Does Image Guidance Need to be Performed Daily in the Treatment of Localized Prostate Cancers? Implications on Treatment Margins

Article Outline

• Purpose/Objective(s)
• Materials/Methods
• Results
• Conclusions
• Appendix. Supplementary data
• Copyright

Back to Article Outline

Purpose/Objective(s) 

To compare the residual localization errors with different image-guidance (IG) strategies utilized in the alignment of prostate cancer patients. Using data from patients who were treated with daily IG, the remaining set-up errors for several different strategies were retrospectively calculated.

Back to Article Outline

Materials/Methods 

A total of 74 patients were treated with daily image guidance with implanted metallic fiducials on a Tomotherapy machine. Daily positional adjustment data were available for a total of 2252 fractions. Given daily positional adjustments, a variety of protocols differing in imaging frequency and methodology were retrospectively studied. The 8 protocols were: 1) No-imaging, 2) Initial fraction only, 3) Mean of initial 3 fractions, 4) Mean of initial 5 fractions, 5) Mean of initial 7 fractions, 6) Weekly imaging, 3 mm threshold, 7) First 5 fractions + weekly imaging, patient specific threshold, 8) Imaging every other fraction, running mean. The imaging frequency ranged from 0 to 49% for the different protocols. On days with no imaging, residual location errors were retrospectively determined. Dimensional treatment margins based on systematic and random errors as determined by van Herk's formula (IJROBP 47, 1121–1135, 2000) were calculated for each scenario for fractions when imaging would not be performed. For ths study, a tight margin plan is utilized when IG is performed, and a larger margin plan is utilized on days when IG is not performed.

Back to Article Outline

Results 

As expected, systematic errors were effectively reduced with increasing imaging frequency. However, the random errors were unaffected and contributed to significant residual errors on days imaging was not performed, regardless of the protocol used. For the 8 different imaging protocols, the data is summarized in Table 1 for the different imaging frequencies. The residual errors exceeding 5 mm and 10 mm is reported and shows a decrease in error with increase in frequency of imaging. Margins were reduced with increasing imaging frequency, following the reduction in the systematic component of the alignment errors.

Back to Article Outline

Conclusions 

With varying imaging frequencies, residual errors in prostate location are significant even with every other day imaging, mostly due to a significant residual random component. Decreasing imaging frequencies result in substantial increases in the size of treatment margins. Since localization strategies other than daily imaging result in significant residual errors requiring large treatment margins, daily imaging and positional adjustments are recommended for the set up of prostate cancer patients during a course of external beam radiotherapy.

Back to Article Outline

Appendix. Supplementary data 

Back to Article Outline

Appendix. Supplementary data 

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ijrobp.2007.07.149