4.4.2. mPERCIST and Other PET-Parameters

Semi-automated measurements of PET-parameters in 18F-FDG-PET of the entire tumor burden in all patients were performed by an experienced nuclear medicine physician using image fusion software (Hybrid Viewer 2.6, Hermes Medical Solutions, Stockholm, Sweden) [29]. First the background was measured in the healthy liver to ensure the

technical comparability of PET studies and to determine if the target tumor lesion shows sufficiently high glucose uptake with a minimum threshold for measurability defined as 1.5 × (mean value of normal liver) + 2 × (standard deviation of liver) or greater at baseline. For measurements of the background activity a fixed 3-cm diameter spherical volume of interest (VOI) was placed in the right side of the liver. Furthermore, a new lesion or unequivocal progression in the follow-up PET scan was rated as progressive disease changes, even if the target lesion at baseline did not show the minimum glucose uptake.

To assess treatment response by mPERCIST change in peak standardized uptake value (SUVpeak) was measured in the tumor region with the highest radiotracer uptake, which describes the average SUV computed in a fixed 1-mL sphere recommended by PERCIST instead of the widely used single-pixel maximum standardized uptake value (SUVmax) to avoid noise errors [18,29]. In a slight modification to the PERCIST 1.0 criteria, the quantitative PET-parameter was adjusted to body weight (SUVpeak in g/mL) rather than the body surface area (SULpeak), as previously described by Fendler et al. for the single most active lesion in the patient at baseline, 3 month and 6 month follow-up [31]. A 1 mL spherical VOI was placed at the focus of the single active lesion and the highest SUVpeak value was computed automatically. The single most active lesion presents the target lesion, from which it is assumed to correspond to the worst behaving portion of the tumor. The percentage changes in SUVpeak from the reference baseline scan to 3 month and 6 month follow-up scans were assessed in all patients. In mPERCIST, decrease greater than or equal to 30% in SUVpeak was considered as PR and increase greater than 30% as PD (see Table 3).

**Table 3.** mPERCIST criteria from Wahl [29], modified by Fendler et al. [31].


Outcome determination is measured on the single most active lesion on each scan (not necessarily the same lesion). Standardized uptake value (SUV).

All further PET-parameters were derived through segmentation of all tumor lesions: mean/maximum standardized uptake value (SUVmean/max), Metabolic Tumor Volume (MTV) and Total Lesion Glycolysis (TLG). MTV was defined by volume delineation and TLG was calculated as (MTV x SUVmean) [23]. All areas with physiological, non-tumoral 18F-FDG-uptake were excluded. Response was determined separately in all PET-parameters using the same criteria for percentage change except for TLG, which had to show an increase of greater than 75% according to Wahl et al. to be assessed as PD (see Table 3) [29].
