ABSTRACT
BACKGROUND: Surgery is the primary treatment for non-small cell lung cancer (NSCLC), but microscopic residual disease may be unavoidable. Preclinical studies have shown that volatile anesthetics might suppress host immunity and promote a pro-malignant environment that supports cancer cell proliferation, migration, and angiogenesis, whereas propofol may preserve cell-mediated immunity and inhibit tumor angiogenesis. However, clinical evidence that propofol-based total intravenous anesthesia (TIVA) can reduce tumor recurrence after curative resection remains inconsistent due to the retrospective observational nature of previous studies. Therefore, we will test the hypothesis that the recurrence-free survival (RFS) after curative resection of NSCLC is higher in patients who received TIVA than volatile anesthetics (GAS) in this multicenter randomized trial. METHODS: This double-blind, randomized trial will enroll patients at 22 international sites, subject to study registration, institutional review board approval, and patient written informed consent. Eligible patients are adult patients undergoing lung resection surgery with curative intent for NSCLC. Exclusion criteria will be contraindications to study drugs, American Society of Anesthesiologists physical status IV or higher, or preexisting distant metastasis or malignant tumor in other organs. At each study site, enrolled subjects will be randomly allocated into the TIVA and GAS groups with a 1:1 ratio. This pragmatic trial does not standardize any aspect of patient care. However, potential confounders will be balanced between the study arms. The primary outcome will be RFS. Secondary outcomes will be overall survival and complications within postoperative 7 days. Enrollment of 5384 patients will provide 80% power to detect a 3% treatment effect (hazard ratio of 0.83) at alpha 0.05 for RFS at 3 years. DISCUSSION: Confirmation of the study hypothesis would demonstrate that a relatively minor and low-cost alteration in anesthetic management has the potential to reduce cancer recurrence risk in NSCLC, an ultimately fatal complication. Rejection of the hypothesis would end the ongoing debate about the relationship between cancer recurrence and anesthetic management. TRIAL REGISTRATION: The study protocol was prospectively registered at the Clinical trials ( https://clinicaltrials.gov , NCT06330038, principal investigator: Hyun Joo Ahn; date of first public release: March 25, 2024) before the recruitment of the first participant.
ABSTRACT
OBJECTIVES: To investigate whether ultrafast sequence improves the diagnostic performance of conventional dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in differentiating additional suspicious lesions (ASLs) on preoperative breast MRI. MATERIALS AND METHODS: A retrospective database search identified 668 consecutive patients who underwent preoperative breast DCE-MRI with ultrafast sequence between June 2020 and July 2021. Among these, 107 ASLs from 98 patients with breast cancer (36 multifocal, 42 multicentric, and 29 contralateral) were identified. Clinical, pathological, conventional MRI findings, and ultrafast sequence-derived parameters were collected. A prediction model that adds ultrafast sequence-derived parameters to clinical, pathological, and conventional MRI findings was developed and validated internally. Decision curve analysis and net reclassification index statistics were performed. A nomogram was constructed. RESULTS: The ultrafast model adding time to peak enhancement, time to enhancement, and maximum slope showed a significantly increased area under the receiver operating characteristic curve compared with the conventional model which includes age, human epidermal growth factor receptor 2 expression of index cancer, size of index cancer, lesion type of index cancer, location of ASL, and size of ASL (0.92 vs. 0.82; p = 0.002). The decision curve analysis showed that the ultrafast model had a higher overall net benefit than the conventional model. The net reclassification index of ultrafast model was 23.3% (p = 0.001). CONCLUSION: A combination of ultrafast sequence-derived parameters with clinical, pathological, and conventional MRI findings can aid in the differentiation of ASL on preoperative breast MRI. CLINICAL RELEVANCE STATEMENT: Our prediction model and nomogram that was based on ultrafast sequence-derived parameters could help radiologists differentiate ASLs on preoperative breast MRI. KEY POINTS: Ultrafast MRI can diminish background parenchymal enhancement and possibly improve diagnostic accuracy for additional suspicious lesions (ASLs). Location of ASL, larger size of ASL, and higher maximum slope were associated with malignant ASL. The ultrafast model and nomogram can help preoperatively differentiate additional malignancies.
ABSTRACT
INTRODUCTION: Preparing appropriate red blood cells (RBCs) before surgery is crucial for improving both the efficacy of perioperative workflow and patient safety. In particular, thoracic surgery (TS) is a procedure that requires massive transfusion with high variability for each patient. Hence, the precise prediction of RBC requirements for individual patients is becoming increasingly important. This study aimed to 1) develop and validate a machine learning algorithm for personalized RBC predictions for TS patients and 2) assess the usability of a clinical decision support system (CDSS) integrating this artificial intelligence model. METHODS: Adult patients who underwent TS between January 2016 and October 2021 were included in this study. Multiple models were developed by employing both traditional statistical- and machine-learning approaches. The primary outcome evaluated the model's performance in predicting RBC requirements through root mean square error and adjusted R2. Surgeons and informaticians determined the precision MSBOS-Thoracic Surgery (pMSBOS-TS) algorithm through a consensus process. The usability of the pMSBOS-TS was assessed using the System Usability Scale (SUS) survey with 60 clinicians. RESULTS: We identified 7,843 cases (6,200 for training and 1,643 for test sets) of TSs. Among the models with variable performance indices, the extreme gradient boosting model was selected as the pMSBOS-TS algorithm. The pMSBOS-TS model showed statistically significant lower root mean square error (mean: 3.203 and 95% confidence interval [CI]: 3.186-3.220) compared to the calculated Maximum Surgical Blood Ordering Schedule (MSBOS) and a higher adjusted R2 (mean: 0.399 and 95% CI: 0.395-0.403) compared to the calculated MSBOS, while requiring approximately 200 fewer packs for RBC preparation compared to the calculated MSBOS. The SUS score of the pMSBOS-TS CDSS was 72.5 points, indicating good acceptability. CONCLUSIONS: We successfully developed the pMSBOS-TS capable of predicting personalized RBC transfusion requirements for perioperative patients undergoing TS.
ABSTRACT
BACKGROUND: Risk factors for cardiovascular disease, including elevated blood pressure, are known to increase risk of Alzheimer's disease. There has been increasing awareness of the relationship between long-term blood pressure (BP) patterns and their effects on the brain. We aimed to investigate the association of repeated BP measurements with Alzheimer's and vascular disease markers. METHODS: We recruited 1,952 participants without dementia between August 2015 and February 2022. During serial clinic visits, we assessed both systolic BP (SBP) and diastolic BP (DBP), and visit-to-visit BP variability (BPV) was quantified from repeated measurements. In order to investigate the relationship of mean SBP (or DBP) with Alzheimer's and vascular markers and cognition, we performed multiple linear and logistic regression analyses after controlling for potential confounders (Model 1). Next, we investigated the relationship of with variation of SBP (or DBP) with the aforementioned variables by adding it into Model 1 (Model 2). In addition, mediation analyses were conducted to determine mediation effects of Alzheimer's and vascular makers on the relationship between BP parameters and cognitive impairment. RESULTS: High Aß uptake was associated with greater mean SBP (ß = 1.049, 95% confidence interval 1.016-1.083). High vascular burden was positively associated with mean SBP (odds ratio = 1.293, 95% CI 1.015-1.647) and mean DBP (1.390, 1.098-1.757). High tau uptake was related to greater systolic BPV (0.094, 0.001-0.187) and diastolic BPV (0.096, 0.007-0.184). High Aß uptake partially mediated the relationship between mean SBP and the Mini-Mental State Examination (MMSE) scores. Hippocampal atrophy mediated the relationship between diastolic BPV and MMSE scores. CONCLUSIONS: Each BP parameter affects Alzheimer's and vascular disease markers differently, which in turn leads to cognitive impairment. Therefore, it is necessary to appropriately control specific BP parameters to prevent the development of dementia. Furthermore, a better understanding of pathways from specific BP parameters to cognitive impairments might enable us to select the managements targeting the specific BP parameters to prevent dementia effectively.
Subject(s)
Alzheimer Disease , Blood Pressure , Humans , Female , Male , Alzheimer Disease/physiopathology , Alzheimer Disease/epidemiology , Blood Pressure/physiology , Aged , Middle Aged , Asian People , Biomarkers/blood , Cognitive Dysfunction/physiopathology , Cognitive Dysfunction/epidemiology , Cognitive Dysfunction/etiology , Risk Factors , Hypertension/physiopathology , Hypertension/epidemiologyABSTRACT
PURPOSE: This study compared the diagnostic performance of quantitative ultrasonography (QUS) with that of conventional ultrasonography (US) in assessing hepatic steatosis among individuals undergoing health screening using magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) as the reference standard. METHODS: This single-center prospective study enrolled 427 participants who underwent abdominal MRI and US. Measurements included the attenuation coefficient in tissue attenuation imaging (TAI) and the scatter-distribution coefficient in tissue scatter-distribution imaging (TSI). The correlation between QUS and MRI-PDFF was evaluated. The diagnostic capabilities of QUS, conventional B-mode US, and their combined models for detecting hepatic fat content of ≥5% (MRI-PDFF ≥5%) and ≥10% (MRI-PDFF ≥10%) were compared by analyzing the areas under the receiver operating characteristic curves. Additionally, clinical risk factors influencing the diagnostic performance of QUS were identified using multivariate linear regression analyses. RESULTS: TAI and TSI were strongly correlated with MRI-PDFF (r=0.759 and r=0.802, respectively; both P<0.001) and demonstrated good diagnostic performance in detecting and grading hepatic steatosis. The combination of QUS and B-mode US resulted in the highest areas under the ROC curve (AUCs) (0.947 and 0.975 for detecting hepatic fat content of ≥5% and ≥10%, respectively; both P<0.05), compared to TAI, TSI, or B-mode US alone (AUCs: 0.887, 0.910, 0.878 for ≥5% and 0.951, 0.922, 0.875 for ≥10%, respectively). The independent determinants of QUS included skinliver capsule distance (ß=7.134), hepatic fibrosis (ß=4.808), alanine aminotransferase (ß=0.202), triglyceride levels (ß=0.027), and diabetes mellitus (ß=3.710). CONCLUSION: QUS is a useful and effective screening tool for detecting and grading hepatic steatosis during health checkups.
ABSTRACT
KRAS plays a crucial role in regulating cell survival and proliferation and is one of the most commonly mutated oncogenes in human cancers. The novel KRASG12D inhibitor, MRTX1133, demonstrates promising antitumor efficacy in vitro and in vivo. However, the development of acquired resistance in treated patients presents a considerable challenge to sustained therapeutic effectiveness. In response to this challenge, we conducted site-specific mutagenesis screening to identify potential secondary mutations that could induce resistance to MRTX1133. We screened a range of KRASG12D variants harboring potential secondary mutations, and 44 representative variants were selected for in-depth validation of the pooled screening outcomes. We identified eight variants (G12D with V9E, V9W, V9Q, G13P, T58Y, R68G, Y96W, and Q99L) that exhibited substantial resistance, with V9W showing notable resistance, and downstream signaling analyses and structural modeling were conducted. We observed that secondary mutations in KRASG12D can lead to acquired resistance to MRTX1133 and BI-2865, a novel pan-KRAS inhibitor, in human cancer cell lines. This evidence is critical for devising new strategies to counteract resistance mechanisms and, ultimately, enhance treatment outcomes in patients with KRASG12D-mutant cancers.
Subject(s)
Drug Resistance, Neoplasm , Mutagenesis, Site-Directed , Mutation , Proto-Oncogene Proteins p21(ras) , Humans , Proto-Oncogene Proteins p21(ras)/genetics , Drug Resistance, Neoplasm/genetics , Cell Line, Tumor , Antineoplastic Agents/pharmacology , Cell Proliferation/drug effectsABSTRACT
BACKGROUND: Food allergy (FA) often occurs in early childhood with and without atopic dermatitis (AD). FA can be severe and even fatal. For primary prevention, it is important to find early biomarkers to predict the future onset of FA before any clinical manifestations. OBJECTIVE: Our aim was to find early predictors of future onset of FA in the stratum corneum (SC). METHODS: Skin tape strips were collected from the forearm of newborns (n = 129) at age 2 months, before any signs of clinical FA or AD. Children were clinically monitored until they reached age 2 years to confirm the presence or absence of FA and AD. Skin tape strips were subjected to lipidomic analyses by liquid chromatography-tandem mass spectrometry and cytokine determination by Meso Scale Discovery U-Plex assay. RESULTS: Overall, 9 of 129 infants (7.0%) developed FA alone and 9 of 129 infants (7.0%) developed FA concomitantly with AD. In the stratum corneum of children with future FA and concomitant AD and FA, absolute amounts of unsaturated (N24:1)(C18-sphingosine)ceramide and (N26:1)(C18-sphingosine)ceramide and their relative percentages within the molecular group were increased compared with the amounts and percentages in healthy children, with P values ranging from less than .01 to less than .05 according to ANOVA. The children with future AD had normal levels of these molecules. IL-33 level was upregulated in those infants with future FA but not in those with future AD, whereas thymic stromal lymphopoietin was upregulated in those with future AD but not in those with future FA. Logistic regression analysis revealed strong FA predicting power for the combination of dysregulated lipids and cytokines, with an odds ratio reaching 101.4 (95% CI = 5.4-1910.6). CONCLUSION: Noninvasive skin tape strip analysis at age 2 months can identify infants at risk of FA in the future.
Subject(s)
Biomarkers , Cytokines , Dermatitis, Atopic , Food Hypersensitivity , Humans , Infant , Food Hypersensitivity/immunology , Food Hypersensitivity/diagnosis , Male , Female , Dermatitis, Atopic/immunology , Dermatitis, Atopic/metabolism , Cytokines/metabolism , Infant, Newborn , Skin/immunology , Skin/metabolism , Child, Preschool , Ceramides/metabolism , Ceramides/analysisABSTRACT
Purpose To develop an artificial intelligence (AI) system for humeral tumor detection on chest radiographs (CRs) and evaluate the impact on reader performance. Materials and Methods In this retrospective study, 14 709 CRs (January 2000 to December 2021) were collected from 13 468 patients, including CT-proven normal (n = 13 116) and humeral tumor (n = 1593) cases. The data were divided into training and test groups. A novel training method called false-positive activation area reduction (FPAR) was introduced to enhance the diagnostic performance by focusing on the humeral region. The AI program and 10 radiologists were assessed using holdout test set 1, wherein the radiologists were tested twice (with and without AI test results). The performance of the AI system was evaluated using holdout test set 2, comprising 10 497 normal images. Receiver operating characteristic analyses were conducted for evaluating model performance. Results FPAR application in the AI program improved its performance compared with a conventional model based on the area under the receiver operating characteristic curve (0.87 vs 0.82, P = .04). The proposed AI system also demonstrated improved tumor localization accuracy (80% vs 57%, P < .001). In holdout test set 2, the proposed AI system exhibited a false-positive rate of 2%. AI assistance improved the radiologists' sensitivity, specificity, and accuracy by 8.9%, 1.2%, and 3.5%, respectively (P < .05 for all). Conclusion The proposed AI tool incorporating FPAR improved humeral tumor detection on CRs and reduced false-positive results in tumor visualization. It may serve as a supportive diagnostic tool to alert radiologists about humeral abnormalities. Keywords: Artificial Intelligence, Conventional Radiography, Humerus, Machine Learning, Shoulder, Tumor Supplemental material is available for this article. © RSNA, 2024.
Subject(s)
Artificial Intelligence , Neoplasms , Humans , Retrospective Studies , Humerus/diagnostic imaging , Radiography , RadiopharmaceuticalsABSTRACT
PURPOSE: We aimed to investigate the mediating factors between maternal anxiety and the development of food allergy (FA) in children until 2 years from birth. METHODS: In this longitudinal cohort of 122 mother-child dyads from pregnancy to 24 months of age, we regularly surveyed maternal psychological states, infant feeding data, and allergic symptoms and collected stool samples at 6 months of age for microbiome analysis. Considering the temporal order of data collection, we investigated serial mediating effects and indirect effects among maternal anxiety, dietary diversity (DD), gut microbial diversity, and FA using structural equation modeling. RESULTS: Among the 122 infants, 15 (12.3%) were diagnosed with FA. Increased maternal anxiety between 3 and 6 months after delivery was associated with a lower DD score. Infants with low DD at 4 months showed low gut microbial richness, which was associated with FA development. When the infants were grouped into 4 subtypes, using consensus clustering of 13 gut bacteria significantly associated with maternal anxiety and DD, Prevotella, Eubacterium, Clostridiales and Lachnospiraceae were more abundant in the group with lower FA occurrence. CONCLUSIONS: Postpartum maternal anxiety, mediated by reduced DD and gut microbial diversity, may be a risk factor for the development of FA in infants during the first 2 years of life.
ABSTRACT
Background/Aims: : The effect of proton pump inhibitors (PPIs) on the lower gastrointestinal (GI) tract is uncertain, with potential to worsen damage. This study aimed to find the best method for protecting the entire GI tract from mucosal damage. Methods: : A retrospective cohort study at Samsung Medical Center (2002-2019) included 195,817 patients prescribed GI mucosa-damaging agents. The primary goal was to assess the effectiveness of GI protective agents in preventing significant hemoglobin drops (>2 g/dL), indicating overall GI mucosal damage. Self-controlled case series and landmark analysis were used to address biases in real-world data. Results: : The incidence rate ratios for rebamipide, PPI, and histamine-2 receptor antagonist (H2RA) were 0.34, 0.33, and 0.52, respectively. Rebamipide showed a significantly lower incidence rate than H2RA and was comparable to PPIs. Landmark analysis revealed significant reductions in hemoglobin drop risk with rebamipide and H2RA, but not with PPI. Conclusions: : Rebamipide, like PPIs, was highly effective in preventing blood hemoglobin level decreases, as shown in real-world data. Rebamipide could be a comprehensive strategy for protecting the entire GI tract, especially when considering PPIs' potential side effects on the lower GI tract.
ABSTRACT
BACKGROUND: In this study, the risk of dementia in patients with a history of herpes simplex virus (HSV) or varicella zoster virus (VZV) infection was evaluated. METHODS: This nationwide cohort study used data from the Korean National Health Insurance Service collected between 2006 and 2017. A total of 752,205 subjects ≥ 45 years of age not diagnosed with dementia until 2006 were included. A multivariate Cox regression model, adjusted for age, sex, and other comorbidities, was used to assess the hazard ratio (HR) for dementia based on VZV or HSV infection. The interaction effects of both viral infections were analysed. Viral infections are classified into four categories: eye, central nervous system (CNS), simple, and complicated. The hazard ratio (HR) of viral infection was analysed based on the type of dementia. RESULTS: In multivariable analysis, both HSV and VZV infection were associated with an increased risk of dementia (HR = 1.38, 95% confidence interval, CI:1.33-1.43) and (HR = 1.41, 95% CI:1.37-1.46), respectively. Patients who experienced both HSV and VZV infections were also at an increased risk of dementia (HR = 1.57, 95% CI:1.50-1.63). The co-infection group showed the shortest time from viral infection to dementia diagnosis (4.09 ± 3.02 years). In the subgroup analysis, all types of HSV and VZV infections were associated with an increased risk of dementia compared to the non-infection group. The eye, CNS, and complicated VZV infections were associated with a significantly higher risk than simple VZV infections. There were no significant differences between the subtypes of HSV infection. Furthermore, HSV, VSV, and co-infection were associated with an increased risk of all dementia types, including Alzheimer's disease (AD) and vascular dementia (VD). CONCLUSIONS: Individual HSV and VZV infections were associated with an increased risk of all types of dementia, including AD and VD. Patients co-infected with HSV and VZV, VZV infection in the eye, CNS, or complicated type were more vulnerable to the development of dementia.
Subject(s)
Coinfection , Dementia , Herpes Simplex , Herpes Zoster , Virus Diseases , Humans , Herpesvirus 3, Human , Simplexvirus , Cohort Studies , Retrospective StudiesABSTRACT
Background: To evaluate the efficacy and optimal timing of local treatment in patients with borderline resectable (BR) or locally advanced pancreatic cancer (LAPC) treated with upfront FOLFIRINOX. Method: Between 2015 and 2020, 258 patients with pancreatic ductal adenocarcinoma (PDAC) were analysed. Treatment outcomes were compared between systemic treatment group (ST) and multimodality treatment groups (MT) using Kaplan-Meier curves and log-rank test. The MT were stratified as follows: FOLFIRINOX + radiation therapy (RT) (MT1), FOLFIRINOX + surgical resection (MT2), and FOLFIRINOX + RT + surgical resection (MT3). Results: With median follow-up period of 18 months, the 2-year overall survival (OS) for the ST was 22.0%, and it was significantly worse than MT (MT1, 46.3%; MT2, 65.7% and MT3; 90.2%; P < .001). The 2-year locoregional progression free survival (LRPFS) and overall PFS in ST were 10.7% and 7.0%, which were also significantly lower than those of MT (2-year LRPFS: MT1, 31.8%; MT2, 45.3%; MT3, 81.0%; 2-year overall PFS: MT1, 23.3%; MT2, 35.0%; MT3, 66.3%; P < .001). In time-varying multivariate Cox proportional hazard model, local treatment contributed to better treatment outcomes, with adjusted hazard ratios of 0.568 (95% confidence interval [CI], 0.398-0.811), 0.490 (95% CI, 0.331-0.726), and 0.656 (95% CI, 0.458-0.940) for OS, LRPFS, and overall PFS, respectively. The time window of 11-17 months after FOLFIRINOX appeared to demonstrate the maximal efficacy of local treatments in OS. Conclusions: Adding local treatment in BR/LAPC patients treated with upfront FOLFIRINOX seemed to contribute in improved treatment outcomes, and it showed maximal efficacy in OS when applied 11-17 months after the initiation of FOLFIRINOX. We suggest that administration of sufficient period of upfront FOLFIRINOX may intensify the efficacy of local treatments, and well controlled prospective trials are expected.
ABSTRACT
BACKGROUND: This study aimed to compare the outcomes of proton beam therapy (PBT) and carbon ion radiotherapy (CIRT) by a systematic review and meta-analysis of the existing clinical evidence. METHODS: A systematic literature search was performed to identify studies comparing the clinical outcomes of PBT and CIRT. The included studies were required to report oncological outcomes (local control [LC], progression-free survival [PFS], or overall survival [OS]) or adverse events. RESULTS: Eighteen articles comprising 1857 patients (947 treated with PBT and 910 treated with CIRT) were included in the analysis. The pooled analysis conducted for the overall population yielded average hazard ratios of 0.690 (95% confidence interval (CI), 0.493-0.967, p = 0.031) for LC, 0.952 (95% CI, 0.604-1.500, p = 0.590) for PFS, and 1.183 (0.872-1.607, p = 0.281) for OS with reference to CIRT. The subgroup analyses included patients treated in the head and neck, areas other than the head and neck, and patients with chordomas and chondrosarcomas. These analyses revealed no significant differences in most outcomes, except for LC in the subgroup of patients treated in areas other than the head and neck. Adverse event rates were comparable in both groups, with an odds ratio (OR) of 1.097 (95% CI, 0.744-1.616, p = 0.641). Meta-regression analysis for possible heterogeneity did not demonstrate a significant association between treatment outcomes and the ratio of biologically effective doses between modalities. CONCLUSION: This study highlighted the comparability of PBT and CIRT in terms of oncological outcomes and adverse events.
Subject(s)
Heavy Ion Radiotherapy , Proton Therapy , Humans , Proton Therapy/adverse effects , Proton Therapy/methods , Heavy Ion Radiotherapy/adverse effects , Heavy Ion Radiotherapy/methods , Treatment Outcome , Neoplasms/radiotherapy , Neoplasms/mortality , Chordoma/radiotherapy , Chordoma/mortality , Progression-Free SurvivalABSTRACT
OBJECTIVES: To use clinical, radiographic, and CT radiomics features to develop and validate a preoperative prediction model for the early recurrence of pancreatic cancer. METHODS: We retrospectively analyzed 190 patients (150 and 40 in the development and test cohort from different centers) with pancreatic cancer who underwent pancreatectomy between January 2018 and June 2021. Radiomics, clinical-radiologic (CR), and clinical-radiologic-radiomics (CRR) models were developed for the prediction of recurrence within 12 months after surgery. Performance was evaluated using the area under the curve (AUC), Brier score, sensitivity, and specificity. RESULTS: Early recurrence occurred in 36.7% and 42.5% of the development and test cohorts, respectively (P = 0.62). The features for the CR model included carbohydrate antigen 19-9 > 500 U/mL (odds ratio [OR], 3.60; P = 0.01), abutment to the portal and/or superior mesenteric vein (OR, 2.54; P = 0.054), and adjacent organ invasion (OR, 2.91; P = 0.03). The CRR model demonstrated significantly higher AUCs than the radiomics model in the internal (0.77 vs. 0.73; P = 0.048) and external (0.83 vs. 0.69; P = 0.038) validations. Although we found no significant difference between AUCs of the CR and CRR models (0.83 vs. 0.76; P = 0.17), CRR models showed more balanced sensitivity and specificity (0.65 and 0.87) than CR model (0.41 and 0.91) in the test cohort. CONCLUSIONS: The CRR model outperformed the radiomics and CR models in predicting the early recurrence of pancreatic cancer, providing valuable information for risk stratification and treatment guidance.
Subject(s)
Pancreatic Neoplasms , Radiomics , Humans , Retrospective Studies , Pancreatic Neoplasms/diagnostic imaging , Pancreatic Neoplasms/surgery , Area Under Curve , Tomography, X-Ray ComputedABSTRACT
We aimed to determine the effect of optic disc tilt on deep learning-based optic disc classification. A total of 2507 fundus photographs were acquired from 2236 eyes of 1809 subjects (mean age of 46 years; 53% men). Among all photographs, 1010 (40.3%) had tilted optic discs. Image annotation was performed to label pathologic changes of the optic disc (normal, glaucomatous optic disc changes, disc swelling, and disc pallor). Deep learning-based classification modeling was implemented to develop optic-disc appearance classification models with the photographs of all subjects and those with and without tilted optic discs. Regardless of deep learning algorithms, the classification models showed better overall performance when developed based on data from subjects with non-tilted discs (AUC, 0.988 ± 0.002, 0.991 ± 0.003, and 0.986 ± 0.003 for VGG16, VGG19, and DenseNet121, respectively) than when developed based on data with tilted discs (AUC, 0.924 ± 0.046, 0.928 ± 0.017, and 0.935 ± 0.008). In classification of each pathologic change, non-tilted disc models had better sensitivity and specificity than the tilted disc models. The optic disc appearance classification models developed based all-subject data demonstrated lower accuracy in patients with the appearance of tilted discs than in those with non-tilted discs. Our findings suggested the need to identify and adjust for the effect of optic disc tilt on the optic disc classification algorithm in future development.
Subject(s)
Deep Learning , Eye Abnormalities , Glaucoma , Optic Disk , Male , Humans , Middle Aged , Female , Optic Disk/diagnostic imaging , Optic Disk/pathology , Tomography, Optical Coherence/methods , Eye Abnormalities/pathology , Glaucoma/diagnosis , Glaucoma/pathologyABSTRACT
BACKGROUND: Robust and accurate prediction of severity for patients with COVID-19 is crucial for patient triaging decisions. Many proposed models were prone to either high bias risk or low-to-moderate discrimination. Some also suffered from a lack of clinical interpretability and were developed based on early pandemic period data. Hence, there has been a compelling need for advancements in prediction models for better clinical applicability. OBJECTIVE: The primary objective of this study was to develop and validate a machine learning-based Robust and Interpretable Early Triaging Support (RIETS) system that predicts severity progression (involving any of the following events: intensive care unit admission, in-hospital death, mechanical ventilation required, or extracorporeal membrane oxygenation required) within 15 days upon hospitalization based on routinely available clinical and laboratory biomarkers. METHODS: We included data from 5945 hospitalized patients with COVID-19 from 19 hospitals in South Korea collected between January 2020 and August 2022. For model development and external validation, the whole data set was partitioned into 2 independent cohorts by stratified random cluster sampling according to hospital type (general and tertiary care) and geographical location (metropolitan and nonmetropolitan). Machine learning models were trained and internally validated through a cross-validation technique on the development cohort. They were externally validated using a bootstrapped sampling technique on the external validation cohort. The best-performing model was selected primarily based on the area under the receiver operating characteristic curve (AUROC), and its robustness was evaluated using bias risk assessment. For model interpretability, we used Shapley and patient clustering methods. RESULTS: Our final model, RIETS, was developed based on a deep neural network of 11 clinical and laboratory biomarkers that are readily available within the first day of hospitalization. The features predictive of severity included lactate dehydrogenase, age, absolute lymphocyte count, dyspnea, respiratory rate, diabetes mellitus, c-reactive protein, absolute neutrophil count, platelet count, white blood cell count, and saturation of peripheral oxygen. RIETS demonstrated excellent discrimination (AUROC=0.937; 95% CI 0.935-0.938) with high calibration (integrated calibration index=0.041), satisfied all the criteria of low bias risk in a risk assessment tool, and provided detailed interpretations of model parameters and patient clusters. In addition, RIETS showed potential for transportability across variant periods with its sustainable prediction on Omicron cases (AUROC=0.903, 95% CI 0.897-0.910). CONCLUSIONS: RIETS was developed and validated to assist early triaging by promptly predicting the severity of hospitalized patients with COVID-19. Its high performance with low bias risk ensures considerably reliable prediction. The use of a nationwide multicenter cohort in the model development and validation implicates generalizability. The use of routinely collected features may enable wide adaptability. Interpretations of model parameters and patients can promote clinical applicability. Together, we anticipate that RIETS will facilitate the patient triaging workflow and efficient resource allocation when incorporated into a routine clinical practice.
Subject(s)
Algorithms , COVID-19 , Triage , Humans , Biomarkers , COVID-19/diagnosis , Hospital Mortality , Neural Networks, Computer , Triage/methods , Republic of KoreaABSTRACT
PURPOSE: To investigate the relationship of body composition metrics with bone mineral density (BMD) and trabecular attenuation in a cohort of healthy individuals. METHODS: We retrospectively analyzed data of consecutively evaluated individuals who underwent dual-energy X-ray absorptiometry (DXA) and abdominopelvic computed tomography (CT) on the same day during routine medical check-ups between January 2021 and December 2021. Trabecular attenuation was measured at L1 level, while body composition metrics, including skeletal muscle index (SMI), skeletal muscle attenuation (SMA), visceral fat index (VFI), and subcutaneous fat index (SFI), were measured at L3 level. The association of body composition metrics with BMD and trabecular attenuation was analyzed using partial correlation analysis. RESULTS: A total of 634 patients (median age, 56 years; range 50-62 years; 392 men) were included. In men, the SMI and SMA were positively correlated with BMD and trabecular attenuation, both before (r, 0.157-0.344; p < 0.05) and after (r, 0.103-0.246; p < 0.05) adjusting for age and body mass index. The VFI showed negative correlations with trabecular attenuation in both men (r, -0.170; p = 0.001) and women (r, -0.394; p < 0.001), which remained significant after adjusting for age and body mass index (r, -0.181 to -0.122; p < 0.05). CONCLUSION: Low skeletal muscle mass and attenuation were significantly correlated with low BMD and trabecular attenuation in men. Visceral adiposity was associated with reduced BMD and trabecular attenuation in both men and women, demonstrating a stronger correlation with trabecular attenuation.
Subject(s)
Body Composition , Bone Density , Male , Humans , Female , Middle Aged , Bone Density/physiology , Retrospective Studies , Absorptiometry, Photon/methods , Tomography, X-Ray Computed/methods , Lumbar VertebraeABSTRACT
BACKGROUND: The incidence of colorectal cancer (CRC) and preinvasive CRC (e.g., early colon cancer and advanced adenoma) is gradually increasing in several countries. AIM: To evaluate the trend in incidence of CRC and preinvasive CRC according to the increase in the number of colonoscopies performed in Korea. METHODS: This retrospective cohort study enrolled Korean patients from 2002 to 2020 to evaluate the incidence of CRC and preinvasive CRC, and assess the numbers of diagnostic colonoscopies and colonoscopic polypectomies. Colonoscopy-related complications by age group were also determined. RESULTS: The incidence of CRC showed a rapid increase, then decreased after 2012 in the 50-75 year-age group. During the study period, the rate of incidence of preinvasive CRC increased at a similar level in patients under 50 and 50-75 years of age. Since 2009, the increase has been rapid, showing a pattern similar to the increase in colonoscopies. The rate of colonoscopic polypectomy in patients aged under 50 was similar to the rate in patients over 75 years of age after 2007. The rate of complications after colonoscopy and related deaths within 3 mo was high for those over 75 years of age. CONCLUSION: The diagnosis of preinvasive CRC increased with the increase in the number of colonoscopies performed. As the risk of colonoscopy-related hospitalization and death is high in the elderly, if early lesions at risk of developing CRC are diagnosed and treated under or at the age of 75, colonoscopy-related complications can be reduced for those aged 76 years or over.
ABSTRACT
We aimed to investigate the age-related glaucoma progression rates in myopic normal tension glaucoma (NTG). In this long-term retrospective cohort (7.2 ± 3.5 years), we grouped patients based on their age at initial presentation: group A (age < 30 years, 60 eyes), group B (30 ≤ age < 40, 66 eyes), and group C (40 ≤ age < 50 years, 63 eyes). We used a linear mixed-effects model to estimate retinal nerve fiber layer (RNFL) defect width enlargement rates. Group A showed a significantly faster rate of RNFL defect progression (3.01 ± 1.74°/year) than those of groups B and C (2.05 ± 1.55°/year and 2.06 ± 1.29°/year, P = 0.004 and 0.002). The difference was more marked when calculated for the first 10 years of follow-up in group A, B, and C (3.95 ± 2.70°/year, 2.39 ± 1.64°/year, and 1.98 ± 1.31°/year), and between the periods of age < 30 years, 30 ≤ age < 40 years, and 40 ≤ age < 50 years within group A. This is the first evidence of rapid glaucoma progression in the young adulthood and stabilization in older age in myopic NTG. Clinicians should consider the potentially aggressive course of glaucoma, especially in younger patients with myopic NTG, in contrast to the general slow progression in adulthood.
Subject(s)
Glaucoma , Low Tension Glaucoma , Myopia , Optic Disk , Humans , Young Adult , Adult , Middle Aged , Retrospective Studies , Visual Fields , Intraocular Pressure , Nerve Fibers , Disease Progression , Tomography, Optical CoherenceABSTRACT
BACKGROUND AND OBJECTIVE: We aimed to investigate the association between glycemic variability (GV) and neuroimaging markers of white matter hyperintensities (WMH), beta-amyloid (Aß), brain atrophy, and cognitive impairment. METHODS: This was a retrospective cohort study that included participants without dementia from a memory clinic. They all had Aß PET, brain MRI, and standardized neuropsychological tests and had fasting glucose (FG) levels tested more than twice during the study period. We defined GV as the intraindividual visit-to-visit variability in FG levels. Multivariable linear regression and logistic regression were used to identify whether GV was associated with the presence of severe WMH and Aß uptake with DM, mean FG levels, age, sex, hypertension, and presence of APOE4 allele as covariates. Mediation analyses were used to investigate the mediating effect of WMH and Aß uptake on the relationship between GV and brain atrophy and cognition. RESULTS: Among the 688 participants, the mean age was 72.2 years, and the proportion of female participants was 51.9%. Increase in GV was predictive of the presence of severe WMH (coefficient [95% CI] 1.032 [1.012-1.054]; p = 0.002) and increased Aß uptake (1.005 [1.001-1.008]; p = 0.007). Both WMH and increased Aß uptake partially mediated the relationship between GV and frontal-executive dysfunction (GV â WMH â frontal-executive; direct effect, -0.319 [-0.557 to -0.080]; indirect effect, -0.050 [-0.091 to -0.008]) and memory dysfunction (GV â Aß â memory; direct effect, -0.182 [-0.338 to -0.026]; indirect effect, -0.067 [-0.119 to -0.015]), respectively. In addition, increased Aß uptake completely mediated the relationship between GV and hippocampal volume (indirect effect, -1.091 [-2.078 to -0.103]) and partially mediated the relationship between GV and parietal thickness (direct effect, -0.00101 [-0.00185 to -0.00016]; indirect effect, -0.00016 [-0.00032 to -0.000002]). DISCUSSION: Our findings suggest that increased GV is related to vascular and Alzheimer risk factors and neurodegenerative markers, which in turn leads to subsequent cognitive impairment. Furthermore, GV can be considered a potentially modifiable risk factor for dementia prevention.