Interfraction variation in prostate cancer - analysis of 11726 cone-beam CT

J BUON. 2015 Jul-Aug;20(4):1081-7.

Abstract

Purpose: To investigate the setup margins in prostate cancer treatment without using daily online repositioning methods.

Methods: We analysed the data from patients treated with curative-intend radiotherapy. Each patient underwent a series of pretreatment online localizations during daily setup using conebeam CT. The skin-prostate shifts and bone-prostate shifts were recorded in anteroposterior (AP), craniocaudal (CC), and laterolateral (LL)direction. The safety margins based on van Herk equation (2.5Σ+0.7σ) were calculated and the correlations between margins and various patient characteristics and prostate locations were investigated.

Results: A total of 307 patients were included, representing 11,726 localisations resulting in 70,356 shifts. The man skin-prostate setup inaccuracy was 0.8 ± 5.4mm in AP, 1.3 ± 4.8mm for CC, and 0.1 ± 5.6mm in LL direction. The man bone-prostate setup inaccuracy was 0.4 ± 3.3mm in AP, 0.1 ± 2.5mm for CC, and 0.1 ± 1.4mm in LL direction. According to van Herk equation, clinical target volume (CTV)-planning target volume (PTV) margins of 11.4, 10.6, and 11.8 mm (AP, CC, and LL, respectively) would be required for setup using skin markers and margins of 7.0, 4.7, and 2.1mm would be necessary for setup using bone structures. The average rectal area < 11cm(2) and volume of bladder > 300 cm(3) were associated with smaller CTV-PTV margins for setup using bone structures. The largest margins (15.8 mm in LL direction) were needed in patients with body mass index (BMI) > 35 using skin markers.

Conclusions: Our results confirm that the commonly used CTV-PTV margins are inadequate.

MeSH terms

  • Cone-Beam Computed Tomography*
  • Humans
  • Male
  • Prostatic Neoplasms / diagnostic imaging*
  • Prostatic Neoplasms / radiotherapy