A clinical prediction model identifies a subgroup with inferior survival within intermediate risk acute myeloid leukemia

Investigation of distinct gene expression profile patterns that can improve the classification of intermediate-risk prognosis in AML patients

Background: Acute myeloid leukemia (AML) is a heterogeneous type of blood cancer that generally affects the elderly. AML patients are categorized with favorable-, intermediate-, and adverse-risks based on an individual's genomic features and chromosomal abnormalities. Despite the risk stratification, the progression and outcome of the disease remain highly variable. To facilitate and improve the risk stratification of AML patients, the study focused on gene expression profiling of AML patients within various risk categories. Therefore, the study aims to establish gene signatures that can predict the prognosis of AML patients and find correlations in gene expression profile patterns that are associated with risk groups. Methods: Microarray data were obtained from Gene Expression Omnibus (GSE6891). The patients were stratified into four subgroups based on risk and overall survival. Limma was applied to screen for differentially expressed genes (DEGs) between short survival (SS) and long survival (LS). DEGs strongly related to general survival were discovered using Cox regression and LASSO analysis. To assess the model's accuracy, Kaplan-Meier (K-M) and receiver operating characteristic (ROC) were used. A one-way ANOVA was performed to assess for differences in the mean gene expression profiles of the identified prognostic genes between the risk subcategories and survival. GO and KEGG enrichment analyses were performed on DEGs. Results: A total of 87 DEGs were identified between SS and LS groups. The Cox regression model selected nine genes CD109, CPNE3, DDIT4, INPP4B, LSP1, CPNE8, PLXNC1, SLC40A1, and SPINK2 that are associated with AML survival. K-M illustrated that the high expression of the nine-prognostic genes is associated with poor prognosis in AML. ROC further provided high diagnostic efficacy of the prognostic genes. ANOVA also validated the difference in gene expression profiles of the nine genes between the survival groups, and highlighted four prognostic genes to provide novel insight into risk subcategories poor and intermediate-poor, as well as good and intermediate-good that displayed similar expression patterns. Conclusion: Prognostic genes can provide more accurate risk stratification in AML. CD109, CPNE3, DDIT4, and INPP4B provided novel targets for better intermediate-risk stratification. This could enhance treatment strategies for this group, which constitutes the majority of adult AML patients.

Re-induction therapy in patients with acute myeloid leukemia not in complete remission after the first course of treatment

A standard salvage regimen for patients with acute myeloid leukemia (AML) who are not in complete remission (CR) after initial induction therapy does not exist. We retrospectively investigated re-induction therapy for 151 patients with AML who did not achieve CR after the initial course between January 2014 and March 2021. The re-induction regimen did not correlate with the CR rate after the second course, whereas patients had similar 5-year overall survival (OS) and event-free survival (EFS) based on different re-induction regimens. Multivariable analysis revealed that International European Leukaemia Net (ELN) risk stratification independently predicted both OS and EFS among patients not in CR after the first course, although the re-induction regimen did not predict prognosis. Urgent salvage alloHSCT may improve the prognosis of patients with refractory AML. In summary, our study showed that the re-induction regimen did not significantly predict the prognosis of patients with AML not in CR after the first course of treatment. The development and selection of an efficient treatment algorithm for the treatment of AML remains a pressing research challenge.

Making Decisions When No Further Diagnostic Testing is Available

In this chapter, we illustrate how evidence about treatments' benefits and harms can be integrated to enable rational decision-making even under considerable clinical uncertainty.

Decision-Making When Diagnostic Testing is Available

When a decision-maker has the option of diagnostic testing, they face a typical dilemma: (1) do not administer treatment and do not test, (2) test and decide to administer treatment based on the test result, and (3) administer treatment without testing. In this chapter, we will discuss the theory behind threshold modeling when diagnostic testing is available; we will illustrate the approach by presenting a case vignette.

Predicting the risk of acute kidney injury after hematopoietic stem cell transplantation: development of a new predictive nomogram

The purpose was to predict the risk of acute kidney injury (AKI) within 100 days after hematopoietic stem cell transplantation (HSCT) in patients with hematologic disease by using a new predictive nomogram. Collect clinical data of patients with hematologic disease undergoing HSCT in our hospital from August 2012 to March 2018. Parameters with non-zero coefficients were selected by the Least Absolute Selection Operator (LASSO). Then these parameters were selected to build a new predictive nomogram model. Receiver operating characteristic (ROC) curve, calibration curve, C-index, and decision curve analysis (DCA) were used for the validation of the evaluation model. Finally, the nomogram was further evaluated by internal verification. According to 2012 Kidney Disease Improving Global Guidelines (KDIGO) diagnostic criteria, among 144 patients, the occurrence of AKI within 100 days after HSCT The rate was 29.2% (42/144). The C-index of the nomogram was 0.842. The C-value calculated by the internal verification was 0.809. The AUC was 0.842, and The DCA range of the predicted nomogram was from 0.01 to 0.71. This article established a high-precision nomogram for the first time for predicting the risk of AKI within 100 days after HSCT in patients with hematologic diseases. The nomogram had good clinical validity and reliability. For clinicians, it was very important to prevent AKI after HSCT.

Comprehensive analysis of ECHDC3 as a potential biomarker and therapeutic target for acute myeloid leukemia: Bioinformatic analysis and experimental verification

Background

Enoyl-CoA hydratase domain containing 3 (ECHDC3) increased in CD34+ progenitor cells of acute myeloid leukemia (AML) cells after chemotherapy. However, the prognostic significance and function of ECHDC3 in AML remain to be clarified.

Methods

In the training cohort, 24 AML (non-acute promyelocytic leukemia, APL) patients were enrolled in Peking University People’s Hospital and tested for ECHDC3 in enriched CD34+ cells at diagnosis. In the validation set, 351 bone marrow RNA-seq data of non-APL AML were obtained by two independent online datasets (TCGA-LAML and BEAT-AML). LASSO regression model was conducted to a new prediction model of ECHDC3-related genes. In addition, the ECHDC3 signature was further explored by GO, KEGG, GSEA, and immuno-infiltration analysis. By RNA interference, the function of ECHDC3 in mitochondrial DNA (mt-DNA) transcriptome and chemoresistance was further explored, and the GSE52919 database re-verified the ECHDC3 chemoresistance feature.

Results

By Kaplan-Meier analysis, patients with ECHDC3high demonstrated inferior overall survival (OS) compared to those with ECHDC3low both in the training (2-year OS, 55.6% vs. 100%, p = 0.011) and validation cohorts (5-year OS, 9.6% vs. 24.3%, p = 0.002). In addition, ECHDC3high predicted inferior OS in the subgroup of patients with ELN 2017 intermediated (int) risk (5-year OS, 9.5% vs. 26.3%, p = 0.039) or FLT3+NPM1− adverse (adv) risk (4-year OS, 6.4% vs. 31.8%, p = 0.003). In multivariate analysis, ECHDC3 was an independent risk factor of inferior OS (HR 1.159, 95% CI 1.013–1.326, p = 0.032). In the prediction model combining ECHDC3 and nine selected genes (RPS6KL1, RELL2, FAM64A, SPATS2L, MEIS3P1, CDCP1, CD276, IL1R2, and OLFML2A) by Lasso regression, patients with high risk showed inferior 5-year OS (9.3% vs. 23.5%, p < 0.001). Bioinformatic analysis suggested that ECHDC3 alters the bone marrow microenvironment by inducing NK, resting mast cell, and monocyte differentiation. Knocking down ECHDC3 in AML cells by RNAi promoted the death of leukemia cells with cytarabine and doxorubicin.

Conclusion

These bioinformatic analyses and experimental verification indicated that high ECHDC3 expression might be a poor prognostic biomarker for non-APL AML, which might be a potential target for reverting chemoresistance.

Comparing machine learning algorithms to predict 5-year survival in patients with chronic myeloid leukemia

INTRODUCTION

Chronic myeloid leukemia (CML) is a myeloproliferative disorder resulting from the translocation of chromosomes 19 and 22. CML includes 15-20% of all cases of leukemia. Although bone marrow transplant and, more recently, tyrosine kinase inhibitors (TKIs) as a first-line treatment have significantly prolonged survival in CML patients, accurate prediction using available patient-level factors can be challenging. We intended to predict 5-year survival among CML patients via eight machine learning (ML) algorithms and compare their performance.

METHODS

The data of 837 CML patients were retrospectively extracted and randomly split into training and test segments (70:30 ratio). The outcome variable was 5-year survival with potential values of alive or deceased. The dataset for the full features and important features selected by minimal redundancy maximal relevance (mRMR) feature selection were fed into eight ML techniques, including eXtreme gradient boosting (XGBoost), multilayer perceptron (MLP), pattern recognition network, k-nearest neighborhood (KNN), probabilistic neural network, support vector machine (SVM) (kernel = linear), SVM (kernel = RBF), and J-48. The scikit-learn library in Python was used to implement the models. Finally, the performance of the developed models was measured using some evaluation criteria with 95% confidence intervals (CI).

RESULTS

Spleen palpable, age, and unexplained hemorrhage were identified as the top three effective features affecting CML 5-year survival. The performance of ML models using the selected-features was superior to that of the full-features dataset. Among the eight ML algorithms, SVM (kernel = RBF) had the best performance in tenfold cross-validation with an accuracy of 85.7%, specificity of 85%, sensitivity of 86%, F-measure of 87%, kappa statistic of 86.1%, and area under the curve (AUC) of 85% for the selected-features. Using the full-features dataset yielded an accuracy of 69.7%, specificity of 69.1%, sensitivity of 71.3%, F-measure of 72%, kappa statistic of 75.2%, and AUC of 70.1%.

CONCLUSIONS

Accurate prediction of the survival likelihood of CML patients can inform caregivers to promote patient prognostication and choose the best possible treatment path. While external validation is required, our developed models will offer customized treatment and may guide the prescription of personalized medicine for CML patients.

Cytogenetic abnormalities in NPM1-mutated acute myeloid leukemia

NPM1^mut acute myeloid leukemia (AML) has been identified as a distinct entity of myeloid neoplasms according to the 2017 European LeukemiaNet (ELN) guidelines. It confers a favorable prognosis regardless of cytogenetic abnormalities. We evaluated 418 newly diagnosed AML patients to test the validity of this hypothesis. Seventy-four patients with NPM1^mut AML showed a good response to induction and a relatively favorable prognosis. Abnormal karyotypes were observed in 15 patients. Chromosomal abnormalities were significantly associated with a worse prognosis in NPM1^mut AML patients (5-year overall survival (OS): 38.9 ± 12.9%, p = .037; event-free survival (EFS): 33.3 ± 12.2%, p = .043, respectively). Four patients with abnormal karyotypes who underwent allogeneic hematopoietic stem cell transplantation (alloHSCT) during CR1 had longer survival than those who received chemotherapy only. Multivariable analysis revealed abnormal karyotypes independently predicted OS and EFS among NPM1^mut AML patients. In summary, cytogenetic abnormalities are strong prognostic indicators in NPM1^mut AML. Therefore, they should be classified accordingly, and alloHSCT should be performed on selected patients during CR1.