Role of Resin Microsphere Y90 Dosimetry in Predicting Objective Tumor Response, Survival and Treatment Related Toxicity in Surgically Unresectable Colorectal Liver Metastasis: A Retrospective Single Institution Study

SIMPLE SUMMARY: Colorectal liver metastases are difficult to treat, with only a minority of patients eligible for surgical resection. Yttrium-90 selective internal radiation therapy is an alternative treatment currently used for patients who have progressed on chemotherapy. A technique called dosimetry allows clinicians to analyze how much radiation was delivered to target lesions post-treatment. The aim of this study is to evaluate the relationship of various dosimetric parameters with objective tumor response, overall survival, and treatment related toxicity with the potential goal of optimizing Yttrium-90 treatment in this patient population. Additionally, other potential predictors of survival outcomes, including clinical and demographic factors, were also evaluated. We found that delivering a mean tumor dose ≥100 Gy when using resin microspheres was significantly associated with objective tumor response and prolonged overall survival. In this study, no mean non-tumoral liver dose threshold was found to predict treatment related toxicity. ABSTRACT: Purpose: To Evaluate the correlation between tumor dosimetric parameters with objective tumor response (OR) and overall survival (OS) in patients with surgically unresectable colorectal liver metastasis (CRLM) undergoing resin-based Ytrrium-90 selective internal radiation therapy (Y90 SIRT). Materials and Methods: 45 consecutive patients with CRLM underwent resin-based Y90 SIRT in one or both hepatic lobes (66 treated lobes total). Dose volume histograms were created with MIM Sureplan(®) v.6.9 using post-treatment SPECT/CT. Dosimetry analyses were based on the cumulative volume of the five largest tumors in each treatment session and non-tumoral liver (NTL) dose. Receiver operating characteristic (ROC) curve was used to evaluate tumor dosimetric factors in predicting OR by Response Evaluation Criteria for Solid Tumors at 3 months post-Y90. Additionally, ROC curve was used to evaluate non-tumoral liver dose as a predictor of grade ≥ 3 liver toxicity and radioembolization induced liver disease (REILD) 3 months post Y90. To minimize for potential confounding demographic and clinical factors, univariate and multivariate analysis of survival with mean tumor dose as one of the factors were also performed. Kaplan-Meier estimation was used for OS analysis from initial Y90 SIRT. Results: 26 out of 45 patients had OR with a median OS of 17.2 months versus 6.8 months for patients without OR (p < 0.001). Mean tumor dose (TD) of the five largest tumors was the strongest predictor of OR with an area under the curve of 0.73 (p < 0.001). Minimum TD, and TD to 30%, 50%, and 70% of tumor volume also predicted OR (p’s < 0.05). Mean TD ≥ 100 Gy predicted a significantly prolonged median OS of 19 vs. 11 months for those receiving TD < 100 Gy (p = 0.016). On univariate analysis, mean TD < 100 Gy, presence of any genomic mutation, presence of MAPK pathway mutation, bilobar hepatic metastases and diffuse metastatic disease (>10 lesions per liver lobe) were found to be predictors of shorter median OS. On multivariate analysis, mean TD < 100 Gy, presence of any genomic mutation, and diffuse hepatic metastatic disease were found to be independent predictors of shorter OS. Overall, six (13.3%) patients developed grade ≥ 3 liver toxicity post Y90 of whom two (4.4%) patients developed REILD. No dose threshold predicting grade ≥ 3 liver toxicity or REILD was identified. Conclusions: Mean TD ≥ 100 Gy in patients with unresectable CRLM undergoing resin-based Y90 SIRT predicts OR and prolonged OS.