# High-Precision Proton Therapy for Liver Cancer: Developing an End-to-end Strategy with Real-time Liver Tumor Localization and On-the-fly Plan Delivery Adaptation

> **NIH NIH R01** · UT SOUTHWESTERN MEDICAL CENTER · 2024 · $615,504

## Abstract

PROJECT SUMMARY
 The recent advancement of stereotactic body radiotherapy (SBRT) enables highly focused dose delivery to
tumors while sparing surrounding normal tissues. Emerging clinical evidence showing superior tumor control
and patient survival suggests that SBRT will play an increasingly critical role in liver tumor management.
However, radiation-induced toxicity, especially of the dose-sensitive normal liver tissues, poses a lingering
challenge to liver radiotherapy and SBRT. Proton therapy, with its unique ‘Bragg peaks’, allows simultaneous
tumor dose escalation and normal tissue sparing, paving the way to boost the efficacy and safety of liver SBRT.
Unfortunately, precisely delivering planned proton doses to liver tumors is significantly challenged by the intra-
treatment liver motion, especially for the increasingly dominant pencil beam scanning (PBS) technique. Current
motion management techniques including abdominal compression, beam re-scanning, and respiratory gating,
however, all suffer from various sources of uncertainties and inaccuracy, resulting in limited and case-
dependent efficacy. Motion tracking, a technique that tracks the real-time tumor motion to accordingly adjust
the beam delivery, can systematically and fundamentally address the motion challenge. However, such a
technique is not yet available, and is currently facing three major challenges for the site of the liver: 1). The
respiration-induced fast liver motion requires real-time imaging within hundreds of milliseconds, which results
in extremely limited sampling that makes 3D deformable motion tracking challenging. 2). The low contrast of
liver tumors against surrounding normal liver parenchyma adds another layer of uncertainty to tumor motion
tracking. 3). The beam-tracking technique is purely ‘geometry-guided’, which requires fast energy switching
that is difficult to achieve in current proton systems. It also fails to consider cumulative tracking errors caused
by system latency, motion prediction error, and dose changes due to anatomical deformation. The overarching
goal of this project is to develop a real-time imaging and proton plan adaptation (RIPA) system, composed of
two sub-systems (MeshBioNet and MAO) to solve real-time 3D deformable motion (MeshBioNet) for
simultaneous dose calculation, accumulation, and on-the-fly proton plan adaptation (MAO). MeshBioNet uses
Artificial Intelligence-driven methods to solve real-time 3D deformable motion from a single x-ray projection.
MAO features the first closed-loop, ‘dosimetry-guided’ framework that actively monitors and adapts the proton
dose to ensure its matching with the planned dose without requiring tracking-related fast energy switching. We
have three Specific Aims: 1) Develop and optimize the real-time imaging sub-system (MeshBioNet), 2)
Develop and optimize the on-the-fly proton plan delivery adaptation sub-system (MAO), and 3). Test RIPA via
a pre-clinical motion-enabled phantom study and a retrospective p...

## Key facts

- **NIH application ID:** 10801300
- **Project number:** 1R01CA280135-01A1
- **Recipient organization:** UT SOUTHWESTERN MEDICAL CENTER
- **Principal Investigator:** Weiguo Lu
- **Activity code:** R01 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2024
- **Award amount:** $615,504
- **Award type:** 1
- **Project period:** 2024-03-01 → 2029-02-28

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10801300

## Citation

> US National Institutes of Health, RePORTER application 10801300, High-Precision Proton Therapy for Liver Cancer: Developing an End-to-end Strategy with Real-time Liver Tumor Localization and On-the-fly Plan Delivery Adaptation (1R01CA280135-01A1). Retrieved via AI Analytics 2026-05-27 from https://api.ai-analytics.org/grant/nih/10801300. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*