In vivo dosimetry with asymmetry and heterogeneity correction

Tsi Chian Chao, Pei Chieh Yu, Chung Chi Lee, Ching Jung Wu, Chuan Jong Tung*

*Corresponding author for this work

Research output: Contribution to journalJournal Article peer-review


In vivo dosimetry, which is commonly practised in Europe and North America, is a purely measurement-based check which can independently verify the entire treatment procedure. In vivo dosimetry can be performed using either one or two detectors (i.e. diodes), and the later one applying two detectors in both the entrance and exit surface can independently verify the patients' dose. However, this type of two-detector measurement can only be implemented with tumours of symmetric geometry at the midline, and can only be applied to opposite large-field irradiation. Nasopharyngeal cancer (NPC) is the most interesting symmetric cancer to medical physicists in Asia since it preferentially occurs in 40-50 year old males who have familial responsibilities. The tissues/structures around the nasopharynx are more complicated than those around the prostate, and may introduce unwanted bias to in vivo dosimetry. Asymmetry and inhomogeneity are the most important systematic uncertainties with in vivo dosimetry. This study demonstrates the bias of in vivo dosimetry caused by tissue asymmetry and heterogeneity, and provides corrections accordingly. We have demonstrated that asymmetry correction reduced the uncertainty of deviation between in vivo dosimetry and the prescribed dose from 3.12% to 2.39% for in vivo dosimetry of 102 patients. Heterogeneity correction reduced the mean deviation between in vivo dosimetry and the prescribed dose from 0.56% to 0.22%.

Original languageEnglish
Pages (from-to)1956-1959
Number of pages4
JournalRadiation Measurements
Issue number12
StatePublished - 12 2011


  • Asymmetry correction
  • Diodes
  • In vivo dosimetry
  • Inhomogeneity correction


Dive into the research topics of 'In vivo dosimetry with asymmetry and heterogeneity correction'. Together they form a unique fingerprint.

Cite this