SEARCH

SEARCH BY CITATION

Keywords:

  • cancer registry;
  • hepatocellular carcinoma;
  • hospital volume;
  • prognosis;
  • radiofrequency ablation

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

BACKGROUND:

Hospital volume for several major operations is associated with treatment outcomes. In this study, the authors explored the influence of hospital radiofrequency ablation (RFA) volume on the prognosis of patients who received RFA for hepatocellular carcinoma (HCC).

METHODS:

The authors searched for all patients who were diagnosed with stage I or stage II HCC from 2004 to 2006 and who received RFA as first-line therapy in a population-based cohort. Overall survival (OS) and liver cancer-specific survival (CSS) were compared according to hospital volume. A Cox proportional hazards model was used for multivariate analysis.

RESULTS:

In total, 661 patients received first-line RFA for stage I and II HCC in 28 hospitals. Among these, there were 480 patients (72.6%) in the high-volume group (those who received RFA at hospitals that treated >10 first-line patients per year), and there were 181 patients (27.4%) in the low-volume group (those who received RFA at hospitals that treated ≤10 first-line patients per year). The sex, age, stage, tumor size, and year of diagnosis for patients in the 2 groups did not differ significantly. Patients in the high-volume group demonstrated significantly longer OS and CSS than those in the low-volume group (5-year OS rate, 58.7% vs 47.2%; P = .001; 5-year CSS rate, 67.1% vs 57.1%; P = .009). After adjusting for covariates, high-volume hospitals remained an independent predictor of longer OS (hazard ratio, 0.57; P < .001) and CSS (hazard ratio, 0.57; P = .003).

CONCLUSIONS:

Patients who received first-line RFA for HCC in high-volume hospitals demonstrated better survival outcomes. Cancer 2013. © 2012 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Hepatocellular carcinoma (HCC) is the sixth most common malignancy and the third leading cause of cancer-related mortality in the world.1 For patients with local disease and acceptable liver function reserve, resection provides the opportunity for a cure and, thus, remains a standard treatment.2 However, for small HCCs, percutaneous radiofrequency ablation (RFA) provides a legitimate alternative for patients who cannot undergo surgery because of either poor liver function or other comorbidity.3-6 In 2 randomized, controlled trials that compared RFA with surgery in patients who had small, solitary HCCs, no significant differences in disease-free survival or overall survival (OS) were observed.7,8

The relation between tumor volume and outcome has been widely discussed in past decades, and previous studies have verified this relation in various procedures.9-11 For procedures that depend greatly on technique, such as cancer surgery (including hepatectomy for HCC), hospital volume not only has been associated with decreased treatment-related mortality and complication rates,12-18 it also has been associated with better long-term survival rates.19-26 A successful percutaneous RFA for HCC relies heavily on physician technique, because the procedure requires the insertion of a needle electrode directly into the lesion using imaging guidance (ultrasound or computed tomography).3 However, the effect of hospital volume of RFA on treatment outcomes remains unclear.

A multihospital database is required to study the effects of hospital volume. The Taiwan Cancer Registry (TCR), which is managed by the Bureau of Health Promotion, Department of Health in Taiwan, offers national data on cancer care, because major cancer care providers in Taiwan are obligated to report information concerning cancer patients to the TCR.27 Consequently, in the current study, this population-based database was used to investigate the influence of hospital RFA volume on survival rates.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Data Source

We identified a population-based cohort of patients who had newly diagnosed HCC from 2004 to 2006 using the TCR database, which includes approximately 60% to 80% of the population in Taiwan. Collected data were approved for release by the Data Release Review Board of the Bureau of Health Promotion, Department of Health, Taiwan. All procedures performed through this study also were approved by the Institutional Review Board of the College of Public Health, National Taiwan University.

Study Population

The study population included patients who had a first, newly diagnosed HCC (International Classification of Diseases for Oncology, Third Revision code C220) reported to the TCR database from 2004 to 2006. The following criteria were used to determine eligible patients: 1) first report of HCC as a primary tumor; 2) diagnosed with stage I or stage II HCC according to the American Joint Committee on Cancer (AJCC) cancer staging system, sixth edition28; 3) received RFA as the first line of treatment; and 4) age ≥18 years. Patients who had the following characteristics were excluded: 1) previously diagnosed with other types of cancer, 2) reported the presence of multiple primary cancers, 3) reported lymphoma or Kaposi sarcoma, and 4) received other treatment before or in conjunction with RFA. Hospital volume was calculated from every patient's record as the number of first-line RFA treatments for HCC performed in that patient's hospital in the same year as the patient's procedure. Patients were divided into high-volume and low-volume groups with a preselected cutoff point based on hospital volume.

Study Variables and Endpoints

Basic clinicopathologic variables, such as age, sex, stage, and tumor size, were retrieved from the cancer registry database. The final study cohort was then linked to the National Death Registry from 2004 to 2009 to determine death events. To comply with personal electronic data privacy regulations, personal identities were encrypted, and all data were anonymously analyzed. OS was measured from the date of diagnosis to the date of death from any cause, and liver cancer-specific survival (CSS) was measured from the date of diagnosis to the date of death from HCC. Data were censored if events did not occur before the last date in the National Death Registry database (December 31, 2009).

Statistical Methods

Statistical analyses were performed using the SAS statistical software package (version 9.1.3; SAS Institute, Inc., Cary, NC). The mean or frequency of patient characteristics for the 2 study groups at the time of HCC diagnosis was compared using a one-way analysis of variance for continuous variables or the chi-square test (or the Fisher exact test, as indicated) for categorical variables. Patient survival based on hospital volume status was estimated using the Kaplan-Meier method and was compared using the log-rank test. A Cox proportional hazards model was used to calculate the hazard ratios (HRs) and associated 95% confidence intervals (CIs) for the effect of hospital volume and other covariates on survival. Two-sided P values ≤ .05 were considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Distribution of Hospital Volume

Between 2004 and 2006, in total, 22,932 patients were reported who had newly diagnosed HCC. Among these patients, 661 satisfied the eligibility criteria for the current study and were enrolled (Fig. 1). Approximately 60% of patients were men, and >50% of patients were aged >65 years. Approximately 33% of patients had AJCC stage II HCC, and >50% of patients had tumors that measured >2 cm (Table 1). Patients were treated in 28 hospitals, including of 17 medical centers and 11 regional hospitals. The median hospital volume for RFA treatment was 3 treatments (range, 0-75 treatments) per hospital-year. After excluding hospitals that did not provide RFA treatment in a specific year, the median volume increased to 4 treatments, and the mean volume was 10 RFA treatments per hospital-year (Fig. 2).

thumbnail image

Figure 1. This is a flow diagram of the study patients. HCC indicates hepatocellular carcinoma; TAE, transarterial embolization; TACE, transarterial chemoembolization; PEI, percutaneous ethanol injection; RFA, radiofrequency ablation.

Download figure to PowerPoint

thumbnail image

Figure 2. This chart illustrates the distribution according to hospital volume of radiofrequency ablation (RFA) treatments. Each column represents the number of first-line RFA treatments performed in a hospital during a specific year (2004, 2005, or 2006). The dashed line separates the columns into 2 groups, high and low volume, which were defined as >10 cases and ≤10 cases, respectively.

Download figure to PowerPoint

Table 1. Baseline Characteristics
 No. of Patients (%) 
  Hospital Volume 
VariableAll PatientsLow-Volume: ≤10 Cases/YearHigh-Volume: >10 Cases/YearP
  • Abbreviations: AJCC, American Joint Committee on Cancer.

  • a

    Disease stage was based on the AJCC Cancer Staging Manual, 6th edition.

  • b

    Data were missing for 3 patients.

  • c

    This P value was determined using the Fisher exact test.

Total661 (100)181480 
Sex
 Men396 (59.9)101 (55.8)295 (61.5).186
 Women265 (40.1)80 (44.2)185 (38.5) 
Age group, years
 18-64303 (45.8)73 (40.3)230 (47.9).081
 ≥65358 (54.2)108 (59.7)250 (52.1) 
AJCC stagea
 I437 (66.1)128 (70.7)309 (64.4).124
 II224 (33.9)53 (29.3)171 (35.6) 
Tumor size, cmb
 ≤2287 (43.4)78 (43.1)209 (43.5).949
 >2371 (56.1)100 (55.2)271 (56.5) 
 Unknown3 (0.5)3 (1.7)0 (0) 
Medical center
 Yes586 (88.7)106 (58.7)480 (100)< .001c
 No75 (11.3)75 (41.4)0 (0) 
Year of diagnosis
 2004131 (19.8)46 (25.4)85 (17.7)0.052
 2005204 (30.9)47 (26)157 (32.7) 
 2006326 (49.3)88 (48.6)238 (49.6) 

According to the RFA volume of hospitals in which patients received treatment, patients were classified into a high-volume group and a low-volume group using the mean volume as the cutoff point. The high-volume group received RFA in hospitals that performed >10 RFAs per hospital-year as first-line treatment for HCC, whereas the low-volume group underwent RFA in hospitals that performed ≤10 RFAs per hospital-year as first-line treatment for HCC. In total, 480 patients (72.6%) were classified in the high-volume group, and 181 patients (27.4%) were classified in the low-volume group. No significant differences were observed in sex, age, stage, tumor size, or year of diagnosis between the 2 groups. More patients in the high-volume group had stage II disease or tumors >2 cm. However, these differences were not statistically significant (Table 1). All patients in the high-volume group received first-line RFA at medical centers.

Effect of Hospital Volume on Survival

After a median follow-up of 41.9 months, 250 patients (37.8%) had died, and 186 of those deaths (74.4%) were attributed to HCC. Patients in the high-volume group, compared with patients in the low-volume group, had a significantly longer OS (5-year OS rate: 58.7% vs 47.2%; P = .001) (Fig. 3A) and CSS (5-year CSS rate: 67.1% vs 57.1%; P = .009) (Fig. 3B). When the same comparisons were analyzed separately for stage I and stage II disease subgroups, higher survival rates for patients in the high-volume group remained (Fig. 3C-F).

thumbnail image

Figure 3. Kaplan-Meier curves for patients in the high-volume group (high) and the low-volume group (low) illustrate (A) overall survival (OS) for all patients, (B) liver cancer-specific survival (CSS) for all patients, (C) OS for patients with stage I disease, (D) CSS for patients with stage I disease, (E) OS for patients with stage II disease, and (F) CSS for patients with stage II disease.

Download figure to PowerPoint

After adjusting for covariates, including sex, age, stage, tumor size, hospital type, and year of diagnosis, using the Cox proportional hazards model, high RFA volume remained an independent predictor of longer OS (HR, 0.57; 95% CI, 0.42-0.79; P < .001) and longer CSS (HR, 0.57; 95% CI, 0.39-0.83; P = .003) (Table 2). Younger patients and patients with lower AJCC stage also had longer OS and CSS. The type of hospital (medical center or other) in which patients received treatment did not influence survival outcomes.

Table 2. Univariate and Multivariate Analysis of Overall Survival and Liver Cancer-Specific Survival Rates
  Overall SurvivalLiver Cancer-Specific Survival
 No. of PatientsUnivariate AnalysisAdjusted AnalysisaUnivariate AnalysisAdjusted Analysisa
VariableHR (95% CI)PHR (95% CI)PHR (95% CI)PHR (95% CI)P
  • Abbreviations: AJCC, American Joint Committee on Cancer; CI, confidence interval; HR, hazard ratio; Ref, referent category.

  • a

    The adjusted analysis included all variables in 1 model.

  • b

    Disease stage was based on the AJCC Cancer Staging Manual, 6th edition.

  • c

    Data were missing for 3 patients.

Hospital volume         
 Low181Ref.001Ref< .001Ref.009Ref.003
 High4800.65 (0.50-0.85) 0.57 (0.42-0.79) 0.67 (0.49-0.91) 0.57 (0.39-0.83) 
Sex         
 Men396Ref.257Ref.048Ref.257Ref.064
 Women2650.86 (0.67-1.11) 0.77 (0.59-1.00) 0.84 (0.63-1.13) 0.75 (0.55-1.02) 
Age6111.03 (1.01-1.04)< .0011.03 (1.01-1.04)< .0011.02 (1.01-1.04).0031.02 (1.01-1.04).003
AJCC stageb         
 I437Ref.002Ref< .001Ref< .001Ref< .001
 II2241.48 (1.15-1.91) 1.62 (1.24-2.10) 1.64 (1.23-2.19) 1.78 (1.32-2.41) 
Tumor size, cmc         
 ≤2287Ref.005Ref.116Ref.004Ref.068
 >23711.44 (1.12-1.86) 1.24 (0.95-1.61) 1.55 (1.15-2.10) 1.33 (0.98-1.82) 
Medical center         
 Yes586Ref.082 .400Ref.253Ref.324
 No751.38 (0.96-1.98) 0.82 (0.53-1.29) 1.28 (0.84-1.97) 0.77 (0.45-1.30) 
Year of diagnosis         
 2004131Ref.897Ref.955Ref.504Ref.614
 20052041.07 (0.76-1.50) 1.05 (0.74-1.48) 0.99 (0.68-1.45) 0.96 (0.65-1.41) 
 20063261.01 (0.72-1.42) 1.02 (0.72-1.44) 0.83 (0.56-1.22) 0.84 (0.56-1.24) 

Subgroup analysis defined by sex, age (ages 18-64 years or ≥65 years), disease stage (stage I or II), tumor size (≤2 cm or >2 cm), hospital type, or year of diagnosis was performed to examine the concordance between different subgroups. These variables were adjusted in the Cox model, and the results revealed the same trend toward better survival rates in the high-volume group for all subgroups (Fig. 4).

thumbnail image

Figure 4. Subgroup analyses compare overall survival and liver cancer-specific survival for patient groups that received treatment in high-volume and low-volume hospitals. HR indicates hazard ratio; CI, confidence interval; AJCC, American Joint Committee on Cancer. A single asterisk indicates the inclusion of all variables in 1 model; double asterisks, based on the AJCC Cancer Staging Manual, sixth edition; cross, data were missing for 3 patients.

Download figure to PowerPoint

Patients in the high-volume group were divided further into 2 subgroups according to hospital volume to ascertain whether hospital volume higher than that used to define the original high-volume group was associated with better survival rates. One subgroup of high-volume patients received RFA in hospitals that performed >30 RFAs per hospital-year as first-line treatment for HCC, and the other subgroup received RFA in hospitals that performed >10 but ≤30 RFAs per hospital-year for HCC. No difference in OS or CSS could be identified between the 2 subgroups (Fig. 5A, B).

thumbnail image

Figure 5. Kaplan-Meier curves illustrate (A) overall survival and (B) liver cancer-specific survival according to different volumes in the high-volume group.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

In this study, we have demonstrated the association between hospital RFA volume and prognosis among patients who receive RFA for early-stage HCC. Patients who received RFA as first-line treatment for early-stage HCC in high-volume hospitals had longer OS and CSS than those who received treatment in low-volume hospitals. Both multivariate and subgroup analyses demonstrated that better survival rates were consistent irrespective of disease stage, tumor size, or hospital type. To our knowledge, this is the first study to address the volume-outcome relation in patients who receive RFA for HCC. The use of government-initiated national databases in this study ensured that data for cancer characteristics and survival were accurate and representative of the general population.

Although RFA is a procedure that depends greatly on technique, morbidity and mortality related to RFA are relatively uncommon compared with other major operations and high-risk procedures.4-6 Previous studies based on high-risk procedures have discovered that operative or in-hospital mortality rates are lower in high-volume hospitals.10-18 The differences for post-RFA morbidity and mortality between high-volume and low-volume hospitals may be limited because of the low complication rates attributable to RFA. The influence of hospital volume on late survival cannot be fully explained using differences in post-RFA morbidity and mortality and should be attributed to other factors.

Several potential factors that may result in better survival rates for patients who are treated in high-volume hospitals exist. In the current study, better patient selection in high-volume hospitals was unlikely, because more patients (although not a statistically significant number) in the high-volume group demonstrated higher stage and larger tumor size. Hospital type also was unrelated to patient survival rates. Although we do not have detailed information, technical differences in RFA may yield different efficacy of tumor ablation. In addition, active surveillance and appropriate treatment of recurrence in hospitals with RFA experience also may have contributed to a better prognoses for patients with HCC.

Whether the effect of hospital volume can be attributed (at least in part) to physician volume warrants further exploration. Several previous studies have analyzed the role of surgeon volume in major operations and discovered that surgeon volume accounted for a substantial proportion of the effect of hospital volume on operative mortality or survival.9, 26, 29 No information concerning practicing physicians was available in the database that used for this study, so the influence of physicians on RFA treatment awaits further investigation.

The current study was affected by certain limitations. First, detailed demographic variables were not available in the TCR database. For example, we did not have information on hepatitis virus infection and subsequent antiviral therapy. Nevertheless, Taiwan is a hyperendemic region for hepatitis B virus (HBV) infection. Approximately 70% of HCCs were associated with HBV, and there was no significant difference in the prevalence of HBV infection among different areas in Taiwan.30 The distribution of patients with different hepatitis virus infections in most hospitals should be similar. In addition, antiviral agents were reimbursed in Taiwan, and the administration of these agents for chronic, active viral hepatitis was typical. We believe these etiologic factors may have limited impact on our analytic results. Second, the severity of cirrhosis was unknown. However, the 5-year survival rate of 50% to 60% for the patients in this study was comparable to that of patients who underwent hepatic resection, which is available only for patients who have minimal cirrhosis and adequate liver function reserve.31 We maintain that the majority of patients in our study were homogenous, possessed relatively preserved liver function, and received RFA with curative intent.

The cutoff point between high-volume and low-volume hospitals in this study was relatively low. This suggests that RFA may be easier to master compared, for instance, with hepatic resection. However, when determining hospital volume, only first-line RFA was included. Hospitals actually performed more RFA treatments per year. To ensure the homogeneity of our patient cohort, we did not include the use of RFA to treat recurrent disease. We believe the proportion of first-line RFA and total RFA procedures performed should be similar among hospitals, and the volume of unmeasured subsequent RFA treatment should not affect our current findings.

In conclusion, according to the results from this population-based cohort study, higher survival rates were demonstrated in patients who received first-line RFA for HCC in high-volume hospitals. Specific factors that caused these higher survival rates warrant further investigation.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

This study was supported by grants from the Bureau of Health Promotion, Department of Health, Taiwan (Taiwan Cancer Registry Project); the Science and Technology Unit, Department of Health, Taiwan (DOH99-TD-B-111-001 and DOH100-TD-B-111-001); and National Taiwan University Hospital, Taipei, Taiwan (NCTRC201208, NTUH.101-N1965).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES