Early allograft dysfunction (EAD) significantly affects liver transplantation prognosis. This study evaluated the effectiveness of artificial intelligence (AI)-assisted methods in accurately diagnosing EAD and identifying its causes. The primary metric for assessing the accuracy was the area under the receiver operating characteristic curve (AUC). Accuracy, sensitivity, and specificity were calculated and analyzed to compare the performance of the AI models with each other and with radiologists. EAD classification followed the criteria established by Olthoff et al. A total of 582 liver transplant patients who underwent transplantation between December 2012 and June 2021 were selected. Among these, 117 patients (mean age 33.5 ± 26.5 years, 80 men) were evaluated. The ultrasound parameters, images, and clinical information of patients were extracted from the database to train the AI model. The AUC for the ultrasound-spectrogram fusion network constructed from four ultrasound images and medical data was 0.968 (95%CI: 0.940, 0.991), outperforming radiologists by 30% for all metrics. AI assistance significantly improved diagnostic accuracy, sensitivity, and specificity (P < 0.050) for both experienced and less-experienced physicians. EAD lacks efficient diagnosis and causation analysis methods. The integration of AI and ultrasound enhances diagnostic accuracy and causation analysis. By modeling only images and data related to blood flow, the AI model effectively analyzed patients with EAD caused by abnormal blood supply. Our model can assist radiologists in reducing judgment discrepancies, potentially benefitting patients with EAD in underdeveloped regions. Furthermore, it enables targeted treatment for those with abnormal blood supply.

Automatic radiology report generation can enhance diagnostic efficiency and accuracy. However, clean open-source imaging scan-report pairs are limited in scale and variety. Moreover, the vast amount of radiological texts available online is often too noisy to be directly employed. To address this challenge, we introduce a novel task called Noisy Report Refinement (NRR), which generates radiology reports from noisy free-texts. To achieve this, we propose a report refinement pipeline that leverages large language models (LLMs) enhanced with guided self-critique and report selection strategies. To address the inability of existing radiology report generation metrics in measuring cleanliness, radiological usefulness, and factual correctness across various modalities of reports in NRR task, we introduce a new benchmark, NRRBench, for NRR evaluation. This benchmark includes two online-sourced datasets and four clinically explainable LLM-based metrics: two metrics evaluate the matching rate of radiology entities and modality-specific template attributes respectively, one metric assesses report cleanliness, and a combined metric evaluates overall NRR performance. Experiments demonstrate that guided self-critique and report selection strategies significantly improve the quality of refined reports. Additionally, our proposed metrics show a much higher correlation with noisy rate and error count of reports than radiology report generation metrics in evaluating NRR.

Breast cancer poses a serious threat to women's health, and its early detection is crucial for enhancing patient survival rates. While deep learning has significantly advanced mammographic image analysis, existing methods struggle to balance between view consistency with input adaptability. Furthermore, current models face challenges in accurately capturing multi-scale features, especially when subtle lesion variations across different scales are involved. To address this challenge, this paper proposes a Hybrid View Learning (HVL) paradigm that unifies traditional Single-View and Multi-View Learning approaches. The core component of this paradigm, our Attention-based Hybrid View Learning (AHVL) framework, incorporates two essential attention mechanisms: Contrastive Switch Attention (CSA) and Selective Pooling Attention (SPA). The CSA mechanism flexibly alternates between self-attention and cross-attention based on data integrity, integrating a pre-trained language model for contrastive learning to enhance model stability. Meanwhile, the SPA module employs multi-scale feature pooling and selection to capture critical features from mammographic images, overcoming the limitations of traditional models that struggle with fine-grained lesion detection. Experimental validation on the INbreast and CBIS-DDSM datasets shows that the AHVL framework outperforms both single-view and multi-view methods, especially under extreme view missing conditions. Even with an 80% missing rate on both datasets, AHVL maintains the highest accuracy and experiences the smallest performance decline in metrics like F1 score and AUC-PR, demonstrating its robustness and stability. This study redefines mammographic image analysis by leveraging attention-based hybrid view processing, setting a new standard for precise and efficient breast cancer diagnosis.

The risk of postoperative pancreatic fistula (POPF), one of the most dreaded complications after pancreatic surgery, can be predicted from preoperative imaging and tabular clinical routine data. However, existing studies suffer from limited clinical applicability due to a need for manual data annotation and a lack of external validation. We propose AutoFRS (automated fistula risk score software), an externally validated end-to-end prediction tool for POPF risk stratification based on multimodal preoperative data. We trained AutoFRS on preoperative contrast-enhanced computed tomography imaging and clinical data from 108 patients undergoing pancreatic head resection and validated it on an external cohort of 61 patients. Prediction performance was assessed using the area under the receiver operating characteristic curve (AUC) and balanced accuracy. In addition, model performance was compared to the updated alternative fistula risk score (ua-FRS), the current clinical gold standard method for intraoperative POPF risk stratification. AutoFRS achieved an AUC of 0.81 and a balanced accuracy of 0.72 in internal validation and an AUC of 0.79 and a balanced accuracy of 0.70 in external validation. In a patient subset with documented intraoperative POPF risk factors, AutoFRS (AUC: 0.84 ± 0.05) performed on par with the uaFRS (AUC: 0.85 ± 0.06). The AutoFRS web application facilitates annotation-free prediction of POPF from preoperative imaging and clinical data based on the AutoFRS prediction model. POPF can be predicted from multimodal clinical routine data without human data annotation, automating the risk prediction process. We provide additional evidence of the clinical feasibility of preoperative POPF risk stratification and introduce a software pipeline for future prospective evaluation.

This study aimed to compare six large language models (LLMs) [Chat Generative Pre-trained Transformer (ChatGPT)o1-preview, ChatGPT-4o, ChatGPT-4o with canvas, Google Gemini 1.5 Pro, Claude 3.5 Sonnet, and Claude 3 Opus] in generating radiology references, assessing accuracy, fabrication, and bibliographic completeness. In this cross-sectional observational study, 120 open-ended questions were administered across eight radiology subspecialties (neuroradiology, abdominal, musculoskeletal, thoracic, pediatric, cardiac, head and neck, and interventional radiology), with 15 questions per subspecialty. Each question prompted the LLMs to provide responses containing four references with in-text citations and complete bibliographic details (authors, title, journal, publication year/month, volume, issue, page numbers, and PubMed Identifier). References were verified using Medline, Google Scholar, the Directory of Open Access Journals, and web searches. Each bibliographic element was scored for correctness, and a composite final score [(FS): 0-36] was calculated by summing the correct elements and multiplying this by a 5-point verification score for content relevance. The FS values were then categorized into a 5-point Likert scale reference accuracy score (RAS: 0 = fabricated; 4 = fully accurate). Non-parametric tests (Kruskal-Wallis, Tamhane's T2, Wilcoxon signed-rank test with Bonferroni correction) were used for statistical comparisons. Claude 3.5 Sonnet demonstrated the highest reference accuracy, with 80.8% fully accurate references (RAS 4) and a fabrication rate of 3.1%, significantly outperforming all other models (<i>P</i> < 0.001). Claude 3 Opus ranked second, achieving 59.6% fully accurate references and a fabrication rate of 18.3% (<i>P</i> < 0.001). ChatGPT-based models (ChatGPT-4o, ChatGPT-4o with canvas, and ChatGPT o1-preview) exhibited moderate accuracy, with fabrication rates ranging from 27.7% to 52.9% and <8% fully accurate references. Google Gemini 1.5 Pro had the lowest performance, achieving only 2.7% fully accurate references and the highest fabrication rate of 60.6% (<i>P</i> < 0.001). Reference accuracy also varied by subspecialty, with neuroradiology and cardiac radiology outperforming pediatric and head and neck radiology. Claude 3.5 Sonnet significantly outperformed all other models in generating verifiable radiology references, and Claude 3 Opus showed moderate performance. In contrast, ChatGPT models and Google Gemini 1.5 Pro delivered substantially lower accuracy with higher rates of fabricated references, highlighting current limitations in automated academic citation generation. The high accuracy of Claude 3.5 Sonnet can improve radiology literature reviews, research, and education with dependable references. The poor performance of other models, with high fabrication rates, risks misinformation in clinical and academic settings and highlights the need for refinement to ensure safe and effective use.

Multi-material decomposition (MMD) enables quantitative reconstruction of tissue compositions in the human body, supporting a wide range of clinical applications. However, traditional MMD typically requires spectral CT scanners and pre-measured X-ray energy spectra, significantly limiting clinical applicability. To this end, various methods have been developed to perform MMD using conventional (i.e., single-energy, SE) CT systems, commonly referred to as SEMMD. Despite promising progress, most SEMMD methods follow a two-step image decomposition pipeline, which first reconstructs monochromatic CT images using algorithms such as FBP, and then performs decomposition on these images. The initial reconstruction step, however, neglects the energy-dependent attenuation of human tissues, introducing severe nonlinear beam hardening artifacts and noise into the subsequent decomposition. This paper proposes JSover, a fundamentally reformulated one-step SEMMD framework that jointly reconstructs multi-material compositions and estimates the energy spectrum directly from SECT projections. By explicitly incorporating physics-informed spectral priors into the SEMMD process, JSover accurately simulates a virtual spectral CT system from SE acquisitions, thereby improving the reliability and accuracy of decomposition. Furthermore, we introduce implicit neural representation (INR) as an unsupervised deep learning solver for representing the underlying material maps. The inductive bias of INR toward continuous image patterns constrains the solution space and further enhances estimation quality. Extensive experiments on both simulated and real CT datasets show that JSover outperforms state-of-the-art SEMMD methods in accuracy and computational efficiency.

To conduct a systematic review and meta-analysis of the performance of radiomics, dosiomics and machine learning in generating toxicity prediction in thoracic radiotherapy. An electronic database search was conducted and dual-screened by independent authors to identify eligible studies for systematic review and meta-analysis. Data was extracted and study quality was assessed using TRIPOD for machine learning studies, RQS for Radiomics and RoB for dosiomics. 10,703 studies were identified, and 5252 entered screening. 106 studies including 23,373 patients were eligible for systematic review. Primary toxicity predicted was radiation pneumonitis (81), followed by esophagitis (12) and lymphopenia (4). Fourty-two studies studying radiation pneumonitis were eligible for meta-analysis, with pooled area-under-curve (AUC) of 0.82 (95% CI 0.79-0.85). Studies with machine learning had the best performance, with classical and deep learning models having similar performance. There is a trend towards an improvement of the performance of models with the year of publication. There is variability in study quality among the three study categories and dosiomic studies scored the highest among these. Publication bias was not observed. The majority of existing literature using radiomics, dosiomics and machine learning has focused on radiation pneumonitis prediction. Future research should focus on toxicity prediction of other organs at risk and the adoption of these models into clinical practice.

Predicting long-term functional outcomes for individuals with stroke is a significant challenge. Solving this challenge will open new opportunities for improving stroke management by informing acute interventions and guiding personalized rehabilitation strategies. The location of the stroke is a key predictor of outcomes, yet no clinically deployed tools incorporate lesion location information for outcome prognostication. This study responds to this critical need by introducing a fully automated, three-stage neuroimaging processing and machine learning pipeline that predicts personalized outcomes from clinical imaging in adult ischemic stroke patients. In the first stage, our system automatically processes raw DICOM inputs, registers the brain to a standard template, and uses deep learning models to segment the stroke lesion. In the second stage, lesion location and automatically derived network features are input into statistical models trained to predict long-term impairments from a large independent cohort of lesion patients. In the third stage, a structured PDF report is generated using a large language model that describes the strokes location, the arterial distribution, and personalized prognostic information. We demonstrate the viability of this approach in a proof-of-concept application predicting select cognitive outcomes in a stroke cohort. Brain-behavior models were pre-trained to predict chronic impairment on 28 different cognitive outcomes in a large cohort of patients with focal brain lesions (N=604). The automated pipeline used these models to predict outcomes from clinically acquired MRIs in an independent ischemic stroke cohort (N=153). Starting from raw clinical DICOM images, we show that our pipeline can generate outcome predictions for individual patients in less than 3 minutes with 96% concordance relative to methods requiring manual processing. We also show that prediction accuracy is enhanced using models that incorporate lesion location, lesion-associated network information, and demographics. Our results provide a strong proof-of-concept and lay the groundwork for developing imaging-based clinical tools for stroke outcome prognostication.

To develop a multiparametric breast MRI radiomics and deep learning-based multimodal model for predicting preoperative Ki-67 expression status in breast cancer, with the potential to advance individualized treatment and precision medicine for breast cancer patients. We included 176 invasive breast cancer patients who underwent breast MRI and had Ki-67 results. The dataset was randomly split into training (70 %) and test (30 %) sets. Features from T1-weighted imaging (T1WI), diffusion-weighted imaging (DWI), T2-weighted imaging (T2WI), and dynamic contrast-enhanced MRI (DCE-MRI) were fused. Separate models were created for each sequence: T1, DWI, T2, and DCE. A multiparametric MRI (mp-MRI) model was then developed by combining features from all sequences. Models were trained using five-fold cross-validation and evaluated on the test set with receiver operating characteristic (ROC) curve area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and F1 score. Delong's test compared the mp-MRI model with the other models, with P < 0.05 indicating statistical significance. All five models demonstrated good performance, with AUCs of 0.83 for the T1 model, 0.85 for the DWI model, 0.90 for the T2 model, 0.92 for the DCE model, and 0.96 for the mp-MRI model. Delong's test indicated statistically significant differences between the mp-MRI model and the other four models, with P values < 0.05. The multiparametric breast MRI radiomics and deep learning-based multimodal model performs well in predicting preoperative Ki-67 expression status in breast cancer.

BackgroundComputed tomography (CT) is widely used in clinical diagnosis of lung diseases. The automatic segmentation of lesions in CT images aids in the development of intelligent lung disease diagnosis.ObjectiveThis study aims to address the issue of imprecise segmentation in CT images due to the blurred detailed features of lesions, which can easily be confused with surrounding tissues.MethodsWe proposed a promptable segmentation method based on an improved U-Net and Segment Anything model (SAM) to improve segmentation accuracy of lung lesions in CT images. The improved U-Net incorporates a multi-scale attention module based on a channel attention mechanism ECA (Efficient Channel Attention) to improve recognition of detailed feature information at edge of lesions; and a promptable clipping module to incorporate physicians' prior knowledge into the model to reduce background interference. Segment Anything model (SAM) has a strong ability to recognize lesions and pulmonary atelectasis or organs. We combine the two to improve overall segmentation performances.ResultsOn the LUAN16 dataset and a lung CT dataset provided by the Shanghai Chest Hospital, the proposed method achieves Dice coefficients of 80.12% and 92.06%, and Positive Predictive Values of 81.25% and 91.91%, which are superior to most existing mainstream segmentation methods.ConclusionThe proposed method can be used to improve segmentation accuracy of lung lesions in CT images, enhance automation level of existing computer-aided diagnostic systems, and provide more effective assistance to radiologists in clinical practice.