PCNet: Prior Category Network for CT Universal Segmentation Model.

Accurate segmentation of anatomical structures in Computed Tomography (CT) images is crucial for clinical diagnosis, treatment planning, and disease monitoring. The present deep learning segmentation methods are hindered by factors such as data scale and model size. Inspired by how doctors identify tissues, we propose a novel approach, the Prior Category Network (PCNet), that boosts segmentation performance by leveraging prior knowledge between different categories of anatomical structures. Our PCNet comprises three key components: prior category prompt (PCP), hierarchy category system (HCS), and hierarchy category loss (HCL). PCP utilizes Contrastive Language-Image Pretraining (CLIP), along with attention modules, to systematically define the relationships between anatomical categories as identified by clinicians. HCS guides the segmentation model in distinguishing between specific organs, anatomical structures, and functional systems through hierarchical relationships. HCL serves as a consistency constraint, fortifying the directional guidance provided by HCS to enhance the segmentation model's accuracy and robustness. We conducted extensive experiments to validate the effectiveness of our approach, and the results indicate that PCNet can generate a high-performance, universal model for CT segmentation. The PCNet framework also demonstrates a significant transferability on multiple downstream tasks. The ablation experiments show that the methodology employed in constructing the HCS is of critical importance. The prompt and HCS can be accessed at https://github.com/YixinChen-AI/PCNet.

A novel relative-equilibrium graphical plot for rapid reversible tracer studies in dynamic PET imaging.

OBJECTIVE: This study aims to address the issue of long scan durations required by traditional graphical analysis methods, such as the Logan plot and its variant, the reversible equilibrium (RE) Logan plot, for dynamic PET imaging of tracer kinetics. Approach: We propose a relative RE Logan model that builds on the principles of the Logan plot and its variant to significantly reduce scan time without compromising the accuracy of tracer kinetics analysis. The model is supported by theoretical evidence and experimental validations, including two computer simulations and one clinical data analysis. Main results: The proposed model demonstrates a significant linear relationship between the variable x and the slope DV_T of the RE Logan plot, and the variable x' and the slope DV_T' of the relative RE Logan plot. The Pearson correlation coefficients (r) of the linear fitting of the x' to the x equal 0.9849 in the simulated data and 0.9912 in the clinical data. Similarly, the r value for the linear fitting of DV_T' to DV_T equal 0.9989 and 0.9988 in the simulated data, and 0.9954 in the clinical data. Significance: These results demonstrate the model's capability to maintain strong linear relationships and produce parametric images comparable to those of the traditional RE Logan plot, but with the considerable advantage of shorter scan durations. This innovation holds significant potential for enhancing the efficiency and feasibility of PET imaging in clinical settings.