Rads tend to agree on PET liver measurements

Two readers had excellent interreader agreement on liver standardized uptake value normalized to lean body mass (SUV_lean) on PET exams, with similar results at three locations in the right lobe of the liver, according to a study published online Jan. 7 in Radiology.

As practice transitions from qualitative PET analysis to quantitative analysis, the need to understand interreader and locational variability arises. For example the liver-to-lesion uptake ratio may help predict benign and malignant mediastinal lymphadenopathy in various cancers. The average value of SUL, or SUV_mean, of the liver may provide a target for response to therapy assessment. However, interreader variability or variability by site could compromise target selection.

Maya Viner, BA, from the department of radiology at Boston University School of Medicine, and colleagues devised a retrospective analysis to evaluate inter-reader agreement and the impact of site volume of interest on the reliability and variability of liver SUV_mean.

“The motivation was to understand errors in measuring liver SUL_mean that can affect standardization and quantitative PET imaging in oncology,” wrote Viner et al.

The researchers assessed 116 patients who underwent a baseline PET/CT for staging of breast, head and neck or thyroid cancer or lymphoma.

Viner and colleagues evaluated intrareader agreement for liver SUL_mean measurements at the upper, lower and portal vein levels of the right lobe of the liver for two individual readers and found “a high degree of intrareader agreement for measurements at different levels.”

The researchers also reviewed interreader agreement for liver SUL_mean at the three locations and reported “high agreement between readers for each one of the three levels.” The greatest precision was in the upper level.

SUL_mean has been proposed as quality control measure for FDG PET/CT and was used in the PET Response Criteria in Solid Tumors (PERCIST) 1.0 study. The current findings carry implications for quantitative imaging in oncology, according to the researchers.

“Consistent placement of the volume of interest within the right lobe of the liver at the upper aspect would reduce the variance of the liver SUL_mean measurements and would allow more consistent SUL_peak cutpoints for target lesions to be used in PERCIST criteria, for quality control purposes and for lesion background FDG uptake quantification,” summed Viner et al.