Does the PERCIST 1.0 minimum glycolytic activity metric allow evaluation of primary breast cancers?

Objectives Untreated breast cancers have been reported to have lower glycolytic activity than many lung cancers. We have recently observed high correlation between automated Computer Aided Detection (CAD) and manually-derived PERCIST 1.0 threshold measurements obtained from the liver. Primary breast cancers were assessed using the PERCIST 1.0 threshold to determine if they were detectable and/or metabolically assessable.

Methods Eighteen baseline FDG PET/CT image datasets of eighteen primary breast cancer patients participating in an ongoing IRB approved multi-center clinical trial were evaluated using CAD software and expert clinician review. A 3cm diameter region of interest placed in the posterior right hepatic lobe was used to calculate the PERCIST 1.0 threshold. The mean of this lesion x 1.5 + 2SD was the threshold for metabolic assessability. SUL-max and SUL-peak tumor metrics were assessed, and results of the automated parameters compared with the manually derived parameters.

Results The most metabolically active lesion was not evaluable in 3 of 18 cases(17%) using the manually-derived SUL max, compared to 2 out of 18 cases (11%) using the automated-derived SUL-max and SUL-peak (the lesions were the same as those not detected manually). On removing the threshold all of the 18 baseline cases were detectable using both manual and automated SUL-max. The automated SUL peak tracked a different node to that identified by the manual SUL-max in 1 of 18 cases (5.6%) - this may be due to CAD being more effective at resolving two adjacent metabolically active nodes as being separate.

Conclusions Despite somewhat lower glucose metabolism than other cancers, the vast majority of breast cancers were metabolically evaluable with current proposed manual and automated derived PERCIST 1.0 threshold. It is not as yet clear which of the SUL metrics (SUL-peak and SUL-max) will be more useful clinically in evaluating the primary tumor. Additional studies using a lower threshold or subset of the peak ROI data set may be appropriate in the future