Radiation Dose-Volume Effects for Liver SBRT

Abstract

Stereotactic body radiation therapy (SBRT) has emerged as an effective, noninvasive treatment option for primary liver cancer and metastatic disease occurring in the liver. Although SBRT can be highly effective for establishing local control in hepatic malignancies, a tradeoff exists between tumor control and normal tissue complications. The objective of the present study was to review the normal tissue dose-volume effects for SBRT-induced liver and gastrointestinal toxicities and derive normal tissue complication probability models.

Fig. 1

Grade ≥3 liver enzyme toxicity as a function of mean liver dose (MLD), with probit model fit and 95% and 68% confidence intervals (CIs). Dose “error bars” (horizontal axis) for each data point represent range of reported doses for each study; toxicity “error bars” (vertical axis) represent binomial 68% CIs. The number of patients who developed toxicity of the total number of patients for each study is displayed next to the data point. All studies excluded the gross tumor volume in the MLD calculation, with the exception of Son et al (20). Probit model fitting failed to establish the upper confidence limit for the dose to 50% of the target, and the null hypothesis of no dose response was not rejected (P = .10); therefore, we could not exclude that the incidence of liver enzyme complications was independent of the dose; the probit model fit is displayed for reference. Abbreviation: LML = log maximum likelihood.

Fig. 2

Grade ≥2 general gastrointestinal (GI) toxicity as a function of the prescription (Rx) physical dose to the target, with the probit model result (maximum likelihood parameter fitting) and 95% and 68% confidence intervals (CIs). General GI toxicities were defined as fatigue, nausea, diarrhea, gastritis, ulcers, GI area pain, and colitis. The target Rx dose definition is provided in Table 2. Each data point was placed at the reported mean or median dose and reported complication rate; horizontal error bars represent the reported ranges, and the vertical error bars represent binomial 68% CIs. The number of patients who developed toxicity of the total number of patients for each study is displayed next to the data point. The study by Andolino et al (14) did not distinguish between grade 1 and 2 general GI toxicities. Abbreviation: LML = log maximum likelihood.

Fig. 3

Grade ≥3 general gastrointestinal (GI) toxicity as a function of the prescription (Rx) physical dose to the target, with the probit model result (maximum likelihood parameter fitting) and 95% and 68% confidence intervals (CIs). General GI toxicities were defined as fatigue, nausea, diarrhea, gastritis, ulcers, GI area pain, and colitis. The target RX dose definition is provided in Table 2. Each data point is placed at the reported mean or median dose and reported complication rate; the horizontal error bars represent the reported ranges, and the vertical error bars represent binomial 68% CIs. The number of patients who developed toxicity out of the total number of patients for each study is displayed next to the data point. Abbreviation: LML = log maximum likelihood.