Four-dimensional CT reconstruction method based on real-time kV image driving

By employing a real-time kV image-driven four-dimensional CT reconstruction method, combined with data preprocessing and a deep learning model, the problem of high temporal resolution and low radiation dose in radiotherapy has been solved, achieving efficient and accurate four-dimensional CT reconstruction and dose calculation, and supporting real-time target area monitoring.

CN122134872APending Publication Date: 2026-06-02CHONGQING UNIV CANCER HOSPITAL
0 Cites 1 Cited by

Patent Information

Application Number
CN202610300146.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-12
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Current radiotherapy imaging technology cannot simultaneously achieve high temporal resolution, low radiation dose, and high-precision three-dimensional/four-dimensional dose calculation, resulting in target localization errors and dose distribution uncertainties.

Method used

A four-dimensional CT reconstruction method based on real-time kV image-driven reconstruction is adopted. High-resolution three-dimensional CT volume data is reconstructed through data preprocessing (respiratory phase synchronization, spatial resampling, geometric consistency registration and clipping) and deep learning model (three-dimensional U-Net architecture). Combined with the dose calculation module, real-time target area monitoring and dynamic dose optimization are realized.

Benefits of technology

It achieves efficient four-dimensional CT reconstruction under low radiation dose, improves reconstruction speed and spatial accuracy, and has the feasibility and practical value for clinical application.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a four-dimensional CT (Computed Tomography) reconstruction method based on real-time kV image driving, which comprises data preprocessing and four-dimensional CT reconstruction, the data preprocessing comprises respiration time phase synchronization, space resampling and geometric consistency registration cutting, the four-dimensional CT reconstruction is realized by adopting a deep learning model based on a three-dimensional U-Net framework, model training adopts a supervised learning normal form, and the model training adopts a three-dimensional U-Net framework. According to the model, normalized prior three-dimensional CT body data and a two-dimensional kV projection image of a corresponding time phase are used as joint input, and a mapping relation between the two-dimensional projection image and the three-dimensional body data is learned through a multi-scale feature coding and decoding structure, so that the three-dimensional CT body data under the current breathing time phase is predicted and output. According to the invention, the end-to-end closed loop of two-dimensional projection-four-dimensional CT-dose evaluation is realized, the radiation dose required by image reconstruction is reduced, the reconstruction speed is improved, the spatial precision and time continuity are ensured, and the method has the feasibility and practical value of clinical popularization and application.
Need to check novelty before this filing date? Find Prior Art