Rapid haemoglobin A and S quantification using Tosoh HLC-723G8 in variant mode for patients with sickle cell disease.

The management of life-threatening complications in patients with sickle cell disease (SCD) requires an accurate and reproducible quantification of haemoglobin A (HbA) and S (HbS) with a short turnaround time and 24-7 availability. We propose a novel method for quantifying HbA and HbS using the glycated haemoglobin (HbA1c) assay on a Tosoh HLC-723G8 (G8) analyser in variant mode. HbA and HbS results obtained using our method highly correlated with results obtained using a reference method (r > 0.99 for 124 samples of patients with SCD or sickle cell trait). Our method met laboratory requirements for linearity (coefficient of variation [CV] and bias <5%), between-run and within-run reproducibility (CV <10%) and carryover (<0.5%) over the range of HbS and HbA values expected in a therapeutic context. Using the G8 analyser in variant mode is viable for monitoring HbA and HbS concentrations in dire situations. This method is easy to use, quick (1.6 min per sample), and automatable and produces highly reproducible results without significant bias. Finally, it does not require modifications to the analytical pipeline recommended by the supplier, enabling a 24-7 availability without disrupting routine monitoring of HbA1c in the laboratory.

Vacuoles in circulating immature myeloid cells in VEXAS syndrome and comparison with UBA1-wild type patients.

We did not identify any vacuole-related differences in circulating immature myeloid cells between VEXAS patients and UBA1-WT 'VEXAS-like' patients. The similar vacuolization of circulating immature myeloid cells between VEXAS and UBA1-WT patients is explained by the main bloodstream passage of late precursors, in which the vacuolization is already similar in bone marrow in both cases.

Combined Metabolipidomic and Machine Learning Approach in a Rat Model of Stroke Reveals a Deleterious Impact of Brain Injury on Heart Metabolism.

Cardiac complications are frequently found following a stroke in humans whose pathophysiological mechanism remains poorly understood. We used machine learning to analyse a large set of data from a metabolipidomic study assaying 630 metabolites in a rat stroke model to investigate metabolic changes affecting the heart within 72 h after a stroke. Twelve rats undergoing a stroke and 28 rats undergoing the sham procedure were investigated. A plasmatic signature consistent with the literature with notable lipid metabolism remodelling was identified. The post-stroke heart showed a discriminant metabolic signature, in comparison to the sham controls, involving increased collagen turnover, increased arginase activity with decreased nitric oxide synthase activity as well as an altered amino acid metabolism (including serine, asparagine, lysine and glycine). In conclusion, these results demonstrate that brain injury induces a metabolic remodelling in the heart potentially involved in the pathophysiology of stroke heart syndrome.

An immuno-lipidomic signature revealed by metabolomic and machine-learning approaches in labial salivary gland to diagnose primary Sjögren's syndrome.

Introduction: Assessing labial salivary gland exocrinopathy is a cornerstone in primary Sjögren's syndrome. Currently this relies on the histopathologic diagnosis of focal lymphocytic sialadenitis and computing a focus score by counting lym=phocyte foci. However, those lesions represent advanced stages of primary Sjögren's syndrome, although earlier recognition of primary Sjögren's syndrome and its effective treatment could prevent irreversible damage to labial salivary gland. This study aimed at finding early biomarkers of primary Sjögren's syndrome in labial salivary gland combining metabolomics and machine-learning approaches. Methods: We used a standardized targeted metabolomic approach involving high performance liquid chromatography coupled with mass spectrometry among newly diagnosed primary Sjögren's syndrome (n=40) and non- primary Sjögren's syndrome sicca (n=40) participants in a prospective cohort. A metabolic signature predictive of primary Sjögren's syndrome status was explored using linear (logistic regression with elastic-net regularization) and non-linear (random forests) machine learning architectures, after splitting the data set into training, validation, and test sets. Results: Among 126 metabolites accurately measured, we identified a discriminant signature composed of six metabolites with robust performances (ROC-AUC = 0.86) for predicting primary Sjögren's syndrome status. This signature included the well-known immune-metabolite kynurenine and five phospholipids (LysoPC C28:0; PCaa C26:0; PCaaC30:2; PCae C30:1, and PCaeC30:2). It was split into two main components: the first including the phospholipids was related to the intensity of lymphocytic infiltrates in salivary glands, while the second represented by kynurenine was independently associated with the presence of anti-SSA antibodies in participant serum. Conclusion: Our results reveal an immuno-lipidomic signature in labial salivary gland that accurately distinguishes early primary Sjögren's syndrome from other causes of sicca symptoms.

Accurate stratification between VEXAS syndrome and differential diagnoses by deep learning analysis of peripheral blood smears.

OBJECTIVES: VEXAS syndrome is a newly described autoinflammatory disease associated with UBA1 somatic mutations and vacuolization of myeloid precursors. This disease possesses an increasingly broad spectrum, leading to an increase in the number of suspected cases. Its diagnosis via bone-marrow aspiration and UBA1-gene sequencing is time-consuming and expensive. This study aimed at analyzing peripheral leukocytes using deep learning approaches to predict VEXAS syndrome in comparison to differential diagnoses. METHODS: We compared leukocyte images from blood smears of three groups: participants with VEXAS syndrome (identified UBA1 mutation) (VEXAS); participants with features strongly suggestive of VEXAS syndrome but without UBA1 mutation (UBA1-WT); participants with a myelodysplastic syndrome and without clinical suspicion of VEXAS syndrome (MDS). To compare images of circulating leukocytes, we applied a two-step procedure. First, we used self-supervised contrastive learning to train convolutional neural networks to translate leukocyte images into lower-dimensional encodings. Then, we employed support vector machine to predict patients' condition based on those leukocyte encodings. RESULTS: The VEXAS, UBA1-WT, and MDS groups included 3, 3, and 6 patients respectively. Analysis of 33,757 images of neutrophils and monocytes enabled us to distinguish VEXAS patients from both UBA1-WT and MDS patients, with mean ROC-AUCs ranging from 0.87 to 0.95. CONCLUSIONS: Image analysis of blood smears via deep learning accurately distinguished neutrophils and monocytes drawn from patients with VEXAS syndrome from those of patients with similar clinical and/or biological features but without UBA1 mutation. Our findings offer a promising pathway to better screening for this disease.

In Reply to Performance of Deep Learning in the Interpretation of Serum Protein Electrophoresis.

We thank He et al. for their comments on our article (1), which gives us the opportunity to clarify some methodological points. 1. Detection of abnormal patterns: mechanics.

A Metabolomic Profiling of Intra-Uterine Growth Restriction in Placenta and Cord Blood Points to an Impairment of Lipid and Energetic Metabolism.

(1) Background: Intrauterine growth restriction (IUGR) involves metabolic changes that may be responsible for an increased risk of metabolic and cardiovascular diseases in adulthood. Several metabolomic profiles have been reported in maternal blood and urine, amniotic fluid, cord blood and newborn urine, but the placenta has been poorly studied so far. (2) Methods: To decipher the origin of this metabolic reprogramming, we conducted a targeted metabolomics study replicated in two cohorts of placenta and one cohort of cord blood by measuring 188 metabolites by mass spectrometry. (3) Results: OPLS-DA multivariate analyses enabled clear discriminations between IUGR and controls, with good predictive capabilities and low overfitting in the two placental cohorts and in cord blood. A signature of 25 discriminating metabolites shared by both placental cohorts was identified. This signature points to sharp impairment of lipid and mitochondrial metabolism with an increased reliance on the creatine-phosphocreatine system by IUGR placentas. Increased placental insulin resistance and significant alteration of fatty acids oxidation, together with relatively higher phospholipase activity in IUGR placentas, were also highlighted. (4) Conclusions: Our results show a deep lipid and energetic remodeling in IUGR placentas that may have a lasting effect on the fetal metabolism.

Cancer/Testis Antigen 55 is required for cancer cell proliferation and mitochondrial DNA maintenance.

Cancer/Testis Antigens (CTAs) represent a group of proteins whose expression under physiological conditions is restricted to testis but activated in many human cancers. Also, it was observed that co-expression of multiple CTAs worsens the patient prognosis. Five CTAs were reported acting in mitochondria and we recently reported 147 transcripts encoded by 67 CTAs encoding for proteins potentially targeted to mitochondria. Among them, we identified the two isoforms encoded by CT55 for whom the function is poorly understood. First, we found that patients with tumors expressing wild-type CT55 are associated with poor survival. Moreover, CT55 silencing decreases dramatically cell proliferation. Second, to investigate the role of CT55 on mitochondria, we first show that CT55 is localized to both mitochondria and endoplasmic reticulum (ER) due to the presence of an ambiguous N-terminal targeting signal. Then, we show that CT55 silencing decreases mtDNA copy number and delays mtDNA recovery after an acute depletion. Moreover, demethylation of CT55 promotor increases its expression, which in turn increases mtDNA copy number. Finally, we measured the mtDNA copy number in NCI-60 cell lines and screened for genes whose expression is strongly correlated to mtDNA amount. We identified CT55 as the second highest correlated hit. Also, we show that compared to siRNA scrambled control (siCtrl) treatment, CT55 specific siRNA (siCT55) treatment down-regulates aerobic respiration, indicating that CT55 sustains mitochondrial respiration. Altogether, these data show for first time that CT55 acts on mtDNA copy number, modulates mitochondrial activity to sustain cancer cell proliferation.

Post-infarct cardiac remodeling predictions with machine learning.

BACKGROUND: We sought to improve the risk prediction of 3-month left ventricular remodeling (LVR) occurrence after myocardial infarction (MI), using a machine learning approach. METHODS: Patients were included from a prospective cohort study analyzing the incidence of LVR in ST-elevation MI in 443 patients that were monitored at Angers University Hospital, France. Clinical, biological and cardiac magnetic resonance (CMR) imaging data from the first week post MI were collected, and LVR was assessed with CMR at 3 month. Data were processed with a machine learning pipeline using multiple feature selection algorithms to identify the most informative variables. RESULTS: We retrieved 133 clinical, biological and CMR imaging variables, from 379 patients with ST-elevation MI. A baseline logistic regression model using previously known variables achieved an AUC of 0.71 on the test set, with 67% sensitivity and 64% specificity. In comparison, our best predictive model was a neural network using seven variables (in order of importance): creatine kinase, mean corpuscular volume, baseline left atrial surface, history of diabetes, history of hypertension, red blood cell distribution width, and creatinine. This model achieved an AUC of 0.78 on the test set, reaching a sensitivity of 92% and a specificity of 55%, outperforming the baseline model. CONCLUSION: These preliminary results show the value of using an unbiased data-driven machine learning approach. We reached a higher level of sensitivity compared to traditional methods for the prediction of a 3-month post-MI LVR.

Early recognition of cardiac surgery-associated acute kidney injury: lack of added value of TIMP2 IGFBP7 over short-term changes in creatinine (an observational pilot study).

BACKGROUND: For the detection of cardiac surgery-associated acute kidney injury (CS-AKI), the performance of urine tissue inhibitor of metalloproteinase 2 insulin-like growth factor-binding protein 7 (TIMP2 IGFBP7) has never been compared with that of very early changes in plasma creatinine (∆pCr). We hypothesized that, in the context of perioperative haemodilution, lack of postoperative decrease in pCr would be of honourable performance for the detection of CS-AKI. We therefore aimed at comparing these biomarkers and their kinetics (primary objective). As secondary objectives, we assessed plasma neutrophil gelatinase-associated lipocalin (pNGAL), cystatin C (pCysC) and urea (pUrea). We also determined the ability of these biomarkers to early discriminate persistent from transient CS-AKI. METHODS: Patients over 75 years-old undergoing aortic valve replacement with cardiopulmonary bypass (CPB) were included in this prospective observational study. Biomarkers were measured before/after CPB and at the sixth postoperative hour (H6). RESULTS: In 65 patients, CS-AKI occurred in 27 (42%). ∆pCr from post-CPB to H6 (∆pCrpostCPB-H6): outperformed TIMP2 IGFBP7 at H6 and its intra- or postoperative changes: area under the receiver operating characteristic curve (AUCROC) of 0.84 [95%CI:0.73-0.92] vs. ≤0.67 [95%CI:0.54-0.78], p ≤ 0.03. The AUCROC of pNGAL, pCysC and pUrea did not exceed 0.72 [95%CI:0.59-0.83]. Indexing biomarkers levels for blood or urine dilution did not improve their performance. Combining TIMP2 IGFBP7 and ∆pCrpostCPB-H6 was of no evident added value over considering ∆pCrpostCPB-H6 alone. For the early recognition of persistent CS-AKI, no biomarker outperformed ∆pCrpostCPB-H6 (AUCROC = 0.69 [95%CI:0.48-0.85]). CONCLUSIONS: In this hypothesis-generating study mostly testing early detection of mild CS-AKI, there was no evident added value of the tested modern biomarkers over early minimal postoperative changes in pCr: despite the common perioperative hemodilution in the setting of cardiac surgery, if pCr failed to decline within the 6 h after CPB, the development of CS-AKI was likely. Confirmatory studies with more severe forms of CS-AKI are required.

Achieving Expert-Level Interpretation of Serum Protein Electrophoresis through Deep Learning Driven by Human Reasoning.

BACKGROUND: Serum protein electrophoresis (SPE) is a common clinical laboratory test, mainly indicated for the diagnosis and follow-up of monoclonal gammopathies. A time-consuming and potentially subjective human expertise is required for SPE analysis to detect possible pitfalls and to provide a clinically relevant interpretation. METHODS: An expert-annotated SPE dataset of 159 969 entries was used to develop SPECTR (serum protein electrophoresis computer-assisted recognition), a deep learning-based artificial intelligence, which analyzes and interprets raw SPE curves produced by an analytical system into text comments that can be used by practitioners. It was designed following academic recommendations for SPE interpretation, using a transparent architecture avoiding the "black box" effect. SPECTR was validated on an external, independent cohort of 70 362 SPEs and challenged by a panel of 9 independent experts from other hospital centers. RESULTS: SPECTR was able to identify accurately both quantitative abnormalities (r ≥ 0.98 for fractions quantification) and qualitative abnormalities [receiver operating characteristic-area under curve (ROC-AUC) ≥ 0.90 for M-spikes, restricted heterogeneity of immunoglobulins, and beta-gamma bridging]. Furthermore, it showed highly accurate at both detecting (ROC-AUC ≥ 0.99) and quantifying (r = 0.99) M-spikes. It proved highly reproducible and resilient to minor variations and its agreement with human experts was higher (κ = 0.632) than experts between each other (κ = 0.624). CONCLUSIONS: SPECTR is an algorithm based on artificial intelligence suitable to high-throughput SPEs analyses and interpretation. It aims at improving SPE reproducibility and reliability. It is freely available in open access through an online tool providing fully editable validation assistance for SPE.

Tear metabolomics highlights new potential biomarkers for differentiating between Sjögren's syndrome and other causes of dry eye.

PURPOSE: The lacrimal exocrinopathy of primary Sjögren's syndrome (pSS) is one of the main causes of severe dry eye syndrome and a burden for patients. Early recognition and treatment could prevent irreversible damage to lacrimal glands. The aim of this study was to find biomarkers in tears, using metabolomics and data mining approaches, in patients with newly-diagnosed pSS compared to other causes of dry eye syndrome. METHODS: A prospective cohort of 40 pSS and 40 non-pSS Sicca patients with dryness was explored through a standardized targeted metabolomic approach using liquid chromatography coupled with mass spectrometry. A metabolomic signature predictive of the pSS status was sought out using linear (logistic regression with elastic-net regularization) and non-linear (random forests) machine learning architectures, after splitting the studied population into training, validation and test sets. RESULTS: Among the 104 metabolites accurately measured in tears, we identified a discriminant signature composed of nine metabolites (two amino acids: serine, aspartate; one biogenic amine: dopamine; six lipids: Lysophosphatidylcholine C16:1, C18:1, C18:2, sphingomyelin C16:0 and C22:3, and the phoshatidylcholine diacyl PCaa C42:4), with robust performances (ROC-AUC = 0.83) for predicting the pSS status. Adjustment for age, sex and anti-SSA antibodies did not disrupt the link between the metabolomic signature and the pSS status. The non-lipidic components also remained specific for pSS regardless of the dryness severity. CONCLUSION: Our results reveal a metabolomic signature for tears that distinguishes pSS from other dry eye syndromes and further highlight nine key metabolites of potential interest for early diagnosis and therapeutics of pSS.

Persistent elevation of plasma vitamin B12 is strongly associated with solid cancer.

Elevated plasma vitamin B12 has been associated with solid cancers, based on a single B12 measurement. We evaluated the incidence of solid cancers following B12 measurement in patients with persistent elevated B12, compared to patients without elevated B12 and to patients with non-persistent elevated B12. The study population included patients with at least two plasma B12 measurements without already known elevated-B12-related causes. Patients with elevated plasma B12 (≥ 1000 ng/L) at first measurement (n = 344) were matched for age and sex with patients having 2 normal B12 measurements (< 1000 ng/L) (NN group, n = 344). The patients with elevated plasma B12 at first measurement were split into 2 groups, according to the presence (EE group, n = 144) or the absence (EN group, n = 200) of persistent elevated plasma B12 at second measurement. We compared the cancer-free survival during 60 months between the groups after adjustment for the other elevated-B12-related causes in a survival competing risk model. Compared to the NN group, a persistent elevated plasma B12 ≥ 1000 ng/mL was strongly associated with the occurrence of solid cancer (HR 5.90 [95% CI 2.79-12.45], p < 0.001), contrary to non-persistent plasma B12 elevation (p = 0.29). These results could help to select patients in whom the screening for solid cancers would be of interest.

Metabolomic Sexual Dimorphism of the Mouse Brain is Predominantly Abolished by Gonadectomy with a Higher Impact on Females.

The importance of sexual dimorphism of the mouse brain metabolome was recently highlighted, in addition to a high regional specificity found between the frontal cortex, the cerebellum, and the brain stem. To address the origin of this dimorphism, we performed gonadectomy on both sexes, followed by a metabolomic study targeting 188 metabolites in the three brain regions. While sham controls, which underwent the same surgical procedure without gonadectomy, reproduced the regional sexual dimorphism of the metabolome previously identified, no sex difference was identifiable after gonadectomy, through both univariate and multivariate analyses. These experiments also made it possible to identify which sex was responsible for the dimorphism for 35 metabolites. The female sex contributed to the difference for more than 80% of them. Our results show that gonads are the main contributors to the brain sexual dimorphism previously observed, especially in females.

Preliminary Metabolomic Profiling of the Vitreous Humor from Hypothermia Fatalities.

The postmortem diagnosis of hypothermia fatalities is often complex due to the absence of pathognomonic lesions and biomarkers. In this study, potential novel biomarkers of hypothermia fatalities were searched in the vitreous humor of known cases of hypothermia fatalities (n = 20) compared to control cases (n = 16), using a targeted metabolomics approach allowing quantitative detection of 188 metabolites. A robust discriminant model with good predictivity was obtained with the supervised OPLS-DA multivariate analysis, showing a distinct separation between the hypothermia and control groups. This signature was characterized by the decreased concentrations of five metabolites (methionine sulfoxide, tryptophan, phenylalanine, alanine, and ornithine) and the increased concentration of 28 metabolites (21 phosphatidylcholines, 3 sphingomyelins, spermine, citrulline, acetylcarnitine, and hydroxybutyrylcarnitine) in hypothermia fatalities compared to controls. The signature shows similarities with already identified features in serum such as the altered concentrations of tryptophan, acylcarnitines, and unsaturated phosphatidylcholines, revealing a highly significant increased activity of methionine sulfoxide reductase, attested by a low methionine sulfoxide-to-methionine ratio. Our results show a preliminary metabolomics signature of hypothermia fatalities in the vitreous humor, highlighting an increased methionine sulfoxide reductase activity.

Deep learning shows no morphological abnormalities in neutrophils in Alzheimer's disease.

INTRODUCTION: Several studies have provided evidence of the key role of neutrophils in the pathophysiology of Alzheimer's disease (AD). Yet, no study to date has investigated the potential link between AD and morphologically abnormal neutrophils on blood smears. METHODS: Due to the complexity and subjectivity of the task by human analysis, deep learning models were trained to predict AD from neutrophil images. Control models were trained for a known feasible task (leukocyte subtype classification) and for detecting potential biases of overfitting (patient prediction). RESULTS: Deep learning models achieved state-of-the-art results for leukocyte subtype classification but could not accurately predict AD. DISCUSSION: We found no evidence of morphological abnormalities of neutrophils in AD. Our results show that a solid deep learning pipeline with positive and bias control models with visualization techniques are helpful to support deep learning model results.