A life-threatening bleeding prediction model for immune thrombocytopenia based on personalized machine learning: a nationwide prospective cohort study

YC-4-3, a Novel Glycogen Synthase Kinase 3β Inhibitor, Alleviates the Endoplasmic Reticulum Stress of Macrophages in Primary Immune Thrombocytopenia

Primary immune thrombocytopenia (ITP) is a heterogeneous autoimmune disease, characterized by decreased platelet count and increased risk of hemorrhage, in which macrophages play an important role in the pathogenesis. This study aims to explore the effects of YC-4-3, the patented chemical synthesis of benzothiazepinone compounds (BTZs), a novel GSK-3β inhibitor (GSK-3βi), on macrophages in ITP. The expressions of GSK-3β in monocytes are tested. The effects of GSK-3βi (YC-4-3) on macrophages of ITP patients are examined and validated in passive and active murine models. Signal pathway enrichment analysis is performed. The interaction proteins of endoplasmic reticulum (ER) stress and GSK-3β are explored. The GSK-3β+ cells in monocytes are increased in newly diagnosed ITP patients and decreased in treatment-response patients. YC-4-3 can restrain the proinflammatory differentiation, phagocytosis, and cytokine generation of macrophages and alleviate thrombocytopenia in ITP. YC-4-3 suppresses the PI3K/mTOR/Akt, NFκB/IκBα, and MAPK pathways, as well as the ER stress signal pathway. YC-4-3 directly interacts with the protein chaperone Bip. YC-4-3, a patented GSK-3βi, can modulate the inflammatory status of macrophages and improve the thrombocytopenia in ITP by directly interacting with ER stress response. YC-4-3 may be a novel potential therapeutic agent for ITP.

Prediction of moderate to severe bleeding risk in pediatric immune thrombocytopenia using machine learning

This study aimed to develop and validate a risk prediction model for moderate to severe bleeding in children with immune thrombocytopenia (ITP). Data from 286 ITP patients were prospectively collected and randomly split into training (80%) and test (20%) sets. Least Absolute Shrinkage and Selection Operator (LASSO) regression was used for feature selection. Among seven machine learning algorithms, the eXtreme Gradient Boosting (XGBoost) model demonstrated the best performance (AUC = 0.886, 95% CI: 0.790-0.982) and was selected as the optimal model. Shapley Additive Explanations (SHAP) were used for model interpretation, identifying child age, age at diagnosis, and initial platelet count as key predictors of moderate to severe bleeding risk.

CONCLUSION

The XGBoost-based prediction model shows strong predictive performance and could assist healthcare providers in identifying high-risk ITP patients, supporting appropriate clinical decision-making.

TRIAL REGISTRATION NUMBER

ChiCTR2100054216, December 11, 2021 What is Known: • Current clinical practice relies solely on platelet counts to guide hospitalization and treatment in ITP children, often overlooking bleeding manifestations, leading to delayed or inappropriate treatment. Existing severe bleeding risk prediction models are primarily designed for adults and lack applicability to children.

WHAT IS NEW

 • This study prospectively collected data, enhancing accuracy. A novel machine learning-based prediction model was developed to assess moderate to severe bleeding risk in pediatric ITP patients.

Doctor, what is my risk of bleeding after cardiac surgery while on combined anticoagulant with antiplatelet therapy? A validated nomogram for risk assessment

Background

Patients with comorbid coronary artery disease and valvular heart disease usually undergo coronary artery bypass grafting alongside valve replacement or ring repair surgeries. Following these procedures, they typically receive a combination of anticoagulation and antiplatelet therapy, which notably heightens their bleeding risk. However, Current scoring systems provide limited predictive capability.

Methods

A total of 500 adult patients treated with anticoagulation plus antiplatelet therapy after cardiac surgery were randomly divided into the training set and the validation set at a ratio of 7:3. Predictive factors were identified using univariate logistic regression, LASSO regression and multivariable analysis. Various models were developed, validated and evaluated by using methods including ROC curves, calibration curves, the Hosmer-Lemeshow test, net reclassification improvement (NRI), integrated discrimination improvement (IDI) index, decision curve analysis (DCA) and clinical impact curves (CIC).

Results

Mod2 showed the best performance (AUC of validation set = 0.863) which consists of 8 independent predictive factors (gender, age > 65 years, diabetes, anemia, atrial fibrillation, cardiopulmonary bypass time, intraoperative bleeding and postoperative drainage), with a significantly higher AUC compared to Mod1 (only preoperative factors) and Mod3 (the HAS-BLED scoring model). NRI and IDI analyses further confirmed the superior predictive ability of Mod2 (NRI < 0.05, IDI < 0.05). Both DCA and CIC indicated that Mod2 exhibited good clinical applicability.

Conclusion

This research established and validated a nomogram model incorporating eight predictive factors to evaluate the bleeding risk in patients who receive anticoagulation combined with antiplatelet therapy following cardiac surgery. The model holds significant potential for clinical applications in bleeding risk assessment, decision-making and personalized treatment strategies.

Integrating chemokines and machine learning algorithms for diagnosis and bleeding assessment in primary immune thrombocytopenia: A prospective cohort study

Primary immune thrombocytopenia (ITP) is an autoimmune bleeding disorder, and chemokines have been shown to be dysregulated in autoimmune disorders. We conducted a prospective analysis to identify potential chemokines that could enhance the diagnostic accuracy and bleeding evaluation in ITP patients. In the discovery cohort, a Luminex-based assay was employed to quantify concentrations of plasma multiple chemokines. These levels were subjected to comparative analysis using a cohort of 60 ITP patients and 17 patients with thrombocytopenia other than ITP (non-ITP). Additionally, comparative evaluation was conducted between a subgroup of 12 ITP patients characterised by bleeding episodes (ITP-B, as defined by an ITP-2016 bleeding grade ≥2) and 33 ITP patients without bleeding episodes (ITP-NB, as defined by an ITP-2016 bleeding grade ≤1). Machine learning algorithms further identified CCL20, interleukin-2, CCL26, CCL25, and CXCL1 as promising indicators for accurate diagnosis of ITP and CCL21, CXCL8, CXCL10, CCL8, CCL3, and CCL15 as biomarkers for assessing bleeding risk in ITP patients. The results were confirmed using enzyme-linked immunosorbent assays in a validation cohort (43 ITP patients and 19 non-ITP patients). Overall, the findings suggest that specific chemokines show promise as potential biomarkers for diagnosis and bleeding evaluation in ITP patients.

Could machine learning revolutionize how we treat immune thrombocytopenia?

The absence of reliable biomarkers in immune thrombocytopenia (ITP) complicates treatment choice, necessitating a trial-and-error approach. Machine learning (ML) holds promise for transforming ITP treatment by analysing complex data to identify predictive factors, as demonstrated by Xu et al.'s study which developed ML-based models to predict responses to corticosteroids, rituximab and thrombopoietin receptor agonists. However, these models require external validation before can be adopted in clinical practice. Commentary on: Xu et al. A novel scoring model for predicting efficacy and guiding individualised treatment in immune thrombocytopenia. Br J Haematol 2024; 205:1108-1120.

Development and validation of prediction models for nosocomial infection and prognosis in hospitalized patients with cirrhosis

BACKGROUND

Nosocomial infections (NIs) frequently occur and adversely impact prognosis for hospitalized patients with cirrhosis. This study aims to develop and validate two machine learning models for NIs and in-hospital mortality risk prediction.

METHODS

The Prediction of Nosocomial Infection and Prognosis in Cirrhotic patients (PIPC) study included hospitalized patients with cirrhosis at the Qingchun Campus of the First Affiliated Hospital of Zhejiang University. We then assessed several machine learning algorithms to construct predictive models for NIs and prognosis. We validated the best-performing models with bootstrapping techniques and an external validation dataset. The accuracy of the predictions was evaluated through sensitivity, specificity, predictive values, and likelihood ratios, while predictive robustness was examined through subgroup analyses and comparisons between models.

RESULTS

We enrolled 1,297 patients into derivation cohort and 496 patients into external validation cohort. Among the six algorithms assessed, the Random Forest algorithm performed best. For NIs, the PIPC-NI model achieved an area under the curve (AUC) of 0.784 (95% confidence interval [CI] 0.741-0.826), a sensitivity of 0.712, and a specificity of 0.702. For in-hospital mortality, the PIPC- mortality model achieved an AUC of 0.793 (95% CI 0.749-0.836), a sensitivity of 0.769, and a specificity of 0.701. Moreover, our PIPC models demonstrated superior predictive performance compared to the existing MELD, MELD-Na, and Child-Pugh scores.

CONCLUSIONS

The PIPC models showed good predictive power and may facilitate healthcare providers in easily assessing the risk of NIs and prognosis among hospitalized patients with cirrhosis.

Can Machine Learning Assist in Diagnosis of Primary Immune Thrombocytopenia? A Feasibility Study

Primary Immune Thrombocytopenia (ITP) is a rare autoimmune disease characterised by the immune-mediated destruction of peripheral blood platelets in patients leading to low platelet counts and bleeding. The diagnosis and effective management of ITP are challenging because there is no established test to confirm the disease and no biomarker with which one can predict the response to treatment and outcome. In this work, we conduct a feasibility study to check if machine learning can be applied effectively for the diagnosis of ITP using routine blood tests and demographic data in a non-acute outpatient setting. Various ML models, including Logistic Regression, Support Vector Machine, k-Nearest Neighbor, Decision Tree and Random Forest, were applied to data from the UK Adult ITP Registry and a general haematology clinic. Two different approaches were investigated: a demographic-unaware and a demographic-aware one. We conduct extensive experiments to evaluate the predictive performance of these models and approaches, as well as their bias. The results revealed that Decision Tree and Random Forest models were both superior and fair, achieving nearly perfect predictive and fairness scores, with platelet count identified as the most significant variable. Models not provided with demographic information performed better in terms of predictive accuracy but showed lower fairness scores, illustrating a trade-off between predictive performance and fairness.

Risk-based and individualised management of bleeding and thrombotic events in adults with primary immune thrombocytopenia (ITP)

Although bleeding is one of the main symptoms of primary immune thrombocytopenia (ITP), risk factors for bleeding have yet to be fully established. Low platelet count (PC; <20-30 × 109 /L) is generally indicative of increased risk of bleeding. However, PC and bleeding events cannot be fully correlated; many other patient- and disease-related factors are thought to contribute to increased bleeding risk. Furthermore, even though ITP patients have thrombocytopenia and are at increased risk of bleeding, ITP also carries higher risk of thrombotic events. Factors like older age and certain ITP treatments are associated with increased thrombotic risk. Women's health in ITP requires particular attention concerning haemorrhagic and thrombotic complications. Management of bleeding/thrombotic risk, and eventually antithrombotic therapies in ITP patients, should be based on individual risk profiles, using a tailored, patient-centric approach. Currently, evidence-based recommendations and validated tools are lacking to support decision-making and help clinicians weigh risk of bleeding against thrombosis. Moreover, evidence is lacking about optimal PC for achieving haemostasis in invasive procedures settings. Further research is needed to fully define risk factors for each event, enabling development of comprehensive risk stratification approaches. This review discusses risk-based and individualised management of bleeding and thrombosis risk in adults with primary ITP.