MPCNN: A Novel Matrix Profile Approach for CNN-based Single Lead Sleep Apnea In Classification Problem.

Sleep apnea (SA) is a significant respiratory condition that poses a major global health challenge. Deep Learning (DL) has emerged as an efficient tool for the classification problem in electrocardiogram (ECG)-based SA diagnoses. Despite these advancements, most common conventional feature extractions derived from ECG signals in DL, such as R-peaks and RR intervals, may fail to capture crucial information encompassed within the complete ECG segments. In this study, we propose an innovative approach to address this diagnostic gap by delving deeper into the comprehensive segments of the ECG signal. The proposed methodology draws inspiration from Matrix Profile algorithms, which generate an Euclidean distance profile from fixed-length signal subsequences. From this, we derived the Min Distance Profile (MinDP), Max Distance Profile (MaxDP), and Mean Distance Profile (MeanDP) based on the minimum, maximum, and mean of the profile distances, respectively. To validate the effectiveness of our approach, we use the modified LeNet-5 architecture as the primary CNN model, along with two existing lightweight models, BAFNet and SE-MSCNN. Our experiment results on the PhysioNet Apnea-ECG dataset (70 overnight recordings), and the UCDDB dataset (25 overnight recordings) revealed that our new feature extraction method achieved per-segment accuracies of up to 92.11% and 81.25%, respectively. Moreover, using the PhysioNet data, we achieved a per-recording accuracy of 100% and yielded the highest correlation of 0.989 compared to state-of-the-art methods. By introducing a new feature extraction method based on distance relationships, we enhanced the performance of certain lightweight models in DL, showing potential for home sleep apnea test (HSAT) and SA detection in IoT devices. The source code for this work is made publicly available in GitHub: https://github.com/vinuni-vishc/MPCNN-Sleep-Apnea.

VinDr-CXR: An open dataset of chest X-rays with radiologist's annotations.

Most of the existing chest X-ray datasets include labels from a list of findings without specifying their locations on the radiographs. This limits the development of machine learning algorithms for the detection and localization of chest abnormalities. In this work, we describe a dataset of more than 100,000 chest X-ray scans that were retrospectively collected from two major hospitals in Vietnam. Out of this raw data, we release 18,000 images that were manually annotated by a total of 17 experienced radiologists with 22 local labels of rectangles surrounding abnormalities and 6 global labels of suspected diseases. The released dataset is divided into a training set of 15,000 and a test set of 3,000. Each scan in the training set was independently labeled by 3 radiologists, while each scan in the test set was labeled by the consensus of 5 radiologists. We designed and built a labeling platform for DICOM images to facilitate these annotation procedures. All images are made publicly available in DICOM format along with the labels of both the training set and the test set.

Directly observed therapy and risk of unfavourable tuberculosis treatment outcomes among an international cohort of people living with HIV in low- and middle-income countries.

INTRODUCTION: Identification of persons living with human immunodeficiency virus (HIV)-associated tuberculosis (TB) at increased risk for unfavourable TB outcomes would inform efforts to improve such outcomes. We sought to identify factors associated with a decreased risk of unfavourable TB treatment outcomes among people living with HIV-infection (PLHIV) in low- and middle-income countries (LMIC), with a specific focus on directly observed therapy (DOT) compared with self-administered therapy (SAT) during the continuation phase of anti-TB therapy. METHODS: We conducted a retrospective cohort study among adults diagnosed with HIV-associated TB in Africa, Asia and the Americas from 2012 to 2013; data were collected from 2012 to 2016. Unfavourable TB treatment outcomes (death during TB treatment, and TB treatment failure or recurrence) were defined according to World Health Organization criteria. Receipt of DOT was obtained at the site level and defined as ≥5 days of DOT per week. The person administering DOT and treatment location varied by site. Lack of receipt of DOT was defined as SAT. Multivariable logistic regression estimated the adjusted odds of unfavourable TB treatment outcomes. RESULTS: Among 1862 adults with HIV-associated TB included, 252 (13.5%) had unfavourable TB outcomes (226 deaths, 26 recurrences/failures). Overall, 1825 (98%) received DOT in the intensive phase and 1617 (87%) received DOT in the continuation phase. DOT in the continuation phase was not significantly associated with unfavourable TB outcomes (aOR 1.43, 95% CI 0.86 to 2.38) compared to SAT. Body mass index (BMI) change during anti-TB treatment (per 2 units increase, aOR 0.74, 95% CI 0.68 to 0.82) and CD4+ count at TB diagnosis (200 vs. 50  cells/µL, aOR 0.54, 95% CI 0.39 to 0.73) were both independently associated with decreased odds of unfavourable TB treatment outcomes. CONCLUSIONS: In this large, international cohort of people living with HIV-associated TB in LMIC who received intensive phase DOT, DOT during the continuation phase of anti-TB therapy was not associated with a decreased odds of unfavourable TB treatment outcomes compared to SAT. Randomized trials evaluating the effect of continuation-phase DOT on TB outcomes among PLHIV are needed.

Gastric cancer: epidemiology, biology, and prevention: a mini review.

Gastric cancer is one of the most common causes of cancer-related mortality worldwide. The objective of this article is to review the epidemiology and biology of gastric cancer risk. This literature review explores the biological, clinical, and environmental factors that influence the rates of this disease and discuss the different intervention methods that may not only increase the awareness of gastric cancer but also increase screening in efforts to reduce the risk of gastric cancer. Helicobacter pylori infection is the primary risk factor for gastric cancer. Additional risk factors include geographical location, age, sex, smoking, socioeconomic status, dietary intake, and genetics. Primary and secondary prevention strategies such as dietary modifications and screenings are important measures for reducing the risk of gastric cancer. Interventions, such as H. pylori eradication through chemoprevention trials, have shown some potential as a preventative strategy. Although knowledge about gastric cancer risk has greatly increased, future research is warranted on the differentiation of gastric cancer epidemiology by subsite and exploring the interactions between H. pylori infection, genetics, and environmental factors. Better understanding of these relationships can help researchers determine the most effective intervention strategies for reducing the risk of this disease.

Routine versus Targeted Viral Load Strategy among Patients Starting Antiretroviral in Hanoi, Vietnam.

INTRODUCTION: HIV viral load (VL) testing is recommended by the WHO as the preferred method for monitoring patients on antiretroviral therapy (ART). However, evidence that routine VL (RVL) monitoring improves clinical outcomes is lacking. METHODS: We conducted a prospective, randomized controlled trial of RVL monitoring every six months versus a targeted VL (TVL) strategy (routine CD4 plus VL testing if clinical or immunological failure) in patients starting ART between April 2011 and April 2014 at Bach Mai Hospital in Hanoi. Six hundred and forty-seven subjects were randomized to RVL (n = 305) or TVL monitoring (n = 342) and followed up for three years. Primary endpoints were death or WHO clinical Stage 4 events between six and thirty-six months of ART and rate of virological suppression at three years. RESULTS: Overall, 37.1% of subjects were female, median age was 33.4 years (IQR: 29.5 to 38.6), and 47% had a CD4 count ≤100 cells/mm3 at time of ART initiation. Approximately 44% of study events (death, LTFU, withdrawal, or Stage 4 event) and 68% of deaths occurred within the first six months of ART. Among patients on ART at six months, death or Stage 4 event occurred in 3.6% of RVL and 3.9% of TVL (p = 0.823). Survival analysis showed no significant difference between the groups (p = 0.825). Viral suppression at 36 months of ART was 97.2% in RVL and 98.9% in TVL (p = 0.206) at a threshold of 400 copies/mL and was 98.0% in RVL and 98.9% in TVL (p = 0.488) at 1000 copies/mL. In ITT analysis, 20.7% in RVL and 21.9% in TVL (p = 0.693) were unsuppressed at 1000 copies/mL. CONCLUSIONS: We found no significant difference in rates of death or Stage 4 events and virological failure in patients with RVL monitoring compared to those monitored with a TVL strategy after three years of follow-up. Viral suppression rates were high overall and there were few study events among patients alive and on ART after six months, limiting the study's power to detect a difference among study arms. Nonetheless, these data suggest that the choice of VL monitoring strategy may have less impact on patient outcomes compared to efforts to reduce early mortality and improve ART retention.

Cotrimoxazole prophylaxis decreases tuberculosis risk among Asian patients with HIV.

INTRODUCTION: Cotrimoxazole (CTX) is recommended as prophylaxis against Pneumocystis jiroveci pneumonia, malaria and other serious bacterial infections in HIV-infected patients. Despite its in vitro activity against Mycobacterium tuberculosis, the effects of CTX preventive therapy on tuberculosis (TB) remain unclear. METHODS: Adults living with HIV enrolled in a regional observational cohort in Asia who had initiated combination antiretroviral therapy (cART) were included in the analysis. Factors associated with new TB diagnoses after cohort entry and survival after cART initiation were analysed using Cox regression, stratified by site. RESULTS: A total of 7355 patients from 12 countries enrolled into the cohort between 2003 and 2016 were included in the study. There were 368 reported cases of TB after cohort entry with an incidence rate of 0.99 per 100 person-years (/100 pys). Multivariate analyses adjusted for viral load (VL), CD4 count, body mass index (BMI) and cART duration showed that CTX reduced the hazard for new TB infection by 28% (HR 0.72, 95% CI l 0.56, 0.93). Mortality after cART initiation was 0.85/100 pys, with a median follow-up time of 4.63 years. Predictors of survival included age, female sex, hepatitis C co-infection, TB diagnosis, HIV VL, CD4 count and BMI. CONCLUSIONS: CTX was associated with a reduction in the hazard for new TB infection but did not impact survival in our Asian cohort. The potential preventive effect of CTX against TB during periods of severe immunosuppression should be further explored.

Cardiovascular disease incidence projections in the TREAT Asia HIV Observational Database (TAHOD).

BACKGROUND: We aimed to project the 10-year future incidence of cardiovascular disease (CVD) and model several intervention scenarios based on a multi-site Asian HIV-positive cohort. METHODS: Analyses were based on patients recruited to the TREAT Asia HIV Observational Database (TAHOD), consisting of 21 sites in 12 countries. Patients on triple antiretroviral therapy (ART) were included if they were alive, without previous CVD, and had data on CVD risk factors. Annual new CVD events for 2019-2028 were estimated with the D:A:D equation, accounting for age- and sex-adjusted mortality. Modelled intervention scenarios were treatment of high total cholesterol, low high-density lipoprotein cholesterol (HDL) or high blood pressure, abacavir or lopinavir substitution, and smoking cessation. RESULTS: Of 3,703 included patients, 69% were male, median age was 46 (IQR 40-53) years and median time since ART initiation was 9.8 years (IQR 7.5-14.1). Cohort incidence rates of CVD were projected to increase from 730 per 100,000 person-years (pys) in 2019 to 1,432 per 100,000 pys in 2028. In the modelled intervention scenarios, most events can be avoided by smoking cessation, abacavir substitution, lopinavir substitution, decreasing total cholesterol, treating high blood pressure and increasing HDL. CONCLUSIONS: Our projections suggest a doubling of CVD incidence rates in Asian HIV-positive adults in our cohort. An increase in CVD can be expected in any ageing population, however, according to our models, this can be close to averted by interventions. Thus, there is an urgent need for risk screening and integration of HIV and CVD programmes to reduce the future CVD burden.

Diabetes mellitus burden among people living with HIV from the Asia-Pacific region.

INTRODUCTION: Comorbidities including diabetes mellitus (DM) among people living with HIV (PLHIV) are of increasing clinical concerns in combination antiretroviral therapy (cART) era. We aimed to determine the incidence and risk factors of new-onset DM among PLHIV in Asian settings. METHODS: PLHIV from a regional observational cohort without DM prior to antiretroviral therapy (ART) initiation were included in the analysis. DM was defined as having a fasting blood glucose ≥126 mg/dL, glycated haemoglobin ≥6.5%, a two-hour plasma glucose ≥200 mg/dL, or a random plasma glucose ≥200 mg/dL. A Cox regression model, stratified by site, was used to identify risk factors associated with DM. RESULTS AND DISCUSSION: Of the 1927 participants included, 127 were diagnosed with DM after ART initiation. Median follow-up time from ART initiation to DM diagnosis was 5.9 years (interquartile range (IQR): 2.8 to 8.9 years). The crude incidence rate of DM was 1.08 per 100 person-years (100 PYS), 95% confidence interval (CI) (0.9 to 1.3). In the multivariate analysis, later years of follow-up (2011 to 2013: HR = 2.34, 95% CI 1.14 to 4.79, p = 0.02; and 2014 to 2017: HR = 7.20, 95% CI 3.27 to 15.87, p < 0.001) compared to <2010, older age (41 to 50 years: HR = 2.46, 95% CI 1.39 to 4.36, p  = 0.002; and >50 years: HR = 4.19, 95% CI 2.12 to 8.28, p < 0.001) compared to <30 years, body mass index (BMI) >30 kg/m2 (HR = 4.3, 95% CI 1.53 to 12.09, p = 0.006) compared to BMI <18.5 kg/m2 , and high blood pressure (HR = 2.05, 95% CI 1.16 to 3.63, p = 0.013) compared to those without high blood pressure, were associated with developing DM. The hazard was reduced for females (HR = 0.47, 95% CI 0.28 to 0.80, p = 0.006). CONCLUSIONS: Type 2 DM in HIV-infected Asians was associated with later years of follow-up, high blood pressure, obesity and older age. This highlights the importance of monitoring and routine screening for non-communicable diseases including DM as PLHIV age.

Melioidosis in Vietnam: Recently Improved Recognition but still an Uncertain Disease Burden after Almost a Century of Reporting.

The first cases of human melioidosis were described in Vietnam in the 1920s, almost a century ago. It was in Vietnam in the thirties that the saprophytic nature of B. pseudomallei was first recognized. Although a significant number of French and U.S. soldiers acquired the disease during the Vietnam wars, indigenous cases in the Vietnamese population were only sporadically reported over many decades. After reunification in 1975, only two retrospective studies reported relatively small numbers of indigenous cases from single tertiary care hospitals located in the biggest cities in the South and the North, respectively. Studies from provincial hospitals throughout the country were missing until the Research Network on Melioidosis and Burkholderia pseudomallei (RENOMAB) project started in 2014. From then on seminars, workshops, and national scientific conferences on melioidosis have been conducted to raise awareness among physicians and clinical laboratory staff. This led to the recognition of a significant number of cases in at least 36 hospitals in 26 provinces and cities throughout Vietnam. Although a widespread distribution of melioidosis has now been documented, there are still challenges to understand the true epidemiology of the disease. Establishment of national guidelines for diagnosis, management, and reporting of the disease together with more investigations on animal melioidosis, genomic diversity of B. pseudomallei and its environmental distribution are required.

Sensitivity and specificity of two dried blood spot methods for HIV-1 viral load monitoring among patients in Hanoi, Vietnam.

The use of dried blood spot (DBS) specimens for HIV viral load (VL) monitoring is recommended to support the roll-out of routine VL monitoring in low and middle income countries (LMICs). To better understand the use of DBS for VL monitoring, we evaluated two DBS testing methods, Roche TaqMan® Free Virus Evolution protocol (DBS-FVE) and Roche TaqMan® SPEX protocol (DBS-SPEX)) in patients receiving ART at an HIV clinic in Hanoi, Vietnam. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each DBS testing method at the thresholds of 1000 and 5000 copies/ml compared to plasma VL. At a threshold of 1000 copies/ml, sensitivity, specificity, PPV and NPV of the DBS-SPEX method were 98.8% (95% CI: 93.3%-100%), 74.3% (95% CI: 70.8%-77.5%), 31.5% (95% CI: 25.8%-37.6%), and 99.8% (95% CI: 98.9%-100%), respectively. Increasing the VL threshold value to 5000 copies/ml improved specificity (97.9% CI: 96.6%-98.9%) and PPV (83.9% CI: 74.5%-90.9%). Using the DBS-FVE method, at the threshold of 1000 copies/ml and with a correction factor of +0.3 log copies/ml, sensitivity was 95.1% (87.8%-98.6%) and specificity was 98.8% (97.7%-99.5%). Sensitivity decreased at the threshold of 5000 copies/ml (65.8%, 95% CI: 54.3%-76.1%). With a correction factor of +0.7 log copies/ml, the sensitivity was 96.3% (89.6%-99.2%) and specificity was 98.2% (96.9%-99.1%) at the threshold of 1000 copies/ml. We found that the Roche DBS-FVE method, with a +0.7 log copies/ml correction factor, performed well with sensitivity and specificity greater than 96% at a VL threshold of 1000 copies/m. These findings add to the growing body of evidence supporting the use of DBS VL testing for ART monitoring. Future research should evaluate the association between VL results by DBS and clinical outcome measures such as HIV drug resistance, morbidity, and mortality.

Regional Differences in the Prevalence of Major Opportunistic Infections among Antiretroviral-Naïve Human Immunodeficiency Virus Patients in Japan, Northern Thailand, Northern Vietnam, and the Philippines.

To identify regional differences in the distribution of opportunistic infections (OIs) among human immunodeficiency virus (HIV)-infected patients in Asia, the medical records of antiretroviral therapy (ART)-naïve patients who attended the following tertiary hospitals from 2003 to 2011 were reviewed: Nagoya Medical Center (NMC, Nagoya, Japan), Lampang Hospital (LPH, Lampang, northern Thailand), Bach Mai Hospital (BMH, Hanoi, northern Vietnam), and Philippine General Hospital (PGH, Manila, Philippines). Logistic regression analyses were performed to identify associations between country of origin and risk of major OIs. In total, 1,505 patients were included: NMC, N = 365; LPH, N = 442; BMH, N = 384; and PGH, N = 314. The median age was 32 years, and 73.3% of all patients were male. The median CD4 count was 200 cells/µL. Most patients at NMC and PGH were men who have sex with men. Injection drug users were most common at BMH (35.7%). Mycobacterium tuberculosis (TB) was most common at PGH (N = 75) but was rare at NMC (N = 4). Pneumocystis pneumonia (PCP) prevalence was highest at NMC (N = 74) and lowest at BMH (N = 13). Multivariable logistic regression showed increased odds of TB at PGH (adjusted odds ratio [aOR] = 42.2, 95% confidence interval [CI] = 14.6-122.1), BMH (aOR = 12.6, CI = 3.9-40.3), and LPH (aOR = 6.6, CI = 2.1-21.1) but decreased odds of PCP at BMH (aOR = 0.1, CI = 0.04-0.2) and LPH (aOR = 0.2, CI = 0.1-0.4) compared with those at NMC. The cryptococcosis risk was increased at LPH (aOR = 6.2, CI = 0.9-41.0) compared with that at NMC. Cytomegalovirus (CMV) retinitis prevalences were similar in all countries. OI prevalence remained high among ART-naïve patients in our cohort. The risks of TB, PCP, and cryptococcosis, but not CMV retinitis, differed between countries. Improved early HIV detection is warranted.

Cigarette smoking is associated with high HIV viral load among adults presenting for antiretroviral therapy in Vietnam.

High HIV viral load (VL >100,000 cp/ml) is associated with increased HIV transmission risk, faster progression to AIDS, and reduced response to some antiretroviral regimens. To better understand factors associated with high VL, we examined characteristics of patients presenting for treatment in Hanoi, Vietnam. We examined baseline data from the Viral Load Monitoring in Vietnam Study, a randomized controlled trial of routine VL monitoring in a population starting antiretroviral therapy (ART) at a clinic in Hanoi. Patients with prior treatment failure or ART resistance were excluded. Characteristics examined included demographics, clinical and laboratory data, and substance use. Logistic regression was used to calculate crude and adjusted odds ratios (aOR) and 95% confidence intervals (95% CI). Out of 636 patients, 62.7% were male, 72.9% were ≥30 years old, and 28.3% had a history of drug injection. Median CD4 was 132 cells/mm3, and 34.9% were clinical stage IV. Active cigarette smoking was reported by 36.3% with 14.0% smoking >10 cigarettes per day. Alcohol consumption was reported by 20.1% with 6.1% having ≥5 drinks per event. Overall 53.0% had a VL >100,000 cp/ml. Male gender, low body weight, low CD4 count, prior TB, and cigarette smoking were associated with high VL. Those who smoked 1-10 cigarettes per day were more likely to have high VL (aOR = 1.99, 95% CI = 1.15-3.45), while the smaller number of patients who smoked >10 cigarettes per day had a non-significant trend toward higher VL (aOR = 1.41, 95% CI = 0.75-2.66). Alcohol consumption was not significantly associated with high VL. Tobacco use is increasingly recognized as a contributor to premature morbidity and mortality among HIV-infected patients. In our study, cigarette smoking in the last 30 days was associated with a 1.5 to 2-fold higher odds of having an HIV VL >100,000 cp/ml among patients presenting for ART. These findings provide further evidence of the negative effects of tobacco use among HIV-infected patients.