PACS showing improvement for PET/CT SUVs

Jun 12, 2006

Bringing PET/CT on board PACS has been problematic, particularly in generating standardized uptake value (SUV) measurements. Generally, most SUV determination has been conducted at dedicated PET or PET/CT workstations. However, radiology information technology (IT) vendors have recently been offering SUV analysis modules that can be added to PACS.

Researchers from Memorial Sloan-Kettering Cancer Center in New York City put one of these PACS modules to the test by comparing it with SUV measurements from two vendors' dedicated PACS/CT workstations. The scientists, from the nuclear medicine service at the institution's department of radiology, presented their results in a poster presentation at the Academy of Molecular Imaging (AMI) scientific conference in Orlando, FL, earlier this year.

The study cohort was comprised of 32 consecutive studies using FDG-PET/CT on 16 men and 16 women (mean age 57.7 years) conducted during the first week of July 2005. In addition, the researchers performed two additional studies on two other patients with the radiotracers C-11 and I-124 to test the propagation of a look-up table in regard to the half-life of the radioisotopes.

Half of the FDG studies were done on a Discovery LS PET/CT system (GE Healthcare, Chalfont St. Giles, U.K.), while the other half were carried out on a Biograph PET/CT system (Siemens Medical Solutions, Malvern, PA). For the Biograph system, the team employed a protocol of a scout view with 130 kVp at 30 mAs, followed by a CT scan at 50 mAs, 130 kVp, 4-mm collimation and 5-mm slice thickness.

For the Discovery device, the group obtained a scout view with 120 kVp at 30 mAs, then a CT scan at 0.8 seconds/rotation at 80 mAs, 140 kVp, and slice thickness of 5 mm. PET emission images were acquired at four minutes per bed position utilizing 3D mode on the Biograph and 2D mode on the Discovery.

"The total acquisition time varied between 25 and 35 minutes per patient," the authors wrote.

The PET/CT images were reviewed by a radiologist at the manufacturers' dedicated workstations for the modality, GE Healthcare's Xeleris and Siemens' Syngo. The images from all the FDG studies were also reviewed on the facility's GE Healthcare Centricity PACS with the Volume Viewer Plus module.

The team had the radiologist draw four similar-size regions of interest (ROI) on transaxial images of the liver, lung, bladder, and the malignant lesion with the highest FDG uptake on all workstations. A z-dimension localization function was common to all the applications, which allowed the ROIs to be placed on the same transaxial slice the researchers reported. The maximum SUV was recorded for the bladder and the most FDG-avid lesion, and mean SUVs were calculated for the liver and lung.

The scientists then analyzed the data from the two dedicated PET/CT workstations and the PACS workstation.

"As far as the maximum SUVs were concerned, Pearson correlation coefficients between the workstations (dedicated and PACS-integrated) were 0.99 for the lesion and 0.96 for the bladder when both systems (GE Healthcare and Siemens) were analyzed together," the authors wrote. "For the average SUVs, when both systems were studied together, Pearson coefficients between the workstations were 0.71 (lung) and 0.81 (liver)."

They also noted that the analysis of the C-11 and I-124 data taken from the Discovery system revealed no significant differences for the measurements of maximum SUVs on the dedicated PET/CT workstation or the PACS-integrated workstation.

The researchers observed that the correlation of the mean SUV was not as good as the correlation of the maximum SUV on the dedicated and PACS-integrated workstations. This discrepancy was attributed to the variability of drawing and placing the ROIs by the radiologists. However, they were satisfied that the PET/CT application on the PACS workstation was correctly interrogating the DICOM header, applying decay corrections, and correctly accounting for variations in patient body weight and injected dose.

The benefits of adopting a PET/CT analysis tool within a PACS, according to the authors, are that it allows a reduction in the number of computers for the practice; it allows direct comparison with prior exams irrespective of the modality; and it centralizes archive procedures, which is beneficial in a multivendor environment. Although the number of dedicated PET/CT workstations can be probably be diminished in a multidevice center, they cannot be eliminated.

"A dedicated workstation has additional functionality to perform region of interest analysis on dynamically acquired PET/CT image sets, and therefore at least one dedicated workstation is probably needed for this purpose," the authors wrote.

By Jonathan S. Batchelor
AuntMinnie.com staff writer
June 13, 2006

Multivendor IT integration improves workflow, patient care, April 28, 2006

Diagnostic imaging and clinical information systems: An integration primer, April 14, 2005

Contracting IHE key for interoperable implementation, March 11, 2005

Patience required for adding PET/CT to PACS, June 25, 2003