Jie Tian’s research while affiliated with Beihang University and other places

In vivo tracking of stem cells is critical to stem cell transplantation and regenerative medicine research. Achieving long-term, non-invasive imaging of transplanted cells demands highly sensitive imaging technology. Magnetic particle imaging (MPI) is an emerging modality capable of visualizing the distribution of superparamagnetic iron oxide nanoparticles (SPIOs). Previous studies have reported that MPI can detect clusters of approximately 2500 SPIO-labeled stem cells. However, more than this sensitivity level is needed to monitor the survival, migration, differentiation, and regeneration of stem cells. To address this limitation, we developed an ultrahigh-sensitivity magnetic particle imaging (US-MPI) device capable of detecting stem cells at scales of hundreds of cells. Through full-chain design and optimization, US-MPI maximizes the capture of SPIO signals from stem cells. Using a commercial SPIO (VivoTrax), mouse mesenchymal stem cells were labeled, and samples with varying cell numbers (ranging from 50 to 1×107) were tested using US-MPI and 9.4 T MRI (uMR 9.4T, United Imaging Life Science Instrument, Wuhan, China) for comparison. US-MPI demonstrated the ability to detect as few as 50 labeled stem cells in vitro and 150 cells in vivo, whereas 9.4 T MRI required at least 1×105 cells for detection in vitro. Thus, US-MPI exhibited a sensitivity 2, 000 times higher than that of 9.4 T MRI. The efficacy of US-MPI for stem cell tracking was further validated in a lipopolysaccharide (LPS)-induced mouse hind limb inflammation model. Eight hours after SPIO-labeled stem cells were injected intravenously, US-MPI detected strong signals on the inflamed limb, which remained significantly higher than those of the healthy limb during a 48-hour monitoring period. Post-mortem Prussian blue staining of inflamed tissues confirmed the presence of blue-stained SPIO particles, verifying that the US-MPI signals originated from labeled stem cells. These results highlight the ultrahigh sensitivity of US-MPI for stem cell tracking, providing a valuable tool for advancing stem cell research. Citation Format Yimeng Li, Yu An, Yang Du, Lin Shen, Jie Tian. Development of ultrahigh sensitivity magnetic particle imaging for stem cell tracking [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2526.

The second near-infrared (NIR-II) fluorescence imaging is a new biomedical imaging method. Compared with the NIR-IIa window (1000–1300 nm), the NIR-IIb window (1500–1700 nm) produces clearer images and better imaging effects. Due to technical limitations, no NIR-IIb molecular probes are used in clinic. In order to get the NIR-IIb images, we translate the given NIR-IIa images into the NIR-IIb images through artificial intelligence. We propose a generative model to complete the above translation process, which is based on the encoder–decoder structure. The decoder is a pre-trained StyleGAN trained only in the NIR-IIb domain. NIR-II images are usually unpaired. In order to achieve unsupervised image translation, we separate the latent code of StyleGAN into two parts: style–content and style–style. The encoder encodes the images into the latent space, and by learning the content representation and style representation in the latent space, the generated images can inherit the content of the NIR-IIa images and the style of the NIR-IIb images. Experiment results show that the proposed model can generate high quality NIR-IIb fluorescence images compared with multiple existing methods.

Background Early detection and treatment are critical for improving the survival and prognosis of patients with cervical cancer. However, there is a notable scarcity of targeted imaging probes specifically designed to detect high-grade squamous intraepithelial lesions (HSIL) and cervical cancer. Our study aimed to address this gap by identifying and validating a targeted imaging probe for these conditions. Results Using bioinformatics data, we identified galectin-7 (GAL7) as highly expressed in patients with cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC). Immunohistochemical staining of biopsy samples from 30 HSIL and cervical cancer patients verified the high and specific expression of GAL7. Further validation was performed using mouse and human CESC cell lines and tumor xenografts, confirming the consistent expression of GAL7. Based on this finding, we synthesized a GAL7-specific antibody conjugated with FITC, creating the GAL7-FITC fluorescence imaging probe. Fluorescence molecular imaging revealed that GAL7-FITC exhibited specific binding to various CESC cell lines and xenograft mouse models. Additionally, the diagnostic capability of GAL7-FITC was demonstrated in fresh HSIL specimens from cervical cone excisions, validated through histopathology and immunohistochemical analysis. Conclusions Our study identified GAL7 as a specific target for CESC and successfully developed the GAL7-FITC fluorescence imaging probe. GAL7-FITC has shown promising potential for clinical application in the early detection of HSIL and CESC, providing rapid fluorescence imaging diagnosis without observable toxicity. This advancement may significantly enhance the accuracy and speed of cervical cancer diagnostics, ultimately improving patient outcomes.

Magnetic particle imaging (MPI) holds immense promise as a non-invasive biomedical imaging modality, but its advancement is hindered by the absence of activatable probes for biomarker-specific imaging. Here, we introduce a ligand-mediated acid-activatable MPI probe (LAMP) composed of magnetic nanoparticles linked via acid-sensitive imine bond-containing ligands. The LAMP exhibits stable assembly in neutral conditions and rapid disassembly in acidic microenvironments. The imine bond crosslinking promotes a compact structure that enhances magnetic dipole interactions, significantly quenching the initial MPI signal. Upon exposure to acidic conditions, the probe disassembles, restoring the MPI signal in a highly controlled, environmentdependent manner. This switchable assembly enables precise, real-time imaging of acidity-associated diseases such as sepsis progression. Our findings demonstrate that the ligand-mediated modulation of magnetic dipole interactions provides a versatile platform for the design of next-generation MPI probes for biomarker-specific MPI imaging.

Patients with diabetes mellitus (DM) have an increased risk of in stent restenosis (ISR). Neovascularization (NV) is considered as a unique pathophysiology factor of ISR in diabetic patients. However, the restructuring patterns of in vivo human coronary NV and their relationship with ISR, especially in diabetic patients remain unclear. In this study, we aimed to investigate the NV structure differentiations between patients with and without DM after coronary stent implantation using optical coherence tomography (OCT). We included 136 patients with ISR (70 patients in DM group and 66 patients in non-DM group) who underwent OCT during coronary angiography follow-up. NVs were manually segmented, after which three-dimensional (3D) rendering of OCT images was conducted. NVs greater than 1 mm in length were classified as longitudinal running or coral tree types based on their 3D structures. NV structures were compared between DM and non-DM patients. The prevalence of the coral tree pattern NV in the DM group was 2.14-fold higher than in the non-DM group(p = 0.012). 47.14% of patients in the DM group and 51.51% of patients in the non-DM group presented longitudinal running NV (p = 0.610). The number of coral tree pattern NV was relatively higher in DM patients than in the non-DM patients (p = 0.019). However, the number of longitudinal running NV showed no difference between the two groups (p = 0.872). The normalized NV volume was significantly larger in the DM group (p = 0.008). Patients with coral tree pattern NV have thinner minimum fibrous cap thickness (p = 0.030). DM was the risk factor for coral tree pattern NV formation in ISR lesions after adjustment for other factors. NV with specific restructuring patterns, such as longitudinal running and coral tree patterns, can be identified in ISR lesions. NV with a coral tree pattern, characterized by higher leakiness and immaturity, is more commonly found in patients with DM and is associated with tissue instability in ISR. Accurate and feasible imaging modalities for NV might offer promising opportunities to evaluate NV and prevent progression of ISR in diabetic patients.

Magnetic particle imaging (MPI) enables real-time, sensitive and quantitative visualization of magnetic tracers' spatial distribution, augmenting the capability of in vivo imaging technologies. Previous tracer studies in MPI have primarily focused on superparamagnetic nanoparticles; however, their non-ideal sigmoidal magnetization response limits the spatial resolution. Here we demonstrate the utilization of magnetically induced magnetosome chain (MAGiC) as a novel superferromagnetic MPI tracer, exhibiting a 25-fold improvement in resolution and a 91-fold enhancement in signal intensity compared to the commercial tracer VivoTrax+. The spatial resolution of MPI was pushed to an unprecedented 80 μm under a 4 T/m gradient field. Additionally, MAGiC can be precisely controlled using magnetic fields, enabling it to function as a MPI trackable microrobot. We provided a theoretical model elucidating MAGiC's unique properties, and validated its imaging and actuation performance through phantom studies and in vivo experiments. As a high-performance MPI tracer and magnetic microrobot with exceptional capabilities, MAGiC holds tremendous potential for diverse applications including cell tracking, targeted drug delivery as well as therapeutic interventions.

The second near-infrared (NIR-II) fluorescence imaging has become a new imaging mode due to its characteristics of real-time intraoperative imaging. The NIR-IIb window (1500–1700 nm) has stronger light penetration and has a clearer imaging effect than the NIR-IIa window (1000–1300 nm). The molecular probe currently used for human imaging (indocyanine green) can only be used for NIR-IIa window imaging, and there is a lack of effective molecular probes for NIR-IIb imaging. In addition, there are various types of NIR-II fluorescence images, and it is difficult for a single neural network model to enhance multiple types of images. To solve these problems, we design a universal NIR-II fluorescence image enhancement model to transfer the style of NIR-IIb images to NIR-IIa images and improve the quality of fluorescence images. The model is based on an encoder-decoder framework and can process multiple types of NIR-II fluorescence images simultaneously. Specifically, it includes a module for fusion of low-quality and high-quality image feature maps, the covariance weighted attention network (CWANet), which improves the attention mechanism through covariance weight, allowing the model to filter the style features that are irrelevant or detrimental to image structure in the attention mechanism and pay attention to the key content details of the image. Furthermore, we propose a multiple high-frequency matching loss, which matches the high-frequency information of the enhanced image with the NIR-IIa image to further maintain the structure. Extensive experiments demonstrate that our model generates results that exceed state-of-the-art models, with impressive image enhancement effect.

Background Early detection and treatment are effective methods for the management of oral squamous cell carcinoma (OSCC), which can be facilitated by the detection of tumor-specific OSCC biomarkers. The epidermal growth factor receptor (EGFR) and programmed death-ligand 1 (PD-L1) are important therapeutic targets for OSCC. Multispectral fluorescence molecular imaging (FMI) can facilitate the detection of tumor multitarget expression with high sensitivity and safety. Hence, we developed Nimotuzumab-ICG and Atezolizumab-Cy5.5 imaging probes, in combination with multispectral FMI, to sensitively and noninvasively identify EGFR and PD-L1 expression for the detection and comprehensive treatment of OSCC. Methods The expression of EGFR and PD-L1 was analyzed using bioinformatics data sources and specimens. Nimotuzumab-ICG and Atezolizumab-Cy5.5 imaging probes were developed and tested on preclinical OSCC cell line and orthotopic OSCC mouse model, fresh OSCC patients’ biopsied samples, and further clinical mouthwash trials were conducted in OSCC patients. Results EGFR and PD-L1 were specifically expressed in human OSCC cell lines and tumor xenografts. Nimotuzumab-ICG and Atezolizumab-Cy5.5 imaging probes can specifically target to the tumor sites in an in situ human OSCC mouse model with good safety. The detection sensitivity and specificity of Nimotuzumab-ICG in patients were 96.4% and 100%, and 95.2% and 88.9% for Atezolizumab-Cy5.5. Conclusions EGFR and PD-L1 are highly expressed in OSCC, the combination of which is important for a precise prognosis of OSCC. EGFR and PD-L1 expression can be sensitively detected using the newly synthesized multispectral fluorescence imaging probes Nimotuzumab-ICG and Atezolizumab-Cy5.5, which can facilitate the sensitive and specific detection of OSCC and improve treatment outcomes. Trial registration Chinese Clinical Trial Registry, ChiCTR2100045738. Registered 23 April 2021, https://www.chictr.org.cn/bin/project/edit?pid=125220

In the domain of medical image analysis, there is a burgeoning recognition and adoption of large models distinguished by their extensive parameter count and intricate neural network architecture that is predominantly due to their outstanding performance. This review article seeks to concisely explore the historical evolution, specific applications, and training methodologies associated with these large models considering their current prominence in medical image analysis. Moreover, we delve into the prevailing challenges and prospective opportunities related to the utilization of large models in the context of medical image analysis. Through a comprehensive analysis of these substantial models, this study aspires to provide valuable insights and guidance to researchers in the field of radiology, fostering further advances and optimizations in their incorporation into medical image analysis practices, in accordance with the submission requirements.