Background: Dose-volume histograms (DVHs) are used for the prediction and calculation of late radiation side effects. In literature the predictive value of rectal DVHs is controversially discussed. Differences in contouring might contribute to the contradicting results. In particular the cranial and caudal border of the contoured organ are not uniformly defined.
Patients and methods: The DVHs of 12 patients who were treated with conformal radiotherapy for prostate cancer were investigated. Six of the patients suffered from mild rectal bleeding as a late side effect of radiotherapy. Six patients without rectal bleeding (minimal follow-up 30 months) matched for age, concomitant disease and treatment concept served as controls. Four different DVHs with 4 different definitions of the cranial and caudal rectal border were generated for each patient. For each of the 48 DVHs the percent volume fractions (V50, V80, V95) and absolute volume fractions (aV50, aV80, aV95) were calculated that received more than 50%, 80% and 95% of the reference dose.
Results: For every patient there were considerable variations in the volume fractions depending on the definition of the rectum borders (Table 1). The mean and median values of the percent and absolute volume fractions of the bleeding patients were higher than those of the non-bleeders no matter how the rectum borders were defined. None of the volume fractions could totally separate bleeding from non-bleeding patients.
Conclusion: There is a high variability of absolute and percent volume fractions of rectal DVHs depending on how the rectal borders were defined. For the comparison and for the interpretation of rectal DVHs a uniform definition would be helpful.