Hepatic imaging response to radioembolization with yttrium-90-labeled resin microspheres for tumor progression during systemic chemotherapy in patients with colorectal liver metastases.

Background: To assess response and the impact of imaging artifacts following radioembolization with yttrium-90-labeled resin microspheres ((90)Y-RE) based on the findings from a central independent review of patients with liver-dominant metastatic colorectal cancer (mCRC).
Methods: Patients with mCRC who received (90)Y-RE (SIR-Spheres(®); Sirtex Medical, Sydney, Australia) at nine US institutions between July 2002 and December 2011 were included in the analysis. Tumor response was assessed at baseline and 3 months using either the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.0 or 1.1. For each lesion, known artifacts affecting the interpretation of response (peri-tumoral edema and necrosis) were documented. Survivals (Kaplan-Meier analyses) were compared in responders [partial response (PR)] and non-responders [stable (SD) or progressive disease (PD)].
Results: Overall, 195 patients (mean age 62 years) received (90)Y-RE after a median of 2 (range, 1-6) lines of prior chemotherapy. Using RECIST 1.0 and RECIST 1.1, 7.6% and 6.9% of patients were partial responders, 47.3% and 48.1% had SD, and 55.0% and 55.0% PD, respectively. RECIST 1.0 and RECIST 1.1 showed excellent agreement {Kappa =0.915 [95% confidence interval (CI): 0.856-0.975]}. Peri-tumoral edema was documented in 32.8%, necrosis in 48.1% and both in 57.3% of cases (using RECIST 1.0). Although baseline characteristics were similar in responders and non-responders (P>0.05), responders survived significantly longer in an analysis according to RECIST 1.0: PR median (95% CI) 25.2 (range, 9.2-49.4) months vs. SD 15.8 (range, 9.3-21.1) months vs. PD 7.1 (range, 6.0-9.5) months (P<0.0001).
Conclusions: RECIST 1.0 and RECIST 1.1 imaging responses provide equivalent interpretations in the assessment of hepatic tumors following (90)Y-RE. Radiologic lesion responses at 3 months must be interpreted with caution due to the significant proportion of patients with peri-tumoral edema and necrosis, which may lead to an under-estimation of PR/SD. Nevertheless, 3-month radiologic responses were predictive of prolonged survival.