Subspecialty knowledge barriers have limited the adoption of large language models (LLMs) in oncology. We introduce Woollie, an open-source, oncology-specific LLM trained on real-world data from Memorial Sloan Kettering Cancer Center (MSK) across lung, breast, prostate, pancreatic, and colorectal cancers, with external validation using University of California, San Francisco (UCSF) data. Woollie surpasses ChatGPT in medical benchmarks and excels in eight non-medical benchmarks. Analyzing 39,319 radiology impression notes from 4002 patients, it achieved an overall area under the receiver operating characteristic curve (AUROC) of 0.97 for cancer progression prediction on MSK data, including a notable 0.98 AUROC for pancreatic cancer. On UCSF data, it achieved an overall AUROC of 0.88, excelling in lung cancer detection with an AUROC of 0.95. As the first oncology specific LLM validated across institutions, Woollie demonstrates high accuracy and consistency across cancer types, underscoring its potential to enhance cancer progression analysis.

The use of deep learning (DL) in medical image analysis has significantly improved the ability to predict lung cancer. In this study, we introduce a novel deep convolutional neural network (CNN) model, named ResNet+, which is based on the established ResNet framework. This model is specifically designed to improve the prediction of lung cancer and diseases using the images. To address the challenge of missing feature information that occurs during the downsampling process in CNNs, we integrate the ResNet-D module, a variant designed to enhance feature extraction capabilities by modifying the downsampling layers, into the traditional ResNet model. Furthermore, a convolutional attention module was incorporated into the bottleneck layers to enhance model generalization by allowing the network to focus on relevant regions of the input images. We evaluated the proposed model using five public datasets, comprising lung cancer (LC2500 $n$=3183, IQ-OTH/NCCD $n$=1336, and LCC $n$=25000 images) and lung disease (ChestXray $n$=5856, and COVIDx-CT $n$=425024 images). To address class imbalance, we used data augmentation techniques to artificially increase the representation of underrepresented classes in the training dataset. The experimental results show that ResNet+ model demonstrated remarkable accuracy/F1, reaching 98.14/98.14\% on the LC25000 dataset and 99.25/99.13\% on the IQ-OTH/NCCD dataset. Furthermore, the ResNet+ model saved computational cost compared to the original ResNet series in predicting lung cancer images. The proposed model outperformed the baseline models on publicly available datasets, achieving better performance metrics. Our codes are publicly available at https://github.com/AIPMLab/Graduation-2024/tree/main/Peng.

The generation of medical images presents significant challenges due to their high-resolution and three-dimensional nature. Existing methods often yield suboptimal performance in generating high-quality 3D medical images, and there is currently no universal generative framework for medical imaging. In this paper, we introduce a 3D Medical Latent Diffusion (3D MedDiffusion) model for controllable, high-quality 3D medical image generation. 3D MedDiffusion incorporates a novel, highly efficient Patch-Volume Autoencoder that compresses medical images into latent space through patch-wise encoding and recovers back into image space through volume-wise decoding. Additionally, we design a new noise estimator to capture both local details and global structural information during diffusion denoising process. 3D MedDiffusion can generate fine-detailed, high-resolution images (up to 512x512x512) and effectively adapt to various downstream tasks as it is trained on large-scale datasets covering CT and MRI modalities and different anatomical regions (from head to leg). Experimental results demonstrate that 3D MedDiffusion surpasses state-of-the-art methods in generative quality and exhibits strong generalizability across tasks such as sparse-view CT reconstruction, fast MRI reconstruction, and data augmentation for segmentationand classification. Source code and checkpoints are available at https://github.com/ShanghaiTech-IMPACT/3D-MedDiffusion.

The early diagnosis of brain tumors is crucial for patient prognosis, and medical imaging techniques such as MRI and CT scans are essential tools for diagnosing brain tumors. However, high-quality medical image data for brain tumors is often scarce and difficult to obtain, which hinders the development and application of medical image analysis models. With the advancement of artificial intelligence, particularly deep learning technologies in the field of medical imaging, new concepts and tools have been introduced for the early diagnosis, treatment planning, and prognosis evaluation of brain tumors. To address the challenge of imbalanced brain tumor datasets, we propose a novel data augmentation technique based on a diffusion model, referred to as the Multi-Channel Fusion Diffusion Model(MCFDiffusion). This method tackles the issue of data imbalance by converting healthy brain MRI images into images containing tumors, thereby enabling deep learning models to achieve better performance and assisting physicians in making more accurate diagnoses and treatment plans. In our experiments, we used a publicly available brain tumor dataset and compared the performance of image classification and segmentation tasks between the original data and the data enhanced by our method. The results show that the enhanced data improved the classification accuracy by approximately 3% and the Dice coefficient for segmentation tasks by 1.5%-2.5%. Our research builds upon previous work involving Denoising Diffusion Implicit Models (DDIMs) for image generation and further enhances the applicability of this model in medical imaging by introducing a multi-channel approach and fusing defective areas with healthy images. Future work will explore the application of this model to various types of medical images and further optimize the model to improve its generalization capabilities. We release our code at https://github.com/feiyueaaa/MCFDiffusion.

Artificial intelligence (AI) has introduced numerous opportunities for human assistance and task automation in medicine. However, it suffers from poor generalization in the presence of shifts in the data distribution. In the context of AI-based computed tomography (CT) analysis, significant data distribution shifts can be caused by changes in scanner manufacturer, reconstruction technique or dose. AI harmonization techniques can address this problem by reducing distribution shifts caused by various acquisition settings. This paper presents an open-source benchmark dataset containing CT scans of an anthropomorphic phantom acquired with various scanners and settings, which purpose is to foster the development of AI harmonization techniques. Using a phantom allows fixing variations attributed to inter- and intra-patient variations. The dataset includes 1378 image series acquired with 13 scanners from 4 manufacturers across 8 institutions using a harmonized protocol as well as several acquisition doses. Additionally, we present a methodology, baseline results and open-source code to assess image- and feature-level stability and liver tissue classification, promoting the development of AI harmonization strategies.

Image registration is a key technique in image processing and analysis. Due to its high complexity, the traditional registration frameworks often fail to meet real-time demands in practice. To address the real-time demand, several deep learning networks for registration have been proposed, including the supervised and the unsupervised networks. Unsupervised networks rely on large amounts of training data to minimize specific loss functions, but the lack of physical information constraints results in the lower accuracy compared with the supervised networks. However, the supervised networks in medical image registration face two major challenges: physical mesh folding and the scarcity of labeled training data. To address these two challenges, we propose a novel few-shot learning framework for image registration. The framework contains two parts: random diffeomorphism generator (RDG) and a supervised few-shot learning network for image registration. By randomly generating a complex vector field, the RDG produces a series of diffeomorphism. With the help of diffeomorphism generated by RDG, one can use only a few image data (theoretically, one image data is enough) to generate a series of labels for training the supervised few-shot learning network. Concerning the elimination of the physical mesh folding phenomenon, in the proposed network, the loss function is only required to ensure the smoothness of deformation (no other control for mesh folding elimination is necessary). The experimental results indicate that the proposed method demonstrates superior performance in eliminating physical mesh folding when compared to other existing learning-based methods. Our code is available at this link https://github.com/weijunping111/RDG-TMI.git.

Low-dose computed tomography (LDCT) imaging employed in lung cancer screening (LCS) programs is increasing in uptake worldwide. LCS programs herald a generational opportunity to simultaneously detect cancer and non-cancer-related early-stage lung disease. Yet these efforts are hampered by a shortage of radiologists to interpret scans at scale. Here, we present TANGERINE, a computationally frugal, open-source vision foundation model for volumetric LDCT analysis. Designed for broad accessibility and rapid adaptation, TANGERINE can be fine-tuned off the shelf for a wide range of disease-specific tasks with limited computational resources and training data. Relative to models trained from scratch, TANGERINE demonstrates fast convergence during fine-tuning, thereby requiring significantly fewer GPU hours, and displays strong label efficiency, achieving comparable or superior performance with a fraction of fine-tuning data. Pretrained using self-supervised learning on over 98,000 thoracic LDCTs, including the UK's largest LCS initiative to date and 27 public datasets, TANGERINE achieves state-of-the-art performance across 14 disease classification tasks, including lung cancer and multiple respiratory diseases, while generalising robustly across diverse clinical centres. By extending a masked autoencoder framework to 3D imaging, TANGERINE offers a scalable solution for LDCT analysis, departing from recent closed, resource-intensive models by combining architectural simplicity, public availability, and modest computational requirements. Its accessible, open-source lightweight design lays the foundation for rapid integration into next-generation medical imaging tools that could transform LCS initiatives, allowing them to pivot from a singular focus on lung cancer detection to comprehensive respiratory disease management in high-risk populations.

Vision Transformers (ViTs) have gained significant popularity in the natural image domain but have been less successful in 3D medical image segmentation. Nevertheless, 3D ViTs are particularly interesting for large medical imaging volumes due to their efficient self-supervised training within the masked autoencoder (MAE) framework, which enables the use of imaging data without the need for expensive manual annotations. intracranial arterial calcification (IAC) is an imaging biomarker visible on routinely acquired CT scans linked to neurovascular diseases such as stroke and dementia, and automated IAC quantification could enable their large-scale risk assessment. We pre-train ViTs with MAE and fine-tune them for IAC segmentation for the first time. To develop our models, we use highly heterogeneous data from a large clinical trial, the third International Stroke Trial (IST-3). We evaluate key aspects of MAE pre-trained ViTs in IAC segmentation, and analyse the clinical implications. We show: 1) our calibrated self-supervised ViT beats a strong supervised nnU-Net baseline by 3.2 Dice points, 2) low patch sizes are crucial for ViTs for IAC segmentation and interpolation upsampling with regular convolutions is preferable to transposed convolutions for ViT-based models, and 3) our ViTs increase robustness to higher slice thicknesses and improve risk group classification in a clinical scenario by 46%. Our code is available online.

High-quality computed tomography (CT) scans are essential for accurate diagnostic and therapeutic decisions, but the presence of metal objects within the body can produce distortions that lower image quality. Deep learning (DL) approaches using image-to-image translation for metal artifact reduction (MAR) show promise over traditional methods but often introduce secondary artifacts. Additionally, most rely on paired simulated data due to limited availability of real paired clinical data, restricting evaluation on clinical scans to qualitative analysis. This work presents CALIMAR-GAN, a generative adversarial network (GAN) model that employs a guided attention mechanism and the linear interpolation algorithm to reduce artifacts using unpaired simulated and clinical data for targeted artifact reduction. Quantitative evaluations on simulated images demonstrated superior performance, achieving a PSNR of 31.7, SSIM of 0.877, and Fréchet inception distance (FID) of 22.1, outperforming state-of-the-art methods. On real clinical images, CALIMAR-GAN achieved the lowest FID (32.7), validated as a valuable complement to qualitative assessments through correlation with pixel-based metrics (r=-0.797 with PSNR, p<0.01; r=-0.767 with MS-SSIM, p<0.01). This work advances DL-based artifact reduction into clinical practice with high-fidelity reconstructions that enhance diagnostic accuracy and therapeutic outcomes. Code is available at https://github.com/roberto722/calimar-gan.

Medical image segmentation is crucial for computer-aided diagnosis and treatment planning, directly influencing clinical decision-making. To enhance segmentation accuracy, existing methods typically fuse local, global, and various other features. However, these methods often ignore the negative impact of noise on the results during the feature fusion process. In contrast, certain regions of the human visual system, such as the inferotemporal cortex and parietal cortex, effectively suppress irrelevant noise while integrating multiple features-a capability lacking in current methods. To address this gap, we propose MS-Net, a medical image segmentation network inspired by human visual perception. MS-Net incorporates a multi-feature compression (MFC) module that mimics the human visual system's processing of complex images, first learning various feature types and subsequently filtering out irrelevant ones. Additionally, MS-Net features a segmentation refinement (SR) module that emulates how physicians segment lesions. This module initially performs coarse segmentation to capture the lesion's approximate location and shape, followed by a refinement step to achieve precise boundary delineation. Experimental results demonstrate that MS-Net not only attains state-of-the-art segmentation performance across three public datasets but also significantly reduces the number of parameters compared to existing models. Code is available at https://github.com/guangguangLi/MS-Net.