The delivered dose assessment in pancreas SBRT with the target position determined using an in-house position monitoring system

Sankar Arumugam, Tony Young, Meredith Johnston, Darren Pavey, Mark LeeFrontiers in Oncology 2022 · DOI

This study, published in Frontiers in Oncology, evaluates a novel in-house software system called SeedTracker designed to provide real-time target tracking for pancreatic cancer patients undergoing Stereotactic Body Radiotherapy (SBRT) on standard clinical hardware.

TL;DR

Researchers developed a real-time, image-based monitoring system to track pancreatic tumor motion during radiation delivery on standard linear accelerators. The study reveals that even when motion stays within a "safe" 3mm tolerance, it can significantly increase the radiation dose to the sensitive duodenum; however, real-time gating and realignment effectively mitigate these risks.

What Problem Does This Paper Address?

Pancreatic SBRT involves delivering high, ablative doses of radiation to a tumor that is notoriously mobile due to respiration, stomach fullness, and peristalsis. Because the pancreas sits immediately adjacent to highly radiosensitive "Organs at Risk" (OARs)—specifically the duodenum, stomach, and bowel—even a few millimeters of unexpected movement can result in under-dosing the tumor or, more dangerously, over-dosing the gut, leading to severe toxicity.

While specialized machines like CyberKnife offer built-in tracking, most clinics use general-purpose C-arm linear accelerators (linacs) that lack integrated real-time target monitoring. The authors addressed this gap by creating a software solution to enable real-time tracking on standard equipment and quantifying exactly how much "delivered dose" deviates from "planned dose" due to internal motion.

Methods

The authors conducted a study on eight patients from the MASTERPLAN pilot trial. Key technical components included:

  • SeedTracker: An in-house developed software that processes monoscopic x-ray images (from the Elekta XVI system) in real-time to identify the 3D position of implanted radiopaque fiducial markers.
  • Motion Management: Patients used various techniques including Exhale Breath Hold (EBH), Abdominal Compression (AC), or Free Breathing (FB) based on their tolerance.
  • Gating Protocol: A position tolerance of ±3mm was set. If the markers strayed beyond this for more than 5 seconds, a "Gating Event" (GE) was triggered, and the patient was realigned.
  • Dose Reconstruction: To calculate the actual delivered dose, the researchers convolved the plan's "control point dose matrices" (the specific instructions for the radiation beams) with the temporal target positions recorded during the actual treatment sessions.

Datasets

The study utilized clinical data from 8 patients treated within the MASTERPLAN pilot trial (ACTRN 12617001642370). The dataset included 4D-CT planning scans, real-time 2D x-ray trajectories, and post-treatment CBCT (Cone Beam CT) verification data.

Key Insights and Ideas

  • The "Safe" 3mm Window is a Simplification: One of the most important takeaways is that maintaining a target within a 3mm tolerance window does not guarantee the planned dose distribution. Cumulative motion within that window still shifts the dose, particularly impacting the high-dose gradients near OARs.
  • Duodenal Vulnerability: The duodenum is the most critical organ at risk in this scenario. The study found that without real-time correction, the maximum dose ($D_{max}$) to the duodenum increased significantly more than the dose to the stomach or small bowel.
  • Software vs. Hardware: The research demonstrates that sophisticated real-time tracking can be "retrofitted" onto standard linacs via software, potentially expanding the availability of high-precision SBRT without requiring specialized, dedicated robotic platforms.

Key Results

  • Frequency of Motion: Gating events (where treatment had to stop for realignment) occurred in 38% of all treatment fractions. In 10% of fractions, the patient moved out of bounds multiple times.
  • Dosimetric Impact (Uncorrected): If the researchers had not realigned the patients when they moved, the mean increase in $D_{max}$ to the duodenum would have been 1.9 Gy (with some cases seeing an increase as high as 4.6 Gy).
  • Effectiveness of Real-time Correction: By using SeedTracker to trigger realignments, the mean duodenal $D_{max}$ increase was reduced to 1.1 Gy.
  • Tumor Coverage: The minimum dose to the tumor was also stabilized. Without correction, deviations ranged from -1.1 to 0.4 Gy; with correction, this was tightened to a mean difference of only 0.1 Gy.

Limitations and Future Work

  • Sample Size: The study was limited to 8 patients. While the data is high-resolution, a larger cohort is needed to generalize the findings.
  • Surrogate Accuracy: The system tracks fiducial markers, not the tumor itself. While markers are standard, they can occasionally migrate or not perfectly represent the deformation of the pancreatic tissue.
  • Deformation: The current model accounts for translational shifts (moving up/down, left/right) but does not fully account for "deformation"—the tumor or OARs changing shape during treatment.
  • Future Direction: The authors suggest that these real-time motion insights should be used to develop "robust planning" techniques, where the initial radiation plan is mathematically optimized to be less sensitive to the specific motion patterns identified in this study.

Summary updated

Abstract

The in-house real-time position monitoring system improved the treatment accuracy of pancreatic SBRT in a general-purpose linac and enabled assessment of delivered dose by incorporating the temporal target position during delivery. The intrafraction motion impacts the dose to tumour even if target position is maintained within a 3mm position tolerance.
BibTeX
@inproceedings{arumugam2022delivered,
  author = {Sankar Arumugam and Tony Young and Meredith Johnston and Darren Pavey and Mark Lee},
  title = {{The delivered dose assessment in pancreas SBRT with the target position determined using an in-house position monitoring system}},
  booktitle = {Frontiers in Oncology},
  year = {2022},
  doi = {10.3389/fonc.2022.1009916},
}
Figure