner quality parameters on intact mangos by near infrared spectroscopy: Impact of spectra pre-processing on prediction accuracy

Near infrared spectroscopy or known as NIRS has been widely employed in many fields including agriculture, especially for sorting and grading of agricultural products. Spectra pre-processing is one of the main factors affecting model accuracy and prediction capabilities of NIRS. The objective of the present study was to study the impact of different spectra corrections namely mean centering (MC), mean normalization (MN), de-trending (DT), multiplicative scatter correction (MSC), standard normal variate (SNV) and orthogonal signal correction (OSC), to the prediction accuracy of quality parameters: titratable acidity (TA) and soluble solids content (SSC) in intact mango. A total of 91 mango samples (cv. Kent) were used as dataset for calibration and external prediction which was separated by means of systematic sampling based on a property (SSBP) approach. Diffuse reflectance spectra (log1/R) were acquired and recorded in wavelength range of 1000 – 2500 nm by Antaris Fourier transform NIR instrument. Judging from calibration and prediction performance, MSC found to be the best spectra pre-processing method prior to prediction model development with R2 prediction are 0.72 for TA and 0.76 for SSC. Although MSC increase the prediction performances based on R2, RMSE, RPD and RER metrics compared to the baseline, the achieved RPD, 1.9 for TA and 1.8 for SSC of this findings are still poor and need improvements to achieve even higher levels of accuracy and reliability necessitates for real-time applications.

Development and Validation of a Clinical Prediction Model for Paclitaxel Hypersensitivity Reaction on the Basis of Real-World Data: Pac-HSR Score.

Paclitaxel is effective chemotherapy against various cancers but can cause hypersensitivity reaction (HSR). This study aimed to identify predictors associated with paclitaxel HSR and develop a clinical prediction model to guide clinical decisions. Data were collected from the medical records database of Rajavithi Hospital. Patients with cancer treated with paclitaxel from 2015 to 2022 were included, and a multivariable logistic regression analysis identified predictors associated with paclitaxel HSR. The scoring system was transformed and calibrated on the basis of diagnostic parameters. Discrimination and calibration performances were assessed. Internal validation was conducted using bootstrap resampling with 1,000 replications. This study involved 3,708 patients with cancer, with an incidence of paclitaxel HSR of 10.11%. An 11-predictor-based Pac-HSR scoring system was developed, involving the following factors: younger age; poor Eastern Cooperative Oncology Group performance status; previous history of paclitaxel HSR; medication allergy history; chronic obstructive airway disease; lung and cervical cancers; high actual dose of paclitaxel; no diphenhydramine premedication; low hemoglobin level; high WBC count; and high absolute lymphocyte count. The C-statistics was 0.73 (95% CI, 0.70 to 0.76), indicating acceptable discrimination. The P value of the Hosmer-Lemeshow goodness-of-fit test was 0.751. The ratio of observed and expected values was 1.00, indicating good calibration. At a cutoff point of 8, specificity was 75.28% and sensitivity was 57.07%. Internal validation indicated good performance with minimal bias, and decision curve analysis demonstrated improved prediction with the use of this scoring system in clinical decision making. This study developed the 11-predictor-based Pac-HSR scoring system for predicting paclitaxel HSR in patients with cancer. High-risk patients identified by this score should be prioritized for close monitoring and early treatment prophylaxis.

Association between programmed cell death ligand-1 expression in patients with cervical cancer and apparent diffusion coefficient values: a promising tool for patient´s immunotherapy selection.

To investigate the associations between apparent diffusion coefficient (ADC) values extracted from three different region of interest (ROI) position approaches and programmed cell death ligand-1 (PD-L1) expression, and evaluate the performance of the nomogram established based on ADC values and clinicopathological parameters in predicting PD-L1 expression in cervical cancer (CC) patients. Through retrospective recruitment, a training cohort of 683 CC patients was created, and a validation cohort of 332 CC patients was prospectively recruited. ROIs were delineated using three different methods to measure the mean ADC (ADCmean), single-section ADC (ADCss), and the minimum ADC of tumors (ADCmin). Logistic regression was employed to identify independent factors related to PD-L1 expression. A nomogram was drawn based on ADC values combined with clinicopathological features, its discrimination and calibration performances were estimated using the area under the curve (AUC) of receiver operating characteristic and calibration curve. The clinical benefits were evaluated by decision curve analysis. The ADCmin independently correlated with PD-L1 expression. The nomogram constructed with ADCmin and other independent clinicopathological-related factors: FIGO staging, pathological grade, parametrial invasion, and lymph node status demonstrated excellent diagnostic performance (AUC = 0.912 and 0.903, respectively), good calibration capacities, and greater net benefits compared to the clinicopathological model in both the training and validation cohorts. ADCmin independently correlated PD-L1 expression, and the nomogram established with ADCmin and clinicopathological independent prognostic factors had a strong predictive performance for PD-L1 expression, thereby serving as a promising tool for selecting cases eligible for immunotherapy. The minimum ADC can serve as a reliable imaging biomarker related to PD-L1 expression; the established nomogram combines the minimum ADC and clinicopathological factors that can assist clinical immunotherapy decisions.

Development and internal validation of prehospital prediction models for identifying intracerebral haemorrhage in suspected stroke patients

IntroductionDistinguishing patients with intracerebral haemorrhage (ICH) from other suspected stroke cases in the prehospital setting is crucial for determining the appropriate level of care and minimising the onset-to-treatment time, thereby...

Adverse Drug Reactions in an Ageing PopulaTion risk Prediction (ADAPTiP) tool: development of a prediction model related to hospital admissions

Abstract Background Older patients are at an increased risk of developing adverse drug reactions (ADRs). The aim of this study was to develop a risk prediction model (ADAPTiP) for ADR-related hospital admissions in older populations, based on predictors available at the time of hospital admission. Methods Using the Adverse Drg reactions in an Ageing PopulaTion (ADAPT) cohort (N=798; 361 ADR-related admissions; 437 non-ADR-related admissions), a cross-sectional and prospective cohort study designed to examine the prevalence of and risk factors for ADR-related hospital admissions in patients aged ≥ 65 years, twenty predictors (categorised as sociodemographic-related, functional ability-related, disease-related and medication-related) were considered in the development of the model. A multivariable logistic regression model was developed using statistically significant univariate associations and/or clinically relevant predictors to estimate adjusted odds ratios and 95% confidence intervals (CI). Calibration and discriminative performance of the model was assessed by the Hosmer-Lemeshow test and by calculating the area under the receiver operator characteristic (AUROC) curve. Results The multivariable model (ADAPTiP) included ten predictors; age, frailty, chronic lung disease, antithrombotic agents, diuretics, renin angiotensin aldosterone system (RAAS) drug classes, primary presenting complaints of respiratory, bleeding, gastrointestinal disorders and syncope on hospital admission. Antithrombotic agents, diuretics and RAAS drug classes and the primary presenting complaints of bleeding disorders, gastrointestinal disorders and syncope and frailty were significantly associated with an ADR-related hospital admission. Whereas, increasing age and presence of chronic lung disease were significantly associated with having a non-ADR-related hospital admission. The AUROC was 0.78 [95%CI:0.75;0.81] with sensitivity and specificity values of 59% and 83%, respectively. Conclusion ADAPTiP has the potential for use as a risk prediction model for ADR-related hospital admissions. Future research will validate this model in other settings.

Development and validation of a deep learning-based survival prediction model for pediatric glioma patients: A retrospective study using the SEER database and Chinese data

ObjectiveDevelop a time-dependent deep learning model to accurately predict the prognosis of pediatric glioma patients, which can assist clinicians in making precise treatment decisions and reducing patient risk. Study designThe study involved pediatric glioma patients from the Surveillance, Epidemiology, and End Results (SEER) Registry (2000–2018) and Tangdu Hospital in China (2010–2018) within specific time frames. For training, we selected two neural network-based algorithms (DeepSurv, neural multi-task logistic regression [N-MTLR]) and one ensemble learning-based algorithm (random survival forest [RSF]). Additionally, a multivariable Cox proportional hazard (CoxPH) model was developed for comparison purposes. The SEER dataset was randomly divided into 80 % for training and 20 % for testing, while the Tangdu Hospital dataset served as an external validation cohort. Super-parameters were fine-tuned through 1000 repeated random searches and 5-fold cross-validation on the training cohort. Model performance was assessed using the concordance index (C-index), Brier score, and Integrated Brier Score (IBS). Furthermore, the accuracy of predicting survival at 1, 3, and 5 years was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and the area under the ROC curves (AUC). The generalization ability of the model was assessed using the C-index of the Tangdu Hospital data, ROC curves for 1, 3, and 5 years, and AUC values. Lastly, decision curve analysis (DCA) curves for 1, 3, and 5-year time frames are provided to assess the net benefits across different models. ResultsA total of 9532 patients with pediatric glioma were included in this study, comprising 9274 patients from the SEER database and 258 patients from Tangdu Hospital in China. The average age at diagnosis was 9.4 ± 6.2 years, and the average survival time was 96 ± 66 months. Through comprehensive performance comparison, the DeepSurv model demonstrated the highest effectiveness, with a C-index of 0.881 on the training cohort. Furthermore, it exhibited excellent accuracy in predicting the 1-year, 3-year, and 5-year survival rates (AUC: 0.903–0.939). Notably, the DeepSurv model also achieved remarkable performance and accuracy on the Chinese dataset (C-index: 0.782, AUC: 0.761–0.852). Comprehensive analysis of DeepSurv, N-MTLR, and RSF revealed that tumor stage, radiotherapy, histological type, tumor size, chemotherapy, age, and surgical method are all significant factors influencing the prognosis of pediatric glioma. Finally, an online version of the pediatric glioma survival predictor based on the DeepSurv model has been established and can be accessed through https://pediatricglioma-tangdu.streamlit.app. ConclusionsThe DeepSurv model exhibits exceptional efficacy in predicting the survival of pediatric glioma patients, demonstrating strong performance in discrimination, calibration, stability, and generalization. By utilizing the online version of the pediatric glioma survival predictor, which is based on the DeepSurv model, clinicians can accurately predict patient survival and offer personalized treatment options.

Development of a unified framework of low-rank approximation and deep neural networks for predicting the spatial variability of SSC in `Spania' watermelons using vis/NIR hyperspectral imaging

Soluble Solids Content (SSC) is an important quality attribute that represents the internal quality of fruits. Visible and near-infrared (Vis/NIR) combined with chemometric algorithms are now popular methods for non-invasive measurement and visualization of SSC. However, in fruits with a thick rind and a large flesh volume, such as watermelon, SSC is not evenly distributed across the fruit. The variability in SSC across the fruit flesh may have an adverse effect on the accuracy of quality analysis algorithms. Thus, this paper presents an accurate and efficient approach for predicting the spatial variation of SSC in watermelons using a combined framework of low-rank approximation and deep neural networks. Watermelon ‘Spania’ was selected for this study, and hyperspectral images of each watermelon were taken from various views, including the top, bottom, and two lateral views. The low-rank property is employed as a constraint in this approach to eliminate the unwanted variations in the spectral data. The low-rank component of the spectral data, free of unwanted variations, is then fed into a fully connected neural network (FNN) for the prediction of watermelon SSC values. The proposed approach obtained optimal performance in calibration and prediction with RC2=0.982, RMSEP = 0.132, and RP2=0.945, RMSEP = 0.195 respectively. Further, the prediction outcomes of the proposed method were compared with state-of-the-art such as Partial Least Square Regression (PLSR), Support Vector Regression (SVR), Multiple Linear Regression (MLR), Decision Tree (DT), and Random Forest (RF). The overall results showed that combining HSI data with a low-rank and deep neural network framework is an efficient method for accurately predicting SSC in watermelons as well as correctly predicting spatial variability of SSC in individual fruits.

Predicting functional outcome in ischemic stroke patients using genetic, environmental, and clinical factors: a machine learning analysis of population-based prospective cohort study.

Ischemic stroke (IS) is a leading cause of adult disability that can severely compromise the quality of life for patients. Accurately predicting the IS functional outcome is crucial for precise risk stratification and effective therapeutic interventions. We developed a predictive model integrating genetic, environmental, and clinical factors using data from 7819 IS patients in the Third China National Stroke Registry. Employing an 80:20 split, we randomly divided the dataset into development and internal validation cohorts. The discrimination and calibration performance of models were evaluated using the area under the receiver operating characteristic curves (AUC) for discrimination and Brier score with calibration curve in the internal validation cohort. We conducted genome-wide association studies (GWAS) in the development cohort, identifying rs11109607 (ANKS1B) as the most significant variant associated with IS functional outcome. We employed principal component analysis to reduce dimensionality on the top 100 significant variants identified by the GWAS, incorporating them as genetic factors in the predictive model. We employed a machine learning algorithm capable of identifying nonlinear relationships to establish predictive models for IS patient functional outcome. The optimal model was the XGBoost model, which outperformed the logistic regression model (AUC 0.818 versus 0.756, P < .05) and significantly improved reclassification efficiency. Our study innovatively incorporated genetic, environmental, and clinical factors for predicting the IS functional outcome in East Asian populations, thereby offering novel insights into IS functional outcome.