SU‐FF‐J‐116

Registration of X‐Ray Portal Images with 4DCT DRRs for Patient Setup Verification

Q. xu, Z. he, Russell J Hamilton, L. Xing

Research output: Contribution to journalArticle

Abstract

Purpose: To provide a noninvasive method for both setup verification of 4DCT treatment planning and isocenter positioning shifts for treatment delivery. Method and Materials: A ten phase 4DCT treatment planning study of the patient was acquired using a respiratory gating system. When the patient was on the treatment couch, multiple on board x‐ray images were taken in the anterior‐posterior (AP) and lateral directions. An intensity based rigid registration algorithm was applied to obtain optimal shifts in the x, y and z directions. First, a digitally reconstructed radiograph (DRR) using ray‐casting algorithm was computed from one of the ten phases of the 4DCT. The correlation coefficient (CC) between the DRR and the x‐ray image was then calculated. The 4DCT volume was shifted and new DRRs were generated until the maximum CC value was reached through an optimization process. This procedure was repeated for all ten 4DCT phases. Results: Registration of each phase of the 4DCT with the x‐ray image provided ten maximum CC values. These values exhibited a single maximum at the phase corresponding to the breathing phase when the x‐ray image was taken. The corresponding isocenter positioning shifts for treatment delivery were also obtained. The robustness of our algorithm was demonstrated by registering x‐ray images taken at five random phases to 4DCT. The resulting isocenter shifts were consistent between all phases. The standard deviations of the shifts determined for the AP x‐rays were 2 mm (lateral), 5 mm (anterior‐posterior), and 2 mm (superior‐inferior). Conclusion: Using only the RPM system, one cannot be certain that the internal anatomy is consistent between 4DCT acquisition and any particular treatment day. Our noninvasive method accounts for internal organ motion and may be used for daily 4DCT treatment setup verification and isocenter positioning.

Original languageEnglish (US)
Pages (from-to)2047
Number of pages1
JournalMedical Physics
Volume33
Issue number6
DOIs
StatePublished - 2006

Fingerprint

X-Rays
Therapeutics
Respiratory System
Anatomy
Respiration
Direction compound

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

SU‐FF‐J‐116 : Registration of X‐Ray Portal Images with 4DCT DRRs for Patient Setup Verification. / xu, Q.; he, Z.; Hamilton, Russell J; Xing, L.

In: Medical Physics, Vol. 33, No. 6, 2006, p. 2047.

Research output: Contribution to journalArticle

@article{f2ee4384d7b3431cac599ecb76b40da6,
title = "SU‐FF‐J‐116: Registration of X‐Ray Portal Images with 4DCT DRRs for Patient Setup Verification",
abstract = "Purpose: To provide a noninvasive method for both setup verification of 4DCT treatment planning and isocenter positioning shifts for treatment delivery. Method and Materials: A ten phase 4DCT treatment planning study of the patient was acquired using a respiratory gating system. When the patient was on the treatment couch, multiple on board x‐ray images were taken in the anterior‐posterior (AP) and lateral directions. An intensity based rigid registration algorithm was applied to obtain optimal shifts in the x, y and z directions. First, a digitally reconstructed radiograph (DRR) using ray‐casting algorithm was computed from one of the ten phases of the 4DCT. The correlation coefficient (CC) between the DRR and the x‐ray image was then calculated. The 4DCT volume was shifted and new DRRs were generated until the maximum CC value was reached through an optimization process. This procedure was repeated for all ten 4DCT phases. Results: Registration of each phase of the 4DCT with the x‐ray image provided ten maximum CC values. These values exhibited a single maximum at the phase corresponding to the breathing phase when the x‐ray image was taken. The corresponding isocenter positioning shifts for treatment delivery were also obtained. The robustness of our algorithm was demonstrated by registering x‐ray images taken at five random phases to 4DCT. The resulting isocenter shifts were consistent between all phases. The standard deviations of the shifts determined for the AP x‐rays were 2 mm (lateral), 5 mm (anterior‐posterior), and 2 mm (superior‐inferior). Conclusion: Using only the RPM system, one cannot be certain that the internal anatomy is consistent between 4DCT acquisition and any particular treatment day. Our noninvasive method accounts for internal organ motion and may be used for daily 4DCT treatment setup verification and isocenter positioning.",
author = "Q. xu and Z. he and Hamilton, {Russell J} and L. Xing",
year = "2006",
doi = "10.1118/1.2240892",
language = "English (US)",
volume = "33",
pages = "2047",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "6",

}

TY - JOUR

T1 - SU‐FF‐J‐116

T2 - Registration of X‐Ray Portal Images with 4DCT DRRs for Patient Setup Verification

AU - xu, Q.

AU - he, Z.

AU - Hamilton, Russell J

AU - Xing, L.

PY - 2006

Y1 - 2006

N2 - Purpose: To provide a noninvasive method for both setup verification of 4DCT treatment planning and isocenter positioning shifts for treatment delivery. Method and Materials: A ten phase 4DCT treatment planning study of the patient was acquired using a respiratory gating system. When the patient was on the treatment couch, multiple on board x‐ray images were taken in the anterior‐posterior (AP) and lateral directions. An intensity based rigid registration algorithm was applied to obtain optimal shifts in the x, y and z directions. First, a digitally reconstructed radiograph (DRR) using ray‐casting algorithm was computed from one of the ten phases of the 4DCT. The correlation coefficient (CC) between the DRR and the x‐ray image was then calculated. The 4DCT volume was shifted and new DRRs were generated until the maximum CC value was reached through an optimization process. This procedure was repeated for all ten 4DCT phases. Results: Registration of each phase of the 4DCT with the x‐ray image provided ten maximum CC values. These values exhibited a single maximum at the phase corresponding to the breathing phase when the x‐ray image was taken. The corresponding isocenter positioning shifts for treatment delivery were also obtained. The robustness of our algorithm was demonstrated by registering x‐ray images taken at five random phases to 4DCT. The resulting isocenter shifts were consistent between all phases. The standard deviations of the shifts determined for the AP x‐rays were 2 mm (lateral), 5 mm (anterior‐posterior), and 2 mm (superior‐inferior). Conclusion: Using only the RPM system, one cannot be certain that the internal anatomy is consistent between 4DCT acquisition and any particular treatment day. Our noninvasive method accounts for internal organ motion and may be used for daily 4DCT treatment setup verification and isocenter positioning.

AB - Purpose: To provide a noninvasive method for both setup verification of 4DCT treatment planning and isocenter positioning shifts for treatment delivery. Method and Materials: A ten phase 4DCT treatment planning study of the patient was acquired using a respiratory gating system. When the patient was on the treatment couch, multiple on board x‐ray images were taken in the anterior‐posterior (AP) and lateral directions. An intensity based rigid registration algorithm was applied to obtain optimal shifts in the x, y and z directions. First, a digitally reconstructed radiograph (DRR) using ray‐casting algorithm was computed from one of the ten phases of the 4DCT. The correlation coefficient (CC) between the DRR and the x‐ray image was then calculated. The 4DCT volume was shifted and new DRRs were generated until the maximum CC value was reached through an optimization process. This procedure was repeated for all ten 4DCT phases. Results: Registration of each phase of the 4DCT with the x‐ray image provided ten maximum CC values. These values exhibited a single maximum at the phase corresponding to the breathing phase when the x‐ray image was taken. The corresponding isocenter positioning shifts for treatment delivery were also obtained. The robustness of our algorithm was demonstrated by registering x‐ray images taken at five random phases to 4DCT. The resulting isocenter shifts were consistent between all phases. The standard deviations of the shifts determined for the AP x‐rays were 2 mm (lateral), 5 mm (anterior‐posterior), and 2 mm (superior‐inferior). Conclusion: Using only the RPM system, one cannot be certain that the internal anatomy is consistent between 4DCT acquisition and any particular treatment day. Our noninvasive method accounts for internal organ motion and may be used for daily 4DCT treatment setup verification and isocenter positioning.

UR - http://www.scopus.com/inward/record.url?scp=85024803146&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024803146&partnerID=8YFLogxK

U2 - 10.1118/1.2240892

DO - 10.1118/1.2240892

M3 - Article

VL - 33

SP - 2047

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 6

ER -