Using artificial intelligence in a primary care setting to identify patients at risk for cancer: a risk prediction model based on routine laboratory tests.

OBJECTIVES: To evaluate the ability of an artificial intelligence (AI) model to predict the risk of cancer in patients referred from primary care based on routine blood tests. Results obtained with the AI model are compared to results based on logistic regression (LR). METHODS: An analytical profile consisting of 25 predefined routine laboratory blood tests was introduced to general practitioners (GPs) to be used for patients with non-specific symptoms, as an additional tool to identify individuals at increased risk of cancer. Consecutive analytical profiles ordered by GPs from November 29th 2011 until March 1st 2020 were included. AI and LR analysis were performed on data from 6,592 analytical profiles for their ability to detect cancer. Cohort I for model development included 5,224 analytical profiles ordered by GP's from November 29th 2011 until the December 31st 2018, while 1,368 analytical profiles included from January 1st 2019 until March 1st 2020 constituted the "out of time" validation test Cohort II. The main outcome measure was a cancer diagnosis within 90 days. RESULTS: The AI model based on routine laboratory blood tests can provide an easy-to use risk score to predict cancer within 90 days. Results obtained with the AI model were comparable to results from the LR model. In the internal validation Cohort IB, the AI model provided slightly better results than the LR analysis both in terms of the area under the receiver operating characteristics curve (AUC) and PPV, sensitivity/specificity while in the "out of time" validation test Cohort II, the obtained results were comparable. CONCLUSIONS: The AI risk score may be a valuable tool in the clinical decision-making. The score should be further validated to determine its applicability in other populations.

Quantifying SARS-CoV-2 nucleocapsid antigen in oropharyngeal swabs using single molecule array technology.

This study aimed to develop a highly sensitive SARS-CoV-2 nucleocapsid antigen assay using the single molecule array (Simoa) technology and compare it with real time RT-PCR as used in routine clinical practice with the ambition to achieve a comparative technical and clinical sensitivity. Samples were available from 148 SARS-CoV-2 real time RT-PCR positive and 73 SARS-CoV-2 real time RT-PCR negative oropharyngeal swabs. For determination of technical sensitivity SARS-CoV-2 virus culture material was used. The samples were treated with lysis buffer and analyzed using both an in-house and a pre-commercial SARS-CoV-2 nucleocapsid antigen assay on Simoa. Both nucleocapsid antigen assays have a technical sensitivity corresponding to around 100 SARS-CoV-2 RNA molecules/mL. Using a cut-off at 0.1 pg/mL the pre-commercial SARS-CoV-2 nucleocapsid antigen assay had a sensitivity of 96% (95% CI 91.4-98.5%) and specificity of 100% (95% CI 95.1-100%). In comparison the in-house nucleocapsid antigen assay had sensitivity of 95% (95% CI 89.3-98.1%) and a specificity of 100% (95% CI 95.1-100%) using a cut-off at 0.01 pg/mL. The two SARS-CoV-2 nucleocapsid antigen assays correlated with r = 0.91 (P < 0.0001). The in-house and the pre-commercial SARS-CoV-2 nucleocapsid antigen assay demonstrated technical and clinical sensitivity comparable to real-time RT-PCR methods for identifying SARS-CoV-2 infected patients and thus can be used clinically as well as serve as a reference method for antigen Point of Care Testing.

Daily monitoring of viral load measured as SARS-CoV-2 antigen and RNA in blood, IL-6, CRP and complement C3d predicts outcome in patients hospitalized with COVID-19.

OBJECTIVES: We hypothesized that the amount of antigen produced in the body during a COVID-19 infection might differ between patients, and that maximum concentrations would predict the degree of both inflammation and outcome for patients. METHODS: Eighty-four hospitalized and SARS-CoV-2 PCR swab-positive patients, were followed with blood sampling every day until discharge or death. A total of 444 serial EDTA plasma samples were analyzed for a range of biomarkers: SARS-CoV-2 nuclear antigen and RNA concentration, complement activation as well as several inflammatory markers, and KL-6 as a lung marker. The patients were divided into outcome groups depending on need of respiratory support and death/survival. RESULTS: Circulating SARS-CoV-2 nuclear antigen levels were above the detection limit in blood in 65 out of 84 COVID-19 PCR swab-positive patients on day one of hospitalization, as was viral RNA in plasma in 30 out of 84. In all patients, complete antigen clearance was observed within 24 days. There were definite statistically significant differences between the groups depending on their biomarkers, showing that the concentrations of virus RNA and antigen were correlated to the inflammatory biomarker levels, respiratory treatment and death. CONCLUSIONS: Viral antigen is cleared in parallel with the virus RNA levels. The levels of antigens and SARS-CoV-2 RNA in the blood correlates with the level of IL-6, inflammation, respiratory failure and death. We propose that the antigens levels together with RNA in blood can be used to predict the severity of disease, outcome, and the clearance of the virus from the body.

Rhodopsin in plasma from patients with diabetic retinopathy - development and validation of digital ELISA by Single Molecule Array (Simoa) technology.

BACKGROUND: Diabetic retinopathy (DR) is the most frequent cause of blindness among younger adults in the western world. No blood biomarkers exist to detect DR. Hypothetically, Rhodopsin concentrations in blood has been suggested as an early marker for retinal damage. The aim of this study was therefore to develop and validate a Rhodopsin assay by employing digital ELISA technology, and to investigate whether Rhodopsin concentrations in diabetes patients with DR are elevated compared with diabetes patients without DR. METHODS: A digital ELISA assay using a Simoa HD-1 Analyzer (Quanterix©, Lexington, MA 02421, USA) was developed and validated and applied on a cohort of diabetes patients characterised with (n=466) and without (n=144) DR. RESULTS: The Rhodopsin assay demonstrated a LOD of 0.26ng/l, a LLOQ of 3ng/l and a linear measuring range from 3 to 2500ng/l. Total CV% was 32%, 23%, 19% and 17% respectively at the following Rhodopsin concentrations: 1, 3, 5 and 13ng/l. Recovery was 17%, 34%, 51% and 55% respectively at Rhodopsin concentrations of 2, 10, 50 and 250ng/l. There was no statistically significant difference in the plasma concentration of Rhodopsin between the diabetes patients with or without DR, but significantly increased number of DR patients having concentrations above the LOD. CONCLUSION: We developed and validated a digital ELISA method for quantification of Rhodopsin in plasma but found no statistically significant difference in the plasma concentration of Rhodopsin between diabetes patients with DR compared to diabetes patients without DR, though significantly more DR patients had values above the LOD.

Vejle Diabetes Biobank - a resource for studies of the etiologies of diabetes and its comorbidities.

AIMS: Carefully designed and established biobanks are considered one of the most essential resources to foster biomedical research as they provide cost-effective and rapid access to a vast variety of biological materials and related anthropometrics allowing for testing of various biomarkers as well as numerous original and pertinent bioclinical hypotheses related to human disease etiology and prognosis. The objective of the present study was to present the baseline data, design, and methods used for the establishment of the Vejle Diabetes Biobank. Further aims included assessment of the prevalence of diabetes and quality of diabetes treatment in a specified Danish region. METHODS: The Vejle Diabetes Biobank was established from 2007 to 2010 as a regional Biobank containing blood, DNA, and urine samples from patients with diabetes and a gender- and age-matched control population aged 25-75 years. Anthropometrics were obtained by physical examination, questionnaires, and interviews at the time of inclusion into the Biobank. The cohort was linked to the Danish Civil Registration System, the Danish National Patient Registry, and the Danish National Prescription Registry. RESULTS: In total, 4,255 nondiabetic individuals and 3,320 patients with diabetes were included. Type 2 diabetes (T2D) patients had a higher body mass index (30 kg/m2) than type 1 diabetes (T1D) patients (25 and 26 kg/m2 in women and men, respectively) and control subjects (25 and 27 kg/m2 in women and men, respectively). Fasting levels of plasma triglycerides and blood pressure were higher in T2D patients (1.5 mmol/L and 148/85 mmHg, respectively) compared with T1D patients (0.9 mmol/L and 139/81 mmHg, respectively), whereas glycated hemoglobin (HbA1c), plasma high density lipoprotein, low density lipoprotein, and total cholesterol were lower in T2D patients (51 mmol/mol, 1.2 mmol/L, 2.2 mmol/L, and 4.2 mmol/L, respectively) compared with findings in T1D patients (61 mmol/mol, 1.6 mmol/L, 2.3 mmol/L, and 4.4 mmol/L, respectively). At the time of inclusion into the Biobank, 56% of the T2D patients and 25% of T1D patients had an HbA1c <7% (53 mmol/mol). Only 28% and 34% of the T2D patients, respectively, reached treatment target for blood pressure and lipids. CONCLUSION: The Vejle Diabetes Biobank represents one of the largest open diabetes case-control cohorts in Denmark. The Biobank invites collaborative investigations of diabetes and diabetes complication etiologies as well as studies of prognostic or predictive biomarkers.

Standardised Resting Time Prior to Blood Sampling and Diurnal Variation Associated with Risk of Patient Misclassification: Results from Selected Biochemical Components.

BACKGROUND: According to current recommendations, blood samples should be taken in the morning after 15 minutes' resting time. Some components exhibit diurnal variation and in response to pressures to expand opening hours and reduce waiting time, the aims of this study were to investigate the impact of resting time prior to blood sampling and diurnal variation on biochemical components, including albumin, thyrotropin (TSH), total calcium and sodium in plasma. METHODS: All patients referred to an outpatient clinic for blood sampling were included in the period Nov 2011 until June 2014 (opening hours: 7am-3pm). Each patient's arrival time and time of blood sampling were registered. The impact of resting time and the time of day for all components was analysed using simple linear regression. The "maximum allowable bias" was used as quality indicator for the change in reference interval. RESULTS: Significant diurnal variation was found for albumin (n = 15,544; p<2×10-16), TSH (n = 20,019; p<2×10-16), calcium (n = 13,588; p = 2.8×10-12) and sodium (n = 51,917; p<2×10-16). Further significant influence for resting time was found for albumin (p = 2.6×10-4), TSH (p = 0.004), calcium (p = 8.9×10-7) and sodium (p = 8.7×10-16). Only TSH and albumin were clinically significantly influenced by diurnal variation. Resting time had no clinically significant effect. CONCLUSIONS: We found no need for resting 15 minutes prior to blood sampling. However, diurnal variation was found to have a significant and considerable impact on TSH and, to a minor degree, albumin. This has to be taken into account to ensure that reference intervals provided by the laboratory are valid on a 24-hour basis.

Circulating microRNAs as biomarkers of adult Crohn's disease.

OBJECTIVE: Previous studies have found a differential expression of microRNAs (miRNAs) in the blood of patients with Crohn's disease (CD) compared with healthy controls. The aim of this study was to identify circulating miRNAs expressed in CD and assess their performance as biomarkers in patients with clinically suspected or known CD. METHODS: The study consisted of two parts: (a) miRNA profiling: The miRNA expression pattern was examined in six patients with CD and six controls using OpenArray miRNA profiling, and the best miRNAs were selected according to their P-value. (b) Validation cohort: In a well-characterized cohort of 102 patients with suspected or known CD, miRNAs identified by miRNA profiling or selected from previously published studies (hsa-miR-16, hsa-miR-21, hsa-miR-106a, and hsa-miR-140-3p) were measured in plasma using reverse transcription PCR. RESULTS: miRNA profiling: hsa-miR-369-3p, hsa-miR-376a, hsa-miR-376, hsa-miR-411#, hsa-miR-411, and mmu-miR-379 were downregulated in CD patients compared with the controls; hsa-miR-200c, hsa-miR-181-2#, and hsa-miR-125a-5p were upregulated (P<0.05). Validation cohort: Only hsa-miR-16 was significantly downregulated in patients with CD compared with patients without CD (fold change 0.83, P=0.02). Receiver operating characteristic analyses showed an area under the curve of 0.65. miRNAs could not discriminate inflammatory from stricturing CD or small bowel CD from CD involving the colon. CONCLUSION: In a clinically relevant cohort of patients, miRNAs in plasma identified in the present and previous studies were inadequate biomarkers for the diagnosis of CD.

Diabetes mortality differs between registers due to various disease definitions.

INTRODUCTION: We evaluated the impact of including haemolobbing A1c (HbA1c) measurements in a regional algorithm for identification of diabetics by comparing the population identified by the regional algorithm with diabetics registered in the National Danish Diabetes Register (NDR) relative to prevalence, co-morbidity and five-year mortality rate. MATERIAL AND METHODS: The regional (County of Vejle) and national diabetes populations were compared per the inclusion date of 31 December 2006 limited to persons residing in four municipalities in the County of Vejle, Denmark. RESULTS: A total of 14,998 diabetics were identified by the regional algorithm, of whom 11,499 (prevalence 4.1%) resided in the four municipalities. The total number of diabetics registered in the NDR was 227,621 in Denmark, of whom 10,976 (prevalence 4.0%) resided in the four municipalities. The regional diabetics (2,802 persons) not identified in the NDR population had a significantly lower mortality rate (57%) than the diabetics (2,279 persons) in the NDR population not identified by the regional algorithm. CONCLUSION: The significantly higher mortality in the NDR population not identified by the regional algorithm may stem from differences between the components of the two algorithms, i.e. frequency of glucose measurements in the NDR versus frequency of HbA1c measurements including elevated values in the regional algorithm. The NDR algorithm, which includes the use of frequency of glucose measurements without a value over the diagnostic threshold, identified about 21% of persons who probably had their glucose measured for other reasons than diabetes. FUNDING: The Danish Council for Research and Innovation, Region of Southern Denmark's PhD Fund, University of Southern Denmark and the Research Fund, Vejle Hospital. TRIAL REGISTRATION: not relevant.

Changing from glucose to HbA1c for diabetes diagnosis: predictive values of one test and importance of analytical bias and imprecision.

BACKGROUND: In Denmark, the use of HbA1c in the diagnosis of diabetes was adopted from March 2012. We evaluated the change in the number of diabetes cases diagnosed by haemoglobin A1c (HbA1c) versus fasting venous plasma glucose (FPG), and estimated the influence of analytical variation and bias on the HbA1c-based prevalence of diabetes. METHODS: The study population constituted 4239 individuals not known to have diabetes randomly selected from all inhabitants aged 25-75 years in the former County of Vejle, Denmark. The number of undiagnosed patients with diabetes in the study population using FPG or HbA1c as the diagnostic criterion was estimated. Furthermore, changes in the analytical bias and coefficient of variation (CV) for HbA1c analysis were simulated and the effect on the number of diabetes cases was observed. RESULTS: Changing the diagnostic test from FPG to HbA1c reduced the number of patients with diabetes by approximately 46% based on one measurement. The predictive value of one test of HbA1c was 91% versus only 66% for one test of FPG. Analytical variation had a much greater impact on the number of patients with diabetes than bias. At a bias of 0%, an increase of CVanalytical from 2.7% to 3.7% increased the number of diabetes cases by 90%. CONCLUSIONS: In the study population, the percentage of undiagnosed patients with diabetes aged 25-75 years was reduced from 3.6% (95% CI 3.0%-4.2%) based on one FPG measurement (FPG ≥7.0 mmol/L) to only 1.9% (95% CI 1.5%-2.3%) if the diagnosis of diabetes was based on the criterion of HbA1c ≥48 mmol/mol (6.5% DCCT).

New national Biobank of The Danish Center for Strategic Research on Type 2 Diabetes (DD2).

Long-term storage of biological samples from patients has become increasingly important in studies of disease control and treatment. The first nationwide Danish diabetes project, ie, The Danish Center for Strategic Research in Type II Diabetes (DD2), aims to improve treatment and the long-term outcome of patients with newly diagnosed type 2 diabetes (T2D). The DD2 project includes establishment of a biobank with samples from 50,000 patients with newly diagnosed T2D. This paper describes how blood and urine samples from 10,000 patients per year are collected, handled, and stored. The biobank includes whole blood, DNA, and plasma and urine samples, all frozen at -80°C. Sampling tubes have been standardized and are sent to hospital outpatient clinics and general practitioners where samples are taken, handled, aliquoted, and returned by mail overnight in standardized cryostorage tubes. When received at the biobank, samples are frozen without further treatment. From each patient, 24 cryostorage tubes are stored. Each tube is labeled with a barcode that links the data to other information available in a clinical databank registry. When patients are enrolled in DD2, a questionnaire is filled out and a quality monitoring system ensures that patients, samples, and questionnaires can be linked together at all times. The biobank is located at Vejle Hospital and the Danish National Biobank at Statens Serum Institut. As of the end of March 2012, samples from 1186 patients have been stored, and currently samples from 8-10 patients arrive per day. We have established the first national biobank in Denmark where blood, DNA, and plasma and urine samples from patients with newly diagnosed T2D are systematically collected and stored. This biobank enables sophisticated analysis of genetic variation and response to treatment, as well as disease marker studies that better classify disease status, progression, and complications.

Home management of oral anticoagulation via telemedicine versus conventional hospital-based treatment.

BACKGROUND AND OBJECTIVE: We have developed an expert computer system for the control of oral anticoagulation therapy, accessible by the patients via their own computer. To investigate if the weekly measurement and dosing of international normalized ratio (INR) at home using the online Internet-based system was superior to conventional treatment, we performed a randomized, controlled trial. PATIENTS AND METHODS: All 669 patients in our anticoagulation clinic were asked to participate in the trial, providing that they had Internet access and could use the CoaguChek XS system. A total of 140 patients were included and randomized to (A) once weekly measurement and report online, (B) twice weekly measurement and report online, and (C) continued conventional treatment with INR measurement in the lab every 4 weeks and dose adjustment by letter. RESULTS: Group A had 79.7% (95% CI 79.0-80.3) of time in therapeutic range (TTR), group B 80.2% (95% CI 79.4-80.9) of TTR, and group C 72.7% (95% CI 71.9-73.4) TTR. Groups A and B perform statistically significantly better than the conventional group C, with a difference of TTR of 7% points (p < 2.2 × 10(-16)), whereas no difference was seen between A and B. CONCLUSION: Home measurement of INR and the reporting and dosing of results online once a week increase TTR from 72% to 79% as compared to conventional computer-assisted monitoring in an anticoagulation clinic.

In vitro potency of CXA-101, a novel cephalosporin, against Pseudomonas aeruginosa displaying various resistance phenotypes, including multidrug resistance.

We describe the activity of a novel cephalosporin, CXA-101 (FR26 4205), against a panel of highly resistant Pseudomonas aeruginosa isolates collected from U.S. hospitals. CXA-101 demonstrated increased potency against this population of resistant isolates, with activity that is 4- to 10-fold higher than that of comparator agents in each phenotypic category. The addition of tazobactam did not improve its activity. CXA-101 appears to be a promising addition to the category of antipseudomonal beta-lactams.

Comparison of the activity of a human simulated, high-dose, prolonged infusion of meropenem against Klebsiella pneumoniae producing the KPC carbapenemase versus that against Pseudomonas aeruginosa in an in vitro pharmacodynamic model.

We have previously demonstrated that a high-dose, prolonged-infusion meropenem regimen (2 g every 8 h [q8h]; 3-hour infusion) can achieve 40% free drug concentration above the MIC against Pseudomonas aeruginosa with MICs of or=3 log CFU reduction against all KPC isolates within 6 h, followed by regrowth in all but two isolates. The targeted %fT>MIC (percent time that free drug concentrations remain above the MIC) exposure was achieved against both of these KPC isolates (100% fT>MIC versus MIC=2 microg/ml, 75% fT>MIC versus MIC=8 microg/ml). Against KPC isolates with MICs of 8 and 16 microg/ml that did regrow, actual meropenem exposures were significantly lower than targeted due to rapid in vitro hydrolysis, whereby targeted %fT>MIC was reduced with each subsequent dosing. In contrast, a >or=3 log CFU reduction was maintained over 24 h for all Pseudomonas isolates with meropenem MICs of 8 and 16 microg/ml. Although KPC and P. aeruginosa isolates may share similar meropenem MICs, the differing resistance mechanisms produce discordant responses to a high-dose, prolonged infusion of meropenem. Thus, predicting the efficacy of an antimicrobial regimen based on MIC may not be a valid assumption for KPC-producing organisms.

In vitro activity and pharmacodynamics of commonly used antibiotics against adult systemic isolates of Escherichia coli and Pseudomonas aeruginosa at Forty US Hospitals.

BACKGROUND: Treatment of infections caused by gram-negative bacilli is increasingly challenging because of emerging resistance. Current surveillance data are informative, but may not discern differences by infection site and clinical setting, and do not incorporate pharmacodynamic (PD) characteristics when determining susceptibility. OBJECTIVES: This study explored the differences in infection site and clinical setting and evaluated dose-optimization strategies using PD principles. The stud focused only on systemic isolates, and targeted a cohort of 40 hospitals in the United States with a nationwide geographic distribution. METHODS: Nonduplicate, nonurine isolates of Escherichia coli (n = 937) and Pseudomonas aeruginosa (n = 1044) collected from adult patients at 40 US hospitals underwent MIC testing by broth microdilution to 15 agents. Results were analyzed by infection site and unit type (ward or intensive care unit [ICU]). PD modeling employing Monte Carlo simulation was used to predict the microbiologic success of varying dosing strategies. RESULTS: E coli were highly susceptible except to fluoroquinolones; 6.8% were multidrug resistant (MDR) and were more likely in ICUs (risk ratio [RR], 2.5; 95% CI, 1.6-4.0). P aeruginosa displayed 90% for most drug classes), with the important exception of the fluoroquinolones; however, susceptibility of P aeruginosa is low enough to warrant concern. Attention to the source of the organism and the patient's location in an ICU or a ward-combined with knowledge of local epidemiology and PD principles--should prove valuable in empiric agent selection. Additionally, reassessment of breakpoints employing PD principles is recommended, particularly for fluoroquinolones and piperacillin/tazobactam.