This study was performed with patients who underwent RT for unresectable advanced HCC without distant metastases. Between July 2003 and June 2006, 80 patients with unresectable HCC underwent local RT to the liver at the Korea University Medical Center. Fifteen of these patients were excluded due to the presence of distant metastases prior to RT.

Diagnosis of HCC was based on either the identification of hypervascular masses by two imaging studies or by one imaging study combined with a serum alpha-fetoprotein (AFP) level > 400 ng/mL. If the vascular profile by dynamic imaging was not characteristic of HCC and the AFP was less than 400 ng/mL a biopsy was performed[21]. Unresectability was determined using accepted surgical criteria[3].

The baseline characteristics of the 65 patients are presented in Table 1. Fifty-six patients (86.2%) were male and 9 were female. The median age was 60 years (range, 42-83) years. Underlying liver diseases included chronic Hepatitis B virus (HBV) infection in 49 patients (75.4%), alcoholic liver cirrhosis in 13 patients (20%) and chronic Hepatitis C virus (HCV) infection in two patients (3.1%). In one patient (1.5%), co-infection with HBV and HCV was noted. Liver cirrhosis was present in 50 patients (76.9%). According to the Child-Pugh classification, 43 patients (66.2%) were classified as grade A and 22 patients (33.8%) were classified as grade B. Patients in class C were not included. Baseline tumor size was 10.8 ± 4.7 cm (median, 9.9 cm). In 31 patients (47.7%), the tumor size was larger than 10 cm. Based on the types of HCC described by Eggel[22], the most frequent tumor type was massive (58.5%), followed by multinodular (36.9%) and single nodular (4.6%). Prior to RT, portal vein thrombosis was observed in 45 patients (69.2%); this was confirmed by CT and/or angiogram. Among these 45 patients, thrombosis was observed in the main portal vein in 20 patients (30.8%), at the first branch level in 23 (35.4%), and at the second branch level in 2 (3.1%). The hepatic vein and bile duct were involved in 8 and 6 patients, respectively. No patients showed evidence of extrahepatic metastasis prior to RT. According to the TNM staging system of the Liver Cancer Study Group of Japan[23], 53 patients (81.5%) fell into stage IV-A, and 12 (18.5%) fell into stage III.

RT was performed as a primary treatment in 40 of the 65 patients (61.5%) due to an overly large tumor size in 20 patients (30.8%), portal vein thrombosis in 12 patients (18.5%), IVC thrombosis in 3 patients (4.6%), bile duct invasion in 3 patients (4.6%), and a massive porto-systemic shunt around the tumor in 2 patients (3.1%). In the remaining 25 patients (38.5%), RT was performed as a salvage treatment after ineffective TACE (21 patients, 32.3%) or vascular inaccessibility to the feeding vessel of the HCC (4 patients, 6.2%). External beam RT at a target dose of 61 Gy/34 fractions was planned, using 10 MV of X-rays. The RT strategy was devised using a CT-based 2-D planning system (CT Port, Toshiba, Tokyo, Japan). To account for respiratory-based liver motion, a 1-1.5 cm margin was added in the craniocaudal direction. The full 61-Gy irradiation dose was feasible in 55 of the 65 patients (84.6%).

During and after RT, TACE was also employed in 57 patients (87.7%; 2.9 ± 1.8 sessions; median, three sessions; range, 1-8 sessions). TACE was performed with an emulsion of doxorubicin at a dose of 10-30 mg and 4-12 mL of mixed solution of lipiodol and contrast agent. TACE was usually combined with embolization using gelfoam particles, except in cases with significant portal vein thrombosis. In 16 patients with portal vein invasion (24.6%), HAI with cisplatin and 5-FU was combined with or without TACE (3.3 ± 2.4 cycles; median, 2.5 cycles; range, 1-8 cycles).

Tumor size was measured by computed tomography (CT) and was calculated as the longest diameter multiplied by the longest perpendicular diameter. CT scans were obtained before RT, 1 and 4 mo after the initiation of RT, and then every 2-3 mo. If a patient had multiple nodules, the extent of the tumor was determined by the sum of the extent of all tumors > 2 cm in diameter.

Treatment response was assessed at four months after initiation of RT. A complete response was defined as the complete disappearance of all clinical and radiographic tumor evidence. A partial response was defined as more than a 50% decrease in tumor size from baseline. Stable disease was defined as less than a 50% decrease or a 25% increase in tumor size. The objective treatment response was calculated based on the complete and partial responses. Progressive disease was defined as a greater than 25% increase in extent of the tumor from the nadir extent of the tumor.

To evaluate the efficacy of using the early tumor response to predict the treatment response, the tumor regression rate at one month after initiation of RT (TRR_1m) was assessed using the following equation: TRR_1m = [(baseline tumor extent - tumor extent at one month after RT)/baseline tumor extent] × 100.

Adverse events were evaluated weekly during RT and one month following the treatment. Adverse hematologic events were evaluated by measuring hemoglobin, white blood cell (WBC) and platelet counts, while hepatic adverse events were evaluated by measuring serum bilirubin, albumin, aspartate aminotransferase (AST), alanine aminotransferase (ALT) and alkaline phosphatase (ALP) levels. Gastrointestinal (GI) bleeding included any bleeding from the esophagus, stomach, duodenum, or liver. All adverse events were graded according to Common Terminology Criteria for Adverse Events V3.0.

All calculations were performed using SPSS 10.0 software for Windows (SPSS, Chicago, IL). Quantitative variables were expressed as mean ± SD or medians. Differences in quantitative and qualitative variables were assessed using the Student’s t-test and chi-square test, respectively. Logistic regression analysis was performed to evaluate predictive factors for tumor response. Survival and progression-free survival were assessed from the initiation of RT according to the Kaplan-Meier method. Differences between variables were assessed using the log-rank test. The Cox regression model was used to detect associations between survival and AFP status, tumor type, the location of tumor thrombi, stage and therapeutic models. For multivariate analysis, variables with P < 0.2 at univariate analysis were entered. Differences with P < 0.05 were considered to be statistically significant.

Fifty-five of 65 patients (84.6%) completed the RT schedule. Ten patients (15.4%) could not complete RT due to HCC aggravation or deterioration of liver function after RT. Interruption of RT was more frequent in Child-Pugh class B patients (7 of 20 patients, 35%) than in class A patients (3 of 45, 6.7%; P = 0.003). Among the 55 patients who completed RT, treatment response was evaluated 4 mo after the initiation of RT. None of our patients had completely responded at this point in the response evaluation, but 37 patients (67.3%) had partially responded. Seventeen patients (30.9%) had stable disease, and 1 (1.8%) had progressive disease. Therefore, the objective treatment response at four months was 67.3%. However, if we label the 10 patients who did not complete RT as non-responders, then the partial response and stable disease rates decreased to 56.9% (37 of 65 patients) and 26.2% (17 of 65 patients), respectively.

Table 2 presents baseline characteristics according to treatment response. Logistic regression analysis was performed to evaluate predictive factors for an objective treatment response. Child-Pugh grade was the only independent predictive factor for an objective treatment response (Child-Pugh grade A vs B; OR, 5.167; 95% CI, 1.643-16.250; P = 0.005). Among the 45 patients with Child-Pugh grade A, 31 patients (68.9%) showed a partial response, as did 6 of the 20 grade B patients (30%, P = 0.003).

During follow-up, 4 of the 37 patients showing a partial response (10.8%) and 3 of the 25 patients showing stable disease (12%) had progressive disease after a median of 6 (range, 3-9) mo. Among the 65 patients, time to progressive disease was 5 ± 3 mo after initiation of RT (median, 4 mo). Duration without progressive disease was longer in patients who met the objective treatment response (14.8 ± 1.4 mo) than in patients who did not (4.6 ± 0.4 mo, P < 0.001; Figure 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Duration without disease progression, according to treatment response. TR: Treatment response; RT: Radiotherapy.

All enrolled patients were followed for 8 ± 6 (median, 6; range, 1-30) mo. During this period, 37 patients died. Fourteen patients died of hepatic failure, 13 died of HCC aggravation, 5 died of gastrointestinal bleeding, 2 died of tumor rupture, and 2 died of sepsis. The cumulative survival rates at 6, 12 and 18 mo were 61.5%, 34.7% and 27.0%, respectively. Patients who showed an objective treatment response (median survival, 346 d) survived longer than those who did not (median survival, 212 d; P = 0.032; Figure 2).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Overall survival of the 65 patients who underwent radiotherapy for advanced hepatocellular carcinoma according to treatment response. Patients with objective treatment responses (median survival, 346 d) survived longer than those without objective treatment responses (median survival, 212 d; P = 0.032). RT: Radiotherapy; TR: Treatment response.

When multivariate Cox-regression analysis was performed with baseline characteristics, large tumor size (≥ 10 cm vs < 10 cm; OR, 2.416; 95% CI, 1.213-4.811; P = 0.012), Child-Pugh grade B vs A (OR, 4.094; 95% CI, 1.977-8.480; P < 0.001) and the presence of tumor thrombi in the main portal vein (OR, 2.315; 95% CI, 1.156-4.634; P = 0.018) were independent predictive factors for mortality. However, when multivariate analysis was performed after inclusion of the objective treatment response, Child-Pugh grade B vs A (OR, 3.706; 95% CI, 1.718-7.996; P = 0.001), high serum AFP level (OR, 2.459; 95% CI, 1.187-5.094; P = 0.015) and a failure to meet the objective treatment response (OR, 5.619; 95% CI, 2.475-12.760; P < 0.001) were independent prognostic factors for mortality (Table 3).

TRR_1m was assessed in all 65 patients. TRR_1m was more than 20% in 41 patients (63.1%). Of the 41 patients with a TRR_1m≥ 20%, 35 patients (85.4%) showed a partial response, while only 2 (8.3%) of the 24 patients with TRR_1m < 20% showed a partial response (P < 0.001; Figure 3). When logistic regression analysis was performed after inclusion of TRR_1m among the variables used to predict objective treatment response, Child-Pugh grade A (OR, 0.121; 95% CI, 0.019-0.784; P = 0.027) and TRR_1m≥ 20% (OR, 158.302; 95% CI, 14.032-1785.827; P < 0.001) were independent predictive factors.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Proportion of patients who achieved an objective treatment response according to the tumor regression rate at one month after the initiation of radiotherapy. RT: Radiotherapy.

Adverse events during and one month after RT are summarized in Table 4. Adverse hematologic events were identified in 78.5% of patients. Although most of them were mild and transient, grade 3 or 4 adverse events were noted in six patients (9.2%). Grade 3 or 4 hematologic adverse events were more frequent in patients who underwent RT combined with HAI (4 of 16 patients, 25%) than in patients with RT alone (2 of 49, 4.1%; P = 0.012). Adverse hepatic events were identified in 51.8% of the patients; the most common were hypoalbuminemia (33.8%) and hyperbilirubinemia (24.6%). Grade 3/4 adverse hepatic events developed in four patients (6.2%), with hyperbilirubinemia or elevation of AST in 1 (1.5%). Grade 3/4 adverse hepatic events were more frequent in patients with TRR_1m < 20% (4 of 24 patients, 16.7%) than in those with TRR_1m≥ 20% (1 of 41, 2.4%; P = 0.038). GI bleeding developed in five patients. The status of two patients with peptic ulcer disease and one patient with variceal bleeding was improved with medical or endoscopic treatment. However, two patients with hemobilia or HCC rupture expired after these events.

Hepatocellular carcinoma (HCC) is one of the most common causes of cancer-related death in the world, especially in Asia. In a large proportion of patients, HCC is diagnosed at an advanced stage and, in this stage, widely performed treatment modalities including surgical resection, local ablation therapy and transarterial chemoembolization are not indicated. Recently, several reports have suggested high-dose radiotherapy (RT) could be an effective treatment option for advanced HCC.

Most previous studies have included only patients who completed the RT schedule. However, according to our experience, some proportion of patients could not complete the whole RT schedule and their prognosis was usually very poor. Therefore, this might lead to a selection bias when analyzing the treatment response and survival of patients. In this study, all patients with HCC who were treated with RT for more than 1 mo during the study period were included. In addition, we evaluated the prognostic significance of early tumor response by follow-up CT at 1 month after the initiation of RT.

RT was effective in patients with advanced HCC with an objective tumor response rate of 56.9%. Early tumor response rate at 1 month after the initiation of RT was shown to be a good prognostic indicator for RT response. In addition, severe adverse events were more frequent in patients with poor early tumor response rates.

RT could be considered as a treatment option for patients with advanced HCC. If the tumor does not decrease to 20% from baseline after one month of RT, interruption of RT can be considered due to the likelihood of a low objective treatment response rate and a high rate of severe adverse hepatic events.

This is an interesting article, which may offer new insights in the treatment of advanced unresectable HCC. The paper is well organized and the results are clearly described and commented.