Ablation, chemoembolization, and RT have been used for the treatment of unresectable locally advanced HCC. Because RT alone cannot achieve a complete response in most cases, other locoregional treatments, such as TACE, RFA, and PEI, have been used repeatedly prior to or after RT. After completion of RT, transient elevation of hepatic enzymes such as AST, ALT, and ALP commonly occurs, but, in many cases, these kinds of hepatic toxicities are recovered in a few months
. However, if hepatic function deteriorates because of radiation, other necessary treatments cannot be performed in a timely manner. Since no specific treatment for this condition exists except conservative care, it is also important to reduce the development of the deterioration of hepatic function. Thus, it is important to identify a parameter that can predict the deterioration of hepatic function and to develop a plan for RT using the parameter and its values.
The relationship between the radiation dose to liver volume and the incidence of hepatic toxicities has been studied previously
[12–15, 17]. Radiation-induced liver disease (RILD) is a traditionally accepted concept of hepatic toxicity
. In the past, classic RILD was a serious manifestation of hepatic toxicity caused by irradiation of 30–35 Gy to the entire liver. However, the incidence of classic RILD has been lowered, since partial volume irradiation has become more common. Other authors have reported the parameters predicting the non-classic RILD or the elevation of hepatic enzymes ≥ grade 2 or 3 according to Radiation Therapy Oncology Group toxicity criteria or Common Terminology Criteria for Adverse Events (CTCAE) and their cut-off values to present a guideline for radiation planning
[12, 14, 15, 17].
Our study differs from previous studies in 3 aspects. First, the presence or absence of an increase of at least 2 points in the CP score as an end-point of RIHT was used for analysis. A few authors have analyzed the elevation of hepatic enzymes (AST, ALT and ALP) as end points to find significant parameters that predict hepatic toxicities
[12–15, 17]. Kim et al. showed that the elevation of hepatic enzymes was transient and recovered within a median of 2 months after the completion of RT
. However, Furuse et al. showed that hypoalbuminemia, hyperbilirubinemia and ascites were important adverse hepatic events that occur after the application of RT to treat advanced HCC, and these events seriously affected survival
. An albumin, bilirubin and ascites were used to calculate the CP score. In our previous study, the progression of CP class was analyzed as a useful radiation dose-limiting factor predicting the deterioration of hepatic function, whereas the elevation of hepatic enzymes according to the CTCAE scale was inappropriate as a useful end point
. Liaw et al. also used an increase of at least 2 points in the CP score to evaluate the deterioration of hepatic function in patients who were treated with lamivudine
. Therefore, the CP score is appropriate for the assessment of hepatic function, and an increase of at least 2 points in the CP score should be considered the dose-limiting factor.
Another difference between our study and previous studies is that only patients treated with helical tomotherapy were included in the present study. Previous studies regarding the dosimetric parameters predicting hepatic toxicity were based on the data from 3D-CRT. The planning and delivery method of helical tomotherapy is different from that of 3D-CRT. Because helical tomotherapy is delivered continuously from all angles around the patient via a ring gantry, in which the linear accelerator is mounted, a low to moderate radiation dose is delivered to a much wider region of liver. Because of this characteristic, the parameter and its cut-off value could be different between 3D-CRT and helical tomotherapy. According to Kim et al., V30 was demonstrated as a significant parameter in patients treated with conventional fractionated RT
, and according to Liang et al., V20 was a significant parameter in patients treated with hypofractionated RT (4–6 Gy per fraction)
. In our study, a significant parameter is V15Gy, which is a parameter that corresponds to lower doses than those of above studies. The cut-off value of 43.2% for V15Gy in our study is lower than that in the above studies (60% for V30 in the study of Kim et al. and 48.5% for V20 in the study of Ling et al.). This result is probably because of the characteristic planning and delivery method of helical tomotherapy and indicates that, to reduce the risk of the deterioration of hepatic function, a wider region of normal liver should be preserved from a low to moderate dose of radiation.
The third differences between our study and previous studies is that V15Gy was confirmed the only significant dosimetric parameter in multivariate analysis. However, because of the significant correlations between dosimetric parameters shown in univariate analysis, consideration of the values of these parameters could be helpful in the treatment planning phase. Based on the estimated probability curves of V5Gy, V10Gy, V15Gy, V20Gy, V25Gy, V30Gy, and V35Gy, which were statistically significant in univariate logistic and ROC curve analyses, we presented the normal liver DVH indicating a 10%, 20%, and 30% risk of the deterioration of hepatic function (Figure
3). This curve could be used as a reference tolerance curve to evaluate treatment plans.
In conclusion, an increase of at least 2 points in the CP score is a radiation dose-limiting factor, and the non-target normal liver receiving a dose more than 15 Gy (V15Gy) should be <43.2% to reduce the risk of the deterioration of hepatic function. Moreover, the proposed normal liver DVH could be useful as a reference curve to evaluate the dose to the liver.