News Release

HPET & CT image fusion study reveals limitations for radiotherapy planning

Peer-Reviewed Publication

Society of Nuclear Medicine and Molecular Imaging

Los Angeles, CA…Image fusion with HPET and CT currently has significant limitations for use in radiotherapy treatment planning according to a study being presented June 17 during the Instrumentation and Data Analysis session for nuclear technologists at the Society of Nuclear Medicine Meeting in Los Angeles. Designed to evaluate the co-registration accuracy of PET images transferred from a hybrid-PET gamma camera system (HPET) to a radiotherapy treatment planning system, the study used a phantom (i.e. simulated body) with heart, lung, and tumor inserts, and markers that would show up both on a CT, and because it used 18F, also on PET.

The phantom was scanned by the HPET system and by a CT scanner. The HPET images were transferred by DICOM (digital communication data transfer) to RTPS (a system for determining where and how much radiation therapy should be delivered)

The study found that while the CT and HPET images could be co-registered within 2mm in centrally located tumor inserts, tumor inserts on the peripheral was only 8-10mm, due to distortion and resolution lost of HPET images. In addition, markers and tumor inserts appeared larger on the HPET images than on the CT. The authors attributed this to settings on the RTPS that were compatible for the CT and not for PET. In addition, the authors found that the DICOM transfer truncated the maximum tumor pixel density in the HPET images, complicating edge definition of the tumors. As a result of their findings the authors determined that it would be difficult for a therapist to refine treatment volume using the co-registered HPET/CT images because the tumor edges on HPET images were not consistently defined, and because of image transfer issues between HPET and RTPS.

###

TECHNIQUES AND PITFALLS FOR FUSING PET AND CT IN RADIATION ONCOLOGY was written by R. A. Sajdak*, J. Halama, N. Mirkovic, and A. McCrum, Loyola University Medical Center, Maywood, IL. (200930)

The Society of Nuclear Medicine Technologist Section, (SNMTS) formally established in 1970 is an integral part of the Society of Nuclear Medicine, a diversified 501(c)(3) organization, headquartered in Reston, Virginia. More than 6500 nuclear technologists are current members of SNMTS, which is the recognized world leader in providing knowledge that advances and promotes quality patient care and nuclear medicine technology. Visit SNMTS online at www.snm.org.

TECHNOLOGIST SECTION ABSTRACT
Society of Nuclear Medicine
No. 1563
TECHNIQUES AND PITFALLS FOR FUSING PET AND CT IN RADIATION ONCOLOGY.R. A. Sajdak*, J. Halama, N. Mirkovic, A. McCrum, Loyola University Medical Center, Maywood, IL. (200930) Purpose: To evaluate co-registration accuracy of PET images transferred from a hybrid-PET gamma camera system (HPET) to a radiotherapy treatment planning system (RTPS). Method: Co-registration accuracy was evaluated using an anthropomorphic phantom of the chest with heart, lung, and tumor (20 and 30 ml vials) inserts. The phantom was set into a molded Styrofoam cradle that has fiduciary markers, opaque on CT and filled with 18F, embedded into it. The Styrofoam cradle with embedded fiducial markers is also used for patients undergoing radiation treatment. The phantom was scanned on a CT scanner in Radiation Therapy and by the HPET system. HPET images were transferred by DICOM to a RTPS. Image registration was performed by manually identifying fiducial markers on each modality, and then the software on the RTPS aligned them. Results: The RTPS co-registered the fiducial markers to within 2 mm. Co-registration of centrally located tumor inserts was within 2mm. Co-registration of tumors at the periphery were within 8-10 mm, due to distortion and resolution loss of HPET images at the periphery. Fiducial markers and tumor inserts appear larger on the HPET images than on the CT images, because window and level (W/L) settings on the RTPS are designed for CT and not for PET. The tumors edges on the HPET images were not identifiable and change as the W/L was changed. Also, the maximum tumor pixel density in the HPET images was truncated during the DICOM transfer further complicating edge definition of the tumors. Conclusion: Image fusion of HPET images with CT images on the RTPS planner is accurate to within 2 mm for central lesions, but is less accurate at the periphery. Tumor edges on HPET images were not consistently defined, making it difficult for the therapist to refine the treatment volume using the co-registered HPET images. Further validation of DICOM image transfer between the two systems is also needed.

Session Title: Instrumentation and Data Analysis I
Date: Monday, June 17
Room 408
9:45 AM - 10:00 AM
Rebecca Sajdak
Loyola University Medical Center
2160 S 1st Ave.
Maywood IL 60153

####


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.