Cell Population Data (CPD) for Early Recognition of Sepsis and Septic Shock in Children: A Pilot Study

Development and validation of a machine learning-based interpretable model for predicting sepsis by complete blood cell parameters

Background

Sepsis, a severe infectious disease, carries a high mortality rate. Early detection and prompt treatment are crucial for reducing mortality and improving prognosis. The aim of this research is to develop a clinical prediction model using machine learning algorithms, leveraging complete blood cell (CBC) parameters, to detect sepsis at an early stage.

Methods

The study involved 572 patients admitted to West China Hospital of Sichuan University between July 2020 and September 2021. Among them, 215 were diagnosed with sepsis, while 357 had local infections. Demographic information was collected, and 57 CBC parameters were analyzed to identify potential predictors using techniques such as the Least Absolute Shrinkage and Selection Operator (LASSO), Random Forest (RF), Support Vector Machine (SVM), and eXtreme Gradient Boosting (XGBoost). The prediction model was built using Logistic Regression and evaluated for diagnostic specificity, discrimination, and clinical applicability including metrics such as the area under the curve (AUC), calibration curve, clinical impact curve, and clinical decision curve. Additionally, the model's diagnostic performance was assessed on a separate validation cohort. Shapley's additive explanations (SHAP), and breakdown (BD) profiles were used to explain the contribution of each variable in predicting the outcome.

Results

Among all the machine learning methods' prediction models, the LASSO-based model (λ = min) demonstrated the highest diagnostic performance in both the discovery cohort (AUC = 0.9446, P < 0.001) and the validation cohort (AUC = 0.9001, P < 0.001). Furthermore, upon local analysis and interpretation of the model, we demonstrated that LY-Z, MO-Z, and PLT-I had the most significant impact on the outcome.

Conclusions

The predictive model based on CBC parameters can be utilized as an effective approach for the early detection of sepsis.

Significant Variability in the Identification and Reporting of Band Neutrophils by Participants Enrolled in the College of American Pathologists Proficiency Testing Program: Time for a Change

CONTEXT.—

Increased band neutrophils in blood smear differential counts ("bandemia") are entrenched in medicine as a flag for sepsis. However, laboratory hematology experts have long advocated for discontinuation of reporting bands separately from segmented neutrophils because of poor sensitivity and specificity, poor interobserver agreement, and availability of alternative biomarkers for sepsis.

OBJECTIVE.—

To describe band neutrophil reporting practices and reproducibility of band classification among laboratories participating in the College of American Pathologists (CAP) proficiency testing (PT) program.

DESIGN.—

A survey questionnaire was distributed to hematology PT participants. A subsequent morphologic challenge included 12 preselected cell identifications of segmented neutrophils, bands, and metamyelocytes, and a 100-cell manual differential count of a digitally scanned blood smear.

RESULTS.—

Among laboratories that reported manual differentials, most respondents reported bands (4554 of 5268; 86.4%). Only 3222 of 4412 respondents (73.0%) provided band reference ranges. Though participants classified "easy" band neutrophils well (78.0%-98.3%), categorization of cell identifications for "moderate" and "difficult" bands was poor (3.1%-39.0% of laboratories), with classification instead as segmented neutrophils. This pattern was seen regardless of laboratory demographic characteristics. Marked variability in band counts was observed on the 100-cell differential count for both CAP PT participants and CAP Hematology and Clinical Microscopy Committee (HCMC) members (coefficients of variation, 55.8% and 32.9%, respectively). Variability was significantly improved when segmented and band neutrophils were grouped together (coefficients of variation, 6.2% and 5.0%, respectively).

CONCLUSIONS.—

Most CAP PT-participating laboratories report band counts, many without reference ranges. The survey confirms significant interlaboratory variability of band enumeration when bands are separately identified from segmented neutrophils. This study reaffirms the CAP Hematology and Clinical Microscopy Committee's strong recommendation to group segmented and band neutrophils together in manual differential counts.

Construction of the prediction model for multiple myeloma based on machine learning

INTRODUCTION

The global burden of multiple myeloma (MM) is increasing every year. Here, we have developed machine learning models to provide a reference for the early detection of MM.

METHODS

A total of 465 patients and 150 healthy controls were enrolled in this retrospective study. Based on the variable screening strategy of least absolute shrinkage and selection operator (LASSO), three prediction models, logistic regression (LR), support vector machine (SVM), and random forest (RF), were established combining complete blood count (CBC) and cell population data (CPD) parameters in the training set (210 cases), and were verified in the validation set (90 cases) and test set (165 cases). The performance of each model was analyzed using receiver operating characteristic (ROC) curve, calibration curves, and decision curve analysis (DCA). Accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the ROC curve (AUC) were applied to evaluate the models. Delong test was used to compare the AUC of the models.

RESULTS

Six parameters including RBC (1012/L), RDW-CV (%), IG (%), NE-WZ, LY-WX, and LY-WZ were screened out by LASSO to construct the model. Among the three models, the AUC of RF model in the training set, validation set, and test set were 0.956, 0.892, and 0.875, which were higher than those of LR model (0.901, 0.849, and 0.858) and SVM model (0.929, 0.868, and 0.846). Delong test showed that there were significant differences among the models in the training set, no significant differences in the validation set, and significant differences only between SVM and RF models in the test set. The calibration curve and DCA showed that the three models had good validity and feasibility, and the RF model performed best.

CONCLUSION

The proposed RF model may be a useful auxiliary tool for rapid screening of MM patients.

Haematology laboratory parameters to assess efficacy of CD19-, CD22-, CD33-, and CD123-directed chimeric antigen receptor T-cell therapy in haematological malignancies

INTRODUCTION

Chimeric antigen receptor (CAR) T cell products are available to treat relapsed/refractory B-lymphoblastic leukaemia/lymphoma (B-ALL), diffuse large B-cell lymphoma, mantle-cell lymphoma, and myeloma. CAR products vary by their target epitope and constituent molecules. Hence, there are no common laboratory assays to assess CAR T cell expansion in the clinical setting. We investigated the utility of common haematology laboratory parameters to measure CAR T cell expansion and response.

METHODS

Archived CellaVision images, absolute lymphocyte counts, and Sysmex CPD parameters spanning 1 month after CD19-CAR, UCAR19, CD22-CAR, CD33-CAR, and UCAR123 therapy were compared against donor lymphocyte infused control patients. Additionally, CellaVision images gathered during acute EBV infection were analysed.

RESULTS

CellaVision images revealed a distinct sequence of three lymphocyte morphologies, common among CD19-CAR, CD22-CAR and UCAR19. This lymphocyte sequence was notably absent in CAR T cell non-responders and stem-cell transplantation controls, but shared some features seen during acute EBV infection. CD19-CAR engraftment kinetics monitored by quantitative PCR show an expansion and persistence phase and mirror CD19-CAR ALC kinetics. We show other novel CAR T cell therapies (UCAR19, CD22-CAR, CD33-CAR and UCAR123) display similar ALC expansion in responders and diminished ALC expansion in non-responders. Furthermore, the CPD parameter LY_WY fluorescence increased within the first week after CD19-CAR infusion, preceding the peak absolute lymphocyte count (ALC) by 3.7 days.

CONCLUSION

Autologous and allogeneic CAR T cell therapy produce unique changes in common haematology laboratory parameters and could be a useful surrogate to follow CAR T-cell expansion after infusion.