Explore the vast range of titles available.

Recently, multi-modal vision-language foundation models have gained significant attention in the medical field. While these models offer great opportunities, they still face crucial challenges, such as the requirement for fine-grained knowledge understanding in computer-aided diagnosis and the capability of utilizing very limited or even no task-specific labeled data in real-world clinical applications. In this study, we present MaCo, a masked contrastive chest X-ray foundation model that tackles these challenges. MaCo explores masked contrastive learning to simultaneously achieve fine-grained image understanding and zero-shot learning for a variety of medical imaging tasks. It designs a correlation weighting mechanism to adjust the correlation between masked chest X-ray image patches and their corresponding reports, thereby enhancing the model’s representation learning capabilities. To evaluate the performance of MaCo, we conducted extensive experiments using 6 well-known open-source X-ray datasets. The experimental results demonstrate the superiority of MaCo over 10 state-of-the-art approaches across tasks such as classification, segmentation, detection, and phrase grounding. These findings highlight the significant potential of MaCo in advancing a wide range of medical image analysis tasks. Multi-modal foundation models are increasingly important in medical applications. Here, authors show a masked contrastive chest X-ray model that achieves fine-grained image understanding and zero-shot capabilities, outperforming existing methods

Pre-training lays the foundation for recent successes in radiograph analysis supported by deep learning. It learns transferable image representations by conducting large-scale fully- or self-supervised learning on a source domain; however, supervised pre-training requires a complex and labour-intensive two-stage human-assisted annotation process, whereas self-supervised learning cannot compete with the supervised paradigm. To tackle these issues, we propose a cross-supervised methodology called reviewing free-text reports for supervision (REFERS), which acquires free supervision signals from the original radiology reports accompanying the radiographs. The proposed approach employs a vision transformer and is designed to learn joint representations from multiple views within every patient study. REFERS outperforms its transfer learning and self-supervised learning counterparts on four well-known X-ray datasets under extremely limited supervision. Moreover, REFERS even surpasses methods based on a source domain of radiographs with human-assisted structured labels; it therefore has the potential to replace canonical pre-training methodologies. To train machine learning models for medical imaging, large amounts of training data are needed. Zhou and colleagues instead propose a method of weak supervision which uses the information of radiology reports to learn visual features without the need for expert labelling.

Multi-modal image analysis using deep learning (DL) lays the foundation for neoadjuvant treatment (NAT) response monitoring. However, existing methods prioritize extracting multi-modal features to enhance predictive performance, with limited consideration on real-world clinical applicability, particularly in longitudinal NAT scenarios with multi-modal data. Here, we propose the Multi-modal Response Prediction (MRP) system, designed to mimic real-world physician assessments of NAT responses in breast cancer. To enhance feasibility, MRP integrates cross-modal knowledge mining and temporal information embedding strategy to handle missing modalities and remain less affected by different NAT settings. We validated MRP through multi-center studies and multinational reader studies. MRP exhibited comparable robustness to breast radiologists, outperforming humans in predicting pathological complete response in the Pre-NAT phase (ΔAUROC 14% and 10% on in-house and external datasets, respectively). Furthermore, we assessed MRP’s clinical utility impact on treatment decision-making. MRP may have profound implications for enrolment into NAT trials and determining surgery extensiveness. Deep learning for medical image analysis is a promising new avenue to predict treatment response, however the clinical application of these methods has been so far limited. Here, the authors propose a model to predict chemotherapy response in breast cancer in real world clinical settings.

Background Studies on risk factors for carbapenem-resistant Klebsiella pneumoniae (CRKP) infection have provided inconsistent results, partly due to the choice of the control group. We conducted a systematic review and meta-analysis to assess the risk factors for CRKP infection by comparing CRKP-infected patients with two types of controls: patients infected with carbapenem-susceptible Klebsiella pneumoniae (comparison 1) or patients not infected with CRKP (comparison 2). Methods Data on potentially relevant risk factors for CRKP infection were extracted from studies indexed in PubMed, EMBASE, Web of Science or EBSCO databases from January 1996 to April 2019, and meta-analyzed based on the outcomes for each type of comparison. Results The meta-analysis included 18 studies for comparison 1 and 14 studies for comparison 2. The following eight risk factors were common to both comparisons: admission to intensive care unit (ICU; odds ratio, OR comparison 1  = 3.20, OR comparison 2  = 4.44), central venous catheter use (2.62, 3.85), mechanical ventilation (2.70, 4.78), tracheostomy (2.11, 8.48), urinary catheter use (1.99, 0.27), prior use of antibiotic (6.07, 1.61), exposure to carbapenems (4.16, 3.84) and exposure to aminoglycosides (1.85, 1.80). Another 10 risk factors were unique to comparison 1: longer length of hospital stay (OR = 15.28); prior hospitalization (within the previous 6 months) (OR = 1.91); renal dysfunction (OR = 2.17); neurological disorders (OR = 1.52); nasogastric tube use (OR = 2.62); dialysis (OR = 3.56); and exposure to quinolones (OR = 2.11), fluoroquinolones (OR = 2.03), glycopeptides (OR = 3.70) and vancomycin (OR = 2.82). Conclusions Eighteen factors may increase the risk of carbapenem resistance in K. pneumoniae infection; eight factors may be associated with both K. pneumoniae infections in general and CRKP in particular. The eight shared factors are likely to be ‘true’ risk factors for CRKP infection. Evaluation of risk factors in different situations may be helpful for empirical treatment and prevention of CRKP infections.

Glioma is the most common primary intracranial tumors. Although great achievements have been made in the treatment, the efficacy is still unsatisfactory, which imposes a hefty burden on patients and society. Therefore, the exploration of new and effective anti-glioma drugs is urgent. Human glioma cell lines U251 and LN229 were included in the study. Cell proliferation was detected by cell counting kit-8 (CCK8), plate clone formation assay, EdU incorporation assay and xCELLigence real-time cell analyzer. Cell apoptosis was evaluated by TUNEL assay and flow cytometry. Then, transwell assay was used for assessing the migration. Moreover, tumor xenograft model was established to examine the effect of scutellarin (SCU) and lidocaine on the growth of glioma in vivo. Lastly, western blot was performed to detect the protein level of epidermal growth factor receptor (EGFR). In present study, we found that SCU and lidocaine suppressed the proliferation and migration, and induced the apoptosis of human glioma cell lines, including U251 and LN229 cells, in a dose-dependent manner in vitro. Moreover, the combination of SCU and lidocaine further restrained the proliferation and migration ability of U251 and LN229 cells, while induced their apoptosis in vitro. Additionally, SCU and lidocaine also inhibited the growth of glioma in vivo, and the effect of the combination was better. Above all, the toxicity of SCU and its combination with lidocaine was low to normal astrocytes and neurons. Mechanistically, the effect of SCU and its combination with lidocaine on glioma cells was partially associated with the repression of EGFR signaling. Scutellarin and lidocaine exerted a synergistic effect on suppressing the proliferation and migration and inducing the apoptosis of glioma cells, which was partly associated with the repression of EGFR signaling.

Biological age is an important indicator of organ functions and health. Although mammograms are widely used in breast cancer screening, the potential of mammogram-based biological age predictors remains underexplored. Here, we propose a deep learning model to estimate the biological age of the breast using healthy mammograms. The model is developed on three large datasets and externally validated on two additional datasets, encompassing 95,826 mammograms from 44,497 women aged 18 to 98 years. It demonstrates accurate age estimation (mean absolute error: 4.2 - 6.1 years) with strong correlation to chronological age. Predicted breast age stratifies breast cancer risk similarly to chronological age. Occlusion analysis, employed for model interpretation, reveals the aging-related pattern of the breast. The breast age gap (the difference between system-bias-corrected breast age and chronological age) may reflect breast health status. Breast cancer patients show higher breast age gaps than the healthy population. In two longitudinal datasets, larger breast age gaps are associated with increased future breast cancer risk, with hazard ratios of 1.013 - 1.022. Furthermore, we finetune the model specifically for downstream breast cancer diagnosis and risk prediction. Our approach outperforms other comparative methods, showing its potential for supporting both early detection and personalized screening strategies.

Lung adenocarcinoma (LUAD) remains one of the most aggressive tumors and the efficacy of conventional treatment has been bleak. Nowadays, gene-targeted therapy has become a new favorite in tumor therapy. Herein, we investigated the effect of platelet derived growth factor BB (PDGFBB) on LUAD. Firstly, PDGFBB was upregulated in LUAD patients and closely linked with poor survival. Furthermore, the expression of PDGFBB and PDGFRα/β in LUAD cells was higher than that in normal lung cells. By loss-of-function with herpes simplex virus (HSV)-PDGFi-shRNA, we found that PDGFBB knockdown caused a significant decrease in proliferation and migration, but evoked apoptosis of LUAD cells in vitro. Conversely, exogenous PDGFBB held adverse effect. Additionally, A549 cells with PDGFBB knockdown had a low probability of tumorigenesis in vivo. Moreover, PDGFBB knockdown restrained the growth of xenografts derived from normal A549 cells. Mechanistically, PDGFBB knockdown suppressed PI3K/AKT and Ras/MAPK signaling, while PDGFBB was the opposite. Therefore, we concluded that PDGFBB might facilitate the tumorigenesis and malignancy of LUAD through its functional downstream nodes—PI3K/AKT and Ras/MAPK signaling, which supported that PDGFBB could serve as a rational therapeutic target for LUAD.

11β-Hydroxysteroid dehydrogenase 1 (11β-HSD1) activates glucocorticoid locally in liver and fat tissues to aggravate metabolic syndrome. 11β-HSD1 selective inhibitor can be used to treat metabolic syndrome. Curcumin and its derivatives as selective inhibitors of 11β-HSD1 have not been reported. Curcumin and its 12 derivatives were tested for their potencies of inhibitory effects on human and rat 11β-HSD1 with selectivity against 11β-HSD2. 200 mg/kg curcumin was gavaged to adult male Sprague-Dawley rats with high-fat-diet-induced metabolic syndrome for 2 months. Curcumin exhibited inhibitory potency against human and rat 11β-HSD1 in intact cells with IC50 values of 2.29 and 5.79 µM, respectively, with selectivity against 11β-HSD2 (IC50, 14.56 and 11.92 µM). Curcumin was a competitive inhibitor of human and rat 11β-HSD1. Curcumin reduced serum glucose, cholesterol, triglyceride, low density lipoprotein levels in high-fat-diet-induced obese rats. Four curcumin derivatives had much higher potencies for Inhibition of 11β-HSD1. One of them is (1E,4E)-1,5-bis(thiophen-2-yl) penta-1,4-dien-3-one (compound 6), which had IC50 values of 93 and 184 nM for human and rat 11β-HSD1, respectively. Compound 6 did not inhibit human and rat kidney 11β-HSD2 at 100 µM. In conclusion, curcumin is effective for the treatment of metabolic syndrome and four novel curcumin derivatives had high potencies for inhibition of human 11β-HSD1 with selectivity against 11β-HSD2.

Background Since the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection exhibits multi-organ damage with diverse complications, the correlation between age, gender, medical history and clinical manifestations of novel coronavirus disease 2019 (COVID-19) patients was investigated. Methods 1640 patients who were infected with SARS-CoV-2 and hospitalized at the First Affiliated Hospital of Ningbo University from 22 December 2022 to 1 March 2023 were categorized and analysed. Normal distribution test and variance homogeneity test were performed. Based on the test results, one-way analysis of variance, Pearson's chi-squared test and logistic regression analysis were conducted in the study. Results According to the ANOVA, there was a significant difference in the age distribution (P = .001) between different clinical presentations, while gender did not (P = .06). And regression analysis showed that age, hypertension, atherosclerosis and cancer were significant hazard factors for the development of predominant clinical manifestations in patients hospitalized with novel COVID-19. Additionally, infection with SARS-CoV-2 has the potential to exacerbate the burden on specific diseased or related organs. Conclusion The elderly who are infected with SARS-CoV-2 ought to be treated with emphasis not only on antiviral therapy but also on individualized treatment that takes their medical history and comorbidities into account.

Background Here we describe the treatments and prognosis for metachronous metastases from gastroenteropancreatic neuroendocrine tumors (GEP-NETs) after initial R0 surgical resection at a large center in China. Methods The clinicopathological data and survival outcomes for 108 patients (median age, 54.0 years) with metachronous hepatic metastatic GEP-NETs disease who were initially treated using R0 surgical resection between August 2003 and July 2014 were analyzed using one-way comparisons, survival analysis, and a predictive nomogram. Results Fifty-five (50.9%) patients had pancreatic NETs and 92 (85.2%) had G2 primary tumors. For treatment of the hepatic metastases, 48 (44.4%) patients received liver-directed local treatment (metastasectomy, radiofrequency ablation, transcatheter arterial chemoembolization, etc.), 15 (13.9%) received systemic treatment (interferon, somatostatin analogs, etc.), and 45 (41.7%) received both treatments. Multivariable analyses revealed that OS was associated with hepatic tumor number ( P  < 0.001), treatment modality ( P  = 0.045), and elevated Ki-67 index between the metastatic and primary lesions ( P  = 0.027). The predictive nomogram C-index was 0.63. Conclusions A higher Ki-67 index in metastases compared to primary tumor was an independent factor for poor prognosis. Local treatment was associated with prolonged survival of hepatic metastatic GEP-NET patients. Optimal treatment strategies based on clinicopathological characteristics should be developed.