Liver perfusion as a predictor of post-radioembolization changes in liver and spleen volumes

PurposeTo investigate the relationship between changes in liver perfusion and hepatic and splenic volumes following targeted yttrium-90-radioembolization (RAE) for hepatocellular carcinoma.Materials13 prospectively recruited patients (67±11yrs, 8 men) with hepatocellular carcinoma and no history of liver directed therapy, transjugular intrahepatic shunt placement, or portal vein thrombosis underwent perfusion imaging using contrast-enhanced MR angiography and 3D T1-weighted gradient recalled echo (GRE) imaging (to assess hepatic and splenic volumes) at 1.5T immediately before and after targeted Y90 RAE (2±1 segments treated). Embolized (tumor/non-tumor) and remote (non-embolized) liver parenchymal perfusion were normalized to patients’ arterial input functions, and treatment volumes were defined by enhancement on one month follow up 3D GRE imaging. Pre-/post-treatment volumes were compared with t-tests and related to perfusion via correlation coefficients and receiver operating characteristic curves.ResultsAverage treatment volume, dose, and follow up were 374±234mL, 200±122Gy, and 266±229d. Overall, treated liver segments decreased in size by 24.0±9.7% (p=0.01), splenic volumes increased by 8.5±3.2% (p=0.03), and significant (>10%) untreated liver hypertrophy was only observed in 23% (3/13) of patients. No correlation was found between treatment dose or volume and subsequent volume changes with the exception of treatment volume and splenic hypertrophy (r=0.58). A post-treatment decrease in tumor perfusion >0.074 or increase in adjacent, embolized parenchymal perfusion >0.295 were ≥75% sensitive and specific for predicting treated liver segment(s) atrophy. Greater baseline tumor perfusion or changes in treated liver perfusion were modestly predictive of splenic hypertrophy: AUCs of 0.75 and 0.76.ConclusionsGreater baseline target segment(s) perfusion and larger changes in embolized liver perfusion may serve as early predictors of successful radiation segmentectomy as well as splenic hypertrophy. Our preliminary results suggest that incorporating MR perfusion imaging into the imaging paradigm before and after Y90 RAE of HCC may enable earlier prediction of successful treatment response. PurposeTo investigate the relationship between changes in liver perfusion and hepatic and splenic volumes following targeted yttrium-90-radioembolization (RAE) for hepatocellular carcinoma. To investigate the relationship between changes in liver perfusion and hepatic and splenic volumes following targeted yttrium-90-radioembolization (RAE) for hepatocellular carcinoma. Materials13 prospectively recruited patients (67±11yrs, 8 men) with hepatocellular carcinoma and no history of liver directed therapy, transjugular intrahepatic shunt placement, or portal vein thrombosis underwent perfusion imaging using contrast-enhanced MR angiography and 3D T1-weighted gradient recalled echo (GRE) imaging (to assess hepatic and splenic volumes) at 1.5T immediately before and after targeted Y90 RAE (2±1 segments treated). Embolized (tumor/non-tumor) and remote (non-embolized) liver parenchymal perfusion were normalized to patients’ arterial input functions, and treatment volumes were defined by enhancement on one month follow up 3D GRE imaging. Pre-/post-treatment volumes were compared with t-tests and related to perfusion via correlation coefficients and receiver operating characteristic curves. 13 prospectively recruited patients (67±11yrs, 8 men) with hepatocellular carcinoma and no history of liver directed therapy, transjugular intrahepatic shunt placement, or portal vein thrombosis underwent perfusion imaging using contrast-enhanced MR angiography and 3D T1-weighted gradient recalled echo (GRE) imaging (to assess hepatic and splenic volumes) at 1.5T immediately before and after targeted Y90 RAE (2±1 segments treated). Embolized (tumor/non-tumor) and remote (non-embolized) liver parenchymal perfusion were normalized to patients’ arterial input functions, and treatment volumes were defined by enhancement on one month follow up 3D GRE imaging. Pre-/post-treatment volumes were compared with t-tests and related to perfusion via correlation coefficients and receiver operating characteristic curves. ResultsAverage treatment volume, dose, and follow up were 374±234mL, 200±122Gy, and 266±229d. Overall, treated liver segments decreased in size by 24.0±9.7% (p=0.01), splenic volumes increased by 8.5±3.2% (p=0.03), and significant (>10%) untreated liver hypertrophy was only observed in 23% (3/13) of patients. No correlation was found between treatment dose or volume and subsequent volume changes with the exception of treatment volume and splenic hypertrophy (r=0.58). A post-treatment decrease in tumor perfusion >0.074 or increase in adjacent, embolized parenchymal perfusion >0.295 were ≥75% sensitive and specific for predicting treated liver segment(s) atrophy. Greater baseline tumor perfusion or changes in treated liver perfusion were modestly predictive of splenic hypertrophy: AUCs of 0.75 and 0.76. Average treatment volume, dose, and follow up were 374±234mL, 200±122Gy, and 266±229d. Overall, treated liver segments decreased in size by 24.0±9.7% (p=0.01), splenic volumes increased by 8.5±3.2% (p=0.03), and significant (>10%) untreated liver hypertrophy was only observed in 23% (3/13) of patients. No correlation was found between treatment dose or volume and subsequent volume changes with the exception of treatment volume and splenic hypertrophy (r=0.58). A post-treatment decrease in tumor perfusion >0.074 or increase in adjacent, embolized parenchymal perfusion >0.295 were ≥75% sensitive and specific for predicting treated liver segment(s) atrophy. Greater baseline tumor perfusion or changes in treated liver perfusion were modestly predictive of splenic hypertrophy: AUCs of 0.75 and 0.76. ConclusionsGreater baseline target segment(s) perfusion and larger changes in embolized liver perfusion may serve as early predictors of successful radiation segmentectomy as well as splenic hypertrophy. Our preliminary results suggest that incorporating MR perfusion imaging into the imaging paradigm before and after Y90 RAE of HCC may enable earlier prediction of successful treatment response. Greater baseline target segment(s) perfusion and larger changes in embolized liver perfusion may serve as early predictors of successful radiation segmentectomy as well as splenic hypertrophy. Our preliminary results suggest that incorporating MR perfusion imaging into the imaging paradigm before and after Y90 RAE of HCC may enable earlier prediction of successful treatment response.