[Definition of prostatic contours using tomodensitometric slices: study of differences among radiotherapists and between examinations]

Cancer Radiother. 1999 Jul-Aug;3(4):333-40. doi: 10.1016/s1278-3218(99)80076-6.
[Article in French]

Abstract

Accuracy of conformal treatment planning for prostatic radiotherapy is based on the contours of target volumes (prostate +/- seminal vesicles) and normal tissues (rectum and bladder), drawn on CT (computed tomography) images by radiation oncologists. The interpretation of a given CT image can be different from one radiation oncologist to another, and may change in time with the state of filling of the bladder and of the rectum during the treatment. In order to quantify these variations, 12 patients treated with conformal radiotherapy for prostate carcinoma (pelvis 40 Gy/20 sessions + prostate 30 Gy/15 sessions) had two series of CT at one month intervals. Contouring of prostate, rectum and bladder were performed independently on each CT by two radiation oncologists. The first CT scan (planning CT) and the first series of contours (planning contours) were used for treatment planning. The contours of the second scan were compared to the planning contours after image fusion based on manual superimposition of bony anatomy of the two sets of CT images. Coherence ratio were defined to measure discrepancies in prostate volumes between radiation oncologists (RCE) and between scans (RCT). The mean RCE was 38 +/- 7% (1 standard deviation). Those discrepancies were primarily located at the prostate apex and at the interface between bladder and prostate and between rectum and prostate. The mean RCT was 42 +/- 8% (1 sigma). Those discrepancies were due to the prostate motion related to the state of filling of the rectum and bladder. For bladder and rectal walls, less important differences were observed between the two radiation oncologists for the same CT (4.5% for rectal volume receiving 65 Gy or more, 3% for bladder volume receiving 65 Gy or more). However, important differences in bladder and rectal volumes receiving 65 Gy or more (16% and 7% respectively) were noted for the same patient from a CT to another due to the variation in bladder or rectal filling. New techniques for planning CT acquisition are needed to decrease the discrepancies due to contouring. The treatment must, as far as possible, be delivered with an empty bladder and rectum in order to ensure a good reproduction of the initially planned treatment.

Publication types

  • Comparative Study

MeSH terms

  • Anthropometry / methods*
  • Carcinoma / diagnostic imaging*
  • Humans
  • Male
  • Observer Variation
  • Prostate / diagnostic imaging
  • Prostatic Neoplasms / diagnostic imaging*
  • Rectum / diagnostic imaging
  • Reproducibility of Results
  • Seminal Vesicles / diagnostic imaging
  • Tomography, X-Ray Computed*
  • Urinary Bladder / diagnostic imaging