Influence of air mapping errors on the dosimetric accuracy of prostate CBCT‐guided online adaptive radiation therapy

Olga M Dona Lemus, Sean Tanny, Michael Cummings, Matthew J. Webster, Joshua Wancura, Hyunuk Jung, Yuwei Zhou, Jihyung Yoon, Matthew Pacella, Dandan ZhengJournal of Applied Clinical Medical Physics 2023 · DOI

TL;DR

This study investigates how inaccuracies in mapping air pockets in the abdomen and pelvis between daily CBCT scans and synthetic CT (sCT) images affect radiation dose calculations for prostate cancer patients undergoing online adaptive radiotherapy (oART). While significant air volume discrepancies were found, the dosimetric impact on the planned radiation dose was minimal, suggesting that current oART workflows are robust enough for prostate treatments despite these mapping errors.

What Problem Does This Paper Address?

Modern radiation therapy increasingly uses online adaptive radiotherapy (oART) to adjust treatment plans daily based on changes in a patient's anatomy. A key component of this process involves generating a synthetic CT (sCT) from daily cone-beam CT (CBCT) images, often using deformable image registration (DIR) algorithms. However, these algorithms struggle to accurately reproduce the variable volumes of gas in the abdominopelvic region, which can significantly change between treatments due to factors like diet and digestion. These "air mapping errors" between the CBCT and the sCT could potentially lead to inaccurate dose calculations for the tumor and surrounding healthy tissues. This paper aims to quantify these air mapping errors and assess their dosimetric consequences in the context of prostate cancer oART.

Methods

The researchers analyzed data from 207 oART sessions for six prostate cancer patients. They focused on the abdominopelvic air volume, segmenting it on both the daily CBCT images and the corresponding sCT images. Air mapping errors were quantified by comparing the volumes and tracking variations over treatment fractions.

To assess the dosimetric impact, two "worst-case" scenarios with the largest air mapping errors were selected. For these cases, the sCT was "corrected" by overriding the air volume with water density and then reintroducing the air volume from the CBCT with a specific density. The radiation dose was then recalculated on both the original and corrected sCTs using the Varian Ethos system's planning software. Dosimetric differences were evaluated using dose-volume histogram (DVH) metrics and 3D gamma passing rates (comparing dose distributions with specific tolerances for dose difference and distance to agreement).

Datasets

The study utilized data from 207 daily adaptive sessions for six prostate cancer patients treated on a Varian Ethos system between July 2021 and July 2022. This dataset represents real-world clinical data from patients undergoing oART for prostate bed and lymph node treatments. No new benchmark datasets were introduced; rather, existing clinical CBCT and sCT images were used for analysis.

Key Insights and Ideas

  • DIR's struggle with gas: Deformable image registration (DIR) algorithms, while effective for many anatomical changes, are inherently limited in accurately reproducing the dynamic and often large volume changes of gas in the abdominopelvic region. This is due to the low contrast of air and the significant deformations involved.
  • Underestimation of air: The synthetic CTs (sCTs) generated by the system tended to underestimate the volume of air present in the daily CBCT images. This suggests that the DIR process, when encountering large air pockets, often smooths them out or fails to fully incorporate them.
  • Dosimetric impact is surprisingly small: Despite observing substantial air volume discrepancies (up to 276.3 cc, a >386% difference), the impact on critical dose metrics for the target volumes was minimal, generally less than 2.6% for key parameters. This indicates a degree of robustness in the adaptive planning process.
  • Prostate treatments are resilient: For prostate cancer treatments, the current oART workflow appears to be sufficiently accurate even with these air mapping errors, as the dosimetric deviations are unlikely to exceed clinically significant thresholds (e.g., 3%).

Key Results

  • Large air volume errors: The largest observed air volume difference between CBCT and sCT for a single patient session was 276.3 cc, representing a difference of over 386% compared to the sCT.
  • Minimal DVH changes: In the two case studies analyzed for dosimetric impact, the largest change in DVH metrics was a 2.6% reduction in the minimum PTV dose. Other DVH metrics varied by less than 1.5%.
  • High gamma passing rates: 3D gamma passing rates, using a stringent 1%/1 mm criterion, remained above 90% when comparing the uncorrected and corrected dose distributions for the two case studies. PTV-specific gamma passing rates were lower in some instances, but overall dose metrics were still within acceptable limits.

Limitations and Future Work

  • Limited cohort size: The study included six patients and 207 sessions, which, while substantial, could be expanded to include a larger and more diverse patient population.
  • Focus on prostate: The findings are specific to prostate cancer treatments. The impact of air mapping errors might differ for other cancer sites with different target volumes and surrounding anatomy.
  • OARs not fully assessed: While the paper mentions that organs at risk (OARs) could be affected, a detailed analysis of OAR dosimetric impact was not performed for all cases.
  • Future work: The authors suggest that while manual correction of air volumes is not recommended due to time constraints and potential for introducing new errors, algorithmic improvements to better handle air cavity mapping in an automated fashion could be beneficial, especially for patients with more extreme air volume fluctuations. They also emphasize the importance of monitoring abdominal gas and implementing strategies to reduce it, as it affects both dose calculation accuracy and image quality.

Summary updated

Abstract

Current CBCT-based oART workflow can lead to inaccuracies in the mapping of abdominopelvic air pockets from daily CBCT to the sCT images used for the optimization and calculation of the adaptive plan. Despite the large observed mapping errors, the dosimetric effects of such differences on the accuracy of the adapted plan dose calculation are unlikely to cause differences greater than 3% for prostate treatments.
BibTeX
@article{lemus2023influence,
  author = {Olga M Dona Lemus and Sean Tanny and Michael Cummings and Matthew J. Webster and Joshua Wancura and Hyunuk Jung and Yuwei Zhou and Jihyung Yoon and Matthew Pacella and Dandan Zheng},
  title = {{Influence of air mapping errors on the dosimetric accuracy of prostate CBCT‐guided online adaptive radiation therapy}},
  journal = {Journal of Applied Clinical Medical Physics},
  year = {2023},
  doi = {10.1002/acm2.14057},
}
Figure