Explore the vast range of titles available.

Tomographic imaging using penetrating waves generates cross-sectional views of the internal anatomy of a living subject. For artefact-free volumetric imaging, projection views from a large number of angular positions are required. Here we show that a deep-learning model trained to map projection radiographs of a patient to the corresponding 3D anatomy can subsequently generate volumetric tomographic X-ray images of the patient from a single projection view. We demonstrate the feasibility of the approach with upper-abdomen, lung, and head-and-neck computed tomography scans from three patients. Volumetric reconstruction via deep learning could be useful in image-guided interventional procedures such as radiation therapy and needle biopsy, and might help simplify the hardware of tomographic imaging systems. A deep-learning model trained to map 2D projection views of a patient to the corresponding 3D anatomy can subsequently generate volumetric tomographic X-ray images of the patient from a single projection view.

Magnetic resonance imaging offers unrivaled visualization of the fetal brain, forming the basis for establishing age-specific morphologic milestones. However, gauging age-appropriate neural development remains a difficult task due to the constantly changing appearance of the fetal brain, variable image quality, and frequent motion artifacts. Here we present an end-to-end, attention-guided deep learning model that predicts gestational age with R 2 score of 0.945, mean absolute error of 6.7 days, and concordance correlation coefficient of 0.970. The convolutional neural network was trained on a heterogeneous dataset of 741 developmentally normal fetal brain images ranging from 19 to 39 weeks in gestational age. We also demonstrate model performance and generalizability using independent datasets from four academic institutions across the U.S. and Turkey with R 2 scores of 0.81–0.90 after minimal fine-tuning. The proposed regression algorithm provides an automated machine-enabled tool with the potential to better characterize in utero neurodevelopment and guide real-time gestational age estimation after the first trimester.

Single cell RNA sequencing is a promising technique to determine the states of individual cells and classify novel cell subtypes. In current sequence data analysis, however, genes with low expressions are omitted, which leads to inaccurate gene counts and hinders downstream analysis. Recovering these omitted expression values presents a challenge because of the large size of the data. Here, we introduce a data-driven gene expression recovery framework, referred to as self-consistent expression recovery machine (SERM), to impute the missing expressions. Using a neural network, the technique first learns the underlying data distribution from a subset of the noisy data. It then recovers the overall expression data by imposing a self-consistency on the expression matrix, thus ensuring that the expression levels are similarly distributed in different parts of the matrix. We show that SERM improves the accuracy of gene imputation with orders of magnitude enhancement in computational efficiency in comparison to the state-of-the-art imputation techniques. Recovering dropout-affected gene expression values is a challenging problem in bioinformatics. Here, the authors propose a data-driven framework, that first learns the underlying data distribution and then recovers the expression values by imposing a self-consistency on the expression matrix.

Biomarkers screening is a benefit approach for early diagnosis of major diseases. In this study, magnetic nanoparticles (MNPs) have been utilized as labels to establish a multi-line immunochromatography (MNP-MLIC) for simultaneous detection of carcinoembryonic antigen (CEA), carbohydrate antigen 199 (CA 19–9), and alpha-fetoprotein (AFP) in a single serum sample. Under the optimal parameters, the three biomarkers can be rapidly and simultaneously qualitative screening within 15 min by naked eye. As for quantitative detection, the MNP-MLIC test strips were precisely positioned and captured by a smartphone, and signals on the test and control lines were extracted by ImageJ software. The signal ratio of test and control lines has been calculated and used to plot quantitative standard curves with the logarithmic concentration, of which the correlation coefficients are more than 0.99, and the limit of detection for CEA, CA 19–9, and AFP were 0.60 ng/mL, 1.21 U/mL, and 0.93 ng/mL, respectively. The recoveries of blank serum were 75.0 ~ 112.5% with the relative standard deviation ranging from 2.5 to 15.3%, and the specificity investigation demonstrated that the MNP-MLIC is highly specific to the three biomarkers. In conclusion, the developed MNP-MLIC offers a rapid, simple, accurate, and highly specific method for simultaneously detecting multiple biomarkers in serum samples, which provides an efficient and accurate approach for the early diagnosis of diseases.

Immunochromatographic assay (ICA) plays an important role in in vitro diagnostics because of its simpleness, convenience, fastness, sensitivity, accuracy, and low cost. The employment of magnetic nanoparticles (MNPs), possessing both excellent optical properties and magnetic separation functions, can effectively promote the performances of ICA. In this study, an ICA based on MNPs (MNP–ICA) has been successfully developed for the sensitive detection of carcinoembryonic antigen (CEA). The magnetic probes were prepared by covalently conjugating carboxylated MNPs with the specific monoclonal antibody against CEA, which were not only employed to enrich and extract CEA from serum samples under an external magnetic field but also used as a signal output with its inherent optical property. Under the optimal parameters, the limit of detection (LOD) for qualitative detection with naked eyes was 1.0 ng/mL, and the quantitative detection could be realized with the help of a portable optical reader, indicating that the ratio of optical signal intensity correlated well with CEA concentration ranging from 1.0 ng/mL to 64.0 ng/mL (R2 = 0.9997). Additionally, method comparison demonstrated that the magnetic probes were beneficial for sensitivity improvement due to the matrix effect reduction after magnetic separation, and the MNP–ICA is eight times higher sensitive than ICA based on colloidal gold nanoparticles. The developed MNP–ICA will provide sensitive, convenient, and efficient technical support for biomarkers rapid screening in cancer diagnosis and prognosis.

Deepening our understanding of human health is more important than ever before for addressing real-world challenges in biomedicine and healthcare, especially with the recent pandemic. My research focuses on AI in medicine, to develop efficient ML models for biomedical imaging and perception for addressing clinically important problems. In this thesis, I will first explore the challenges in this emerging field and then present the two majors lines of my PhD research work.First, I will introduce my work on AI in biomedical imaging. Specifically, I will discuss how to integrate different kinds of prior knowledge to develop reliable data-efficient ML models, by exploiting the personalized priors, population priors and physics priors. With the prior-informed ML models, the proposed approaches can be applied to various applications including sparse-sampling CT and MRI image reconstruction, X-ray projection synthesis, and Cryo-EM imaging. These techniques show significant potential to impact cancer imaging and treatment.Second, I will introduce my work on AI in image perception. I will discuss how to develop ML-driven perception models that can adapt to the unique characteristics of biomedical data. Specifically, I will present a self-attention-guided ML model for quantitative image perception of fetal brain MRI images. This includes a cross-institute validation involving four U.S. clinical centers and a Turkish institute. This work demonstrates the ability of the developed ML model to characterize in utero neurodevelopment for anomaly detection in real-world clinical deployments.

Cryogenic electron tomography (cryoET) is a powerful tool in structural biology, enabling detailed 3D imaging of biological specimens at a resolution of nanometers. Despite its potential, cryoET faces challenges such as the missing wedge problem, which limits reconstruction quality due to incomplete data collection angles. Recently, supervised deep learning methods leveraging convolutional neural networks (CNNs) have considerably addressed this issue; however, their pretraining requirements render them susceptible to inaccuracies and artifacts, particularly when representative training data is scarce. To overcome these limitations, we introduce a proof-of-concept unsupervised learning approach using coordinate networks (CNs) that optimizes network weights directly against input projections. This eliminates the need for pretraining, reducing reconstruction runtime by 3–20× compared to supervised methods. Our in silico results show improved shape completion and reduction of missing wedge artifacts, assessed through several voxel-based image quality metrics in real space and a novel directional Fourier Shell Correlation (FSC) metric. Our study illuminates benefits and considerations of both supervised and unsupervised approaches, guiding the development of improved reconstruction strategies.

Precision medicine, such as patient-adaptive treatments assisted by medical image analysis, poses new challenges for segmentation algorithms in adapting to new patients, due to the large variability across different patients and the limited availability of annotated data for each patient. In this work, we propose a data-efficient segmentation algorithm, namely Part-aware Prompted Segment Anything Model (\\(P^2SAM\\)). Without any model fine-tuning, \\(P^2SAM\\) enables seamless adaptation to any new patients relying only on one-shot patient-specific data. We introduce a novel part-aware prompt mechanism to select multiple-point prompts based on the part-level features of the one-shot data, which can be extensively integrated into different promptable segmentation models, such as SAM and SAM 2. Moreover, to determine the optimal number of parts for each specific case, we propose a distribution-guided retrieval approach that further enhances the robustness of the part-aware prompt mechanism. \\(P^2SAM\\) improves the performance by +8.0% and +2.0% mean Dice score for two different patient-adaptive segmentation applications, respectively. In addition, \\(P^2SAM\\) also exhibits impressive generalizability in other adaptive segmentation tasks in the natural image domain, e.g., +6.4% mIoU within personalized object segmentation task. The code is available at: https://github.com/Zch0414/p2sam

Precision medicine, such as patient-adaptive treatments utilizing medical images, poses new challenges for image segmentation algorithms due to (1) the large variability across different patients and (2) the limited availability of annotated data for each patient. In this work, we propose a data-efficient segmentation method to address these challenges, namely Part-aware Personalized Segment Anything Model (P^2SAM). Without any model fine-tuning, P^2SAM enables seamless adaptation to any new patients relying only on one-shot patient-specific data. We introduce a novel part-aware prompt mechanism to select multiple-point prompts based on part-level features of the one-shot data. To further promote the robustness of the selected prompt, we propose a retrieval approach to handle outlier prompts. Extensive experiments demonstrate that P^2SAM improves the performance by +8.0% and +2.0% mean Dice score within two patient-specific segmentation settings, and exhibits impressive generality across different application domains, e.g., +6.4% mIoU on the PerSeg benchmark. Code will be released upon acceptance.

Epidemic forecasting models typically rely on surveillance data reported over administrative regions, treating them as atomic units, thereby obscuring sub-regional spatial structure that shapes disease dynamics. We introduce a spatially structured multimodal epidemic forecasting setting that integrates region-level temporal surveillance data with spatially localized auxiliary signals that are misaligned in resolution and structure, reflecting realistic public health reporting constraints. Building on this formulation, we propose M-SPICE (Multimodal SPatIal Context for Epidemic Forecasting), a structure-aware spatiotemporal forecasting framework that performs joint reasoning over temporal disease dynamics and spatial context via attention-based multimodal fusion, allowing spatial signals to selectively condition temporal representations across forecast horizons. We evaluate our approach on real-world COVID-19, influenza, and influenza-like illness (ILI) forecasting tasks under realistic real-time evaluation protocols. Across all forecasting settings, our method consistently outperforms state-of-the-art multivariate time-series, multimodal, and epidemiological forecasting baselines while maintaining strong probabilistic forecasting performance. Finally, interpretability analyses reveal when, where, and how spatial signals are leveraged, highlighting settings in which purely temporal, region-aggregated models are most likely to fail.