ABSTRACT
The aim of this paper was to explore the role of artificial intelligence (AI) applied to ultrasound imaging in gynecology oncology. Web of Science, PubMed, and Scopus databases were searched. All studies were imported to RAYYAN QCRI software. The overall quality of the included studies was assessed using QUADAS-AI tool. Fifty studies were included, of these 37/50 (74.0%) on ovarian masses or ovarian cancer, 5/50 (10.0%) on endometrial cancer, 5/50 (10.0%) on cervical cancer, and 3/50 (6.0%) on other malignancies. Most studies were at high risk of bias for subject selection (i.e., sample size, source, or scanner model were not specified; data were not derived from open-source datasets; imaging preprocessing was not performed) and index test (AI models was not externally validated) and at low risk of bias for reference standard (i.e., the reference standard correctly classified the target condition) and workflow (i.e., the time between index test and reference standard was reasonable). Most studies presented machine learning models (33/50, 66.0%) for the diagnosis and histopathological correlation of ovarian masses, while others focused on automatic segmentation, reproducibility of radiomics features, improvement of image quality, prediction of therapy resistance, progression-free survival, and genetic mutation. The current evidence supports the role of AI as a complementary clinical and research tool in diagnosis, patient stratification, and prediction of histopathological correlation in gynecological malignancies. For example, the high performance of AI models to discriminate between benign and malignant ovarian masses or to predict their specific histology can improve the diagnostic accuracy of imaging methods.
ABSTRACT
BACKGROUND: Little research has focused on the relationship between gut microbiome and chemotherapy-induced toxicity. METHODS: This prospective study involves 301 patients with breast cancer who had prechemotherapy stool samples collected. Gut microbiome was sequenced by shotgun metagenomics; associations with chemotherapy-induced toxicities during first-line treatment by gut microbial diversity, composition, and metabolic pathways with severe (i.e., grade ≥3) hematological and gastrointestinal toxicities were evaluated via multivariable logistic regression. RESULTS: High prechemotherapy α-diversity was associated with a significantly reduced risk of both severe hematological toxicity (odds ratio [OR] = 0.94; 95% CI, 0.89-0.99; p = .048) and neutropenia (OR = 0.94; 95% CI, 0.89-0.99; p = .016). A high abundance of phylum Synergistota, class Synergistia, and order Synergistales were significantly associated with a reduced risk of severe neutropenia; conversely, enrichment of phylum Firmicutes C, class Negativicutes, phylum Firmicutes I, and class Bacilli A, order Paenibacillales were significantly associated with an increased risk of severe neutropenia (p range: 0.012-2.32 × 10-3; false discovery rate <0.1). Significant positive associations were also observed between severe nausea/vomiting and high Chao1 indexes, ß-diversity (p < .05), 20 species belonging to the family Lachnospiraceae, Oscillospiraceae, and Ruminococcaceae (p value range: 6.14 × 10-3 to 1.33 × 10-5; false discovery rate <0.1), and three metabolic pathways involved in reductive tricarboxylic acid cycle I and cycle II, and an incomplete reductive tricarboxylic acid cycle (p < .01). Conversely, a high abundance of species Odoribacter laneus and the pathway related to the L-proline biosynthesis II were inversely associated with severe nausea/vomiting. CONCLUSIONS: Our study suggests that gut microbiota may be a potential preventive target to reduce chemotherapy-induced toxicity.
Subject(s)
Breast Neoplasms , Gastrointestinal Microbiome , Humans , Breast Neoplasms/drug therapy , Gastrointestinal Microbiome/drug effects , Female , Middle Aged , Prospective Studies , Aged , Adult , Neutropenia/chemically induced , Neutropenia/microbiology , Metagenomics/methods , Feces/microbiology , Antineoplastic Combined Chemotherapy Protocols/adverse effects , Antineoplastic Combined Chemotherapy Protocols/therapeutic use , Antineoplastic Agents/adverse effectsABSTRACT
Linking meteorology and air pollutants is a key challenge. The study investigated meteorological effects on PM2.5 concentration using the advanced convergent cross mapping method, utilizing hourly PM2.5 concentration and six meteorological factors across eight provinces and cities in Vietnam. Results demonstrated that temperature (ρ = 0.30) and radiation (ρ = 0.30) produced the highest effects, followed by humidity (ρ = 0.28) and wind speed (ρ = 0.24), while pressure (ρ = 0.22) and wind direction (ρ = 0.17) produced the weakest effects on PM2.5 concentration. Comparing the ρ values showed that temperature, wind speed, and wind direction had greater impacts on PM2.5 concentration during the dry season whereas radiation had a more influence during the wet season; Southern stations experienced larger meteorological effects. Temperature, humidity, pressure, and wind direction had both positive and negative influences on PM2.5 concentration, while radiation and wind speed mostly had negative influences. During PM2.5 pollution episodes, there was more contribution of meteorological effects on PM2.5 concentration indicated by ρ values. At contaminated levels, humidity (ρ = 0.45) was the most dominant factor affecting PM2.5 concentration, followed by temperature (ρ = 0.41) and radiation (ρ = 0.40). Pollution episodes were pointed out to be more prevalent under higher humidity, higher pressure, lower temperature, lower radiation, and lower wind speed. The ρ calculation also revealed that lower temperature, lower radiation, and higher humidity greatly accelerated each other under pollution episodes, further enhancing PM2.5 concentration. The findings contributed to the literature on meteorology and air pollution interaction.
Subject(s)
Air Pollutants , Air Pollution , Cities , Environmental Monitoring , Particulate Matter , Vietnam , Particulate Matter/analysis , Air Pollutants/analysis , Air Pollution/statistics & numerical data , Meteorological Concepts , Seasons , WindABSTRACT
We analyzed the risk-benefit of COVID-19 vaccine using a causal model to explain and weigh up possible risk factors of blood clots after vaccination. A self-controlled case series method was used to examine the association between blood clots and COVID-19 vaccination. To avoid bias due to the under-reported infection among non-hospitalized subjects, a case-control study was used to compare the risk of blood clots in infected subjects to control subjects who were hospitalized due to physical injury. We found increased risks of blood clots after vaccination (incidence rate ratio is 1.13, 95% CI: [1.03,1.24] after the first dose and 1.23, 95% CI: [1.13,1.34] after the second dose). Furthermore, vaccination attenuated the increased risk of blood clots associated with infection (odds ratio is 2.16, 95% CI: [1.93,2.42] in unvaccinated versus 1.46, 95% CI: [1.25,1.70] in vaccinated). After accounting for vaccine efficacy against infection and the protection against infection-associated blood clots, receiving the COVID-19 vaccines decreases the risk of blood clots, especially during high infection rate period.
ABSTRACT
AIM: Sleep disorders are common in cancer patients and have negative consequences for patient well-being and treatment outcomes. This study aimed to investigate sleep quality and related factors in Vietnamese middle-aged cancer patients. METHODS: A cross-sectional study was conducted on 246 middle-aged in-patient cancer patients at Vietnam National Cancer Hospital (VNCH) from 1/2021 to 7/2021. Sleep was measured by the Pittsburgh Sleep Quality Index (PSQI), with a cutoff of 5. RESULTS: The results showed a male/female ratio of 0.85 with an average age of 52. The five most prevalent cancer types were breast, colorectal, lung, and esophagus-stomach cancer, primarily in the late stage and treated with chemotherapy. The prevalence of sleep disturbances was 58.5%. The mean PSQI score was 7.5, with sleep duration and latency of 5.4 h and 1 h, respectively. Approximately 44% of participants reported poor sleep quality, nearly 9% had daytime dysfunction, and 10.6% used sleep medication. The multivariate logistic regression results indicate that people with depression were 8.89 times more likely to have poor sleep than those without depression (95% CI:2.63-28.27, p < .001). CONCLUSION: Sleep problems are common among middle-aged people with cancer in Vietnam, especially individuals with depression. It is necessary to have more effective approaches to sleep management for cancer patients with limited resources.
Subject(s)
Neoplasms , Sleep Wake Disorders , Middle Aged , Humans , Male , Female , Cross-Sectional Studies , Vietnam/epidemiology , Cancer Care Facilities , Sleep , Neoplasms/epidemiology , Sleep Wake Disorders/epidemiology , Sleep Wake Disorders/etiologyABSTRACT
Objectives and importance of study: This study analyses Australian healthy hospital retail policies to identify the similarities and differences in the policies and policy implementation processes. The potential impact of the different policy components on dietary behaviours were examined via a scoping review. STUDY TYPE: Policy analysis and scoping review. METHODS: Healthy retail policy documents and policy implementation guidelines were identified via a grey literature search on Department of Health websites of all Australian jurisdictions. Policy components and policy implementation processes were extracted and analysed for similarities and differences. The potential effectiveness of the different policy components on purchasing and/or dietary behaviour were identified via a scoping review of the academic literature, conducted in March 2023 across seven electronic databases and Google Scholar. The scoping review included studies reporting the impacts of healthy food retail interventions implemented in hospitals. No timeframe restriction was applied for both the grey literature search and the scoping review. RESULTS: All Australian jurisdictions, except Tasmania, have implemented jurisdiction specific healthy retail policies in public hospital settings. There are similarities and difference in the policy components and implementation design across the jurisdictions. Similarities included the policy scope, use of a traffic light system to classify the nutritional healthiness of food and beverages for sale, and the standards used to determine the mix of healthy and unhealthy food availability. These similarities allowed sharing of resources across some jurisdictions. There is limited evaluation of policy impacts on purchase and/or consumption behaviours. Twenty of 27 studies identified via the scoping literature review examined interventions similar to the Australian policies and showed that these policies could result in increased purchase of healthier products among staff and visitors. Key implementation success factors include strong support for the policy from all stakeholders, practical implementation support resources, and impacts on retailer profitability. CONCLUSIONS: The healthy hospital retail policies implemented across Australian jurisdictions could encourage healthier food and beverage purchases among staff and visitors. Evaluation of the policies could facilitate further refinement to enhance effectiveness and translation of learnings to international contexts.
Subject(s)
Diet , Health Policy , Humans , Australia , Policy Making , HospitalsABSTRACT
Objective and importance of study: Overweight and obesity are the second leading risk factors for death and non-communicable disease in Australia. This study aimed to examine the Australian Federal Government funding landscape for population-level obesity prevention from 2013 to 2022. STUDY TYPE: A retrospective analysis and narrative synthesis of publicly available data on obesity prevention funding from the Federal Government and major federally funded Australian research organisations. METHODS: Searches were conducted of Australian Federal Government Budget documents and funding announcements from the National Health and Medical Research Council (NHMRC), Australian Research Council (ARC) and Medical Research Future Fund (MRFF). Funding allocations targeting obesity prevention, or the prevention of risk factors associated with obesity, were included. These were determined by the presence of keywords related to obesity, unhealthy diet, physical activity and sedentary behaviour. Data were extracted verbatim, coded and narratively synthesised by funding source. RESULTS: From 2013 to 2022, 186 funding allocations for obesity prevention in Australia were identified, totalling approximately A$778 million. The proportion of funding allocated to obesity prevention compared to the total annual budget of each funding source was relatively low: NHMRC = 1.1%; ARC = 0.2%; MRFF = 0.8%; Federal Government = 0.1% (of health budget). Funding for obesity prevention initiatives fluctuated over time. CONCLUSIONS: Findings underscore the need for strategic and ongoing funding allocation to support obesity prevention research, implementation and sustainment of evidence-based obesity prevention initiatives in Australia.
Subject(s)
Health Services Research , Obesity , Humans , Federal Government , Retrospective Studies , Australia , Obesity/prevention & controlABSTRACT
Age and gender are 2 important factors in the treatment of end-stage chronic kidney disease with hemodialysis. Understanding the influence of these 2 factors can help optimize treatment for this population. This study evaluated gender and age differences in achievement of Kidney Disease Improving Global Outcomes (KDIGO) treatment targets. A cross-sectional study was conducted on 324 chronic hemodialysis patients at a tertiary referral hospital in Ho Chi Minh City, Vietnam. KDIGO treatment targets included treatment time, prescribed Qb, treated blood volume, urea reduction ratio, spKt/V, hemoglobin, albumin, phosphorus, calcium, and parathyroid hormone. Men had significantly higher treatment time (Pâ =â .003), prescribed Qb (Pâ =â .037) and hemoglobin (Pâ =â .031) than women. However, women had significantly higher treated blood volume (Pâ <â .001), spKt/V (Pâ <â .001) and URR (Pâ <â .001). No significant difference between men and women was found in albumin, calcium, phosphorus, and parathyroid hormone. Based on KDIGO treatment targets, women had a significantly higher rate of achievement of spKt/Vâ >â 1.2 (91.4% vs 80.7%, Pâ =â .005) and URRâ ≥â 70% (77.0% vs 54.7%, Pâ <â .001) than men. A significantly higher rate of treated volume ofâ ≥â 1 L/kg/BW, and phosphorus 2.5 to 4.6 mg/dL was found in women (90.0% and 40.2%) compared to men (68.7% and 27.3%). In contrast, men had a significantly higher rate of prescribed Qbâ ≥â 300 mL/min (26.7% vs 12.6%, Pâ =â .001), albuminâ ≥â 40 g/L (36.7% vs 26.4%, Pâ =â .047), and Hbâ >â 12 g/dL (22.0% vs 11.5%, Pâ =â .011) than women. There was no significant difference between men and women in the rate of calcium 8.4 to 10.4 mg/dL, and parathyroid hormone 150 to 600 pg/mL. These differences were not the same across 4 age categories (<60, 60-69, 70-79, andâ ≥â 80). Most of the differences above were among patients agedâ <â 60 and 60 to 69 years. Although men had higher satisfactory treatment parameters than women, based on KDIGO treatment targets, women received hemodialysis more effectively than men. Treatment targets for patients on hemodialysis should consider gender and age differences.
Subject(s)
Kidney Diseases , Kidney Failure, Chronic , Male , Humans , Female , Calcium , Tertiary Care Centers , Cross-Sectional Studies , Vietnam , Renal Dialysis , Kidney Failure, Chronic/therapy , Parathyroid Hormone , Phosphorus , Hemoglobins/analysis , AlbuminsABSTRACT
Increasingly complex modern gas-monitoring scenarios necessitate advanced sensing capabilities to detect and identify a diverse range of gases under varying conditions. There is a rising demand for individual sensors with multiple responses capable of recognizing gases, identifying components in mixtures, and providing stable responses. Inspired by gas sensors employing multivariable response principles, we develop a nanoporous anodic alumina high-order microcavity (NAA-HOµCV) gas sensor with multiple optical outputs for discriminative gas detection. The NAA-HOµCV architecture, formed by a Fabry-Pérot microcavity with distributed Bragg reflector (DBR) mirrors and an extended-length microcavity layer supporting multiple resonant modes, serves as an effective solid-state fingerprint platform for distinguishing volatile organic compound (VOC) gases. Our research reveals that the coupling strength of light into resonant modes and their evolution depend on the thickness of the DBR mirrors and the dimension of the microcavity layer, which allows us to optimize the discriminative sensing capability of the NAA-HOµCV sensor through structural engineering of the microcavity and photonic crystal mirrors. Gas-sensing experiments conducted on the NAA-HOµCV sensor demonstrate real-time discrimination between physiosorbed VOC gases (isopropanol, ethanol, or acetone) in reversible gas sensing. It also achieves superior ppb-level sensing in irreversible gas sensing of model silane molecules. Our study presents promising avenues for designing compact, cost-effective, and highly efficient gas sensors with tailored properties for discriminative gas detection.
ABSTRACT
STUDY OBJECTIVE: This study assessed the overall satisfaction with oncological care, including barriers to care, and identified its associated predictors among adult cancer patients in Vietnam. METHODS: In this cross-sectional study, we enrolled 300 adult cancer patients receiving inpatient care at a large urban oncological hospital between June and July 2022. Multivariable linear regression analyses examined associations between patient experiences and overall satisfaction ratings with cancer care. RESULTS: The mean overall satisfaction with oncological care was 8.82 out of 10, with 98.0% recommending this facility to their friends and family. In an adjusted model, being female (ß = 0.29, 95%CI: 0.04, 0.53), endorsing satisfaction with patient-nurse communication (ß = 0.33, 95%CI: 0.13, 0.53), patient-doctor communication (ß = 0.40, 95%CI: 0.11, 0.70), and psychoeducation about oncological medication management (ß = 0.30, 95%CI: 0.14, 0.45) were positively associated with overall ratings. In contrast, individuals with delays in treatment scheduling reported lower overall satisfaction with oncological care (ß = -0.38, 95%CI: -0.64, -0.13). Patients perceived health system, social/environmental, and individual barriers to care: worries about income loss due to attending treatment (43.3%); fear, depression, anxiety, and distress (36.8%); concerns about affordability of treatment (36.7%) and transportation problems (36.7%); and excessive waiting times for appointments (28.8%). CONCLUSION: This study showed high overall patient satisfaction with cancer care quality. Patient-centered communication strategies and psychoeducation about oncological medication management may be targeted to further enhance the cancer inpatient experience. Raising awareness about treatment options and services, and integrating mental health awareness into oncological care may ameliorate patient distress and facilitate greater satisfaction with oncological treatment processes.
Subject(s)
Neoplasms , Patient Satisfaction , Humans , Female , Male , Vietnam , Middle Aged , Patient Satisfaction/statistics & numerical data , Neoplasms/therapy , Neoplasms/psychology , Adult , Cross-Sectional Studies , Aged , Cancer Care Facilities , Health Services AccessibilityABSTRACT
INTRODUCTION: Cancer cells induce hypercoagulability in the tumoral microenvironment by expressing Tissue Factor (TF). We aimed to study the impact of the procoagulant signature of cancer cells on the quality and structure of fibrin network. We also studied the impact of fibrin clot shield (FCS) on the efficiency of anticancer agents and the migration of cancer cells. MATERIALS AND METHODS: Pancreatic cancer cells BXPC3 and breast cancer cells MDA-MB231 and MCF7, were cultured in the presence of normal Platelet Poor Plasma (PPP), diluted 10 % in conditioning media. Their potential to induce thrombin generation and their fibrinolytic activity were assessed. The structure of fibrin network was analyzed with Scanning Electron Microscopy (SEM). Cancer cells' mobility with fibrin clot and their interactions with fibrin were observed. Cancer cells were treated with paclitaxel (PTX) or 4-hydroxy-tamoxifen (4OHTam) in the presence or absence of FCS. RESULTS: Cancer cells, in presence of PPP, induced fibrin network formation. High TF-expressing cancer cells (BXPC3 and MDA-MB23 cells), led to dense fibrin network with fine fibers. Low TF expressing cells MCF7 led to thick fibers. Exogenous TF enhanced the density of fibrin network formed by MCF7 cells. Cancer cells through their inherent profibrinolytic potential migrated within the fiber scaffold. The BXPC3 and MCF7 cells moved in clusters whereas the MDA-MB231 cells moved individually within the fibrin network. FCS decreased the efficiency of PTX and 4OHTam on the viability of cancer cells. CONCLUSIONS: The procoagulant signature of cancer cells is determinant for the quality and structure of fibrin network in the microenvironment. Original SEM images show the architecture of "bird's nest"-like fibrin network being in touch with the cell membranes and surrounding cancer cells. Fibrin network constructed by triggering thrombin generation by cancer cells, provides a scaffold for cell migration. Fibrin clot shields protect cancer cells against PTX and 4OHTam.
Subject(s)
Antineoplastic Agents , Cell Movement , Fibrin , Tumor Microenvironment , Humans , Cell Movement/drug effects , Fibrin/metabolism , Tumor Microenvironment/drug effects , Antineoplastic Agents/pharmacology , Antineoplastic Agents/therapeutic use , Cell Line, Tumor , Drug Resistance, Neoplasm/drug effects , MCF-7 Cells , Female , Breast Neoplasms/pathology , Breast Neoplasms/metabolism , Breast Neoplasms/drug therapy , Paclitaxel/pharmacology , Paclitaxel/therapeutic use , Pancreatic Neoplasms/pathology , Pancreatic Neoplasms/drug therapy , Pancreatic Neoplasms/metabolism , Blood Coagulation/drug effectsABSTRACT
BACKGROUND: Patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer may have poor prognoses and short overall and disease-free survival. Most previous studies focused on assessing the quality of life and health-state utility of the general population of breast cancer patients. The number of studies for HER2-positive breast cancer patients is negligible. This study investigated the health-state utility and its associated factors among Vietnamese patients with HER2-positive breast cancer. METHODS: We conducted face-to-face interviews with 301 HER2-positive breast cancer patients to collect data. Their health-state utility was measured via the EQ-5D-5L instrument. The Mann-Whitney U and Kruskal-Wallis tests were employed to compare the differences in utility scores between two groups and among three groups or more, respectively. Factors associated with patients' heath-state utility were identified via Tobit regression models. RESULTS: Pain/discomfort (56.1%) and anxiety/depression (39.5%) were the two issues that patients suffered from the most, especially among metastatic breast cancer patients. The severity of distress (depression, anxiety, and stress) in patients was relatively mild. Of 301 patients, their average utility score was 0.86±0.17 (range: 0.03-1.00), and the average EQ-visual analogue scale (VAS) score was 69.12±12.60 (range: 30-100). These figures were 0.79±0.21 and 65.20±13.20 for 102 metastatic breast cancer patients, significantly lower than those of 199 non-metastatic cancer patients (0.89±0.13 and 71.13±11.78) (p<0.001), respectively. Lower health-state utility scores were significantly associated with older age (p = 0.002), lower education level (p = 0.006), lower monthly income (p = 0.036), metastatic cancer (p = 0.001), lower EQ-VAS score (p<0.001), and more severe level of distress (p<0.001). CONCLUSIONS: Our findings showed a significant decrement in utility scores among metastatic breast cancer patients. Patients' health-state utility differed by their demographic characteristics (age, education level, and income) and clinical characteristics (stage of cancer and distress). Their utility scores may support further cost-effectiveness analysis in Vietnam.
Subject(s)
Breast Neoplasms , Quality of Life , Receptor, ErbB-2 , Humans , Breast Neoplasms/pathology , Breast Neoplasms/psychology , Female , Vietnam/epidemiology , Middle Aged , Cross-Sectional Studies , Receptor, ErbB-2/metabolism , Adult , Aged , Depression/epidemiology , Anxiety/epidemiologyABSTRACT
BACKGROUND: Childhood obesity prevention initiatives are complex interventions that aim to improve children's obesity-related behaviors and provide health promoting environments. These interventions often impact individuals, communities, and outcomes not primarily targeted by the intervention or policy. To accurately capture the effectiveness and cost-effectiveness of childhood obesity prevention interventions, an understanding of the broader impacts (or spillover effects) is required. This systematic review aims to assess the spillover effects of childhood obesity prevention interventions. METHODS: Six academic databases and two trial registries were searched (2007-2023) to identify studies reporting quantifiable obesity-related and other outcomes in individuals or communities not primarily targeted by an obesity prevention intervention. Critical appraisal was undertaken for studies that reported statistically significant findings, and a narrative synthesis of the data was undertaken. RESULTS: Twenty academic studies and 41 trial records were included in the synthesis. The most commonly reported spillovers were diet or nutrition-related, followed by BMI and physical activity/sedentary behavior. Spillovers were mostly reported in parents/caregivers followed by other family members. Nine of the 20 academic studies reported statistically significant spillover effects. CONCLUSION: Limited evidence indicates that positive spillover effects of childhood obesity prevention interventions can be observed in parents/caregivers and families of targeted participants.
Subject(s)
Exercise , Pediatric Obesity , Humans , Pediatric Obesity/prevention & control , Child , Health Promotion/methods , DietABSTRACT
The fields of plasmonics and photonic crystals (PCs) have been combined to generate model light-confining Tamm plasmon (TMM) cavities. This approach effectively overcomes the intrinsic limit of diffraction faced by dielectric cavities and mitigates losses associated with the inherent properties of plasmonic materials. In this study, nanoporous anodic alumina PCs, produced by two-step sinusoidal pulse anodization, are used as a model dielectric platform to establish the methodology for tailoring light confinement through TMM resonances. These model dielectric mirrors feature highly organized nanopores and narrow bandwidth photonic stopbands (PSBs) across different positions of the spectrum. Different types of metallic films (gold, silver, and aluminum) were coated on the top of these model dielectric mirrors. By structuring the features of the plasmonic and photonic components of these hybrid structures, the characteristics of TMM resonances were studied to elucidate effective approaches to optimize the light-confining capability of this hybrid TMM model system. Our findings indicate that the coupling of photonic and plasmonic modes is maximized when the PSB of the model dielectric mirror is broad and located within the midvisible region. It was also found that thicker metal films enhance the quality of the confined light. Gas sensing experiments were performed on optimized TMM systems, and their sensitivity was assessed in real time to demonstrate their applicability. Ag films provide superior performance in achieving the highest sensitivity (S = 0.038 ± 0.001 nm ppm-1) based on specific binding interactions between thiol-containing molecules and metal films.
ABSTRACT
PURPOSE: Manual recontouring of targets and Organs At Risk (OARs) is a time-consuming and operator-dependent task. We explored the potential of Generative Adversarial Networks (GAN) to auto-segment the rectum, bladder and femoral heads on 0.35T MRIs to accelerate the online MRI-guided-Radiotherapy (MRIgRT) workflow. METHODS: 3D planning MRIs from 60 prostate cancer patients treated with 0.35T MR-Linac were collected. A 3D GAN architecture and its equivalent 2D version were trained, validated and tested on 40, 10 and 10 patients respectively. The volumetric Dice Similarity Coefficient (DSC) and 95th percentile Hausdorff Distance (HD95th) were computed against expert drawn ground-truth delineations. The networks were also validated on an independent external dataset of 16 patients. RESULTS: In the internal test set, the 3D and 2D GANs showed DSC/HD95th of 0.83/9.72 mm and 0.81/10.65 mm for the rectum, 0.92/5.91 mm and 0.85/15.72 mm for the bladder, and 0.94/3.62 mm and 0.90/9.49 mm for the femoral heads. In the external test set, the performance was 0.74/31.13 mm and 0.72/25.07 mm for the rectum, 0.92/9.46 mm and 0.88/11.28 mm for the bladder, and 0.89/7.00 mm and 0.88/10.06 mm for the femoral heads. The 3D and 2D GANs required on average 1.44 s and 6.59 s respectively to generate the OARs' volumetric segmentation for a single patient. CONCLUSIONS: The proposed 3D GAN auto-segments pelvic OARs with high accuracy on 0.35T, in both the internal and the external test sets, outperforming its 2D equivalent in both segmentation robustness and volume generation time.
Subject(s)
Image Processing, Computer-Assisted , Organs at Risk , Male , Humans , Organs at Risk/diagnostic imaging , Tomography, X-Ray Computed , Pelvis/diagnostic imaging , Magnetic Resonance ImagingABSTRACT
Purpose: Magnetic resonance imaging (MRI)-guided radiotherapy enables adaptive treatment plans based on daily anatomical changes and accurate organ visualization. However, the bias field artifact can compromise image quality, affecting diagnostic accuracy and quantitative analyses. This study aims to assess the impact of bias field correction on 0.35 T pelvis MRIs by evaluating clinical anatomy visualization and generative adversarial network (GAN) auto-segmentation performance. Materials and methods: 3D simulation MRIs from 60 prostate cancer patients treated on MR-Linac (0.35 T) were collected and preprocessed with the N4ITK algorithm for bias field correction. A 3D GAN architecture was trained, validated, and tested on 40, 10, and 10 patients, respectively, to auto-segment the organs at risk (OARs) rectum and bladder. The GAN was trained and evaluated either with the original or the bias-corrected MRIs. The Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD95th) were computed for the segmented volumes of each patient. The Wilcoxon signed-rank test assessed the statistical difference of the metrics within OARs, both with and without bias field correction. Five radiation oncologists blindly scored 22 randomly chosen patients in terms of overall image quality and visibility of boundaries (prostate, rectum, bladder, seminal vesicles) of the original and bias-corrected MRIs. Bennett's S score and Fleiss' kappa were used to assess the pairwise interrater agreement and the interrater agreement among all the observers, respectively. Results: In the test set, the GAN trained and evaluated on original and bias-corrected MRIs showed DSC/HD95th of 0.92/5.63 mm and 0.92/5.91 mm for the bladder and 0.84/10.61 mm and 0.83/9.71 mm for the rectum. No statistical differences in the distribution of the evaluation metrics were found neither for the bladder (DSC: p = 0.07; HD95th: p = 0.35) nor for the rectum (DSC: p = 0.32; HD95th: p = 0.63). From the clinical visual grading assessment, the bias-corrected MRI resulted mostly in either no change or an improvement of the image quality and visualization of the organs' boundaries compared with the original MRI. Conclusion: The bias field correction did not improve the anatomy visualization from a clinical point of view and the OARs' auto-segmentation outputs generated by the GAN.
ABSTRACT
PURPOSE: Gut microbiota play an important role in human health, including cancer. Cancer and its treatment, in turn, may alter the gut microbiome. To understand this complex relationship, we profiled the gut microbiome of 356 Vietnamese patients with breast cancer. MATERIALS AND METHODS: Stool samples were collected before chemotherapy, with 162 pre- and 194 postsurgery. The gut microbiome was measured by shotgun metagenomic sequencing. Associations of gut microbial diversity, taxa abundance, and gut microbiome health index (GMHI) with sociodemographic, clinical factors, and tumor characteristics were evaluated. RESULTS: Postsurgery samples were associated with significantly lower α- and ß-diversities (P < .001) and showed significant differences in the abundance of 15% of 2,864 investigated taxa (false discovery rate [FDR] <0.1) compared with presurgery samples. An unhealthy gut microbiome was prevalent among patients with breast cancer, with a mean GMHI of -0.79 and -2.81 in pre- and postsurgery stool samples, respectively. In an analysis of 162 presurgery stool samples, diagnosis delay was significantly associated with lower α-diversity, variation in ß-diversity, an increased abundance of species Enorma massiliensis, and a decreased abundance of Faecalicoccus pleomorphus. High intake of fiber was significantly associated with lower α-diversity and a higher abundance of species belonging to Bifidobacterium, Prevotella, and Bacteroides gena (FDR < 0.1). We did not find that cancer stage and subtype, menopausal status, comorbidity, antibiotic use during 3 months before stool collection, or physical activity was significantly associated with α- and ß-diversities or GMHI although a few significant differences were observed in taxa abundance. CONCLUSION: Our study revealed that diagnosis delay, high fiber intake, and breast cancer surgery, which is always followed by antibiotic prophylaxis in Vietnam, led to a less diverse and unhealthy gut microbiome among patients with breast cancer.