This study was conducted as a retrospective cohort study. Between January 2009 and December 2011, a total of 104 patients received HAIC with 5-FU (750 mg/m²) and cisplatin (25 mg/m²) on days 1-4 every four weeks via an implantable port system for advanced HCC at Yeungnam University Hospital. Of these 104 patients, a total of 54 patients who received HAIC exclusively as a palliative therapy were selected for this study (Fig. 1). All of the selected subjects had advanced HCC at the time of diagnosis (Barcelona Clinic Liver Cancer (BCLC) stage C: n=54), and were older than 18 years of age.13,14 The diagnosis of HCC was made either by histopathologic confirmation or typical radiologic appearance on two dynamic imaging modalities, or via one dynamic imaging examination with elevated serum alpha-fetoprotein (>400 ng/mL). We performed a retrospective review of all medical records, investigated the clinical efficacy of HAIC for advanced HCC divided into Child-Pugh class A and Child-Pugh class B/C, and evaluated safety through the toxicities or adverse events of HAIC.

This study was conducted in accordance with the 1975 Declaration of Helsinki. All subjects submitted written informed consent prior to enrollment in the study. Details of this study were approved by the Institutional Review Board of Yeungnam University Hospital (YUH-13-0341-O19).

Under local anesthesia, the Seldinger method was used to gain access to the right femoral artery. Arteriography of the celiac trunk and superior mesenteric artery was performed in order to confirm portal vein patency, and arterial vascularization of the liver. After detection of tumor location and the main supplying hepatic artery, a 5.8-F indwelling intraarterial catheter was inserted, and the tip of the catheter was positioned at the common hepatic or proper hepatic artery under fluoroscopic guidance. The gastroduodenal artery was embolized using a coil to avoid formation of an ulcer as a result of efflux of the chemocytotoxic agents into the stomach or duodenum. The proximal end of the catheter was placed at the injection port and the device was subcutaneously implanted in the right thigh. To prevent occlusion of the catheter, 10 mL (10,000 units) of a heparin solution was infused via the port system for every cycle of chemotherapy.

Liver cirrhosis was defined by histologic, clinical, or radiological evidence. Clinical signs of liver cirrhosis included lower platelet count (<1.4×10⁵ cells/µL), the existence of varix or ascites, and the occurrence of encephalopathy. The gross type of tumor was classified by enhanced abdominal CT and angiography as diffuse, massive, and nodular type according to Eggel's classification. Responder was defined as a patient who achieved complete or partial response by HAIC according to radiologic examination, and non-responder was defined as a patient who achieved stable disease, progressive disease or no evaluation after HAIC.

The patients received treatment with 5-FU and cisplatin. 5-FU (750 mg/m²) was diluted with 5% dextrose in water and 200 mL was administered over two hours through a portable infusion pump from days 1 to 4. Cisplatin (25 mg/m²) was diluted with normal saline and 200 mL was infused over one hour using an intra-arterial catheter from days 1 to 4. The cycle of chemotherapy was every four weeks. To prevent cisplatin-associated nephrotoxicity, 1,000 mL of half-saline solution mixed with 20 mEq KCl and 8 mEq MgSO₄ was given over three hours before and after administration of cisplatin, respectively. This treatment was continued until radiologic and clinical disease progression or development of intolerable adverse events.

Assessment of treatment for HCC was performed after every two cycles of HAIC through dynamic contrast-enhanced abdominal CT scan or gadolinium-enhanced magnetic resonance imaging using the modified Response Evaluation Criteria in Solid Tumors (mRECIST) for HCC. Concomitant antiviral therapy was allowed. If the adverse events by HAIC were grade 3 or 4, the dosage of sequential chemotherapeutic agents was reduced to 75% of the original dosage. In addition, after reduction of the dose, in the case of additional development of grade 3 or 4 toxicities, the patients were regarded as having intolerance to HAIC and were withdrawn from this therapy.

The primary outcomes of this study were to evaluated OS and PFS after HAIC for advanced HCC according to Child-Pugh class. OS was measured as the time interval from the initiation of treatment of HAIC to death or the last follow-up visit. PFS was measured as the time interval between the date of the first cycle of HAIC and the date that disease progression or any cause of death was first observed. Secondary outcomes included the response rate, and safety or adverse events developed by HAIC. Response rate was defined as the proportion of patients who had a best response rating of complete or partial responses over total responses after HAIC according to mRECIST. Safety was assessed in all patients using the National Cancer Institute's Common Terminology Criteria for Adverse Events (NCI-CTC AE) version 4.0. All of the selected patients were followed up at least every four weeks, and the evaluation for adverse events was performed prior to initiation of a new cycle. The evaluation included checking bone marrow reserve through complete blood count and differential cell count, liver function test, and renal function test, as well as physical examination, vital signs, and clinical manifestations.

Categorical data or frequency data were presented as absolute values and percentages, whereas continuous data were presented as mean±standard deviation. Categorical data were analyzed by Pearson χ² test or Fisher's exact test, and continuous data were analyzed using Student t-test; if it was parametric, or not, the Mann-Whitney U test was used. Univariate and multivariate analyses of independent factors of the response to HAIC were assessed using binary logistic regression analysis. The Kaplan-Meier method was used for univariate analysis of OS and PFS and differences among groups were analyzed using the log-rank test. Multivariate analysis of the factors that influenced survival was performed using the Cox proportional hazards model. All P-values were two-sided, and a P-value <0.05 was considered statistically significant. Microsoft Excel 2007 (Microsoft Corp., Redmond, Washington, USA) was used for organization of all data, and PAWS statistics version 18.0 for Windows (SPSS Inc., Chicago, Illinois, USA) was used for analysis of data.

Median age of selected patients was 59.0 years (range: 39-78 years). Male was predominant (85.2%). Hepatitis B virus infection was the most common cause of HCC (70.4%). Other parameters of baseline characteristics are shown in Table 1. At the time of diagnosis of HCC, BCLC stage of enrolled patients was all stage C. However, after implantation of the drug port system, the liver function of some patients showed deterioration. As a result, BCLC stage of selected patients was changed into stage C (94.4%), and stage D (5.6%). In addition, Child-Pugh class was changed as shown in Table 1. All the analyses were performed according to staging and classification changed after implantation of the port system or at the time of initiation of HAIC. Mean interval between the time of diagnosis of HCC and the time of initiation of HAIC was 7.3±5.2 days.

Mean cycles of HAIC was 4.0±2.8 times. Mean dose of 5-FU and cisplatin for each cycle was 3,683.9±1,306.3 mg and 126.5±43.1 mg, respectively. The cumulative total dose of anticancer agents was 14,895.6±12,855.7 mg (range: 457-54,838 mg) for 5-FU, and 512.2±440.1 mg (range: 18-1,824 mg) for cisplatin, respectively. Mean follow-up duration of the selected subjects was 8.1±6.7 months. Median OS and PFS of enrolled patients were 5.1 months (95% confidence interval [CI]: 2.4-7.8) and 4.4 months (95% CI: 3.5-5.3), respectively.

The selected patients were divided into two groups according to Child-Pugh classification; Child-Pugh A group (n=24) and Child-Pugh B/C group (n=30). A summary of baseline characteristics of each group is shown in Table 2. The median OS was 8.7 months in the Child-Pugh A group (95% CI: 4.7-12.7), and 3.7 months in the Child-Pugh B/C group (95% CI: 2.0-5.3; Fig. 2A). The median PFS was 7.1 months in the Child-Pugh A group (95% CI: 3.8-10.4), and 3.6 months in the Child-Pugh B/C group (95% CI: 2.0-5.2; Fig 2B), respectively. All median OS and PFS in both groups were not statistically significant (P=0.079, P=0.196 by log-rank test, Table 3). In addition, other parameters such as portal vein invasion and the existence of ascites were not significantly related to the OS and PFS after HAIC (P>0.05 by log-rank test).

In total, although 12 patients (22.2%) showed partial responses, no complete response after HAIC for advanced HCC was observed. The best objective therapeutic responses according to mRECIST were not significantly different between the Child-Pugh A group and Child-Pugh B/C group (P=0.979). According to other parameters as well as Child-Pugh class, the best therapeutic responses by HAIC were analyzed in comparison (Table 4). Although other parameters including BCLC stage were not statistically different from the response rates, significant difference was exclusively observed in modified Union for International Cancer Control (UICC) stage (P=0.025). In addition, median OS was not statistically longer in patients who achieved complete response or partial response (responders), compared with other patients (non-responders) in the Child-Pugh A group (13.1 vs. 7.1 months, P=0.120) and the Child-Pugh B/C group (12.5 vs. 2.8 months, P=0.075). Median PFS was not significantly different between responders and non-responders in the Child-Pugh A group (11.0 vs. 5.7 months, P=0.089). However, statistical difference was observed between patients in the Child-Pugh B/C group (10.9 vs. 2.7 months, P=0.042). Considering all of the selected subjects, median OS and PFS were statistically different between responders and non-responders (13.1 vs. 4.4 months, P=0.019; 11.0 vs. 3.6 months, P=0.022). In addition, according to the Cox proportional hazards model, achievement of complete or partial response after HAIC was an independent prognostic factor of OS and PFS (hazard ratio [HR]: 0.4, 95% CI: 0.2-0.8, P=0.011, HR: 0.4, 95% CI: 0.2-0.8, P=0.008) (Fig. 3). Subgroup analyses were performed by dividing the selected subjects into responders and non-responders in order to investigate independent factors related to achievement of a complete or partial response after HAIC. Results of comparative analyses between each group are shown in Table 5. Dose of 5-FU and cisplatin per cycle showed significant difference between the two groups, respectively (P=0.007, P=0.001). According to results of binary logistic regression analysis, dose of cisplatin per cycle was an exclusively independent factor for achievement of a complete or partial response after HAIC (odds ratio: 1.039, 95% CI: 1.0-1.1, P=0.008).

HAIC-related adverse events such as neutropenia (79.2 vs. 83.3%), anemia (95.8 vs. 100%), thrombocytopenia (91.7 vs. 90.0%), elevated aminotransferase (100 vs. 100%), hyperbilirubinemia (91.7 vs. 96.7%), elevated creatinine (33.3 vs. 60.0%), nausea/vomiting (54.2 vs. 63.3%), diarrhea (25.0 vs. 33.3%), and fever (58.3 vs. 63.3%) occurred between the Child-Pugh A group and the Child-Pugh B/C group, irrespective of NCI-CTC AE grading (Table 6). The incidence rate of common adverse events was not significantly different between the two groups (P>0.05). In addition, the incidence rate of serious adverse events over grade 3 for each category of the toxicities did not statistically differ between the two groups (P>0.05). Serious adverse events ≥grade 3 occurred frequently in both groups (83.3 vs. 96.7%, P=0.159). Although dose reduction rates by adverse events were not statistically significant between the two groups (33.3 vs. 26.7%, P=0.765), withdrawal rate significantly differed between the each group (16.7 vs. 43.3%, P=0.044). In total, seventeen patients withdrew from HAIC because of general weakness (64.7%), catheter-related complications (17.6%), and other causes (17.6%). Mean cycles of the patients who withdrew from HAIC was 2.5±2.0 times. Median OS of these patients was only 3.2 months (95% CI: 1.3-5.0).

In this study, 28 patients (51.9%) experienced catheter-related complications. Of these, 22 patients (78.6%) experienced bleeding or hematoma after implantation of the drug delivery system, which did not require a transfusion; and catheter occlusion occurred in three patients (10.7%), which was resolved by thrombolytics or catheter exchange. The remaining two patients and one patient developed catheter-related infection, and extravasation, respectively. Catheter-related infection or cellulitis around the infusion port was cured completely with administration of antibiotics, wound dressing, and removal of the drug port system.