Purpose: To test the hypothesis that mid-treatment measures of the retention of indocyanine green after 15 minutes (ICGR15) would improve the prediction of posttreatment liver function in the setting of hepatocellular carcinoma.
Methods and materials: Between 2006 and 2015, 144 patients with hepatocellular carcinoma received 175 courses of stereotactic body radiation therapy (SBRT). Patient data, such as age, sex, portal vein thrombosis, cirrhosis, Child-Pugh (CP) score, prior liver-directed therapies, and liver function tests, including albumin-bilirubin (ALBI) and ICG clearance, and dosimetric data, such as tumor volume and radiation dose, were collected. Toxicity was evaluated as a 2-point increase in CP score or a change in ALBI score at 3 months from start of SBRT. Logistic or linear regression was used to build toxicity prediction models based on patient and tumor characteristics and ICG clearance variables. Performance of the models for the binary CP outcome was summarized using area under the curve and receive operating characteristic curves. Likelihood ratio tests were used to evaluate whether the model fit improved after incorporating the ICG variable information.
Results: In multivariable analysis age, baseline ICGR15, and change in ICGR15 were associated with toxicity defined by increased CP score. For the continuous ALBI outcome, being female, having cirrhosis, and increasing radiation dose were associated with increased toxicity. When incorporating ICGR15 into these models, an increase in ICGR15 from baseline to mid-treatment was most consistently significantly associated with an increase in toxicity.
Conclusions: Incorporation of ICGR15 variables significantly improves the prediction of post-SBRT liver function. The use of ICGR15 can facilitate the delivery of the maximum safe dose of radiation for patients with hepatocellular carcinoma and has the potential to improve uncomplicated tumor control and survival.
Copyright © 2017 Elsevier Inc. All rights reserved.