Abstract
494
Aim: We aimed to retrospectively examine the applicability of the PET Response Criteria in Solid Tumors (PERCIST) when using FDG-PET/CT for response monitoring in patients with metastatic breast cancer (MBC) and to compare findings with visual assessments reported in the clinic. Methods: Women with biopsy-verified MBC undergoing treatment response monitoring by FDG-PET/CT at our institution from September to December 2017 were eligible. Scans, descriptions, and oncological medical records were analyzed retrospectively. All scans were screened for comparability with regard to a modified standardization set of the PERCIST 1.0 criteria. Scans were non-comparable if scan matrices at follow-up differed from baseline matrices, and if liver SULmean reference volume differed more than 20%. The SULpeak of the hottest metastatic lesion was registered, and analyses were made according to PERCIST 1.0. Response was categorized as complete metabolic response (CMR), partial metabolic response (PMR), stable metabolic disease (SMD), and progressive metabolic disease (PMD). PERCIST response categories were compared to those proposed by visual assessment in the clinic. Results: A total of 315 scans from 43 patients were eligible; exclusion of non-comparable scans reduced this number to 187 (59.4%) scans in 37 patients. Metastatic lesions were measurable according to PERCIST 1.0 in 175 of 187 (96.3%) scans. One patient contributed with 10 of 12 scans with non-measurable lesions; her histological type was a lobular carcinoma. The SULpeak fluctuation and percentage of change in SULpeak could be visualized for patients with measurable lesions, which gave the advantage of visualization of a nadir, the lowest SULpeak, as illustrated for one patient in Figure 1. Response categorization could be performed in 145 follow-up scans; respective categories by PERCIST 1.0 and by visual assessment are presented in Table 1. Large discrepancies were noticed when PERCIST 1.0 suggested SMD or PMD, while visual assessment suggested PMR. The proportion agreement between PERCIST 1.0 and visual assessment was moderate: 0.64 (95% CI=0.56-0.71). Conclusions: A large proportion of scans were excluded from analysis due to lack of comparability with modified PERCIST 1.0 criteria. However, PERCIST 1.0 was applicable in almost all included scans of patients with MBC and provided a semi-quantitative response categorization with a useful variable, SULpeak, allowing monitoring of disease fluctuation. The moderate agreement between PERCIST 1.0 and visual assessment was mainly due to cases where visual assessment suggested response, while PERCIST 1.0 suggested non-response. Thus, there is a need for major prospective studies comparing promising novel quantitative PET/CT variables with semi-quantitative PERCIST 1.0 measurements to identify the optimal way of monitoring treatment response in patients with MBC.