The aim of this study was to investigate the role of diffusion weighted magnetic resonance imaging (DWI) as an early imaging biomarker for the assessment of overall survival (OS) in patients with liver-dominant metastatic colorectal cancer (CRC) after radioembolization with 90-Yttrium microspheres (RE).The use of RE is increasingly considered in patients with advanced stages of metastatic CRC, but is limited due to its potential hepatotoxicity. An early response assessment after RE is thus essential to avoid an ineffective and potentially hepatotoxic second lobar treatment. Currently, imaging response assessment to local interventional therapies is mainly based on anatomical changes, e.g., according to Response and Evaluation Criteria in Solid Tumors (RECIST), which suffer from limitations in assessing early therapy-induced tissue transformation¹. Therefore, functional imaging has been proposed as an expedient marker for early response assessment. Alongside with other functional imaging methods like positron emission tomography, the DWI-derived apparent diffusion coefficient (ADC) was shown to highly correlate with tumor size after RE^2,3,4. Furthermore, ADC was capable of predicting response of liver metastases to local and systemic treatment before significant anatomic changes became apparent⁴. The question to what extent ADC could also provide prognostic information after RE has, however, not yet been evaluated.Forty-four patients (28 men and 16 women, mean age 61±11 years, age range 45-82) with unresectable liver metastases from CRC underwent RE and were retrospectively analyzed in this review board-approved study. All patients were included after failure of at least two lines of standard chemotherapy and had neither received other intra-arterial local therapies before RE nor adjuvant chemotherapy within the follow-up period. A routine clinical magnetic resonance examination which included DWI in axial orientation was performed 19±16 days before and 36±10 days after RE on the same 1.5T scanner system (Intera, Philips Healthcare, Best, The Netherlands): TR=1630 ms; TE=63 ms; b values, 0,50,800 s/mm2; acquisition matrix, 256x256; FOV, 380x380 mm2; section thickness, 7 mm; parallel imaging factor (SENSE), 2; spectral inversion recovery fat-saturation (SPIR). Imaging-based treatment response was evaluated according to RECIST and by quantification of the minimal ADC values (Figure 1). Three target lesions with a diameter of at least 1 cm were defined in each patient. Target lesions were manually contoured along the outer border of viable metastatic tissue. Minimal ADC values were determined for each lesion on both baseline and follow-up DWI. These were averaged for each patient for the 3 predefined target lesions. Response to treatment was defined as post-therapeutic increase in minimal ADC. Overall survival (OS) was assessed from the first RE session and death of patients was considered as an event for OS irrespective of the cause. Survival assessment and stratification was performed with the Kaplan-Meier method and log-rank-test comparing various variables: age ≥60 years, sex, hepatic tumor burden ≥50%, extrahepatic metastases, uni- vs. bilobar tumor extent, progress according to RECIST and administered activity ≥2GBq. A p-value <0.05 was considered statistically significant. Significant variables in univariate analysis were further evaluated using the multivariate Cox proportional hazards model to obtain hazard ratio estimates (HR) and 95% confidence intervals (CI). Changes in ADC and tumor size were analyzed using the paired Wilcoxon test.The mean treatment activity administered was 1.5±0.9 GBq. 28 patients had extrahepatic disease, 13 had a hepatic tumor load ≥50% and 28 presented with bilobar involvement. A total of 132 target lesions were analyzed with mean diameters of 5.44±2.30 before vs. 5.5±2.44 cm after RE (p=0.788). According to RECIST, 1 patient was diagnosed with partial remission whereas 5 patients were diagnosed with progressive disease. Overall, ADC was increased by 21.1±38.9% after RE (0.609±0.316 vs. 0.730±0.415 x 10-3mm2/s; p<0.001) with 26 patients being classified as responders. The median OS after RE was 8 months (95% CI 6-10). Patients with changes in ADC ≥0% had a significantly longer OS (median 15 months, 95%CI 9-21) than nonresponders (median 4 months, 95%CI 2-6) (Figure 2). Among the remaining variables, the following were associated with a significantly shorter OS: progressive disease according to RECIST (8 months, 95%CI 5-11 vs. 3 months, 95%CI n.a.; p=0.001), hepatic tumor burden ≥50% (8 months, 95%CI 4-12 vs. 5 months, 95%CI 1-9; p=0.018, and administered activity≥2GBq (10 months, 95%CI 4-15 vs. 5 months, 95%CI 1-8; p=0.033). In multivariate analysis, the absence of response remained the only independent predictor with significant impact on OS (HR=7.56, 95%CI 3.09-18.51; p<0.001).The post-therapeutic increase in the DWI-derived ADC provides prognostic information in CRC patients shortly after RE by predicting an improved OS and might thus guide future treatment decisions in sequential RE approaches.