Image Guidance for External Beam Partial Breast Irradiation

Reviewer: Eric Shinohara MD, MSCI
Abramson Cancer Center of the University of Pennsylvania
Last Modified: October 30, 2007

Presenter: S.H. Wahab, MD
Presenter's Affiliation: Washington University
Type of Session: Scientific

Background

Partial Breast Irradiation (PBI) is currently being investigated in several studies, and at present, the ideal system to use for proper setup is still not established. PBI is more susceptible to random setup errors compared with whole breast radiation, thereby increasing the risk for geographic misses.
This study investigates whether laser alignment provides sufficient accuracy for highly conformal breast treatment. Additionally, this study evaluates the use of 3D video surface mapping (3DSM), kilovoltage on-board imaging (OBI), and fiducial markers (IFM) as aids for more accurate setup.

Materials and Methods

The present intramural study was comprised of 37 patients treated with external beam PBI.
All patients were treated with nine fractions of 4 Gy each, over the course of one week.
At least six surgical clips were placed during lumpectomy and used as FM in approximately half of the patients in this study. FMs were used as the "gold standard" method, and thus served as the baseline reference to determine the error with other setup modalities.
Laser alignment was performed on all patients initially.
Subsequently, orthogonal imaging was obtained using OBI (Trilogy, Varian Medical Systems) based on skeletal landmarks. A second set of orthogonal images was obtained using FM if available. Isocenter shifts from the laser alignment based on these orthogonal images were recorded, but the patient was not actually shifted.
3DSM was also performed (AlignRT, Vision RT Ltd). An initial surface map was created based on surface contours from CT simulation. A region of interest was then defined on the initial surface map. The patient then underwent daily 3DSM with automatic topographic shifts based on a comparison of the daily map with the initial surface map.
Residual setup error was defined as the difference between the laser alignment and that of the other setup modality used.

Results

Based on 3DSM setup, there was an average shift of 3.4 to 4.6 mm (depending on which axis the shift was in) from the isocenter that was defined by laser alignment.
Based on orthogonal imaging and using skeletal anatomy, there was an average shift of 3.5 to 4.2 mm.
Laser alignment, orthogonals based on skeletal anatomy, and 3DSM were compared with FM (residual error).
- The residual error for laser alignment was ≤ 5 mm
- The residual error for orthogonals was ≤ 5 mm
- The residual error for 3DSM was ≤ 2 mm
Reproducibility of these setup modalities was also investigated. The difference in the first and second setup isocenters was determined and compared (using the FM reference).
- There was a difference of 0.9 mm ± 1.1 mm using 3DSM.
- There was a difference of 1.9 mm ± 2.4 mm using orthogonals.

Author's Conclusions

Laser alignment alone is inadequate for PBI.
3DSM was the most accurate and reproducible setup modality of the modalities studied.
If laser alignment or OBI is used the authors recommended the use of FM in conjunction with them. However, if 3DSM is used the authors suggested that their accuracy (to ≤ 2 mm) was adequate and FM were not necessary.
Breathing motion was not taken into account in the present study but the 3DSM system is capable of respiratory gating and these studies are ongoing.

Clinical/Scientific Implications

PBI remains investigational, but studies of its use are expanding. In particular, the use of external beam PBI is increasing and compared with brachytherapy PBI the daily setup can be more difficult. Geographical misses could seriously compromise local control and optimal setup is required.
The highest accuracy for setup was with surgical clips. Currently, there is not a way that surgical clips can be used for an automated alignment like 3DCM, but there may be a way to develop this from orthogonal imaging that is taken based on the surgical clips. Additionally, systems like radiofrequency FM (i.e. Calypso^TM) may be of value as they permit real-time imaging to account for respiration, and automated tables for shift may soon be available.

Partially funded by an unrestricted educational grant from Bristol-Myers Squibb.