Risk Classification Model Research Articles

Psychosocial and neurocognitive correlates of suicidal thoughts and behaviours amongst siblings of persons with and without neurodevelopmental conditions

ObjectiveSiblings of individuals with neurodevelopmental conditions (NDCs) have greater incidence of neuropsychiatric diagnoses and neurocognitive difficulties compared to siblings of persons without NDCs. Despite suicidality being labelled a global health crisis (WHO, 2014) and NDC siblings experiencing risk factors implicated in suicidality, no previous studies examined suicidality amongst adolescent and young adult siblings of persons with NDCs. Our study aimed to bridge this gap. MethodThe present study used Bayesian analyses and risk classification models to examine individual and environmental risk factors associated with suicidal thoughts and behaviours amongst siblings of persons with and without NDCs (n = 267; 132 NDC, 135 control group, mean age 20.61, range 14–27, 76.40% female, 76.78% White Caucasian), as measured using self-report survey data and remote self-administered cognitive tests. ResultsNDC siblings had higher rates of current nonsuicidal self-injury (NSSI; 18.94% versus 14.07%, δ = −0.32), suicidal ideation (25.76% versus 8.89%, δ = −0.40) and history of suicide attempts (18.18% versus 4.44%, δ = −0.43) compared to controls. Classification models using boosting and random forest demonstrated adequate performance: positive predictive value 0.86–0.91, negative predictive value 0.81–0.90, false negative rates 0.11–0.24. Cognitive inflexibility, alexithymia, inattention, bullying, depression, NSSI, and eating or psychotic disorder history had the highest relative importance in predicting lifetime suicidality. Poorer executive functioning (measured by the Wisconsin Card Sorting Task, Sustained Attention to Response Task, Stop Signal Task, and N-Back 2-back task) was strongly correlated with suicidality. ConclusionsScreening for proximal and modifiable risk factors is critical to inform suicidal behaviour intervention and prevention programs for at-risk siblings.

Analysis of Baseline Serum Lipid Profile for Predicting Clinical Outcomes of Patients with Extensive-Stage Small Cell Lung Cancer.

Serum lipids were reported to be the prognostic factors of various cancers, but their prognostic value in small cell lung cancer (SCLC) patients remains unclear. This study investigated the relationship between lipid profiles and clinical outcomes in extensive-stage (ES) SCLC by establishing a predictive risk classification model. We retrospectively analyzed the prognostic values of pretreatment serum lipids and their derivatives in patients with a confirmed diagnosis ES-SCLC. Independent factors of progression-free survival (PFS) were determined by univariate and multivariate cox analysis. Then, prognostic nomograms were established, of which predictive performance was evaluated by concordance index (C-index), calibration curves, receiver operating characteristic (ROC) curves, and decision curve analyses (DCA). A total of 158 patients was included in this study. Four optimal PFS-related factors, total cholesterol (TC) ≥ 5.30, high-density lipoprotein cholesterol (HDL-C) > 1.30, triglycerides (TG)/HDL-C > 2.18, and ki67 expression > 70%, were included to construct the predictive nomogram. The C-indexes in training and validation sets were 0.758 and 0.792, respectively. ROC curves, calibration plots, and DCA all suggested favorable discrimination and predictive ability. Besides, the nomogram also performed better predictive ability than ki67 expression. Nomogram-related risk score divided the patients into two groups with significant progression disparities. The promising prognostic nomogram based on lipid parameters could help clinicians to conveniently and accurately evaluate the prognosis of ES-SCLC patients and identify high-risk groups, so as to formulate individualized therapeutic regimens and follow-up strategies in time.

Development and Validation of Multi-Omics Thymoma Risk Classification Model Based on Transfer Learning.

The paper aims to develop prediction model that integrates clinical, radiomics, and deep features using transfer learning to stratifying between high and low risk of thymoma. Our study enrolled 150 patients with thymoma (76 low-risk and 74 high-risk) who underwent surgical resection and pathologically confirmed in Shengjing Hospital of China Medical University from January 2018 to December 2020. The training cohort consisted of 120 patients (80%) and the test cohort consisted of 30 patients (20%). The 2590 radiomics and 192 deep features from non-enhanced, arterial, and venous phase CT images were extracted and ANOVA, Pearson correlation coefficient, PCA, and LASSO were used to select the most significant features. A fusion model that integrated clinical, radiomics, and deep features was developed with SVM classifiers to predict the risk level of thymoma, and accuracy, sensitivity, specificity, ROC curves, and AUC were applied to evaluate the classification model. In both the training and test cohorts, the fusion model demonstrated better performance in stratifying high and low risk of thymoma. It had AUCs of 0.99 and 0.95, and an accuracy of 0.93 and 0.83, respectively. This was compared to the clinical model (AUCs of 0.70 and 0.51, accuracy of 0.68 and 0.47), the radiomics model (AUCs of 0.97 and 0.82, accuracy of 0.93 and 0.80), and the deep model (AUCs of 0.94 and 0.85, accuracy of 0.88 and 0.80). The fusion model integrating clinical, radiomics and deep features based on transfer learning was efficient for noninvasively stratifying high risk and low risk of thymoma. The models could help to determine surgery strategy for thymoma cancer.

A risk identification model for ICT supply chain based on network embedding and text encoding

Public bidding for the information and communication technology (ICT) products offers an external regulatory perspective to identify endogenous risks of ICT supply chain. Owing to the multi-level outsourcing, the modeling methods of the first-order topological similarity are prone to misjudgment. Moreover, only using basic structural attributes (such as enterprise type, registered capital, etc.) is difficult to accurately quantify the enterprise' supply capacity, resulting in the models cannot identify the risky nodes with insufficient supply capacity. In this work, a risk identification Model for ICT supply chain based on Network Embedding and Text Encoding (NETEM) is proposed. In addition to the basic business features of enterprises, the DeepWalk is utilized to obtain the high-order topological representation of the ICT supply chain. Meanwhile, the TextCNN combined with keyword sequence extraction is utilized to supplement text semantic information, and excavate the association between business scope and supply capacity of enterprises. Finally, the basic business features, high-level topological features and text semantic features are spliced to obtain the final enterprise node representation vector, then the node risk classification model is trained. In addition, an ontology is developed for ICT supply chain integration and risk knowledge reuse. The experimental results show that NETEM improves the F1 value of the optimal baseline algorithm by 7.3% on the ICT risk identification data set, and the case study in financial industry demonstrates that the model can be practically applied to identify the risk of supply interruption.

Occupational health risk assessment of workplace solvents and noise in the electronics industry using three comprehensive risk assessment models.

Occupational hazards such as solvents and noise in the electronics industry are serious. Although various occupational health risk assessment models have been applied in the electronics industry, they have only been used to assess the risks of individual job positions. Few existing studies have focused on the total risk level of critical risk factors in enterprises. Ten electronics enterprises were selected for this study. Information, air samples and physical factor measurements were collected from the selected enterprises through on-site investigation, and then the data were collated and samples were tested according to the requirements of Chinese standards. The Occupational Health Risk Classification and Assessment Model (referred to as the Classification Model), the Occupational Health Risk Grading and Assessment Model (referred to as the Grading Model), and the Occupational Disease Hazard Evaluation Model were used to assess the risks of the enterprises. The correlations and differences between the three models were analyzed, and the results of the models were validated by the average risk level of all of the hazard factors. Hazards with concentrations exceeding the Chinese occupational exposure limits (OELs) were methylene chloride, 1,2-dichloroethane, and noise. The exposure time of workers ranged from 1 to 11 h per day and the frequency of exposure ranged from 5 to 6 times per week. The risk ratios (RRs) of the Classification Model, the Grading Model and the Occupational Disease Hazard Evaluation Model were 0.70 ± 0.10, 0.34 ± 0.13, and 0.65 ± 0.21, respectively. The RRs for the three risk assessment models were statistically different (P < 0.001), and there were no correlations between them (P > 0.05). The average risk level of all of the hazard factors was 0.38 ± 0.18, which did not differ from the RRs of the Grading Model (P > 0.05). The hazards of organic solvents and noise in the electronics industry are not negligible. The Grading Model offers a good reflection of the actual risk level of the electronics industry and has strong practicability.

The Risk-Level Change of Dam Break Due to the Population Growth in the Dam Downstream

This research intends to evaluate the risk-level change of dam break due to the number of population growth rates in the dam downstream. The number of population growth rates in the dam downstream area causes the number of populations that is impacted by disaster risk. In this research, there is conducted the danger class classification of dam failure based on the increasing number of population growth rates. In this research, there is conducted a class classification about the dam failure based on the increasing of population growth rate. The methodology is started from data collecting such as PMF design flood that is used for analyzing the dam break to be obtained the flood boundary of the dam downstream area, Map of the Digital Elevation Model that is for knowing the topography of the dam downstream area, population data which is obtained from BPS that is for knowing the population growth rate so it can predict the population number until the life time of the reservoir, and the land use data that is for knowing the population growth rate in the downstream dam area so it can be known the development of residence until the life time of reservoir. Then there is carried out the evaluation and assessment, which is based on the number of risk populations. The result show that there are 6 dams with very high danger level of the downstream damage level and the classification assessment of dams is extreme such as the dams of Bintang Bano, Rotiklot, Napungette, Lolak, Kuwil, and Pandanduri. However, the other four dams also show a very high danger level but in the assessment classification of dam is high (not extreme) such as the dams of Batu Nampar, Kengkang, Sepit, and Jangkih Jawa. The result is hoped can predict the danger classification of dam failure during the life time of the reservoir. Additionally, it can give input to the policymaker in determining the risk classification of dam failure. Moreover, it can be as one of the inputs in developing the risk classification model of dam failure.

Evolving Risk Classifications in AML in a Real-Life Scenario: After Changes upon Changes, Is It More and More Adverse?

Acute myeloid leukemia (AML) is a heterogeneous disease classified into three risk categories (favorable, intermediate and adverse) with significant differences in outcomes. Definitions of risk categories evolve overtime, incorporating advances in molecular knowledge of AML. In this study, we analyzed the impacts of evolving risk classifications in 130 consecutive AML patients in a single-center real-life experience. Complete cytogenetic and molecular data were collected using conventional qPCR and targeted Next Generation Sequencing (NGS). Five-year OS probabilities were consistent among all classification models (roughly 50-72%, 26-32% and 16-20% for favorable, intermediate and adverse risk groups, respectively). In the same way, the medians of survival months and prediction power were similar in all models. In each update, around 20% of patients were re-classified. The adverse category consistently increased over time (31% in MRC, 34% in ELN2010, 50% in ELN2017), reaching up to 56% in the recent ELN2022. Noteworthily, in multivariate models, only age and the presence of TP53 mutations remained statistically significant. With updates in risk-classification models, the percentage of patients assigned to the adverse group is increasing, and so will the indications for allogeneic stem cell transplantation.

Variability Among Breast Cancer Risk Classification Models When Applied at the Level of the Individual Woman

BackgroundBreast cancer risk models guide screening and chemoprevention decisions, but the extent and effect of variability among models, particularly at the individual level, is uncertain.ObjectiveTo quantify the accuracy and disagreement between commonly used risk models in categorizing individual women as average vs. high risk for developing invasive breast cancer.DesignComparison of three risk prediction models: Breast Cancer Risk Assessment Tool (BCRAT), Breast Cancer Surveillance Consortium (BCSC) model, and International Breast Intervention Study (IBIS) model.SubjectsWomen 40 to 74 years of age presenting for screening mammography at a multisite health system between 2011 and 2015, with 5-year follow-up for cancer outcome.Main MeasuresComparison of model discrimination and calibration at the population level and inter-model agreement for 5-year breast cancer risk at the individual level using two cutoffs (≥ 1.67% and ≥ 3.0%).Key ResultsA total of 31,115 women were included. When using the ≥ 1.67% threshold, more than 21% of women were classified as high risk for developing breast cancer in the next 5 years by one model, but average risk by another model. When using the ≥ 3.0% threshold, more than 5% of women had disagreements in risk severity between models. Almost half of the women (46.6%) were classified as high risk by at least one of the three models (e.g., if all three models were applied) for the threshold of ≥ 1.67%, and 11.1% were classified as high risk for ≥ 3.0%. All three models had similar accuracy at the population level.ConclusionsBreast cancer risk estimates for individual women vary substantially, depending on which risk assessment model is used. The choice of cutoff used to define high risk can lead to adverse effects for screening, preventive care, and quality of life for misidentified individuals. Clinicians need to be aware of the high false-positive and false-negative rates and variation between models when talking with patients.

Advances in Imaging-Based Biomarkers in Renal Cell Carcinoma: A Critical Analysis of the Current Literature.

Cross-sectional imaging is the standard diagnostic tool to determine underlying biology in renal masses, which is crucial for subsequent treatment. Currently, standard CT imaging is limited in its ability to differentiate benign from malignant disease. Therefore, various modalities have been investigated to identify imaging-based parameters to improve the noninvasive diagnosis of renal masses and renal cell carcinoma (RCC) subtypes. MRI was reported to predict grading of RCC and to identify RCC subtypes, and has been shown in a small cohort to predict the response to targeted therapy. Dynamic imaging is promising for the staging and diagnosis of RCC. PET/CT radiotracers, such as 18F-fluorodeoxyglucose (FDG), 124I-cG250, radiolabeled prostate-specific membrane antigen (PSMA), and 11C-acetate, have been reported to improve the identification of histology, grading, detection of metastasis, and assessment of response to systemic therapy, and to predict oncological outcomes. Moreover, 99Tc-sestamibi and SPECT scans have shown promising results in distinguishing low-grade RCC from benign lesions. Radiomics has been used to further characterize renal masses based on semantic and textural analyses. In preliminary studies, integrated machine learning algorithms using radiomics proved to be more accurate in distinguishing benign from malignant renal masses compared to radiologists' interpretations. Radiomics and radiogenomics are used to complement risk classification models to predict oncological outcomes. Imaging-based biomarkers hold strong potential in RCC, but require standardization and external validation before integration into clinical routines.

Incidence and risk factors of dural ossification in patients with thoracic ossification of the ligamentum flavum.

Dural ossification (DO) is a common clinical feature in patients with thoracic ossification of the ligamentum flavum (OLF) and associated with the increased risk of perioperative complications. However, few studies have been conducted to determine the incidence and independent risk factors of DO in patients with thoracic OLF. The aim of this retrospective study was to determine the incidence and independent risk factors of DO in patients with thoracic OLF. A total of 107 patients with thoracic OLF who were admitted to the authors' hospital from December 2020 to December 2021 were included in this study. The independent risk factors of DO in patients with thoracic OLF were determined through univariate analysis followed by multivariate logistic regression analysis with p < 0.05. The diagnostic efficacy of the DO in OLF (DO-OLF) risk classification model was determined on the basis of independent risk factors and evaluated on the basis of sensitivity, specificity, and agreement rate. The incidence of DO in patients with thoracic OLF was 35% (37/107 patients). The tuberous type according to the Sato classification (OR 9.75, p < 0.01) and larger (≥ 9°) supine local kyphosis angle (LKA) (OR 8.13, p < 0.01) were two independent risk factors of DO in thoracic OLF. The DO-OLF risk classification, a novel approach for the diagnosis of DO in patients with thoracic OLF, was established on the basis of the combination of the tuberous type according to the Sato classification and larger supine LKA. The sensitivity, specificity, and agreement rate of this approach for distinguishing between patients with thoracic OLF at high and low risk of DO were 87%, 93%, and 91%, respectively. The incidence of DO in patients with thoracic OLF was 35%. The tuberous type according to the Sato classification and larger supine LKA (≥ 9°) were independent risk factors of DO in patients with thoracic OLF. The novel DO-OLF risk classification approach could serve as an efficient method for predicting DO in patients with thoracic OLF.

Effect of serum lipid subfractions on nuclear magnetic resonance spectroscopy's ability to detect coronary atherosclerosis early

Nuclear magnetic resonance spectroscopy, also known as NMRS, is a cutting-edge method based on nuclear magnetic induction theory and quantum optics that has found extensive usage in the domains of chemical analysis, life science, and other disciplines. The method is quantitative and repeatable, making it particularly effective for detecting and analyzing tiny compounds in complicated solutions like blood and urine. With a high degree of standardization and automation, it can assess 114 hemo lipoprotein subfractions and particles in human serum/plasma in a targeted way (exact quantification) and produce precise and dependable detection findings. All processes may be tracked back to fulfill our requirements for non-targeted analysis, which is suited for biological fluid specimens and water-soluble specimens from tissue or cell extracts. The 114 lipoprotein subfractions and particles can be targeted using nuclear magnetic resonance spectroscopy (NMRS), which can detect the serum lipid profile in its entirety. The analysis of the serum lipid profile, the correlation between the levels of lipid subfractions and lipid particles and the degree of coronary stenosis, and the development of a coronary stenosis early warning system can therefore improve the current risk model on the inadequacy of the risk stratification of the patients. These studies are based on a cohort of individuals with varying degrees of coronary atherosclerosis. The development of a coronary artery stenosis early warning system might enhance the patient risk classification models already in use. To analyze the correlation of lipid subfractions and particle levels with cardiovascular events and therapeutic efficacy, to explore more effective biomarkers for lipid-lowering monitoring, and to evaluate the efficacy of classical Chinese medicinal preparations.

CHA2DS2-VASc Score as a Predictor of New Onset Atrial Fibrillation in Patients With Non-ST Segment Elevation Myocardial Infarction Who Underwent Percutaneous Coronary Intervention

Background: New-onset atrial fibrillation (NOAF) often complicates acute coronary syndromes (ACSs) with adverse short- and long-term consequences. However, the current risk classification model for estimating NOAF during non-ST-segment elevation myocardial infarction (NSTEMI) remains unclear. Some parameters in the CHA2DS2-VASc score are closely related to the development of atrial fibrilation (AF). Methods: This retrospective study was conducted among 670 consecutive NSTEMI patients who applied to our cardiovascular center between June 2020 and June 2022 and underwent percutaneous coronary intervention (PCI). Results: NOAF developed during hospitalization in 55 patients (12.5%). NOAF patients were older and had higher high sensitivity C-reactive protein (hs-CRP), left atrial volume index, Post PCI thrombolysis in myocardial infarction (TIMI) grade &lt;3, CHA2DS2-VASc score, peak troponin I (ng/ml), and SYNTAX score (SS). After univariate logistic regression analysis for the predictors of NOAF development, CHA2DS2-VASc score, post PCI TIMI grade &lt;3, hemoglobin, hsCRP, SS alone were predictors of NOAF, but after multivariate analysis, CHA2DS2-VASc score, post PCI TIMI grade &lt;3, hemoglobin had a determining effect. Conclusions: The CHA2DS2-VASc score can be evaluated as a predictor of NOAF after PCI in NSTEMI. Except for the components of the CHA2DS2-VASc score, post PCI TIMI grade &lt;3 and low hemoglobin levels are independent risk factors for NSTEMI-NOAF.

The use of multilayer perceptron and radial basis function: an artificial intelligence model to predict progression of oral cancer.

The use of multilayer perceptron and radial basis function: an artificial intelligence model to predict progression of oral cancer.

Comparison of CURB-65 Pneumonia Severity Score, Quick COVID-19 Severity Index, and Brescia-COVID Respiratory Severity Scale in Emergently Hospitalized COVID-19 Patients with Pneumonia.

This study aimed to assess the performance of the CURB-65, the quick COVID-19 severity index (qCSI), and the Brescia-COVID respiratory severity scale (BCRSS) scores in predicting ICU (intensive care unit) hospitalization and in-hospital mortality in emergently hospitalized patients with COVID-19 pneumonia. We retrospectively reviewed the emergently hospitalized 258 patients with COVID-19 pneumonia consecutively. The required sample size was calculated to compare the areas under the two ROC (receiver operating characteristic) curves (AUC) using the MedCalc 20.0 program (MedCalc Software Ltd., Ostend, Belgium). In addition, we actualized ROC analyses of the CURB-65, the qCSI, and the BCRSS scores and compared the ROC curves of these three scores. The median age of the patients was 73, and 63.6% (n=164) were male. Of 258 patients, 29.5% (n=76) were hospitalized in the intensive care unit (ICU), and 15.9% (n=41) died. The CURB-65 and the qCSI scores predicted ICU admission at a moderate level (p≤0.001; AUC values were 0.743 and 0.723, respectively). However, the predictive effect of the BCRSS score for ICU admission was lower (p≤0.001; AUC value was 0.667). The CURB-65 predicted in-hospital mortality at a moderate level ( p≤0.001; AUC value was 0.762). However, the predictive effect of the qCSI and the BCRSS scores for in-hospital mortality were lower ( p≤0.001 and p=0.012, respectively; AUC values were 0.655 and 0.612, respectively). The CURB-65 score predicted ICU hospitalization and in-hospital mortality better than the qCSI and the BCRSS scores. Also, the qCSI score predicted ICU admission better than the BCRSS score.The predictive effect of the BCRSS score was the lowest. We recommend future studies to evaluate the value and utility of COVID-19 risk classification models.

Modelling polycystic liver disease progression using age-adjusted liver volumes and targeted mutational analysis

Background & AimsPolycystic liver disease (PLD) manifests as numerous fluid-filled cysts scattered throughout the liver parenchyma. PLD most commonly develops in females, either as an extra-renal manifestation of autosomal-dominant polycystic kidney disease (ADPKD) or as isolated autosomal-dominant polycystic liver disease (ADPLD). Despite known genetic causes, clinical variability challenges patient counselling and timely risk prediction is hampered by a lack of genotype-phenotype correlations and prognostic imaging classifications.MethodsWe performed targeted next-generation sequencing and multiplex ligation-dependent probe amplification to identify the underlying genetic defect in a cohort of 80 deeply characterized patients with PLD. Identified genotypes were correlated with total liver and kidney volume (assessed by CT or MRI), organ function, co-morbidities, and clinical endpoints.ResultsMonoallelic diagnostic variants were identified in 60 (75%) patients, 38 (48%) of which pertained to ADPKD-gene variants (PKD1, PKD2, GANAB) and 22 (27%) to ADPLD-gene variants (PRKCSH, SEC63). Disease severity defined by age at waitlisting for liver transplantation and first PLD-related hospitalization was significantly more pronounced in mutation carriers compared to patients without genetic diagnoses. While current imaging classifications proved unable to differentiate between severe and moderate courses, grouping by estimated age-adjusted total liver volume progression yielded significant risk discrimination.ConclusionThis study underlines the predictive value of providing a molecular diagnosis for patients with PLD. In addition, we propose a novel risk-classification model based on age- and height-adjusted total liver volume that could improve individual prognostication and personalized clinical management.Lay summaryPolycystic liver disease (PLD) is a highly variable condition that can be asymptomatic or severe. However, it is currently difficult to predict clinical outcomes such as hospitalization, symptom burden, and need for transplantation in individual patients. In the current study, we aimed to investigate the clinical value of genetic confirmation and an age-adjusted total liver volume classification for individual disease prediction. While genetic confirmation generally pointed to more severe disease, estimated age-adjusted increases in liver volume could be useful for predicting clinical outcomes.

A nomogram and risk classification model predicts prognosis in Chinese esophageal squamous cell carcinoma patients.

BackgroundA nomogram model based on gene mutations for predicting the prognosis of patients with resected esophageal squamous cell carcinoma (ESCC) has not been established. We sought to develop a risk classification system.MethodsIn total, 312 patients with complete clinical and genome mutation landscapes in our previous study were chosen for the present study. Public International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA) data of ESCC were also used as an external validation set.ResultsUsing the least absolute shrinkage and selection operator (LASSO) method, we successfully built a 9-gene mutation-based prediction model for overall survival (OS) and a 21-gene mutation model for progression-free survival (PFS). High- and low-risk groups were stratified using the gene mutation-based classifier. Patients in the high-risk group witnessed poorer 3- and 5-year OS and PFS in both the training and validation sets (P<0.01). Moreover, calibration curves and decision curve analyses (DCAs) were used to confirm the independence and potential translational value of this predictive model. In the nomogram analysis, the risk classification model was shown to be a reliable prognostic tool. All results showed better consistency in the external ICGC and TCGA validation sets.ConclusionsWe developed and validated a predictive risk model for ESCC. This practical prognostic model may help doctors make different follow-up decisions in the clinic.

An examination of the fall risk classification model in short-term intensive inpatient rehabilitation for elderly patients with Parkinson's disease

To examine the predictive model that classifies the fluctuation in fall risk two weeks after an evaluation at the beginning of short-term intensive inpatient rehabilitation for elderly patients with Parkinson's disease (PD). Of the patients with PD ≥65 years old who underwent intensive rehabilitation, 17 patients (mean age: 76.5±6.1 years old) who had a fall risk detected by the Timed "Up & Go" including a cognitive task (TUG-cognitive) were enrolled in this study. The patients were divided into two groups depending on the TUG-cognitive two weeks after starting intensive rehabilitation: reduced fall risk or residual fall risk. The differences in the evaluated items between the two groups were examined, and those items found to have significant differences were used as independent variables in the logistic regression analysis with the fluctuation in fall risk after two weeks as the dependent variable. A logistic regression analysis revealed that the fast-walking speed was a factor affecting the fluctuation in fall risk. In addition, the incidence of residual fall risk was calculated from the logistic function, and the risk was determined to be high when the fast-walking speed was ≤0.84 m/sec. The predictive model consisted of the TUG-cognitive and fast-walking speed at the beginning of intensive rehabilitation; it was suggested that the fluctuation in fall risk after two weeks could be classified based on these findings. Improvement of the fast-walking speed is important for reducing the fall risk in elderly patients with PD.

S130: NPM1 MUTATED AML: IMPACT OF CO-MUTATIONAL PATTERNS - RESULTS OF THE EUROPEAN HARMONY ALLIANCE

Background: Acute myeloid leukemia (AML) is a heterogeneous disease in terms of clinical features, outcomes and genetics. While mutations of NPM1 are usually considered as a favorable prognostic marker, the vast majority of the patients carry several co-mutations that might influence the prognosis. Therefore, a better understanding of the NPM1mut AML mutational landscape is warranted. The large cohort of AML patients collected within the European HARMONY Alliance provides an excellent basis for this purpose. Aims: To identify clinically significant co-mutational patterns in NPM1mut AML in order to establish a revised risk stratification model. Methods: From the HARMONY Alliance AML database, a total of 1001 NPM1mut intensively treated patients were selected. Clinically significant co-mutations were evaluated using graphical patterns created with the Gephi tool and confirmed by detailed survival analysis using Kaplan-Meier and Cox regression models. Finally, a novel multi-state risk stratification model for NPM1mut AML was established. Results: The study population of 1001 NPM1mut AML patients included 57% females and median age was 53 years. Regarding ELN2017 classification, 68% of patients were classified into the favorable, 29% intermediate and 3% adverse risk groups. The most frequent co-mutations were DNMT3A (54%), followed by FLT3-ITD (38%). In total, 24% of patients presented with a high allelic mutant-to-wildtype ratio ≥0.5 (FLT3-ITDhigh) while 14% had low allelic ratio <0.5 (FLT3-ITDlow). Other frequent co-mutations were NRAS (21%), TET2 (20%) and PTPN11 (15%). The triple mutation pattern of NPM1mut + FLT3-ITDhigh + DNMT3Amut identified a subgroup with adverse prognosis (2-year OS of 25%), similar to NPM1mut + TP53mut. The combination of FLT3-ITDlow + DNMT3Amut or FLT3-ITDhigh + DNMT3Awt was associated with intermediate prognosis (2-year OS of 45% and 53% respectively). Notably, mutations of NRAS, KRAS, PTPN11 or RAD21 were identified to be associated with better OS. However, in the context of NPM1mut + DNMT3Amut these mutations did not affect the prognosis when a FLT3-ITD was present. This information is summarized in a 3-category risk classification model (Figure 1). The revised NPM1mut favorable group presented with a 2-year OS of 73%, while for intermediate and adverse groups the OS was 54% and 27% respectively (p<0.001). Regarding relapse free survival (RFS), the median was not reached in the favorable group, while it was 23 months for intermediate and 6 months for adverse group (p<0.001). It should be noted that 171 patients in the NPM1mut intermediate group would be considered as favorable according to the ELN2017 criteria, as well as 162 patients in the NPM1mut adverse group were previously classified as intermediate risk. Therefore, our model was able to reclassify 33% of NPM1mut AML patients in comparison to ELN2017 criteria. Multivariate analysis of OS in NPM1mut AML identified the following independent prognostic factors: NPM1mut model (taking favorable group as reference, HR 1.6 for intermediate and HR 2.7 for adverse group, p<0.001); secondary or therapy-related AML (HR 1.8, p<0.001), WBC at diagnosis >100x103/μL (HR 1.5, p<0.001) and age >60 years (HR 1.4, p<0.001). Image:Summary/Conclusion: Analysis of large NPM1mut AML cohorts allows the discovery of co-mutation patterns associated with prognostic outcome. In accordance, we propose a new genetic stratification model for NPM1mut AML that identifies 3 groups with different OS and RFS. This model improves ELN2017 criteria as it is able to correctly reclassify 33% of NPM1mut AML patients.

A risk classification model for prioritising the management of quality issues relating to substandard medicines in Singapore

Monitoring for substandard medicines by regulatory agencies is a key post-market surveillance activity. It is important to prioritise critical product defects for review to ensure that prompt risk mitigation actions are taken. A regulatory risk impact prioritisation model for product defects (RISMED) with 11 factors considering the seriousness and extent of impact of a defect was developed. The model generated an overall score that categorised cases into high, medium or low impact. The model was further developed into a statistical risk scoring model (stat-RISMED) using multivariate logistic regression that classified cases into high and non-high impact. Both models were evaluated against an expert-derived gold standard annotation corpus and tested on an independent dataset. Product defect cases received from January 2011 to June 2020 (n= 660) were used to train stat-RISMED and cases from July 2020 to June 2021 (n= 220) for validation. The stat-RISMED identified four factors associated with high impact cases, namely defect classification based on MedDRA-HSA terms, therapeutic indication of product, detectability of defect and whether any overseas regulatory actions were performed. Compared to RISMED, stat-RISMED achieved an improved sensitivity (94% vs 42%) and positive predictive value (47% vs 43%) for the identification of high impact cases, against the gold standard labels. This study reported characteristics that predicts cases with high impact, and the use of a statistical model to identify such cases. The model may potentially be applied to prioritise product defect issues and strengthen overall surveillance efforts of substandard medicines.

Distinct genomic landscape of Chinese pediatric acute myeloid leukemia impacts clinical risk classification

Studies have revealed key genomic aberrations in pediatric acute myeloid leukemia (AML) based on Western populations. It is unknown to what extent the current genomic findings represent populations with different ethnic backgrounds. Here we present the genomic landscape of driver alterations of Chinese pediatric AML and discover previously undescribed genomic aberrations, including the XPO1-TNRC18 fusion. Comprehensively comparing between the Chinese and Western AML cohorts reveal a substantially distinct genomic alteration profile. For example, Chinese AML patients more commonly exhibit mutations in KIT and CSF3R, and less frequently mutated of genes in the RAS signaling pathway. These differences in mutation frequencies lead to the detection of previously uncharacterized co-occurring mutation pairs. Importantly, the distinct driver profile is clinical relevant. We propose a refined prognosis risk classification model which better reflected the adverse event risk for Chinese AML patients. These results emphasize the importance of genetic background in precision medicine.