QUANTITATIVE PRETREATMENT 99 mTc-MAA SPECT / CT AND FDG PET / CT VOI ANALYSIS OF LIVER METASTASES : RELATIONSHIP WITH TREATMENT RESPONSE TO SIRT

PURPOSE: In this study, using quantitive VOI analysis, the percentage Tc-MAA uptake and SUVmax and mean values of liver metastases obtained prior to SIRT by means of Theraspheres were related to treatment response. METHODS: Pretreatment FDG PET/CT scans and Tc-MAA SPECT/CT scans as well as post-treatment FDG PET/CT scans were coregistered and spherical VOI’s drawn around FDG positive liver lesions on the pre-treatment scan using CT for contourdelineation. VOIs’ were than copied on the Tc-MAA SPECT/CT scan and on post-treatment FDG PET/CT scans. VOI SUV mean and max values were obtained and the VOI counts derived from the TcMAA SPECT/CT images divided by the total Tc-MAA counts in the liver lobe or whole liver, dependingon the procedure used, yielding the % of Tc-MAA activity injected that ended up in the lesions. Based on the VOI % of Tc-MAA activity, the estimated YTherasphere activity/cc (MBq/cc) was calculated from the effective dose of Y-Therasphere injected and by dividing with the VOI-volume in cc. Baseline VOI SUVmean and max values and estimated MBq/cc values were related to treatment response using a clinical dichotomous approach as well as a lesion based approach (% change in SUVmean > 50 %). RESULTS: Ninety-one lesions (18 therapeutic sessions in 16 patients) were analyzed; 57 responded and 34 did not (dichotomous approach). VOI volumes and estimated 90YTherasphere activity (MBq/cc) did not differ between responders and non responders; 24 cc (SD: 27) versus 21 cc (sd:21 cc) (p=0.4) and (1.95 MBq/cc (SD: 1.1 MBq/cc) versus 1.90 MB/cc (SD:2.7 MBq/cc) (p=0.92). Also, activity in MBq/cc was not related to lesion size (p=0.32). Contrariwise, SUVmax and mean values proved significantly different between responders and non-responders; 23.7 (SD 9.8) versus 9.4 (SD: 3.8 ) for SUVmax (p =0.0001) and 13.1 (SD:8.1) versus 4.9(SD:1.4) for SUVmean. ROC curve analysis revealed an AUC for SUVmax and mean of respectively 0.87 and 0.81 (p=0.6) for separating responders from non-responders. Using the lesional approach and a cut-off of 50 % or more reduction in SUVmean for a responding lesion, 40 out of 77 lesions responded and 37 did not. SUVmax and mean values also proved significantly different between non-responding and responding lesions; 18.6 (SD 10.8) versus 13.5 (SD: 8.4 ) for SUVmax (p =0.02) and 11.4 (SD:3.8) versus 6.3(SD:4.5) for SUVmean (p=0.002). ROC curve analysis revealed and AUC for SUVmax and SUVmean for separating responding from non-responding lesions of 0.7 and 0.65 (p= 0.6). CONCLUSION: Baseline FDG SUVmax and mean values of non-responders/nonresponding liver metastases to SIRT are significantly higher when compared to responders/responding lesions for a comparable estimated Y-Therasphere activity/cc. Thus, in patients presenting with high baseline SUVmax and mean values, the administration of higher activities or alternatively, other potentially more useful treatment options might be considered.