Dose Reduction in 4-Dimensional Computed Tomography Imaging: Breathing Signal-Guided Deep Learning-Driven Data Acquisition

Lukas Wimmert; Tobias Gauer; Jannis Dickmann; Christian Hofmann; Thilo Sentker; Rene Werner

doi:10.1016/j.ijrobp.2025.08.047

Dose Reduction in 4-Dimensional Computed Tomography Imaging: Breathing Signal-Guided Deep Learning-Driven Data Acquisition

Int J Radiat Oncol Biol Phys. 2025 Sep 18:S0360-3016(25)06194-2. doi: 10.1016/j.ijrobp.2025.08.047. Online ahead of print.

Authors

Lukas Wimmert¹, Tobias Gauer², Jannis Dickmann³, Christian Hofmann³, Thilo Sentker¹, Rene Werner⁴

Affiliations

¹ Institute for Applied Medical Informatics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; Institute of Computational Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; Center for Biomedical Artificial Intelligence (bAIome), University Medical Center Hamburg-Eppendorf, Hamburg, Germany.
² Department of Radiotherapy and Radiation Oncology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany.
³ Siemens Healthineers AG, Forchheim, Germany.
⁴ Institute for Applied Medical Informatics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; Institute of Computational Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; Center for Biomedical Artificial Intelligence (bAIome), University Medical Center Hamburg-Eppendorf, Hamburg, Germany. Electronic address: r.werner@uke.de.

PMID: 40975131
DOI: 10.1016/j.ijrobp.2025.08.047

Abstract

Purpose: Four-dimensional computed tomography (4D CT) imaging is essential for radiation therapy planning in thoracic tumors. However, current protocols tend to acquire more projection data than is strictly necessary for reconstructing the 4D CT, potentially leading to unnecessary radiation exposure and a misalignment with the ALARA (As Low As Reasonably Achievable) principle. We propose a deep learning (DL)-driven approach that uses the patient's breathing signal to guide data acquisition, aiming to acquire only necessary projection data.

Methods and materials: This retrospective study analyzed 1415 breathing signals from 294 patients, with a 75/25 training/validation split at the patient level. On the basis of the signals, a DL model was trained to predict optimal beam-on events for projection data acquisition. Model testing was performed on 104 independent clinical 4D CT scans. The performance of the model was assessed by measuring temporal alignment between predicted and optimal beam-on events. To assess the impact on the reconstructed images, each 4D CT data set was reconstructed twice: using all clinically acquired projections (reference) and using only the model-selected projections (dose-reduced). Reference and dose-reduced images were compared using Dice coefficients for organ segmentations, deformable image registration-based displacement fields, artifact frequency, and tumor segmentation agreement, the latter evaluated in terms of Hausdorff distance and tumor motion ranges.

Results: The proposed approach reduced beam-on time and imaging dose by a median of 29% (IQR, 24%-35%), corresponding to 11.6 mGy dose reduction for a standard 4D CT CTDIvol of 40 mGy. Temporal alignment between predicted and optimal beam-on events showed marginal differences. Similarly, reconstructed dose-reduced images showed only minimal differences to the reference images, demonstrated by high lung and liver segmentation Dice values, small-magnitude (deformable image registration) displacement fields, and unchanged artifact frequency. Minor deviations of tumor segmentation and motion ranges compared with the reference suggest only minimal impact of the proposed approach on treatment planning.

Conclusions: The proposed DL-driven data acquisition approach has the ability to reduce radiation exposure during 4D CT imaging while preserving diagnostic quality, offering a clinically viable, ALARA-adhering solution for 4D CT imaging.