Enhancing diagnostic accuracy of multiple myeloma through ML-driven analysis of hematological slides: new dataset and identification model to support hematologists.

Multiple Myeloma (MM) is a hematological malignancy characterized by the clonal proliferation of plasma cells within the bone marrow. Diagnosing MM presents considerable challenges, involving the identification of plasma cells in cytology examinations on hematological slides. At present, this is still a time-consuming manual task and has high labor costs. These challenges have adverse implications, which rely heavily on medical professionals' expertise and experience. To tackle these challenges, we present an investigation using Artificial Intelligence, specifically a Machine Learning analysis of hematological slides with a Deep Neural Network (DNN), to support specialists during the process of diagnosing MM. In this sense, the contribution of this study is twofold: in addition to the trained model to diagnose MM, we also make available to the community a fully-curated hematological slide dataset with thousands of images of plasma cells. Taken together, the setup we established here is a framework that researchers and hospitals with limited resources can promptly use. Our contributions provide practical results that have been directly applied in the public health system in Brazil. Given the open-source nature of the project, we anticipate it will be used and extended to diagnose other malignancies.

An Evaluation of Fit Indices Used in Model Selection of Dichotomous Mixture IRT Models.

A Monte Carlo simulation study was conducted to compare fit indices used for detecting the correct latent class in three dichotomous mixture item response theory (IRT) models. Ten indices were considered: Akaike's information criterion (AIC), the corrected AIC (AICc), Bayesian information criterion (BIC), consistent AIC (CAIC), Draper's information criterion (DIC), sample size adjusted BIC (SABIC), relative entropy, the integrated classification likelihood criterion (ICL-BIC), the adjusted Lo-Mendell-Rubin (LMR), and Vuong-Lo-Mendell-Rubin (VLMR). The accuracy of the fit indices was assessed for correct detection of the number of latent classes for different simulation conditions including sample size (2,500 and 5,000), test length (15, 30, and 45), mixture proportions (equal and unequal), number of latent classes (2, 3, and 4), and latent class separation (no-separation and small separation). Simulation study results indicated that as the number of examinees or number of items increased, correct identification rates also increased for most of the indices. Correct identification rates by the different fit indices, however, decreased as the number of estimated latent classes or parameters (i.e., model complexity) increased. Results were good for BIC, CAIC, DIC, SABIC, ICL-BIC, LMR, and VLMR, and the relative entropy index tended to select correct models most of the time. Consistent with previous studies, AIC and AICc showed poor performance. Most of these indices had limited utility for three-class and four-class mixture 3PL model conditions.

Association of cardiac biomarkers with long-term cardiovascular events in a community cohort.

MATERIALS AND METHODS: The study assessed major adverse cardiac events (MACE) (myocardial infarction, coronary artery bypass graft, percutaneous intervention, stroke, and death. Cox proportional hazards models assessed apolipoprotein AI (ApoA1), apolipoprotein B (ApoB), ceramide score, cystatin C, galectin-3 (Gal3), LDL-C, Non-HDL-C, total cholesterol (TC), N-terminal B-type natriuretic peptide (NT proBNP), high-sensitivity cardiac troponin (HscTnI) and soluble interleukin 1 receptor-like 1. In adjusted models, Ceramide score was defined by from N-palmitoyl-sphingosine [Cer(16:0)], N-stearoyl-sphingosine [Cer(18:0)], N-nervonoyl-sphingosine [Cer(24:1)] and N-lignoceroyl-sphingosine [Cer(24:0)]. Multi-biomarker models were compared with C-statistics and Integrated Discrimination Index (IDI). RESULTS: A total of 1131 patients were included. Adjusted NT proBNP per 1 SD resulted in a 31% increased risk of MACE/death (HR = 1.31) and a 31% increased risk for stroke/MI (HR = 1.31). Adjusted Ceramide per 1 SD showed a 13% increased risk of MACE/death (HR = 1.13) and a 29% increased risk for stroke/MI (HR = 1.29). These markers added to clinical factors for both MACE/death (p = 0.003) and stroke/MI (p = 0.034). HscTnI was not a predictor of outcomes when added to the models. DISCUSSION: Ceramide score and NT proBNP improve the prediction of MACE and stroke/MI in a community primary prevention cohort.

In a community cohort, where a wide range of biomarkers were evaluated, Ceramide score provided additive value over traditional cardiac risk factors alone for predicting stroke/MI. NT ProBNP provided additive value in prediction of MACE/death. Other biomarkers failed to improve the discrimination of these models.

High Baseline High-Sensitivity Cardiac Troponin T Concentrations and Risk of Index Acute Myocardial Infarction.

OBJECTIVE: To evaluate the diagnostic performance of the previously recommended baseline high-sensitivity cardiac troponin T (hs-cTnT) thresholds of 52 and 100 ng/L in identifying patients at high risk of acute myocardial infarction (AMI). PATIENTS AND METHODS: This study compared the positive predictive value (PPV) for index AMI of these high-risk hs-cTnT thresholds in adult patients in the emergency department undergoing hs-cTnT measurement. RESULTS: The adjudicated MAyo Southwest Wisconsin 5th Gen Troponin T ImplementatiON cohort included 2053 patients, with 157 (7.6%) who received a diagnosis of AMI. The hs-cTnT concentrations of greater than 52 and greater than 100 ng/L resulted in PPVs of 41% (95% CI, 35%-48%) and 57% (95% CI, 48%-66%). In patients with chest discomfort, hs-cTnT concentrations greater than 52 ng/L resulted in a PPV of 66% (95% CI, 56%-76%) and hs-cTnT concentrations greater than 100 ng/L resulted in a PPV of 77% (95% CI, 65%-87%). The CV Data Mart Biomarker cohort included 143,709 patients, and 3003 (2.1%) received a diagnosis of AMI. Baseline hs-cTnT concentrations greater than 52 and greater than 100 ng/L resulted in PPVs of 12% (95% CI, 11%-12%) and 17% (95% CI, 17%-19%), respectively. In patients with chest pain and hs-cTnT concentrations greater than 52 ng/L, the PPV for MI was 17% (95% CI, 15%-18%) and in those with concentrations greater than 100 ng/L, only 22% (95% CI, 19%-25%). CONCLUSION: In unselected patients undergoing hs-cTnT measurement, the hs-cTnT thresholds of greater than 52 and greater than 100 ng/L provide suboptimal performance for identifying high-risk patients. In patients with chest discomfort, an hs-cTnT concentration of greater than 100 ng/L, but not the European Society of Cardiology-recommended threshold of greater than 52 ng/L, provides an acceptable performance but should be used only with other clinical features.

Association of virtual end-of-life care with healthcare outcomes before and during the COVID-19 pandemic: A population-based study.

The use of virtual care for people at the end-of-life significantly increased during the COVID-19 pandemic, but its association with acute healthcare use and location of death is unknown. The objective of this study was to measure the association between the use of virtual end-of-life care with acute healthcare use and an out-of-hospital death before vs. after the introduction of specialized fee codes that enabled broader delivery of virtual care during the COVID-19 pandemic. This was a population-based cohort study of 323,995 adults in their last 90 days of life between January 25, 2018 and December 31, 2021 using health administrative data in Ontario, Canada. Primary outcomes were acute healthcare use (emergency department, hospitalization) and location of death (in or out-of-hospital). Prior to March 14, 2020, 13,974 (8%) people received at least 1 virtual end-of-life care visit, which was associated with a 16% higher rate of emergency department use (adjusted Rate Ratio [aRR] 1.16, 95%CI 1.12 to 1.20), a 17% higher rate of hospitalization (aRR 1.17, 95%CI 1.15 to 1.20), and a 34% higher risk of an out-of-hospital death (aRR 1.34, 95%CI 1.31 to 1.37) compared to people who did not receive virtual end-of-life care. After March 14, 2020, 104,165 (71%) people received at least 1 virtual end-of-life care visit, which was associated with a 58% higher rate of an emergency department visit (aRR 1.58, 95%CI 1.54 to 1.62), a 45% higher rate of hospitalization (aRR 1.45, 95%CI 1.42 to 1.47), and a 65% higher risk of an out-of-hospital death (aRR 1.65, 95%CI 1.61 to 1.69) compared to people who did not receive virtual end-of-life care. The use of virtual end-of-life care was associated with higher acute healthcare use in the last 90 days of life and a higher likelihood of dying out-of-hospital, and these rates increased during the pandemic.

Attending physicians' annual service volume and use of virtual end-of-life care: A population-based cohort study in Ontario, Canada.

IMPORTANCE: Physicians and their practice behaviors influence access to healthcare and may represent potentially modifiable targets for practice-changing interventions. Use of virtual care at the end-of-life significantly increased during the COVID-19 pandemic, but its association with physician practice behaviors, (e.g., annual service volume) is unknown. OBJECTIVE: Measure the association of physicians' annual service volume with their use of virtual end-of-life care (EOLC) and the magnitude of physician-attributable variation in its use, before and during the pandemic. DESIGN, SETTING AND PARTICIPANTS: Population-based cohort study using administrative data of all physicians in Ontario, Canada who cared for adults in the last 90 days of life between 01/25/2018-12/31/2021. Multivariable modified Poisson regression models measured the association between attending physicians' use of virtual EOLC and their annual service volume. We calculated the variance partition coefficients for each regression and stratified by time period before and during the pandemic. EXPOSURE: Annual service volume of a person's attending physician in the preceding year. MAIN OUTCOMES AND MEASURES: Delivery of ≥1 virtual EOLC visit by a person's attending physician and the proportion of variation in its use attributable to physicians. RESULTS: Among the 35,825 unique attending physicians caring for 315,494 adults, use of virtual EOLC was associated with receiving care from a high compared to low service volume attending physician; the magnitude of this association diminished during the pandemic (adjusted RR 1.25 [95% CI 1.14, 1.37] pre-pandemic;1.10 (95% CI 1.08, 1.12) during the pandemic). Physicians accounted for 36% of the variation in virtual EOLC use pre-pandemic and 12% of this variation during the pandemic. CONCLUSIONS AND RELEVANCE: Physicians' annual service volume was associated with use of virtual EOLC and physicians accounted for a substantial proportion of the variation in its use. Physicians may be appropriate and potentially modifiable targets for interventions to modulate use of EOLC delivery.

Raman spectroscopy and one-dimensional convolutional neural network modeling as a real-time monitoring tool for in vitro transaminase-catalyzed synthesis of a pharmaceutically relevant amine precursor.

Raman spectroscopy has been used to measure the concentration of a pharmaceutically relevant model amine intermediate for positive allosteric modulators of nicotinic acetylcholine receptor in a ω-transaminase-catalyzed conversion. A model based on a one-dimensional convolutional neural network was developed to translate raw data augmented Raman spectra directly into substrate concentrations, with which the conversion from ketone to amine by ω-transaminase could be determined over time. The model showed very good predictive capabilities, with R2 values higher than 0.99 for the spectra included in the modeling and 0.964 for an independent dataset. However, the model could not extrapolate outside the concentrations specified by the model. The presented work shows the potential of Raman spectroscopy as a real-time monitoring tool for biocatalytic reactions.

Effects of soil on the development, survival, and oviposition of Culex quinquefasciatus (Diptera: Culicidae) mosquitoes.

BACKGROUND: Water quality is known to influence the development and survival of larval mosquitoes, which affects mosquito-borne pathogen transmission as a function of the number of mosquitoes that reach adulthood and blood feed. Although water properties are known to affect mosquito development, few studies have investigated the link among soil properties, water quality, and mosquito development. Given the large number of ground-breeding mosquito species, this linkage is a potentially important factor to consider in mosquito ecology. In this study, we explored the effects of different soils on multiple life history parameters of the ground-breeding mosquito species Culex quinquefasciatus (Diptera: Culicidae). METHODS: Cx. quinquefasciatus larvae were reared in water combined with different soil substrates (sandy, silt, or clay loam textures) at increasing soil to water volume ratios, with and without the addition of organic matter (fish food). Gravid mosquitoes were offered different soil-water extracts to investigate soil effects on oviposition preference. RESULTS: Without the addition of organic matter, larval survival and development differed significantly among waters with different soil textures and volumes of substrate. Mosquitoes in water with clay loam soil survived longer and developed further than mosquitoes in other soil waters. Larvae survived for longer periods of time with increased volumes of soil substrate. Adding organic matter reduced the differences in larval survival time, development, and pupation among soil-water extracts. Adult female mosquitoes oviposited more frequently in water with clay loam soil, but the addition of organic matter reduced the soil effects on oviposition preference. CONCLUSIONS: This study suggests soil composition affects larval mosquito survival and development, as well as the oviposition preference of gravid females. Future studies could differentiate abiotic and biotic soil features that affect mosquitoes and incorporate soil variation at the landscape scale into models to predict mosquito population dynamics and mosquito-borne pathogen transmission.

Cardiac Troponin in Patients With Light Chain and Transthyretin Cardiac Amyloidosis: JACC: CardioOncology State-of-the-Art Review.

Cardiac amyloidosis (CA) is an infiltrative disease caused by amyloid fibril deposition in the myocardium; the 2 forms that most frequently involve the heart are amyloid light chain (AL) and amyloid transthyretin (ATTR) amyloidosis. Cardiac troponin (cTn) is the biomarker of choice for the detection of myocardial injury and is frequently found to be elevated in patients with CA, particularly with high-sensitivity assays. Multiple mechanisms of myocardial injury in CA have been proposed, including cytotoxic effect of amyloid precursors, interstitial amyloid fibril infiltration, coronary microvascular dysfunction, amyloid- and non-amyloid-related coronary artery disease, diastolic dysfunction, and heart failure. Regardless of the mechanisms, cTn values have relevant prognostic (and potentially diagnostic) implications in both AL and ATTR amyloidosis. In this review, the authors discuss the significant aspects of cTn biology and measurement methods, potential mechanisms of myocardial injury in CA, and the clinical application of cTn in the management of both AL and ATTR amyloidosis.

Imprecision of high-sensitivity cardiac troponin assays at the female 99th-percentile.

BACKGROUND: An analytical benchmark for high-sensitivity cardiac troponin (hs-cTn) assays is to achieve a coefficient of variation (CV) of ≤ 10.0 % at the 99th percentile upper reference limit (URL) used for the diagnosis of myocardial infarction. Few prospective multicenter studies have evaluated assay imprecision and none have determined precision at the female URL which is lower than the male URL for all cardiac troponin assays. METHODS: Human serum and plasma matrix samples were constructed to yield hs-cTn concentrations near the female URLs for the Abbott, Beckman, Roche, and Siemens hs-cTn assays. These materials were sent (on dry ice) to 35 Canadian hospital laboratories (n = 64 instruments evaluated) participating in a larger clinical trial, with instructions for storage, handling, and monthly testing over one year. The mean concentration, standard deviation, and CV for each instrument type and an overall pooled CV for each manufacturer were calculated. RESULTS: The CVs for all individual instruments and overall were ≤ 10.0 % for two manufacturers (Abbott CVpooled = 6.3 % and Beckman CVpooled = 7.0 %). One of four Siemens Atellica instruments yielded a CV > 10.0 % (CVpooled = 7.7 %), whereas 15 of 41 Roche instruments yielded CVs > 10.0 % at the female URL of 9 ng/L used worldwide (6 cobas e411, 1 cobas e601, 4 cobas e602, and 4 cobas e801) (CVpooled = 11.7 %). Four Roche instruments also yielded CVs > 10.0 % near the female URL of 14 ng/L used in the United States (CVpooled = 8.5 %). CONCLUSIONS: The number of instruments achieving a CV ≤ 10.0 % at the female 99th-percentile URL varies by manufacturer and by instrument. Monitoring assay precision at the female URL is necessary for some assays to ensure optimal use of this threshold in clinical practice.

Serial high sensitivity troponin sampling in patients with SARS-CoV-2 infection.

INTRODUCTION: Multiple studies have investigated the role of cardiac troponin (cTn) in the risk stratification of patients with COVID-19. Most of these investigations are based on cTn values at presentation and do not consider the prognostic significance of cTn changes over time. This study aimed to investigate the prognostic role of serial cTn measurements in patients hospitalized with COVID-19 with samples that were not obtained for clinical indications. METHODS: Patients hospitalized between April 2020 and March 2021 with PCR-confirmed SARS-CoV-2 infection were evaluated. Blood samples collected for any reason were stored for subsequent analysis. If clinical high sensitivity hs-cTnT (Roche) was not measured, samples were tested separately in batches. Hs-cTnI (Abbott) was also evaluated. RESULTS: There were 228 unique patients. There were 21 (9.2 %) deaths. No patient with a low hs-cTnT (<6 ng/L) died and 1 patient with low hs-cTnI (<5 ng/L) died. Myocardial injury was associated with higher odds of death, when defined by hs-cTnT (OR: 7.88, 95 % CI: 2.04-30.40, p = 0.003) or hs-cTnI (OR: 7.46, 95 % CI: 2.68-20.77, p < 0.001). This association remained after propensity weighting. An increasing pattern was associated with higher odds of death compared to a stable pattern for hs-cTnT (OR: 5.45, 95 % CI: 1.81-16.40, p = 0.003) and hs-cTnI (OR: 4.49, 95 % CI: 1.02-19.81, p = 0.048). Among patients with myocardial injury defined by hs-cTnT, an increasing pattern was associated with higher odds of death compared to a decreasing pattern (OR: 4.80, 95 % CI: 1.16-19.97, p = 0.031). CONCLUSIONS: Patients hospitalized with COVID-19 with myocardial injury have higher odds of death. Serial hs-cTn testing provides additional risk stratification in these patients.

Real-Time Tracking of Carbon Dioxide Concentration Using an Optical Microsphere Resonator Sensor.

Whispering gallery mode resonator sensors are nondisruptive optical sensors that can detect and monitor perturbations in a gaseous environment. Through its resonant properties of peak wavelength, amplitude, and quality factor (Q factor), changes in concentration can be quantified within seconds and monitored over days with great stability. In addition, the small footprint, low cost, and high sensitivity are ideal properties for a disposable sensor that can be utilized in extreme environments. The large Q factor of the resonant cavity enables long interaction lengths and amplifies the effect of small changes in the background refractive index, which is detectable in picometer shifts of the resonance wavelength. However, this measurement is susceptible to changes in other environmental factors such as temperature, pressure, and humidity, which manifest on the picometer wavelength scale, reinforcing the need to decouple the variables. In this work, we compare the spectral response of different diameter resonators to carbon dioxide, nitrogen, and its mixtures, observing the spectral shifting and broadening of the cavity resonance near 1550 nm. In addition, the effect of environmental temperature on spectral shifting due to the thermo-optic effect is characterized and quantified. Lastly, the gas concentrations are changed in real time to showcase the tracking and recovery capabilities of the resonator sensor.

An improved method for estimating low LDL-C based on the enhanced Sampson-NIH equation.

BACKGROUND: The accurate measurement of Low-density lipoprotein cholesterol (LDL-C) is critical in the decision to utilize the new lipid-lowering therapies like PCSK9-inhibitors (PCSK9i) for high-risk cardiovascular disease patients that do not achieve sufficiently low LDL-C on statin therapy. OBJECTIVE: To improve the estimation of low LDL-C by developing a new equation that includes apolipoprotein B (apoB) as an independent variable, along with the standard lipid panel test results. METHODS: Using ß-quantification (BQ) as the reference method, which was performed on a large dyslipidemic population (N = 24,406), the following enhanced Sampson-NIH equation (eS LDL-C) was developed by least-square regression analysis: [Formula: see text] RESULTS: The eS LDL-C equation was the most accurate equation for a broad range of LDL-C values based on regression related parameters and the mean absolute difference (mg/dL) from the BQ reference method (eS LDL-C: 4.51, Sampson-NIH equation [S LDL-C]: 6.07; extended Martin equation [eM LDL-C]: 6.64; Friedewald equation [F LDL-C]: 8.3). It also had the best area-under-the-curve accuracy score by Regression Error Characteristic plots for LDL-C < 100 mg/dL (eS LDL-C: 0.953; S LDL-C: 0.920; eM LDL-C: 0.915; F LDL-C: 0.874) and was the best equation for categorizing patients as being below or above the 70 mg/dL LDL-C treatment threshold for adding new lipid-lowering drugs by kappa score analysis when compared to BQ LDL-C for TG < 800 mg/dL (eS LDL-C: 0.870 (0.853-0.887); S LDL-C:0.763 (0.749-0.776); eM LDL-C:0.706 (0.690-0.722); F LDL-C:0.687 (0.672-0.701). Approximately a third of patients with an F LDL-C < 70 mg/dL had falsely low test results, but about 80% were correctly reclassified as higher (≥ 70 mg/dL) by the eS LDL-C equation, making them potentially eligible for PCSK9i treatment. The M LDL-C and S LDL-C equations had less false low results below 70 mg/dL than the F LDL-C equation but reclassification by the eS LDL-C equation still also increased the net number of patients correctly classified. CONCLUSIONS: The use of the eS LDL-C equation as a confirmatory test improves the identification of high-risk cardiovascular disease patients, who could benefit from new lipid-lowering therapies but have falsely low LDL-C, as determined by the standard LDL-C equations used in current practice.

Determination and correction of spectral phase from principal component analysis of coherent phonons.

Measuring the spectral phase of a pulse is key for performing wavelength resolved ultrafast measurements in the few femtosecond regime. However, accurate measurements in real experimental conditions can be challenging. We show that the reflectivity change induced by coherent phonons in a quantum material can be used to infer the spectral phase of an optical probe pulse with few-femtosecond accuracy.

Development and application of a second-generation multilingual tool for invasion risk screening of non-native terrestrial plants.

Under the increasing threat to native ecosystems posed by non-native species invasions, there is an urgent need for decision support tools that can more effectively identify non-native species likely to become invasive. As part of the screening (first step) component in non-native species risk analysis, decision support tools have been developed for aquatic and terrestrial organisms. Amongst these tools is the Weed Risk Assessment (WRA) for screening non-native plants. The WRA has provided the foundations for developing the first-generation WRA-type Invasiveness Screening Kit (ISK) tools applicable to a range of aquatic species, and more recently for the second-generation ISK tools applicable to all aquatic organisms (including plants) and terrestrial animals. Given the most extensive usage of the latter toolkits, this study describes the development and application of the Terrestrial Plant Species Invasiveness Screening Kit (TPS-ISK). As a second-generation ISK tool, the TPS-ISK is a multilingual turnkey application that provides several advantages relative to the WRA: (i) compliance with the minimum standards against which a protocol should be evaluated for invasion process and management approaches; (ii) enhanced questionnaire comprehensiveness including a climate change component; (iii) provision of a level of confidence; (iv) error-free computation of risk scores; (v) multilingual support; (vi) possibility for across-study comparisons of screening outcomes; (vii) a powerful graphical user interface; (viii) seamless software deployment and accessibility with improved data exchange. The TPS-ISK successfully risk-ranked five representative sample species for the main taxonomic groups supported by the tool and ten angiosperms previously screened with the WRA for Turkey. The almost 20-year continuous development and evolution of the ISK tools, as opposed to the WRA, closely meet the increasing demand by scientists and decision-makers for a reliable, comprehensive, updatable and easily deployable decision support tool. For terrestrial plant screening, these requirements are therefore met by the newly developed TPS-ISK.