Research activity shows PET/CT postprocessing coming of age

Three-D image postprocessing of data sets acquired from multislice CT and PET/CT scanners has matured to the point that papers have begun to appear comparing 3D performance to more conventional 2D acquisition.

The trend seems to indicate that postprocessing is morphing from a secondary into a primary imaging technique. Lower dose is one major benefit. Because postprocessing requires only a single acquisition, patients can avoid the extra radiation needed to acquire multiple views of interest.

An Italian study that evaluated lesion detectability in 2D PET/CT and 3D PET/CT with GE's Discovery-ST scanner (Q J Nucl Med Mol Imaging 2007 Apr 30; [Epub ahead of print]) found 3D PET/CT exhibited better image quality, data accuracy, and consistency. These improvements contributed to shorter scan times and higher patient throughput. The study assessed a National Electrical Manufacturers Association 2001 image quality phantom with 11 lesions between 7 mm and 37 mm. Ten-minute 2D and 3D PET scans were sequentially acquired in list mode and then unlisted into four-, three-, and two-minute studies to facilitate evaluation of the effect of different acquisition times and count rates. The research also included studies of 10 cancer patients. Each underwent a 3D PET/CT whole-body scan, followed by a 2D and a 3D PET scan, both in four-minute list mode, over a single field-of-view. Both 2D and 3D scans were unlisted into three- and two-minute scans.

Two physicians compared lesion detectability to statistical variations in background activity and to image quality characteristics-smoothness, edge definition, contrast definition of tissue types-to assess data quality in the phantom series and patient scans, respectively.

"Both quantitative and qualitative evaluations showed the superiority of 3D over 2D across all measures of quality," said lead author Dr. Valentino Bettinardi of the San Raffaele Hospital Scientific Institute at the University of Milano-Bicocca in Milan.

Bettinardi's team found lesion detectability to be better in 3D than 2D when scan times were equal. They also found that 3D acquisition produced images comparable in quality to 2D in approximately half the time, he said.

The Bettinardi study tends to confirm findings from an earlier M.D. Anderson Cancer Center PET/CT evaluation (J Nucl Med 2004;45[10]:1734-1742) that evaluated spatial resolution, scatter fraction, sensitivity, counting rate, image quality, and accuracy (as defined by NU94 and NU01 NEMA protocols) for 2D and 3D acquisition modes, also using the Discovery-ST.

The Anderson results showed "a nearly uniform and isotropic resolution across the FOV and a high sensitivity and counting rate performance, making this PET scanner very suitable for clinical whole-body imaging," according to the study's conclusion.

A new Belgian study (Ultrasound Obstet Gynecol 2007;29[5]:537-543) compared diagnostic accuracy of 3D CT to 2D ultrasound for the diagnosis of fetal skeletal anomalies.

When 11 pregnant women were examined with 2D ultrasound, 10 fetuses presented skeletal anomalies; one looked normal on exam but had a family history of osteopetrosis (abnormally dense bones). The women also underwent 3D CT exams.

"We compared retrospectively the diagnoses established on 2D ultrasound and 3D CT with the neonatal and/or postmortem workups, which were used as the gold standard," said Dr. Marie Cassart, assistant senior registrar in the medical imaging department at Erasme University Hospital in Brussels.

The results showed 2D ultrasound provided the correct diagnosis in only two of the 11 cases and was inconclusive in six, whereas 3D CT yielded the correct diagnosis in eight cases.

Cassart said 3D CT was more accurate than 2D ultrasound in visualizing vertebral anomalies such as abnormal shape of the vertebral bodies, abnormal interpedicular distance, pelvic bone malformations including delayed ossification of the pubic bones, abnormal acetabular shape, and an enlarged metaphysis or synostoses in long bones.

In three cases, neither modality provided the correct diagnosis.

"The main advantage of 3D CT in the assessment of fetal skeletal anomalies is that in one acquisition physicians can visualize the whole fetal skeleton," Cassart said "We can examine different bone segments, isolate them, and turn them around, making CT diagnosis more accurate than 2D ultrasound."

A Stanford study (J Nucl Med 2006;47[7]:1081-1087) has taken the next step by devising a new imaging acquisition and processing technique that produces PET/CT images rendered in 3D volume. These images can then be reviewed in such 3D formats as "flythroughs," in virtual bronchoscopy and colonoscopy, and external "flyarounds."

The study demonstrated the feasibility of rendering PET/CT images into 3D volumes. It also found the technique may be useful clinically for diagnostic characterization of lesions and for preprocedural planning, according to lead author Dr. Andrew Quon, of Stanford's molecular imaging program.

Mr. Page is a freelance writer based in Pine Mountain, CA.