Explore the vast range of titles available.

Automatic medical image segmentation plays a critical role in scientific research and medical care. Existing high-performance deep learning methods typically rely on large training datasets with high-quality manual annotations, which are difficult to obtain in many clinical applications. Here, we introduce Annotation-effIcient Deep lEarning (AIDE), an open-source framework to handle imperfect training datasets. Methodological analyses and empirical evaluations are conducted, and we demonstrate that AIDE surpasses conventional fully-supervised models by presenting better performance on open datasets possessing scarce or noisy annotations. We further test AIDE in a real-life case study for breast tumor segmentation. Three datasets containing 11,852 breast images from three medical centers are employed, and AIDE, utilizing 10% training annotations, consistently produces segmentation maps comparable to those generated by fully-supervised counterparts or provided by independent radiologists. The 10-fold enhanced efficiency in utilizing expert labels has the potential to promote a wide range of biomedical applications. Existing high-performance deep learning methods typically rely on large training datasets with high-quality manual annotations, which are difficult to obtain in many clinical applications. Here, the authors introduce an open-source framework to handle imperfect training datasets.

Medical imaging can assess the tumor and its environment in their entirety, which makes it suitable for monitoring the temporal and spatial characteristics of the tumor. Progress in computational methods, especially in artificial intelligence for medical image process and analysis, has converted these images into quantitative and minable data associated with clinical events in oncology management. This concept was first described as radiomics in 2012. Since then, computer scientists, radiologists, and oncologists have gravitated towards this new tool and exploited advanced methodologies to mine the information behind medical images. On the basis of a great quantity of radiographic images and novel computational technologies, researchers developed and validated radiomic models that may improve the accuracy of diagnoses and therapy response assessments. Here, we review the recent methodological developments in radiomics, including data acquisition, tumor segmentation, feature extraction, and modelling, as well as the rapidly developing deep learning technology. Moreover, we outline the main applications of radiomics in diagnosis, treatment planning and evaluations in the field of oncology with the aim of developing quantitative and personalized medicine. Finally, we discuss the challenges in the field of radiomics and the scope and clinical applicability of these methods.

Breast cancer remains a global challenge, causing over 600,000 deaths in 2018 (ref. 1 ). To achieve earlier cancer detection, health organizations worldwide recommend screening mammography, which is estimated to decrease breast cancer mortality by 20–40% (refs. 2 , 3 ). Despite the clear value of screening mammography, significant false positive and false negative rates along with non-uniformities in expert reader availability leave opportunities for improving quality and access 4 , 5 . To address these limitations, there has been much recent interest in applying deep learning to mammography 6 – 18 , and these efforts have highlighted two key difficulties: obtaining large amounts of annotated training data and ensuring generalization across populations, acquisition equipment and modalities. Here we present an annotation-efficient deep learning approach that (1) achieves state-of-the-art performance in mammogram classification, (2) successfully extends to digital breast tomosynthesis (DBT; ‘3D mammography’), (3) detects cancers in clinically negative prior mammograms of patients with cancer, (4) generalizes well to a population with low screening rates and (5) outperforms five out of five full-time breast-imaging specialists with an average increase in sensitivity of 14%. By creating new ‘maximum suspicion projection’ (MSP) images from DBT data, our progressively trained, multiple-instance learning approach effectively trains on DBT exams using only breast-level labels while maintaining localization-based interpretability. Altogether, our results demonstrate promise towards software that can improve the accuracy of and access to screening mammography worldwide. A generalizable and interpretable artificial-intelligence system achieves clinical accuracy for screening and early breast-cancer detection on 2D and 3D mammograms.

Distant metastasis (DM) is the main cause of treatment failure in locally advanced rectal cancer. Adjuvant chemotherapy is usually used for distant control. However, not all patients can benefit from adjuvant chemotherapy, and particularly, some patients may even get worse outcomes after the treatment. We develop and validate an MRI-based radiomic signature (RS) for prediction of DM within a multicenter dataset. The RS is proved to be an independent prognostic factor as it not only demonstrates good accuracy for discriminating patients into high and low risk of DM in all the four cohorts, but also outperforms clinical models. Within the stratified analysis, good chemotherapy efficacy is observed for patients with pN2 disease and low RS, whereas poor chemotherapy efficacy is detected in patients with pT1–2 or pN0 disease and high RS. The RS may help individualized treatment planning to select patients who may benefit from adjuvant chemotherapy for distant control. Distant metastasis (DM) is the main cause of treatment failure in locally advanced rectal cancer. Here, the authors developed and validated a radiomic signature (RS) for prediction of DM within a multicenter dataset, and suggest that it may help with stratification of patients who might benefit from adjuvant chemotherapy for DM.

PurposeThis study was conducted in order to investigate the value of magnetic resonance imaging (MRI)-based radiomics signatures for the preoperative prediction of hepatocellular carcinoma (HCC) grade.MethodsData from 170 patients confirmed to have HCC by surgical pathology were divided into a training group (n = 125) and a test group (n = 45). The radiomics features of tumours based on both T1-weighted imaging (WI) and T2WI were extracted by using Matrix Laboratory (MATLAB), and radiomics signatures were generated using the least absolute shrinkage and selection operator (LASSO) logistic regression model. The predicted values of pathological HCC grades using radiomics signatures, clinical factors (including age, sex, tumour size, alpha fetoprotein (AFP) level, history of hepatitis B, hepatocirrhosis, portal vein tumour thrombosis, portal hypertension and pseudocapsule) and the combined models were assessed.ResultsRadiomics signatures could successfully categorise high-grade and low-grade HCC cases (p < 0.05) in both the training and test datasets. Regarding the performances of clinical factors, radiomics signatures and the combined clinical and radiomics signature (from the combined T1WI and T2WI images) models for HCC grading prediction, the areas under the curve (AUCs) were 0.600, 0.742 and 0.800 in the test datasets, respectively. Both the AFP level and radiomics signature were independent predictors of HCC grade (p < 0.05).ConclusionsRadiomics signatures may be important for discriminating high-grade and low-grade HCC cases. The combination of the radiomics signatures with clinical factors may be helpful for the preoperative prediction of HCC grade.Key Points• The radiomics signature based on non-contrast-enhanced MR images was significantly associated with the pathological grade of HCC.• The radiomics signatures based on T1WI or T2WI images performed similarly at predicting the pathological grade of HCC.• Combining the radiomics signature and clinical factors (including age, sex, tumour size, AFP level, history of hepatitis B, hepatocirrhosis, portal vein tumour thrombosis, portal hypertension and pseudocapsule) may be helpful for the preoperative prediction of HCC grade.

Globally, vaccine hesitancy is a growing public health problem. It is detrimental to the consolidation of immunization program achievements and elimination of vaccine-targeted diseases. The objective of this study was to estimate the prevalence of COVID-19 vaccine hesitancy in China and explore its contributing factors. A national cross-sectional online survey among Chinese adults (≥18 years old) was conducted between August 6, 2021 and August 9 via a market research company. We collected sociodemographic information; lifestyle behavior; quality of life; the knowledge, awareness, and behavior of COVID-19; the knowledge, awareness, and behavior of COVID-19 vaccine; willingness of COVID-19 vaccination; accessibility of COVID-19 vaccination services; skepticism about COVID-19 and COVID-19 vaccine; doctor and vaccine developer scale; and so on. Odds ratios (OR) with 95% confidence intervals (CI) were used to estimate the associations by using logistic regression models. A total of 29,925 residents (48.64% men) were enrolled in our study with mean age of 30.99 years. We found an overall prevalence of COVID-19 vaccine hesitancy at 8.40% (95% CI, 8.09–8.72) in primary vaccination and 8.39% (95% CI, 8.07–8.70) in booster vaccination. In addition, after adjusting for potential confounders, we found that women, higher educational level, married residents, higher score of health condition, never smoked, increased washing hands, increased wearing mask, increased social distance, lower level of vaccine conspiracy beliefs, disease risks outweigh vaccine risk, higher level of convenient vaccination, and higher level of trust in doctor and developer were more willing to vaccinate than all others (all p < 0.05). Age, sex, educational level, marital status, chronic disease condition, smoking, healthy behaviors, the curability of COVID-19, the channel of accessing information of COVID-19 vaccine, endorsement of vaccine conspiracy beliefs, weigh risks of vaccination against risks of the disease, making a positive influence on the health of others around you, and lower trust in healthcare system may affect the variation of willingness to take a COVID-19 vaccine (all  p  < 0.05). The prevalence of COVID-19 vaccine hesitancy was modest in China, even with the slight resulting cascade of changing vaccination rates between the primary and booster vaccination. Urgent action to address vaccine hesitancy is needed in building trust in medical personnel and vaccine producers, promoting the convenience of vaccination services, and spreading reliable information of COVID-19 vaccination via the Internet and other media.

Objectives To compare the value of amide proton transfer-weighted imaging (APTWI) and diffusion kurtosis imaging (DKI) in differentiating benign and malignant breast lesions and analyze the correlations between the derived parameters and prognostic factors of breast cancer. Methods One hundred thirty-five women underwent breast APTWI and DKI. The magnetization transfer ratio asymmetry (MTRasym (3.5 ppm)), apparent kurtosis coefficient ( K app ), and non-Gaussian diffusion coefficient ( D app ) were calculated according to the histological subtype, grade, and prognostic factors (Ki-67, estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor-2 (HER-2), lymph node metastasis, and maximum lesion diameter). The differences, efficacy, and correlation between the parameters were determined. Results The K app value was higher and the D app and MTRasym (3.5 ppm) values were lower in the malignant group than in the benign group (all p < 0.001; AUC ( K app ) = 0.913, AUC ( D app ) = 0.910, and AUC (MTRasym (3.5 ppm)) = 0.796). The differences in the AUC between K app and MTRasym (3.5 ppm) and between D app and MTRasym (3.5 ppm) were significant ( p = 0.023, 0.046). K app was moderately correlated with the pathological grade (| r | = 0.724) and mildly correlated with Ki-67 and HER-2 expression (| r | = 0.454, 0.333). D app was moderately correlated with the pathological grade (| r | = 0.648) and mildly correlated with Ki-67 expression (| r | = 0.400). MTRasym (3.5 ppm) was only mildly correlated with the pathological grade (| r | = 0.468). Conclusion DKI is superior to APTWI in differentiating between benign and malignant breast lesions. Each parameter is correlated with some prognostic factors to a certain extent. Key Points • DKI and APTWI provide valuable information regarding lesion characterization. • K app , D app , and MTRasym (3.5 ppm) are valid parameters for the characterization of tissue microstructure. • DKI is superior to APTWI in the study of breast cancer.

•Graph-based learning on individual amyloid-PET scans provides meaningful insights into Alzheimer’s disease (AD) staging and prediction.•MR-guided graph learning framework effectively leverages individual graph features from 18F-florbetapir PET, outperforming composite SUVR in staging with a novel topological biomarker.•This interpretable approach enhances conventional amyloid-PET quantification, supporting early AD detection and timely intervention. Subtle structural and molecular brain changes make noninvasive early detection and staging of Alzheimer's disease (AD) challenging and critical for effective intervention. This study develops a novel graph convolutional network (GCN) learning framework that integrates amyloid-β PET imaging and MRI structural features, aiming for improved early detection and accurate staging of AD. The retrospective study utilized 18F-florbetapir PET scans from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) as the training dataset (323 scans from 196 subjects - 45 normal control, 80 mild cognitive impairment/MCI, 71 AD) and two independent datasets for testing (99 scans from 85 subjects - 31 normal control, 15 MCI, 44 AD). Individual brain graphs were constructed for each PET scan, and graph learning framework was designed to extract molecular features from PET while integrating structural features from MRI. Performance was evaluated using receiver operating characteristic (ROC) analysis, comparing results against cortical SUVR. Additionally, a biomarker GCN_score was defined based on identified salient regions-of-interest, with its effectiveness assessed using the Kruskal-Wallis test and Cohen's effect size. The framework achieved AUCs of 89.8 % (specificity 83.6 %, sensitivity 81.6 %) for distinguishing MCI from normal controls and 88.3 % (specificity 81.6 %, sensitivity 80.6 %) for MCI from AD in the ADNI dataset, with comparable performance in external testing. All results significantly outperformed cortical SUVR (DeLong test p < 0.001). The GCN_score demonstrated superior group differentiation (Cohen's effect sizes 1.744 and 1.32) compared to cortical SUVR (0.309 and 0.641). The proposed graph-based learning framework effectively integrates PET and MRI features for accurate AD stage distinction, showing significant promise for early detection and facilitating timely intervention. [Display omitted]

ObjectiveTo retrospectively analyze the chest computed tomography (CT) features in patients with coronavirus disease 2019 (COVID-19) pneumonia.MethodsFrom January 9, 2020, to February 26, 2020, totally 56 laboratory-confirmed patients with COVID-19 underwent chest CT. For 40 patients, follow-up CT scans were obtained. The CT images were evaluated for the number, type and distribution of the opacity, and the affected lung lobes. Furthermore, the initial CT scan and the follow-up CT scans were compared.ResultsForty patients (83.6%) had two or more opacities in the lung. Eighteen (32.7%) patients had only ground-glass opacities; twenty-nine patients (52.7%) had ground-glass and consolidative opacities; and eight patients (14.5%) had only consolidation. A total of 43 patients (78.2%) showed two or more lobes involved. The opacities tended to be both in peripheral and central (30/55, 54.5%) or purely peripheral distribution (25/55, 45.5%). Fifty patients (90.9%) had the lower lobe involved. The first follow-up CT scans showed that twelve patients (30%) had improvement, 26 (65%) patients had mild-moderate progression, and two patients (5%) had severe progression with “white lungs.” The second follow-up CT showed that 22 patients (71%) showed improvement compared with the first follow-up CT, four patients (12.9%) had aggravated progression, and five patients (16.1%) showed unchanged radiographic appearance.ConclusionsThe common CT features of COVID-19 pneumonia are multiple lung opacities, multiple types of the opacity (ground-glass, ground-glass and consolidation, and consolidation alone), and multiple lobes especially the lower lobe involved. Follow-up CT could demonstrate the rapid progression of COVID-19 pneumonia (either in aggravation or absorption).Key Points• The predominant CT features of COVID-19 pneumonia are multiple ground-glass opacities with or without consolidation and, with both lungs, multiple lobes and especially the lower lobe affected.• CT plays a crucial role in early diagnosis and assessment of COVID-19 pneumonia progression.• CT findings of COVID-19 pneumonia may not be consistent with the clinical symptoms or the initial RT-PCR test results.

Previous reviews have been conducted to evaluate the association between maternal use of folic acid supplements during pregnancy and risk of autism spectrum disorders (ASD) in children, with no definitive conclusion. We therefore conducted a more comprehensive meta-analysis to reassess the relationship between folic acid and the risk of ASD. The electronic databases PubMed, Web of Knowledge, and Wanfang Data were carefully searched to find eligible studies as recent as March 2017. A random effects model was used to combine the relative risk (RR) with 95% confidence intervals (CI). Sensitivity analysis and publication bias were conducted. A total of 12 articles with 16 studies comprising 4514 ASD cases were included in this report. It was found that supplementation with folic acid during pregnancy could reduce the risk of ASD [RR = 0.771, 95% CI = 0.641–0.928, I 2  = 59.7%, P heterogeneity  = 0.001] as compared to those women without folic acid supplementation. The associations were significant among Asian, European, and American populations. In summary, this comprehensive meta-analysis suggested that maternal use of folic acid supplements during pregnancy could significantly reduce the risk of ASD in children regardless of ethnicity, as compared to those women who did not supplement with folic acid.