Between June 2005 and June 2009, 30 patients with intractable and advanced HCC were recruited to this pilot trial of low dose metronomic chemotherapy. Intractable and advanced HCC was defined as HCC with portal vein invasion, infiltrative type and/or refractory to surgical resection or nonsurgical intervention (transcatheter arterial chemoembolization, radiofrequency ablation and percutaneous ethanol injection). A diagnosis of HCC was based on either pathologic confirmation or radiologic findings with an elevated level of serum alpha-fetoprotein (AFP) (>400 ng/mL) in the setting of liver cirrhosis. Initially, for inclusion into the study, patients eligible patients met the following criteria: >18-years-of-age, Eastern Cooperative Oncology Group (ECOG) performance status of 0-1, adequate liver function defined as Child-Pugh score of ≤7, acceptable blood cell counts (absolute neutrophil count of ≥1,500 cells/mm³, platelet count ≥60,000 cells/mm³, white blood cell count ≥4,000 cells/mm³), serum creatinine level ≤1.5 mg/dL, and vascular accessible lesion for the implantation of a chemoport. However, because there was few treatment option for patients with advanced HCC, patient with good performance status and favorable liver function was enrolled even if they had low blood cell counts. Portal vein invasion was defined as having a tumor thrombus in the main trunk (Vp4) and the first-order branch (Vp3) of the portal vein. Patients were excluded if they had one or more of the following: evidence of hepatic decompensation including uncontrolled ascites, gastrointestinal bleeding or hepatic encephalopathy; total bilirubin level ≥3 mg/dL or prolongation of prothrombin time by more than 3 seconds; concurrent serious medical condition(s) such as underlying cardiac or renal diseases; or other concurrent primary malignancy. The trial was conducted according to the current declaration of Helsinki, and the protocol was approved by the institutional ethics committee of the Catholic University of Korea. Written informed consent was obtained from each patient.

Chemotherapeutic agents were delivered through a percutaneous hepatic arterial catheter (chemoport) implanted at the time of initial hepatic arterial angiography. The catheter insertion and placement were performed as previously described.10 The metronomic chemotherapeutic agents were cisplatin and 5-fluorouracil (5-FU). Cisplatin and 5-FU was infused weekly for 3 weeks with intervening 1 week break. A single dose of cisplatin and 5-FU was 15 mg/m² and 50 mg/m², respectively. Along with the metronomic chemotherapy, epirubicin (30 mg/m²) was infused monthly without use of lipiodol and gelfoam embolization. This protocol was repeated every 4 weeks and continued until disease progression or unacceptable toxicity was evident, or the patient refused to continue (Fig. 1). In case of cisplatin infusion, intravenous hydration was performed to prevent nephrotoxicity and an appropriate anti-emetic agent was given to all patients. Dose adjustments were made depending on the toxicity observed in the preceding treatment cycle. The following cycle of treatment was reduced by 30% in the case of repeated grade 2 toxicity during the preceding cycle. Treatment was stopped in the case of any grade 3 or grade 4 toxicity until resolution. If a patient required a delay of >4 weeks for recovery, they were excluded from the study.

The primary endpoint of the current study was feasibility. Feasibility was assessed by the proportion of patient who could undergo at least two complete sessions of metronomic chemotherapy without adverse event. The secondary endpoints were an objective response rate (complete response plus partial response [PR]), time to disease progression and overall survival. Tumor response was evaluated using abdominal dynamic computed tomography after every two complete sessions of chemotherapy. The responses were evaluated using Response Evaluation Criteria In Solid Tumors criteria on an intention-to-treat basis.11 The tumor responses were categorized as CR, PR, stable disease (SD), or progressive disease (PD). On computed tomography scans obtained after treatment, the sums of the longest diameter for all target HCCs was measured. Complete response was defined as the disappearance of any intratumoral arterial enhancement in all target lesions. PR was defined as at least a 30% decrease in the sum of the longest diameter of viable (contrast enhancement in the arterial phase) target lesions taking as reference the baseline sum longest diameter. PD was defined as at least a 20% increase in the sum of the longest diameter of viable (enhancing) target lesions or the appearance of one or more new lesions. SD was defined as a tumor status that did not meet the above three response criteria. For the evaluation of endpoint, the best response for each patient was selected. Liver function, renal function, complete blood counts, serum AFP, and chest X-ray were also evaluated at each visit. Toxicity was scored using the Common Terminology Criteria for Adverse Events (CTCAE; version 3.0)12 at every visit. Survival and time to disease progression was estimated from the date of treatment initiation.

Continuous data were expressed as medians and ranges, and categorical data as percentages. Comparison between groups was performed using the Chi-square and/or Fisher's exact tests, and Mann-Whitney U-test, as appropriate. Cumulative survival and time to disease progression rates were calculated using the Kaplan-Meier method and the difference between groups were compared by the log-rank test. A Cox proportional-hazards model was used to reveal independent clinical factors or groups affecting response rate and overall survival. Factors less than 0.1 in univariate analysis was further analyzed as cofactor in multivariate analysis. All data was analyzed using SPSS version 14.0 software (SPSS, Chicago, IL, USA).

Thirty patients were enrolled into the study between June 2005 and June 2009. Table 1 summarizes the patient's baseline characteristics. The median age was 53 years and 93% of the patients were male. The underlying liver disease was predominantly hepatitis B virus infection (90%) and hepatic function was mostly Child-Pugh class A (80%). All patients had portal vein tumor thrombosis (PVT) and 12 (40%) patients had a liver tumor burden ≥50%. At enrollment, 15 patients (50%) had extrahepatic metastasis in the lung (n=9), lymph node (n=1), bone (n=1), abdominal wall (n=1), both bone and lymph node (n=2) and both bone and lung (n=1). Fifteen patients (50%) displayed progressive disease to other treatments for HCC before metronomic chemotherapy.

One hundred sixteen cycles of metronomic chemotherapy with a median of three cycles (mean 3.8 cycles; range, 1-15 cycles) for each patient was administered to the 30 patients. Twenty two (73%) patients received at least two complete sessions of metronomic chemotherapy, The remained 8 patients received one complete session but did not complete two session of chemotherapy due to adverse event (neutropenia, n=4, GI bleeding, n=1) and aggravated liver function due to tumor progression (n=1) or ingestion of herb (n=2). The median dose of anti-cancer agents per cycle was 58 mg (range, 35-74 mg) for epirubicin, 75 mg (range, 45-105 mg) for cisplatin, and 256.95 mg (range, 150-450 mg) for 5-FU. The cumulative median dose for each drug during the entire treatment was 162.5 mg (range, 35-960 mg) for epirubicin, 225 mg (range, 60-1215 mg) for cisplatin, and 757 mg (range, 225-4050 mg) for 5-FU. In 15 patients with extrahepatic metastasis, 11 received concurrent radiation therapy for the metastatic lesion, while the other four patients did not because metronomic chemotherapy was prematurely terminated due to adverse events or progression.

After start of metronomic chemotherapy, tumor response could be evaluated in all 30 patients at a median of 36 days (range, 22-121 days). Of 22 patients who received at least two complete sessions of metronomic chemotherapy, PR was achieved in 23% (5/22), SD was observed in 23% (5/22) and PD was noted in 54% (12/22). Of 8 patients who did not complete two session of chemotherapy, PR was achieved in 12.5% (1/8), SD was observed in 12.5% (1/8) and PD was noted in 75% (6/8). In overall, PR was achieved in 20% (6/30), SD was observed in 20% (6/30) and PD was noted in 60% (18/30). Of the 12 patients with PR or SD initially, six patients showed PD during metronomic chemotherapy at a median of 102 days (range, 72-414 days) and overall progression rate was 80% (24/30). Total duration of metronomic chemotherapy was median 78 days (range, 27-489 days). After metronomic chemotherapy, only 6 patients out of 24 patients with progression underwent further treatment and remained patient did not receive further treatment due to aggravated liver function or performance status. Meanwhile, 1 patient out of 6 patients with objective response received further treatment and other patients did not receive further treatment due to stable tumor status after the end of metronomic chemotherapy. In overall, the median time to progression was 63 days (range, 26-631 days). After the treatment, the median AFP level changed from 381 (range, 1.93-92320) to 371 (range, 1.62-439980) ng/mL. According to the treatment response, the median AFP level changed from 230.44 (range, 5.39-3955) to 18.41 (range, 6.73-2703) ng/mL in patients with PR, from 51.05 (5.15-374) to 106.57 (4.35-658) ng/mL in SD, and from 1455.50 (range, 1.93-92320) to 2988.00 (range, 1.62-439980) ng/ml in PD. Of the various baseline factors, history of previous treatment for HCC, baseline Child-Pugh class, and baseline serum AFP level were significantly associated with tumor response on univariate analysis. Only baseline serum AFP level remained significant on multivariate analysis (Table 2). None of the 15 patients who experienced PD after a previous treatment showed an objective response (four SD and 11 PD) while six of the 15 patients (40%) who received metronomic treatment as the initial HCC therapy displayed PR (P=0.023). Tumor response was not affected by the presence of extrahepatic metastasis. Five patients (33%) with extrahepatic metastatsis showed PR or SD. Figure 2 depicts a sample case with extrahepatic metastasis at baseline, in which marked tumor shrinkage was evident on a follow-up computed tomography scan following three cycles of metronomic chemotherapy. Blood samples were obtained from this HCC patient with lung metastasis by venipuncture before each weekly chemotherapy. Solid phase-enzyme linked immunosorbent assay (ELISA) was used to determine the serum values of VEGF (R & D Systems, Minneapolis, MN, USA).

The total follow-up time after start of metronomic chemotherapy was median 151 days (mean 239 days; range, 33-721 days) and the follow-up period after the end of metronomic chemotherapy was 60 days (mean 131 days; range, 6-605 days). At the end of follow-up, 12 of 30 (40%) patients were still alive. The median overall survival was 162 days (95% confidence interval 62-262 days) and 6-month and 1-year cumulative survival rates were 48.7% and 34.8%, respectively (Fig. 3). On univariate analysis, baseline serum AFP level and tumor response were significantly associated with survival rate (Table 3). Overall survival was significantly longer in patients with baseline AFP ≤400 ng/mL (558 days) than at >400 ng/mL (121 days; P<0.001). However, in multivariate analysis, no factor was found to be significantly associated with survival probably due to small number of enrolled patients.

During treatment, 36 adverse events occurred, and dose reduction or treatment delay was necessary in 21 patients (Table 4). The causes of dose reduction or treatment delay were neutropenia (n=14), aminotransferase elevation (n=1), aggravated ECOG score (n=2), creatine elevation (n=1), hyperbilirubinemia (n=2) and varix bleeding (n=1). Of these, one patient with aggravated ECOG score and one patient with varix bleeding started metronomic chemotherapy with a reduced dose. Most cases of neutropenia were manageable with conservative care and granulocyte colony-stimulating factor injection, while four patients with neutropenic fever had to permanently discontinue the treatment due to PD (n=3) and concern with recurrent neutropenia (n=1), and were managed with broad spectrum antibiotics. Meanwhile, only one of the 30 patients continued the treatment at the end of follow-up. Others stopped the treatment permanently due to tumor progression (n=7), hepatic failure associated with tumor progression (n=11), hyperbilirubinemia (n=2), noncompliance (n=4), femoral catheter dysfunction (n=1), aggravated ECOG score (n=2), varix bleeding (n=1), and neutropenia (n=1).