LAUSR

Purpose To investigate possible correlation between absorbed dose (AD) and tumour response by means of Tumour Control Probability (TCP) in liver metastases treated with radioembolization (RE) with resin 90 Y-microspheres. To assess response, the analysis of parameters from FDG-PET/CT has been preferred to RECIST criteria as metabolic response has proven to anticipate morphologic response. Methods Patients with chemo-refractory liver metastases from solid tumours scheduled to receive RE underwent FDG-PET/CT scan before and 6 weeks after RE. 99m Tc-MAA (75–111 MBq) were injected 2 weeks before RE to perform dosimetry based on SPECT fused to contrast CT. Response assessment was based on PERCIST criteria. Variation (%) of PET parameters versus basal examination were evaluated to establish Complete-Response (CR), Partial-Response (PR), Stable-Disease (SD), Progressive-Disease (PD). Results 22 patients with hepatic lesions from colon-rectal (11), breast (7), ovary (1), endometrial (1), parotid (1) cancer, cholangiocarcinoma (1) were suitable for analysis. All patients received a single RE treatment (median activity: 1.7 GBq; range: 0.6–2.9 GBq). Median (range) tumour AD: 100(30–443) Gy. Metabolic response rate of lesions based on PERCIST were: CR = 31%; PR = 28%; SD = 24%; PD = 17%. Two different TCP curves were obtained by probit regression when considering: i) endpoint-1: PR or CR; ii) endpoint-2: CR only (p   = 220 Gy only CR were observed. TCP of 50%, 75%, 100% were obtained at: i) 108, 120, 165 Gy; ii) 135, 160, 225 Gy. Conclusions Despite the variety of primary tumours, relatively low cohort of patients, and uncertainty of dosimetry with 99mTc-MAA, our preliminary data provided evidence of correlation between response based on PET/CT and AD. These encouraging results need to be confirmed with more ample dataset and, possibly, tumour type differentiation. Other PET/CT parameters such as the metabolic tumour volume and tumour lesion glycolysis are being considered for comparison with PERCIST and possible improved correlations.