Machine Learning-Based Prediction of Hemoglobinopathies Using Complete Blood Count Data.

BACKGROUND: Hemoglobinopathies, the most common inherited blood disorder, are frequently underdiagnosed. Early identification of carriers is important for genetic counseling of couples at risk. The aim of this study was to develop and validate a novel machine learning model on a multicenter data set, covering a wide spectrum of hemoglobinopathies based on routine complete blood count (CBC) testing. METHODS: Hemoglobinopathy test results from 10 322 adults were extracted retrospectively from 8 Dutch laboratories. eXtreme Gradient Boosting (XGB) and logistic regression models were developed to differentiate negative from positive hemoglobinopathy cases, using 7 routine CBC parameters. External validation was conducted on a data set from an independent Dutch laboratory, with an additional external validation on a Spanish data set (n = 2629) specifically for differentiating thalassemia from iron deficiency anemia (IDA). RESULTS: The XGB and logistic regression models achieved an area under the receiver operating characteristic (AUROC) of 0.88 and 0.84, respectively, in distinguishing negative from positive hemoglobinopathy cases in the independent external validation set. Subclass analysis showed that the XGB model reached an AUROC of 0.97 for ß-thalassemia, 0.98 for α0-thalassemia, 0.95 for homozygous α+-thalassemia, 0.78 for heterozygous α+-thalassemia, and 0.94 for the structural hemoglobin variants Hemoglobin C, Hemoglobin D, Hemoglobin E. Both models attained AUROCs of 0.95 in differentiating IDA from thalassemia. CONCLUSIONS: Both the XGB and logistic regression model demonstrate high accuracy in predicting a broad range of hemoglobinopathies and are effective in differentiating hemoglobinopathies from IDA. Integration of these models into the laboratory information system facilitates automated hemoglobinopathy detection using routine CBC parameters.

Analytical assay validation for acute myeloid leukemia measurable residual disease assessment by multiparametric flow cytometry.

BACKGROUND: Measurable residual disease (MRD) assessed by multiparametric flow cytometry (MFC) has gained importance in clinical decision-making for acute myeloid leukemia (AML) patients. However, complying with the recent In Vitro Diagnostic Regulations (IVDR) in Europe and Food and Drug Administration (FDA) guidance in the United States requires rigorous validation prior to their use in investigational clinical trials and diagnostics. Validating AML MRD-MFC assays poses challenges due to the unique underlying disease biology and paucity of patient specimens. In this study, we describe an experimental framework for validation that meets regulatory expectations. METHODS: Our validation efforts focused on evaluating assay accuracy, analytical specificity, analytical and functional sensitivity (limit of blank (LoB), detection (LLoD) and quantitation (LLoQ)), precision, linearity, sample/reagent stability and establishing the assay background frequencies. RESULTS: Correlation between different MFC methods was highly significant (r = 0.99 for %blasts and r = 0.93 for %LAIPs). The analysis of LAIP specificity accurately discriminated from negative control cells. The assay demonstrated a LoB of 0.03, LLoD of 0.04, and LLoQ of 0.1%. Precision experiments yielded highly reproducible results (Coefficient of Variation <20%). Stability experiments demonstrated reliable measurement of samples up to 96 h from collection. Furthermore, the reference range of LAIP frequencies in non-AML patients was below 0.1%, ranging from 0.0% to 0.04%. CONCLUSION: In this manuscript, we present the validation of an AML MFC-MRD assay using BM/PB patient specimens, adhering to best practices. Our approach is expected to assist other laboratories in expediting their validation activities to fulfill recent health authority guidelines.

Guideline development for prevention of transfusion-associated graft-versus-host disease: reduction of indications for irradiated blood components after prestorage leukodepletion of blood components.

Transfusion-associated graft-versus-host disease (TA-GVHD) is a rare, commonly fatal complication of transfusion preventable by irradiation of blood units. The revision of the Dutch transfusion guideline addressed the question whether irradiation is still necessary if blood components are prestorage leukodepleted. We searched for published cases of TA-GVHD following transfusion of prestorage leukodepleted blood and through contacting haemovigilance systems. Six presumed cases were found, dating from 1998 to 2013. Four out of six patients had received one or more non-irradiated units despite recognised indications for irradiated blood components. In the countries providing information, over 50 million prestorage leukodepleted, non-irradiated, non-pathogen-reduced cellular components were transfused in a 10-year period. Potential benefits of lifting indications for irradiation were considered. These include reduced irradiation costs (€ 1.5 million annually in the Netherlands) and less donor exposure for neonates. Findings were presented in an invitational expert meeting. Recommendations linked to human leukocyte antigen similarity between donor and recipient or intra-uterine transfusion were left unchanged. Indications linked to long-lasting deep T-cell suppression were defined with durations of 6 or 12 months after end of treatment (e.g. autologous or allogeneic stem cell transplantation). Need for continued alertness to TA-GVHD and haemovigilance reporting of erroneous non-irradiated transfusions was emphasised.

Applicability and reproducibility of acute myeloid leukaemia stem cell assessment in a multi-centre setting.

Leukaemic stem cells (LSC) have been experimentally defined as the leukaemia-propagating population and are thought to be the cellular reservoir of relapse in acute myeloid leukaemia (AML). Therefore, LSC measurements are warranted to facilitate accurate risk stratification. Previously, we published the composition of a one-tube flow cytometric assay, characterised by the presence of 13 important membrane markers for LSC detection. Here we present the validation experiments of the assay in several large AML research centres, both in Europe and the United States. Variability within instruments and sample processing showed high correlations between different instruments (Rpearson  > 0·91, P < 0·001). Multi-centre testing introduced variation in reported LSC percentages but was found to be below the clinical relevant threshold. Clear gating protocols resulted in all laboratories being able to perform LSC assessment of the validation set. Participating centres were nearly unanimously able to distinguish LSChigh (>0·03% LSC) from LSClow (<0·03% LSC) despite inter-laboratory variation in reported LSC percentages. This study proves that the LSC assay is highly reproducible. These results together with the high prognostic impact of LSC load at diagnosis in AML patients render the one-tube LSC assessment a good marker for future risk classification.

CD34+CD38- leukemic stem cell frequency to predict outcome in acute myeloid leukemia.

Current risk algorithms are primarily based on pre-treatment factors and imperfectly predict outcome in acute myeloid leukemia (AML). We introduce and validate a post-treatment approach of leukemic stem cell (LSC) assessment for prediction of outcome. LSC containing CD34+CD38- fractions were measured using flow cytometry in an add-on study of the HOVON102/SAKK trial. Predefined cut-off levels were prospectively evaluated to assess CD34+CD38-LSC levels at diagnosis (n = 594), and, to identify LSClow/LSChigh (n = 302) and MRDlow/MRDhigh patients (n = 305) in bone marrow in morphological complete remission (CR). In 242 CR patients combined MRD and LSC results were available. At diagnosis the CD34+CD38- LSC frequency independently predicts overall survival (OS). After achieving CR, combining LSC and MRD showed reduced survival in MRDhigh/LSChigh patients (hazard ratio [HR] 3.62 for OS and 5.89 for cumulative incidence of relapse [CIR]) compared to MRDlow/LSChigh, MRDhigh/LSClow, and especially MRDlow/LSClow patients. Moreover, in the NPM1mutant positive sub-group, prognostic value of golden standard NPM1-MRD by qPCR can be improved by addition of flow cytometric approaches. This is the first prospective study demonstrating that LSC strongly improves prognostic impact of MRD detection, identifying a patient subgroup with an almost 100% treatment failure probability, warranting consideration of LSC measurement incorporation in future AML risk schemes.

Management and outcome of 35 cases with foetal/neonatal alloimmune neutropenia.

AIM: The aim of this study was to provide an overview of foetal/neonatal alloimmune neutropenia (FNAIN), together with advice on the clinical management. METHODS: Neutrophil serology in the Netherlands is centralised at Sanquin Diagnostic Services. We examined FNAIN cases between January 1, 1991, and July 1, 2013, to determine the number of cases diagnosed, the relationship with human neutrophil antigen (HNA) antibody, the clinical presentation and therapeutic interventions. RESULTS: We identified 35 FNAIN cases. The detected HNA antibodies were as follows: anti-HNA-1a (n = 7), anti-HNA-1b (n = 12), anti-HNA-1c (n = 2), anti-HNA-2 (n = 8), anti-HNA-3a (n = 1), anti-HNA-5a (n = 1) and anti-FcγRIIIb (n = 4). No infections were diagnosed in 14 neonates, and the other 21 neonates suffered from omphalitis (n = 6), urinary tract infection (n = 1), candida mucositis (n = 1), fever of unknown origin (n = 6) and sepsis (n = 7, 20%). Parity, gestational age, birthweight, neutrophil counts and antibody specificity were not significantly different for cases with, and without, infections. All the infected children were treated with antibiotics. No children died. CONCLUSION: More than half (21) of the 35 cases of FNAIN presented with infections and most implicated were HNA-1a, HNA-1b and HNA-2. Treatment with antibiotics seemed adequate. A neonatal neutropenia workflow model for use in neonatal intensive care units is presented.

A rapid single-tube multiplex polymerase chain reaction assay for the seven most prevalent alpha-thalassemia deletions and alphaalphaalpha(anti 3.7) alpha-globin gene triplication.

alpha-Globin gene triplications may exacerbate the alpha chain and beta chain imbalance in beta-thalassemia (beta-thal) and may compensate for the effect of alpha-globin gene deletion in alpha-thal. Identification of an alpha-globin gene triplication is, therefore, valuable in predicting the clinical phenotype of the thalassemias. To be able to detect alpha-globin gene triplications, we have modified an existing multiplex polymerase chain reaction (PCR) assay for the seven most prevalent alpha-globin gene deletions by incorporating two triplication-specific primers and concurrently substituting one of the original primers by a newly designed primer. This modified multiplex PCR assay was evaluated by performing the assay on archival DNA samples and on peripheral blood samples from 163 suspected thalassemia cases. It was found to function properly. Our assay thereby represents the first multiplex PCR assay that can detect both the seven most prevalent alpha-globin gene deletions and the alphaalphaalpha(anti 3.7) alpha-globin gene triplication in a single-tube reaction.

Circulating cerebral S100B protein is associated with depressive symptoms following myocardial infarction.

BACKGROUND: Prevalence of depressive symptoms in the post-myocardial infarction (MI) period varies from 8 to 30%. Cerebral damage after MI, caused by transient ischemia, an inflammatory response or both, may contribute to development of post-MI depression. S100B is an established protein marker of cerebral damage. In a pilot study, the authors assessed whether S100B serum levels are: (1) increased during the week after MI, and (2) related to depressive symptoms during index hospital admission and the year following MI. METHODS: This pilot study is a substudy of the Myocardial Infarction and Depression Intervention Trial (MIND-IT). In 48 patients, serum levels of S100B were available at 1, 2, 3, 4 and 8 days following MI. Subsequently, in 27 patients, depressive symptoms were measured at 0, 3, 6, 9 and 12 months following MI with the Beck Depression Inventory (BDI). In 21 of the initial 48 patients, BDI data were lacking due to refusals to fill out BDI forms or missing data. RESULTS: Significant and transient increases in serum S100B were observed in 81.3% of the 48 patients: 37.5% reached S100B serum levels comparable to serum levels found in acute brain injury (>0.20 microg/l) and 43.8% reached mildly elevated S100B serum levels comparable to serum levels found in depressive disorder (0.10-0.20 microg/l). In 18.7%, no S100B was detected in serum. Using non-parametric Spearman rank correlation tests, a trend towards an association was found between serum S100B and depressive symptoms during the post-MI year (rho values between 0.16 and 0.53) in 27 patients who completed both the S100B serum study and the BDI study. CONCLUSION: Transiently elevated levels of S100B are suggestive of minor acute cerebral damage in the first days following MI and associated with depressive symptoms in the year following MI. Cerebral damage could be an important mechanism in the pathogenesis in a subtype of post-MI depression.

Analyzing DNA from buccal cells is a reliable method for the exclusion of cystic fibrosis. Results of a pilot study.

PURPOSE: In children there is frequently a reason to exclude cystic fibrosis. Sweat testing is used for this. Because sweat testing has some disadvantages we investigated whether analyzing DNA for the local most common CFTR mutations, harvested from buccal cells, is reliable as a method to exclude cystic fibrosis. METHODS: In patients in whom a sweat test had been ordered during the period January 1, 2002 to December 31, 2004, we harvested buccal cell DNA for analysis. When blood was available, DNA from leukocytes was also analyzed. RESULTS: A total of 73 sweat tests were ordered during the two-year study period, mostly because of recurrent pulmonary infections (36; 49%), failure to thrive (20; 27%) and chronic diarrhea (10, 14%). In 70, children the results of the sweat test were normal, in three patients the results were borderline. Sixty buccal smears were analyzed and no patient was homozygous for cystic fibrosis, two were heterozygous for cystic fibrosis. In none of the children the diagnosis of cystic fibrosis was established. CONCLUSION: Analyzing DNA in cells, harvested from the buccal cells, is a reliable alternative to exclude cystic fibrosis. It is safe, simple, and child-friendly.

Albumin-adjusted calcium is not suitable for diagnosis of hyper- and hypocalcemia in the critically ill.

OBJECTIVE: To evaluate whether calcium adjusted for albumin can be used to monitor calcium homeostasis in critically ill patients. DESIGN: Prospective single-single center observational study. SETTING: Clinical laboratory and critical care unit of a regional teaching hospital. PATIENTS: Fifty-three paired samples were from 36 patients requiring intensive care treatment. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Total calcium, albumin-adjusted calcium, and ionized calcium were measured in critically ill patients during an 8-wk period. Calcium was adjusted for albumin using the formula that is most frequently used in The Netherlands. Using ionized calcium as the gold standard, albumin-adjusted calcium overestimated hypercalcemia and totally missed hypocalcemia. The same seemed to be true for other formulas used for albumin or protein adjustment of calcium concentrations. CONCLUSIONS: Albumin-adjusted calcium cannot be used in an intensive care setting to monitor reliably the calcium levels in critically ill patients and should be replaced by measurement of ionized calcium.

Adequate sampling in cryoglobulinaemia: recommended warmly.

Propersample handling is important for the accurate and prompt diagnosis of cryoglobulinaemia as well as for obtaining correct results of other laboratory tests. In this case report, we present two patients for whom improper sampling delayed the diagnosis of cryoglobulinaemia. The routine sampling procedure was not sufficient for detecting these cryoglobulins and we demonstrated that the temperature dropped below 37 degrees C at different steps during the sample preparation. Adjusting the preanalytical procedure revealed that both patients possessed a cryoglobulin with high thermal insolubility. We conclude that in cryoglobulinaemia strict adherence to guidelines to keep the temperature >37 degrees C is crucial for sample collection and must be strongly recommended.