A methodology to determine margins by EPID measurements of patient setup variation and motion as applied to immobilization devices

Joann I. Prisciandaro, Christina M. Frechette, Michael G. Herman, Paul D. Brown, Yolanda I. Garces, Robert L. Foote

Research output: Contribution to journalArticlepeer-review

24 Scopus citations


Assessment of clinic and site specific margins are essential for the effective use of three-dimensional and intensity modulated radiation therapy. An electronic portal imaging device (EPID) based methodology is introduced which allows individual and population based CTV-to-PTV margins to be determined and compared with traditional margins prescribed during treatment. This method was applied to a patient cohort receiving external beam head and neck radiotherapy under an IRB approved protocol. Although the full study involved the use of an EPID-based method to assess the impact of (1) simulation technique, (2) immobilization, and (3) surgical intervention on inter- and intrafraction variations of individual and population-based CTV-to-PTV margins, the focus of the paper is on the technique. As an illustration, the methodology is utilized to examine the influence of two immobilization devices, the UON™ thermoplastic mask and the Type-S™ head/ neck shoulder immobilization system on margins. Daily through port images were acquired for selected fields for each patient with an EPID. To analyze these images, simulation films or digitally reconstructed radiographs (DRR's) were imported into the EPID software. Up to five anatomical landmarks were identified and outlined by the clinician and up to three of these structures were matched for each reference image. Once the individual based errors were quantified, the patient results were grouped into populations by matched anatomical structures and immobilization device. The variation within the subgroup was quantified by calculating the systematic and random errors (Σsub and σsub). Individual patient margins were approximated as 1.65 times the individual-based random error and ranged from 1.1 to 6.3 mm (A-P) and 1.1 to 12.3 mm (S-I) for fields matched on skull and cervical structures, and 1.7 to 10.2 mm (L-R) and 2.0 to 13.8 mm (S-I) for supraclavicular fields. Population-based margins ranging from 5.1 to 6.6 mm (A-P) and 3.7 to 5.7 mm (S-I) were calculated for the corresponding skull/cervical field and 9.3 to 10.0 mm (L-R) and 6.3 to 6.6 mm (S-I) for the supraclavicular fields, respectively. The reported CTV-to-PTV margins are comparable to a value 7-15 mm based on traditional Mayo margins, but in some cases exceed the default values established in RTOG Head and Neck studies. The data suggests that the population-based margins provide sufficient coverage for the majority of the patients. However, the population-derived margins are excessive for some patients and insufficient for others, suggesting that a re-evaluation of current treatment margins for individual patients is warranted. Finally, this methodology provides direct evidence of treatment variation and thus can demonstrate with confidence, the superiority of one technique over another.

Original languageEnglish (US)
Pages (from-to)2978-2988
Number of pages11
JournalMedical physics
Issue number11
StatePublished - Nov 2004


  • Electronic portal imaging
  • Head and neck cancer
  • Treatment margins

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging


Dive into the research topics of 'A methodology to determine margins by EPID measurements of patient setup variation and motion as applied to immobilization devices'. Together they form a unique fingerprint.

Cite this