PT - JOURNAL ARTICLE AU - Theresa Devasia AU - Matthew Cousins AU - Justin Mikell AU - Ravi Kaza AU - Matthew Schipper AU - Kyle Cuneo AU - Yuni Dewaraja TI - Progression and toxicity following liver Y90 radioembolization: impact of dose metrics, clinical factors, and biomarkers DP - 2021 May 01 TA - Journal of Nuclear Medicine PG - 21--21 VI - 62 IP - supplement 1 4099 - http://jnm.snmjournals.org/content/62/supplement_1/21.short 4100 - http://jnm.snmjournals.org/content/62/supplement_1/21.full SO - J Nucl Med2021 May 01; 62 AB - 21Objectives: We characterized response and toxicity outcomes in primary and metastatic liver cancer patients treated with Y90 glass microsphere radioembolization (RE) with the goal of developing models for future treatment planning. Methods: Patients with primary liver cancer or liver metastases who underwent RE were imaged with Y90 PET/CT (77 treatments, 59 patients, 200 lesions) for voxel-level dosimetry. A subset (45 patients) had baseline blood cytokine measurements (CD40L, HGF, and TNFR1; selected based on prior findings in liver cancer patients treated with external beam radiation [EBRT]). Evaluated baseline clinical factors included age, primary vs. metastatic, prior liver therapies, tumor volume, cirrhosis, and liver function measures. Data from all patients were used for overall survival (OS). For progression, tumors < 2 cm3 posed challenges for dosimetry and were excluded. Progression was assessed using RECIST criteria at baseline and regular follow-up. Time to progression (TTP) was measured at the tumor level, with censoring at date of last follow-up or additional treatment post-Y90. The log-rank test was used to test separation between Kaplan-Meier survival curves, and the c-index was used to assess predictive accuracy of Cox models. Liver toxicity was measured using the 6-month change in ALBI from baseline, where ALBI is a quantitative measure of liver function that serves as an alternative to the Child-Pugh score. Linear models using non-tumor liver dose, log-transformed baseline cytokine values, and clinical factors were created. Results: The median follow-up time was 10.9 months [range: 1.0-51.3]. Median OS was 13.9 months [95% CI: 10.9, 28.0]. Minimum or average tumor absorbed dose [median: 219 Gy, range: 1-2203 Gy] was not a significant predictor of survival. Higher age and baseline ALBI score were associated with increased risk of death. Twelve tumors had progression, and median TTP was not achieved. The probability of progression was significantly higher in tumors with mean absorbed dose ≤ 219 Gy (p = 0.031): 3/52 (5.8%) receiving > 219 Gy progressed, whereas 9/51 (17.6%) receiving ≤ 219 Gy progressed. Log-transformed mean tumor dose was a significant predictor of progression (p < 0.001; c-index > 0.80), but clinical factors were not significant for progression. Higher baseline serum HGF (p = 0.002) and baseline plasma TNFR1 (p = 0.029) were significantly associated with greater liver toxicity in univariate and multivariable models adjusted for mean dose to non-tumoral liver, prior liver therapies, and cirrhosis. Mean dose to normal liver [median: 49 Gy, range: 4-169 Gy] was not significantly associated with toxicity. Conclusions: Higher tumor mean absorbed doses were associated with lower probability of progression. Baseline biomarkers HGF and TNFR1, alone and in combination with normal liver dose, showed strong positive associations with liver toxicity, as noted in prior studies in EBRT. Use of these cytokines in combination with personalized dosimetry could potentially maximize Y90 RE treatment benefit while minimizing associated toxicity. Research Support: NIH R01EB022075 awarded by the National Institute of Biomedical Imaging and Bioengineering.