SU‐C‐WAB‐04: Auto‐Contouring Accuracy in the Presence of a Rectal Balloon

A. Godley, A. Stockham, K. Stephans

Research output: Contribution to journalArticle

Abstract

Purpose: To determine if the presence of rectal balloons improves the accuracy of prostate and rectum auto‐contouring on kV cone‐beam CTs (CBCTs). Methods: Seven patients undergoing five fraction prostate SBRT were selected. Rectal balloons were placed and consistently inflated at simulation and treatment to increase organ stability and reproducibility. All five pre‐treatment CBCTs were contoured using ABAS (Elekta) which deformably registers the plan CT or previous CBCTs to the current CBCT and then morphs their contours according to this deformation field. Six patients undergoing 40 fraction prostate IMRT without rectal balloons were also chosen. Nine fractions each were contoured similarly with ABAS. All scans were manually contoured by a physician and Dice Coefficient (DC) calculated between the physician and auto‐contours as our measure of accuracy. Results: Average accuracy of auto‐to physician rectal contours significantly improved with rectal balloons from DC of 0.85 to 0.91 (P<0.0001). The prostate contour accuracy also improved from DC 0.84 to 0.88 (P=0.0016). Balloons reduced variability between the plan and treatment image, allowing the registration to successfully recover the deformation and morph the contours. Bladder filling is the overriding factor in its auto‐contouring, so its accuracy is not affected by the rectal balloon, DC 0.91 with, 0.92 without. Prostate auto‐contouring was more accurate than the intra‐observer agreement DC of 0.83, and consistent with the bladder agreement of 0.92. The rectal balloon significantly boosted intra‐observer agreement from 0.85 to 0.97 (P<0.001). This is higher than auto‐contouring accuracy, indicating it could be further improved. Visually the tissue air boundary is clear but CBCTs have a scan dependent, ∼5 mm border of rapidly rising intensity that currently confounds the deformation. Auto‐contours are produced in ∼20 seconds. Conclusion: The use of rectal balloons significantly improves the accuracy of prostate and rectum auto‐contours on daily CBCTs, enabling online adaptive therapy. A research version of ABAS was provided by Elekta AB, Stockholm, Sweden.

Original languageEnglish (US)
Number of pages1
JournalMedical physics
Volume40
Issue number6
DOIs
StatePublished - Jun 2013
Externally publishedYes

Fingerprint

Prostate
Physicians
Rectum
Urinary Bladder
Sweden
Therapeutics
Air
Research

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

SU‐C‐WAB‐04 : Auto‐Contouring Accuracy in the Presence of a Rectal Balloon. / Godley, A.; Stockham, A.; Stephans, K.

In: Medical physics, Vol. 40, No. 6, 06.2013.

Research output: Contribution to journalArticle

@article{6d2718ee21944e91a27788c7c65989c1,
title = "SU‐C‐WAB‐04: Auto‐Contouring Accuracy in the Presence of a Rectal Balloon",
abstract = "Purpose: To determine if the presence of rectal balloons improves the accuracy of prostate and rectum auto‐contouring on kV cone‐beam CTs (CBCTs). Methods: Seven patients undergoing five fraction prostate SBRT were selected. Rectal balloons were placed and consistently inflated at simulation and treatment to increase organ stability and reproducibility. All five pre‐treatment CBCTs were contoured using ABAS (Elekta) which deformably registers the plan CT or previous CBCTs to the current CBCT and then morphs their contours according to this deformation field. Six patients undergoing 40 fraction prostate IMRT without rectal balloons were also chosen. Nine fractions each were contoured similarly with ABAS. All scans were manually contoured by a physician and Dice Coefficient (DC) calculated between the physician and auto‐contours as our measure of accuracy. Results: Average accuracy of auto‐to physician rectal contours significantly improved with rectal balloons from DC of 0.85 to 0.91 (P<0.0001). The prostate contour accuracy also improved from DC 0.84 to 0.88 (P=0.0016). Balloons reduced variability between the plan and treatment image, allowing the registration to successfully recover the deformation and morph the contours. Bladder filling is the overriding factor in its auto‐contouring, so its accuracy is not affected by the rectal balloon, DC 0.91 with, 0.92 without. Prostate auto‐contouring was more accurate than the intra‐observer agreement DC of 0.83, and consistent with the bladder agreement of 0.92. The rectal balloon significantly boosted intra‐observer agreement from 0.85 to 0.97 (P<0.001). This is higher than auto‐contouring accuracy, indicating it could be further improved. Visually the tissue air boundary is clear but CBCTs have a scan dependent, ∼5 mm border of rapidly rising intensity that currently confounds the deformation. Auto‐contours are produced in ∼20 seconds. Conclusion: The use of rectal balloons significantly improves the accuracy of prostate and rectum auto‐contours on daily CBCTs, enabling online adaptive therapy. A research version of ABAS was provided by Elekta AB, Stockholm, Sweden.",
author = "A. Godley and A. Stockham and K. Stephans",
year = "2013",
month = "6",
doi = "10.1118/1.4813956",
language = "English (US)",
volume = "40",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "6",

}

TY - JOUR

T1 - SU‐C‐WAB‐04

T2 - Auto‐Contouring Accuracy in the Presence of a Rectal Balloon

AU - Godley, A.

AU - Stockham, A.

AU - Stephans, K.

PY - 2013/6

Y1 - 2013/6

N2 - Purpose: To determine if the presence of rectal balloons improves the accuracy of prostate and rectum auto‐contouring on kV cone‐beam CTs (CBCTs). Methods: Seven patients undergoing five fraction prostate SBRT were selected. Rectal balloons were placed and consistently inflated at simulation and treatment to increase organ stability and reproducibility. All five pre‐treatment CBCTs were contoured using ABAS (Elekta) which deformably registers the plan CT or previous CBCTs to the current CBCT and then morphs their contours according to this deformation field. Six patients undergoing 40 fraction prostate IMRT without rectal balloons were also chosen. Nine fractions each were contoured similarly with ABAS. All scans were manually contoured by a physician and Dice Coefficient (DC) calculated between the physician and auto‐contours as our measure of accuracy. Results: Average accuracy of auto‐to physician rectal contours significantly improved with rectal balloons from DC of 0.85 to 0.91 (P<0.0001). The prostate contour accuracy also improved from DC 0.84 to 0.88 (P=0.0016). Balloons reduced variability between the plan and treatment image, allowing the registration to successfully recover the deformation and morph the contours. Bladder filling is the overriding factor in its auto‐contouring, so its accuracy is not affected by the rectal balloon, DC 0.91 with, 0.92 without. Prostate auto‐contouring was more accurate than the intra‐observer agreement DC of 0.83, and consistent with the bladder agreement of 0.92. The rectal balloon significantly boosted intra‐observer agreement from 0.85 to 0.97 (P<0.001). This is higher than auto‐contouring accuracy, indicating it could be further improved. Visually the tissue air boundary is clear but CBCTs have a scan dependent, ∼5 mm border of rapidly rising intensity that currently confounds the deformation. Auto‐contours are produced in ∼20 seconds. Conclusion: The use of rectal balloons significantly improves the accuracy of prostate and rectum auto‐contours on daily CBCTs, enabling online adaptive therapy. A research version of ABAS was provided by Elekta AB, Stockholm, Sweden.

AB - Purpose: To determine if the presence of rectal balloons improves the accuracy of prostate and rectum auto‐contouring on kV cone‐beam CTs (CBCTs). Methods: Seven patients undergoing five fraction prostate SBRT were selected. Rectal balloons were placed and consistently inflated at simulation and treatment to increase organ stability and reproducibility. All five pre‐treatment CBCTs were contoured using ABAS (Elekta) which deformably registers the plan CT or previous CBCTs to the current CBCT and then morphs their contours according to this deformation field. Six patients undergoing 40 fraction prostate IMRT without rectal balloons were also chosen. Nine fractions each were contoured similarly with ABAS. All scans were manually contoured by a physician and Dice Coefficient (DC) calculated between the physician and auto‐contours as our measure of accuracy. Results: Average accuracy of auto‐to physician rectal contours significantly improved with rectal balloons from DC of 0.85 to 0.91 (P<0.0001). The prostate contour accuracy also improved from DC 0.84 to 0.88 (P=0.0016). Balloons reduced variability between the plan and treatment image, allowing the registration to successfully recover the deformation and morph the contours. Bladder filling is the overriding factor in its auto‐contouring, so its accuracy is not affected by the rectal balloon, DC 0.91 with, 0.92 without. Prostate auto‐contouring was more accurate than the intra‐observer agreement DC of 0.83, and consistent with the bladder agreement of 0.92. The rectal balloon significantly boosted intra‐observer agreement from 0.85 to 0.97 (P<0.001). This is higher than auto‐contouring accuracy, indicating it could be further improved. Visually the tissue air boundary is clear but CBCTs have a scan dependent, ∼5 mm border of rapidly rising intensity that currently confounds the deformation. Auto‐contours are produced in ∼20 seconds. Conclusion: The use of rectal balloons significantly improves the accuracy of prostate and rectum auto‐contours on daily CBCTs, enabling online adaptive therapy. A research version of ABAS was provided by Elekta AB, Stockholm, Sweden.

UR - http://www.scopus.com/inward/record.url?scp=85024774437&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024774437&partnerID=8YFLogxK

U2 - 10.1118/1.4813956

DO - 10.1118/1.4813956

M3 - Article

AN - SCOPUS:85024774437

VL - 40

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 6

ER -