The analysis of factors affecting the threshold on repeated 18F-FDG-PET/CT investigations measured by the PERCIST protocol in patients with esophageal carcinoma.

Background: When applying the PET Response Criteria in Solid Tumors protocol, a threshold value based on standardized uptake value corrected to lean body mass (SUL) in liver parenchyma, or in the blood pool, is used: to metabolically specify a measurable lesion; to calculate metabolic tumor volume (mTV) and its product total lesion glycolysis (TLG); and as a limit for response measurement. The problem with using changes in glucose metabolism as a marker for response to therapy is its reproducibility on test-retest examinations. Therefore, before the evaluation of tumor treatment response, we verified our diagnostic protocol for homogeneity using the PET Response Criteria in Solid Tumors quality parameters. In addition, we analyzed the effect of the time span between examinations on the average value of SUL (SUL MEAN) in liver parenchyma at three different points: first at baseline (BL), after the first course of chemotherapy (ChT1), and finally after finishing therapy (ChT3). We also analyzed the influence of SUL MEAN variation on mTV and TLG.
Methods: Eighty-four patients with esophageal cancer were prospectively examined at BL using 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG)-PET/CT; 53 of 84 patients were examined after ChT1, 47 of 84 after ChT3, and 41 of 84 underwent all three examinations. Coefficient of variance (CV) and relative differences (RDw) were assessed for test-retest liver SUL values. The influence of liver SUL MEAN to mTV and TLG was modeled on BL examinations by artificial changes in liver SUL MEAN by ± 20%.
Results: No significant differences were found in test-retest liver SUL MEAN values. Comparing BL with ChT1, BL with ChT3, and ChT1 with ChT3, the CV of the liver SUL MEAN was 10.4, 10.7, and 10.3%; nevertheless, in 34.0, 38.3, and 36.6% of these examinations, respectively, the liver average SUL MEAN values exceeded the limit for inclusion in the study; that is, the difference was less than ± 0.3 U and ± 20%. The corresponding CV of blood background was 14.9, 16.5, and 17.2%. The artificial decrease of -20% in the liver SUL MEAN resulted in an increase of +43.6% in mTV and of +20.4% in TLG, whereas an increase of +20% in the liver SUL MEAN resulted in a decrease of -20.6% in mTV and -11.9% in TLG.
Conclusion: SUL MEAN values in reference tissues (liver parenchyma or descending aorta) measured before chemotherapy did not differ significantly from those measured during chemotherapy. The CV of liver SUL MEAN was comparable to that seen in published data, but some patients had to be excluded from the study because of the individual variability of their mean liver SUL MEAN, which consequently hinders the clinical usage of mTV and TLG. Even in the standardized protocol, all potential sources of variability should be minimized.