ORIGINAL ARTICLE
Hepatocellular Carcinoma Surveillance Are We Utilizing It? Nowlan Selvapatt, MBBS, BSc, MRCP,* Henry House, BSc,* and Ashley Brown, MBChB, MD, FRCPw
Background and Goals: International guidelines recommend 6monthly ultrasound scan surveillance in cirrhotic patients for hepatocellular carcinoma (HCC) screening. The aim of this study was to evaluate HCC surveillance and to provide the outcomes for the largest reported cirrhosis cohort in the United Kingdom. Study: We retrospectively reviewed all cirrhotic patients during the 6 months before April 1, 2014 at 3 teaching hospitals within the Imperial College Healthcare NHS Trust. All patients with cirrhosis eligible for HCC screening in the cohort were reviewed for evidence of screening and were characterized as either “routine” or “overdue.” Reasons for failure to meet guidelines were identified in overdue patients. Univariate analyses were conducted to determine clinical and sociodemographic predictors of surveillance with a view to performing multivariate analysis. Results: Of the 898 patients eligible for inclusion, 65% patients had HCC surveillance performed in the past 6 months. During the observation period, 61 (6.8%) cases of HCC were detected. Thirtyeight were picked up on surveillance, 9 incidentally and in 14 cases it could not be determined from the notes. Within this cohort HCCs diagnosed on surveillance did not demonstrate significantly superior outcomes compared with those picked up incidentally. Conclusions: Surveillance of patients with cirrhosis within this cohort is suboptimal. Although disappointing at a local level, this is likely to be reflective of practice elsewhere. Key Words: hepatocellular carcinoma, HCC, surveillance, screening, utilization
(J Clin Gastroenterol 2016;50:e8–e12)
H
epatocellular carcinoma (HCC) is the sixth most common cancer worldwide and the third commonest cause of cancer mortality.1,2 Although most of the burden of disease is in developing countries due to hepatitis B virus, it is fast developing into an increasing problem in western Countries due to rising trends of hepatitis C virus, alcoholic liver disease, and nonalcoholic steatohepatitis (NASH).1,3 In the United States approximately 20,000 new cases of HCC are reported per year and in the United Kingdom HCC is responsible for roughly 1500 deaths per year.1,4 Cirrhosis is associated with 80% to 90% of HCC.1
Received for publication October 27, 2014; accepted April 20, 2015. From the *Imperial College; and wImperial College Healthcare NHS Trust, London, UK. The authors declare that they have nothing to disclose. Reprints: Nowlan Selvapatt, MBBS, BSc, MRCP, Department of Hepatology, 10th Floor QEQM Building, St Mary’s Hospital Campus, Imperial College, London W2 1NY, UK (e-mail: [email protected]). Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Regular surveillance increases the likelihood of detecting HCC at an early stage where it is more likely to be amenable to curative treatment.5 This strategy has been demonstrated to be cost-effective.6 The average 5-year survival rate for advanced tumors remains low at approximately 12%.1 Yet 5year survival rates as high as 70% have been demonstrated in tumors picked-up at an early stage that are suitable for curative therapies.7 In a United States retrospective cohort of 4482 patients with HCC listed for liver transplantation the overall 5-year survival was 62%, however, in those with more advanced cancers 5-year survival was 38%.8 International guidelines have been introduced to guide clinicians in the practice of HCC surveillance. The European Association for the Study of the Liver (EASL) and American Association for the Study of Liver Diseases currently recommend screening of high-risk individuals using abdominal ultrasound scan (USS), at 6-monthly intervals.9,10 A recent meta-analysis of 47 studies with 15,158 patients suggested only 41% of HCCs are detected by surveillance. Outcomes of patients detected on surveillance were twice as likely to have curative therapy and live longer.7 Thus far, the small number of studies looking directly at HCC surveillance adherence suggest low utilization rates for HCC surveillance in high-risk individuals.11 Extremes of age, male sex, and etiology have been implicated in negative likelihood for adherence to HCC screening.11 To date, there have been relatively few large studies assessing surveillance practice in Europe. This study aims to provide the largest cohort analysis into surveillance utilization, from a UK perspective, with the aim of providing the international community more insight into the factors affecting surveillance utilization.
MATERIALS AND METHODS Data Collection The electronic clinic letter database at Imperial College Healthcare NHS Trust was interrogated to retrospectively screen all hepatology patients seen at St Mary’s, Hammersmith and Charing Cross hospitals to identify patients diagnosed with cirrhosis before October 1, 2013. The search term (CIRRHO) was used to detect all clinic letters containing the descriptors “cirrhosis” and “cirrhotic.” Only letters from after January 2012 were inspected. Each clinic letter was examined individually to confirm a diagnosis of cirrhosis. Hospital radiology reporting software was used to identify whether these patients had a screening scan between October 1, 2013 and April 1, 2014. Scans performed within this time frame were categorized as “routine” and if not “overdue.” We attempted to identify possible explanations for patients falling into the “overdue” category from hospital
e8 | www.jcge.com J Clin Gastroenterol Volume 50, Number 1, January 2016 Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Clin Gastroenterol
Volume 50, Number 1, January 2016
Hepatocellular Carcinoma Surveillance
records and grouped these into patient-related and hospitalrelated factors. Data on liver disease etiology, Child-Pugh status, presence/absence of HCC, and patient demographics were also documented.
The median age of the cohort was 60 years (range, 20 to 91 y) and the majority of patients (67%) were male. Table 1 summarizes patient demographics in this study group.
Study Population
Outcomes of the Surveillance Programme
Inclusion into the study database was dependent on documented evidence of cirrhosis by the responsible clinician within the clinic letters. It was permissible for patients to be included without histologic confirmation if the responsible clinician believed that there was enough clinical evidence for cirrhosis based on Fibroscan results or evidence of decompensation. Patients were excluded if their medical records could not confirm or exclude cirrhosis. Duplications were also excluded as a small number of patients would have been seen at more than 1 of the 3 surveillance centers during the study period. Patients that deceased before October 1, 2013 were excluded.
Of the 898 patients eligible for inclusion, 532 (59%) had an abdominal USS for HCC surveillance within the 6 months study period. When only including investigations where the intention was explicitly for surveillance purposes alone, an additional 55 patients were found to have received surveillance MRI or CT scans (25 MRI and 30 CT). Hence, 65% of patients overall underwent HCC screening appropriate to guidelines. Twelve patients were identified within the cohort to have clinical comorbidities that precluded them from undergoing surveillance. Allowing for the fact that some decompensated patients with cirrhosis within the cohort may have been too unwell to undergo surveillance, only 444 of the 587 (76%) Child-Pugh A patients with cirrhosis had HCC surveillance as per guidelines. Of the 587 patients who received appropriate routine HCC surveillance, 467 (80%) also had a combined serum AFP measurement taken at the time of liver scan. Three hundred eleven patients were classified as overdue for HCC surveillance. Reasons for failure to complete surveillance within the recommended 6-month period are summarized in Figure 1. The majority of overdue cases (42%, n = 131) were found to be due to patient-related factors with the most common being unexplained patient nonattendance for scheduled scans (n = 119). The remaining 12 patients were not surveyed due to documented clinical reasons by the responsible clinician that the patient was not for on-going screening due to medical comorbidities. Hospital-related factors accounted for a further 40% (n = 125) of patients being overdue, with 116 due to scans not being booked by physicians and 9 due to cancellations by the radiology department. It is possible that some of the 116 scans not booked by clinicians may have due to appropriate medical reasons that were not clearly stated in the notes. In 8 (3%) cases an appropriate surveillance scan was booked but later than guideline recommendations. Three overdue patients were found to have died during the study period. Across the whole cohort we identified 638 (71%) patients who had been under HCC surveillance when the interval was increased to 1 year.
HCC Detection We reviewed information on clinic letters and investigation request forms for each patient to establish the intention of each scan to ensure that all scans were correctly attributed to screening or for other indications. Computed tomography (CT) and magnetic resonance imaging (MRI) scans that were explicitly requested for screening were therefore also included for analysis. Where cases of HCC were detected, each was noted as being either “whilst on surveillance,” as an “incidental” finding from an unrelated investigation or as “unknown.” All patients with suspected tumors were assessed by the Imperial College Healthcare Liver Cancer Multidisciplinary Team. HCC diagnosis was based on EASL criteria using a 4-phase multidetector CT scan or dynamic contrastenhanced MRI with typical hallmark of HCC (hypervascular in the arterial phase with venous washout) or histology in indeterminate cases.8 We defined radiofrequency ablation, surgical resection, and liver transplant as curative treatments in line with Barcelona Clinic Liver Cancer classifications (BCLC).12 Childs-Pugh classification and BCLC staging at the time of diagnosis were ascertained retrospectively using information gathered from clinical letters and hospital reporting software.
Statistical Analysis Univariate analysis was performed using Wald w2 tests to assess the significance of potential predictor categorical variables. We used independent samples t test to compare the mean ages between “routine” and “overdue” groups and a Mann-Whitney U test to compare distances travelled from home to hospital. Comparison between HCC treatments between those found on “surveillance” and those with “incidental” findings was made using 2-tailed Fisher exact testing. Multivariate regression analysis to investigate the influence of these variables was considered in all factors with a P < 0.1 on univariate analysis.
RESULTS Patient Characteristics A total of 1721 patients were identified using the search term to identify potential cirrhotic patients. Postexclusion, demonstrated in Table 1, we identified 898 patients eligible for HCC surveillance in the study period. Copyright
r
Predictors of Surveillance Potential predictors of surveillance were investigated from a range of sociodemographic data retrieved from patient clinic letters. We found no statistically significant relationship between age (P = 0.521) or sex (P = 0.947) and the likelihood of a patient being overdue HCC surveillance. Although there were significantly more patients with Child-Pugh B cirrhosis in the routine screening group (P = 0.0002), the overall hepatic decompensation did not reveal any significant difference (P = 0.767). The distance travelled to surveillance center (P = 0.647) and etiology of underlying liver disease [including alcoholic (P = 0.509) viral (P = 0.166), autoimmune (P = 0.624), metabolic (P = 0.150), nonalcoholic fatty liver (P = 0.736), and mixed etiology cirrhosis (P = 0.120)] were not significantly associated with rates of HCC surveillance. Multivariate analysis was not performed due to the lack of significance on univariate analysis.
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
e9
J Clin Gastroenterol
Selvapatt et al
Volume 50, Number 1, January 2016
TABLE 1. Study Population Derivation and Comparison of Patient Characteristics and Demographics Between “Routine” and “Overdue” Screening Groups Screened Excluded by study criteria Inclusion for analysis Reasons for exclusion No evidence of cirrhosis Incomplete clinical documentation Duplicated cases (between 2 hospitals) Death (before study period) Demographic Variables Age (y)* < 40 [n (%)] 40-55 [n (%)] 56-70 [n (%)] > 71 [n (%)] Male sex [n (%)] Distance to center (miles)w < 10.0 [n (%)] 10.1-50.0 [n (%)] > 51.0 [n (%)] Etiology of liver disease [n (%)] ALD Viral hepatitis Autoimmune Metabolic Non alcoholic fatty liver disease Other Mixed Hepatic compensation [n (%)] Child-Pugh A Child-Pugh B Child-Pugh C Indeterminate
1721 823 898 655 66 57 45 All Patients (n = 898)
Routine (n = 587)
Overdue (n = 311)
P
59.7 (20.6-91.6) 44 (4.9) 287 (32.0) 395 (44.0) 172 (19.1) 599 (66.7) 3.6 (0.9-232.0) 773 (86.1) 105 (11.7) 19 (2.2)
59.5 (23.5-91.6) 29 (4.9) 198 (33.7) 248 (42.3) 112 (19.1) 392 (66.7) 3.6 (0.9-232.0) 506 (86.2) 69 (11.8) 12 (2.0)
60.0 (20.6-91.1) 15 (4.8) 89 (28.6) 147 (47.3) 60 (19.3) 207 (66.6) 3.5 (0.9-140.0) 267 (85.9) 36 (11.6) 8 (2.5)
0.52
0.95 0.65
305 338 42 20 52 32 109
(34.0) (37.6) (4.7) (2.2) (5.8) (3.6) (12.1)
201 235 28 9 31 15 68
(34.2) (40.0) (4.8) (1.5) (5.3) (2.6) (11.6)
104 103 14 11 21 17 41
(33.4) (33.1) (4.5) (3.5) (6.8) (5.5) (13.2)
0.51 0.17 0.62 0.15 0.73
682 124 63 29
(75.9) (13.8) (7.0) (3.3)
444 99 40 4
(75.6) (16.9) (6.8) (0.7)
238 25 23 25
(76.5) (8.0) (7.4) (8.0)
0.81 0.0002 0.78 < 0.0001
0.12
*Expressed as mean (range). wExpressed as median (range). Values in bold are statistically significant. ALD indicates alcoholic liver disease.
HCC Detection and Outcomes
FIGURE 1. Diagrammatic representation for reasons for nonadherence to 6-monthly hepatocellular carcinoma surveillance within the study cohort.
Sixty-one (6.8%) patients were diagnosed with HCC during the study period. Of these, we noted that 38 cases had been picked up as a direct result of a surveillance scan compared with 9 cases detected incidentally. Method of detection could not be identified from the notes for the remaining 14 cases. Liver biopsy was required for histologic confirmation of HCC in 2 cases. Figure 2A demonstrates treatment outcomes between screening and incidentally picked HCCs. Despite a numerical trend toward more curative outcomes in the screening group, statistical significance was not achieved. Out of the 38 cases picked up on surveillance 21 had surgical resection, radiofrequency ablation, or transplant; 12 transarterial chemoembolization (TACE)—of which 10 were for palliation and 2 were neoadjuvant; 5 patients were treated with palliative therapy of which 2 received Sorafenib. In the incidentally diagnosed group, 2 had curative surgery or ablation, 4 had TACE with palliative intent, and 2 further patients were treated with palliative therapy of which 1 received Sorafenib. In this group 1 patient’s treatment outcome could not be accurately ascertained from the data available. Figure 2B demonstrates the BCLC staging and ChildPugh scores at the time of diagnosis. In the incidental
e10 | www.jcge.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Clin Gastroenterol
Volume 50, Number 1, January 2016
FIGURE 2. A, Hepatocellular carcinoma (HCC) outcomes comparison between those diagnosed after screening and those diagnosed incidentally outside of the screening programme. B, Comparison of BCLC and Child-Pugh scores between HCCs identified through surveillance and those identified incidentally.
diagnosed group the clinical notes were insufficient to ascertain the BCLC stage of 1 patient. No statistical difference was noted in BCLC nor Child-Pugh Scores between those diagnosed with HCC on surveillance or incidental diagnosis.
DISCUSSION The overall surveillance utilization within this cirrhotic cohort is disappointing, when considering the intention to screen was for all but 12 patients (1.3%). Although literature in this area is limited, these results seem to support the reported experiences of many countries that HCC surveillance may not be practiced as efficiently as expected. In a previous UK audit performed in 184 cirrhotic patients, 6monthly surveillance was reported in 52% of the patients.13 However, in the United Kingdom and Europe little information exists on HCC surveillance utilization. In the United States, differences in the health care system structure mean that nongastroenterologists may be responsible for HCC surveillance. In a cohort of 1873 patients identified with HCC with prior diagnosis of cirrhosis, only 29% of patients were identified to have undergone regular surveillance. Being seen by a gastroenterologist/hepatologist ensured a 4.5-fold increased chance of receiving regular surveillance and 2.8-fold if seen by a physician with academic affiliation.14 In our cohort, all patients with cirrhosis were under the care of gastroenterologists or hepatologists in university teaching hospitals, which may in part explain the superior utilization rates in comparison. In a single-center study at the University of Michigan in 2010 out of 160 cirrhotic patients routine Copyright
r
Hepatocellular Carcinoma Surveillance
surveillance was performed in 75%. However, it should be noted that the authors deemed that routine ultrasound surveillance be performed yearly as opposed to 6 monthly as used in our study.15 When using these parameters 71% of our study patients were surveyed. Indeed, a worrying population-based systematic review of HCC screening literature in 2012 suggested that fewer than 20% of patients at high risk of developing HCC undergo any form of regular surveillance.11 Where possible we obtained reasons for cases failing to adhere to surveillance guidelines. As discovered in previous studies, our analysis suggested that potential points of failure in the service appeared to fall into 2 main groups: patient-related factors and hospital-related factors.16,17 We found that nonattendance at scheduled USS appointments was the most common patient-related reason for individuals being overdue surveillance (42% of cases). This may be because patients are wilfully choosing not to attend but could also occur if patients are forgetting appointments. Comorbities such as encephalopathy, cognitive impairment, and depression, which are all associated with cirrhosis, may in part contribute to this.18 However, continued patient reinforcement and education can only further improve chances of improved patient participation that has been shown to have a positive impact on HCC surveillance adherence.15 The main system-related factor appeared to be failure to schedule USS appointments. A 3-year audit cycle looking into HCC surveillance in 22 viral hepatitis patients in Adelaide, Australia identified doctor education as 1 of the 3 key areas for improvement and demonstrated a 100% improvement in adherence to routine surveillance at completion compared with baseline.19 Improving physician awareness, especially of nonhepatologists, of international guidelines may therefore reduce the number of missed scans. Given 25% of UK gastroenterologists surveyed were unaware of HCC surveillance guidelines this demonstrates the on-going need to reinforce the guidelines among those required to be responsible for surveillance.6 Similar issues of limitations in health care provider attitudes impeding adequate HCC surveillance have also been recognized in the United States. A web-based survey of 131 primary care providers in Dallas demonstrated only half believed that HCC screening reduced mortality. Furthermore, incorrect perceptions of appropriate surveillance were demonstrated. In total, 45% believed clinical examination was effective, 59% liver enzyme testing, and 89% believed AFP alone were effective tests.20 It must be appreciated as well that unbooked scans by physicians may also be reflective of time pressures or clinical decisions that may not have been captured within the remit of this study. Our study found no association between any sociodemographic factors and the likelihood of receiving routine guideline-recommended surveillance. Whereas previous studies have demonstrated possible associations between etiologies such as viral hepatitis and increased likelihood for utilization of HCC screening no such association was noted between the various etiologies in this study.15 To date cost-effective analyses has suggested that HCC surveillance is cost-effective when pick up incidence is at least 1.5% to 6% per year.21 Within this study cohort 61 (6.8%) new HCC cases were picked up overall within the 6-month surveillance period studied. Thirty-eight out of 587 (6.5%) patients that underwent routine screening were diagnosed with HCC. At least 62% of HCCs within this study period
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
e11
J Clin Gastroenterol
Selvapatt et al
were identified as a direct result of screening which compares favourably to previously published literature.7 HCCs picked up during surveillance were associated with a 61% chance of curative therapy compared with 25% in those picked up incidentally which supports the findings elsewhere.22 Although this study did not demonstrate a statistically improved chance of curative therapy with HCCs picked up through screening it must be emphasized that this was not the primary endpoint of the study and the total number of HCCs were too small to make any categorical conclusions. To our knowledge this is the largest study to date on HCC surveillance utilization in cirrhotic patients within the United Kingdom. However, limitations of the study are related to the retrospective nature of data collection. Missing or incomplete data capture from clinical letters introduces the risk of measurement bias. Furthermore, data were not collected on comorbidities or performance status as this could not be consistently measured in a standardized way between individual clinic letters. This may explain why clinicians, who may have failed to make their intentions clear in the clinic letters, did not survey those patients. We may therefore have underrepresented the surveillance utilization from a true intention to screen perspective. This study only looked at 1 cycle of HCC screening, it is possible that consistent surveillance rates (ie, over multiple 6monthly cycles) may be lower. The study population, being inner city London teaching hospitals, may not truly reflect both national or European practice and patient demographics per se. Further collaborative work with multicenter registry studies may better answer the question of how effectively HCC screening is being performed at an international level. In terms of assessment of outcomes no randomized clinical trials have been performed evaluating whether patients with cirrhosis that undergo surveillance versus no surveillance have a superior outcome, but this study certainly supports the notion that it does. HCC surveillance in patients with cirrhosis is suboptimally adhered to despite evidence that HCC detection during surveillance is more likely to be at an early enough stage to be treated curatively. Although patient nonattendance is a significant factor, physician-related and hospital-related factors are equally important to be addressed. To date the literature has demonstrated that we pick up less than half of HCC on surveillance. This study suggests that a major reason is that the guidelines are not being adhered to and that probably more formalized processes and infrastructure are required to ensure that surveillance is performed more adequately. REFERENCES 1. El-Serag HB. Hepatocellular carcinoma. N Engl J Med. 2011; 365:1118–1127. 2. El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemiology and molecular carcinogenesis. Gastroenterology. 2007;132:2557–2576. 3. El-Serag HB. Hepatocellular carcinoma: an epidemiologic view. J Clin Gastroenterol. 2002;35(suppl 2):S72–S78.
Volume 50, Number 1, January 2016
4. Ryder SD. British Society of G. Guidelines for the diagnosis and treatment of hepatocellular carcinoma (HCC) in adults. Gut. 2003;52(suppl 3):iii1–iii8. 5. Yuen MF, Cheng CC, Lauder IJ, et al. Early detection of hepatocellular carcinoma increases the chance of treatment: Hong Kong experience. Hepatology. 2000;31:330–335. 6. Coon JT, Rogers G, Hewson P, et al. Surveillence of cirrhosis for hepatocellular carcinoma: systematic review and economic analysis. Health Technol Assess. 2007;11:34. 7. Singal AG, Pillai A, Tiro J. Early detection, curative treatment, and survival rates for hepatocellular carcinoma surveillance in patients with cirrhosis: a meta-analysis. PLoS Med. 2014;11: e1001624. 8. Pelletier SF, Thyagarajan V, Romeo-Marrero C, et al. An intention-to-treat analysis of liver transplantation for hepatocellular carcinoma using organ procurement transplant network data. Liver Transpl. 2009;15:859–868. 9. European Association For The Study Of The L, European Organisation For R, Treatment Of C. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol. 2012;56:908–943. 10. Bruix J, Sherman M. American Association for the Study of Liver D. Management of hepatocellular carcinoma: an update. Hepatology. 2011;53:1020–1022. 11. Singal AG, Yopp A, Skinner C, et al. Utilization of hepatocellular carcinoma surveillance among American patients: a systematic review. J Gen Intern Med. 2012;27:861–867. 12. Pons F, Varela M, Llovet JM. Staging systems in hepatocellular carcinoma. HPB. 2005;7:35–41. 13. Townsend SA, Ch’ng CL. PWE-283 Surveillance for hepatocellular carcinoma: a clinical audit. Gut. 2012;61(suppl 2): A413. 14. Davila JA, Morgan RO, Richardson PA, et al. Use of surveillance for hepatocellular carcinoma among patients with cirrhosis in the United States. Hepatology. 2010;52:131–141. 15. Singal AG, Volk ML, Rakoski MO, et al. Patient involvement in healthcare is associated with higher rates of surveillance for hepatocellular carcinoma. J Clin Gastroenterol. 2011;45: 727–732. 16. Singal AG, Yopp AC, Gupta S, et al. Failure rates in the hepatocellular carcinoma surveillance process. Cancer Prev Res. 2012;5:1124–1130. 17. Davila JA, Henderson L, Kramer JR, et al. Utilization of surveillance for hepatocellular carcinoma among hepatitis C virus-infected veterans in the United States. Ann Intern Med. 2011;154:85–93. 18. Bajaj JS, Wade JB, Sanyal AJ. Spectrum of neurocognitive impairment in cirrhosis: implications for the assessment of hepatic encephalopathy. Hepatology. 2009;50:2014–2021. 19. Kennedy NA, Rodgers A, Altus R, et al. Optimisation of hepatocellular carcinoma surveillance in patients with viral hepatitis: a quality improvement study. Intern Med J. 2013; 43:772–777. 20. Dalton-Fitzgerald E, Tiro J, Kandunoori P, et al. Practice patterns and attitudes of primary care providers and barriers to surveillance of hepatocellular carcinoma in patients with cirrhosis. Clin Gastroenterol Hepatol. 2014;12:1958–1963. 21. Sarasin FP, Giostra E, Hadengue A. Cost-effectiveness of screening for detection of small hepatocellular carcinoma in western patients with Child-Pugh class A cirrhosis. Am J Med. 1996;101;422–434. 22. Van Meer S, de Man RA, Siersema OD, et al. Surveillence for hepatocellular carcinoma in chronic liver disease: evidence and controversies. World J Gastroenterol. 2013;19:6744–6756.
e12 | www.jcge.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Hepatocellular Carcinoma Surveillance Are We Utilizing It? Nowlan Selvapatt, MBBS, BSc, MRCP,* Henry House, BSc,* and Ashley Brown, MBChB, MD, FRCPw
Background and Goals: International guidelines recommend 6monthly ultrasound scan surveillance in cirrhotic patients for hepatocellular carcinoma (HCC) screening. The aim of this study was to evaluate HCC surveillance and to provide the outcomes for the largest reported cirrhosis cohort in the United Kingdom. Study: We retrospectively reviewed all cirrhotic patients during the 6 months before April 1, 2014 at 3 teaching hospitals within the Imperial College Healthcare NHS Trust. All patients with cirrhosis eligible for HCC screening in the cohort were reviewed for evidence of screening and were characterized as either “routine” or “overdue.” Reasons for failure to meet guidelines were identified in overdue patients. Univariate analyses were conducted to determine clinical and sociodemographic predictors of surveillance with a view to performing multivariate analysis. Results: Of the 898 patients eligible for inclusion, 65% patients had HCC surveillance performed in the past 6 months. During the observation period, 61 (6.8%) cases of HCC were detected. Thirtyeight were picked up on surveillance, 9 incidentally and in 14 cases it could not be determined from the notes. Within this cohort HCCs diagnosed on surveillance did not demonstrate significantly superior outcomes compared with those picked up incidentally. Conclusions: Surveillance of patients with cirrhosis within this cohort is suboptimal. Although disappointing at a local level, this is likely to be reflective of practice elsewhere. Key Words: hepatocellular carcinoma, HCC, surveillance, screening, utilization
(J Clin Gastroenterol 2016;50:e8–e12)
H
epatocellular carcinoma (HCC) is the sixth most common cancer worldwide and the third commonest cause of cancer mortality.1,2 Although most of the burden of disease is in developing countries due to hepatitis B virus, it is fast developing into an increasing problem in western Countries due to rising trends of hepatitis C virus, alcoholic liver disease, and nonalcoholic steatohepatitis (NASH).1,3 In the United States approximately 20,000 new cases of HCC are reported per year and in the United Kingdom HCC is responsible for roughly 1500 deaths per year.1,4 Cirrhosis is associated with 80% to 90% of HCC.1
Received for publication October 27, 2014; accepted April 20, 2015. From the *Imperial College; and wImperial College Healthcare NHS Trust, London, UK. The authors declare that they have nothing to disclose. Reprints: Nowlan Selvapatt, MBBS, BSc, MRCP, Department of Hepatology, 10th Floor QEQM Building, St Mary’s Hospital Campus, Imperial College, London W2 1NY, UK (e-mail: [email protected]). Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Regular surveillance increases the likelihood of detecting HCC at an early stage where it is more likely to be amenable to curative treatment.5 This strategy has been demonstrated to be cost-effective.6 The average 5-year survival rate for advanced tumors remains low at approximately 12%.1 Yet 5year survival rates as high as 70% have been demonstrated in tumors picked-up at an early stage that are suitable for curative therapies.7 In a United States retrospective cohort of 4482 patients with HCC listed for liver transplantation the overall 5-year survival was 62%, however, in those with more advanced cancers 5-year survival was 38%.8 International guidelines have been introduced to guide clinicians in the practice of HCC surveillance. The European Association for the Study of the Liver (EASL) and American Association for the Study of Liver Diseases currently recommend screening of high-risk individuals using abdominal ultrasound scan (USS), at 6-monthly intervals.9,10 A recent meta-analysis of 47 studies with 15,158 patients suggested only 41% of HCCs are detected by surveillance. Outcomes of patients detected on surveillance were twice as likely to have curative therapy and live longer.7 Thus far, the small number of studies looking directly at HCC surveillance adherence suggest low utilization rates for HCC surveillance in high-risk individuals.11 Extremes of age, male sex, and etiology have been implicated in negative likelihood for adherence to HCC screening.11 To date, there have been relatively few large studies assessing surveillance practice in Europe. This study aims to provide the largest cohort analysis into surveillance utilization, from a UK perspective, with the aim of providing the international community more insight into the factors affecting surveillance utilization.
MATERIALS AND METHODS Data Collection The electronic clinic letter database at Imperial College Healthcare NHS Trust was interrogated to retrospectively screen all hepatology patients seen at St Mary’s, Hammersmith and Charing Cross hospitals to identify patients diagnosed with cirrhosis before October 1, 2013. The search term (CIRRHO) was used to detect all clinic letters containing the descriptors “cirrhosis” and “cirrhotic.” Only letters from after January 2012 were inspected. Each clinic letter was examined individually to confirm a diagnosis of cirrhosis. Hospital radiology reporting software was used to identify whether these patients had a screening scan between October 1, 2013 and April 1, 2014. Scans performed within this time frame were categorized as “routine” and if not “overdue.” We attempted to identify possible explanations for patients falling into the “overdue” category from hospital
e8 | www.jcge.com J Clin Gastroenterol Volume 50, Number 1, January 2016 Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Clin Gastroenterol
Volume 50, Number 1, January 2016
Hepatocellular Carcinoma Surveillance
records and grouped these into patient-related and hospitalrelated factors. Data on liver disease etiology, Child-Pugh status, presence/absence of HCC, and patient demographics were also documented.
The median age of the cohort was 60 years (range, 20 to 91 y) and the majority of patients (67%) were male. Table 1 summarizes patient demographics in this study group.
Study Population
Outcomes of the Surveillance Programme
Inclusion into the study database was dependent on documented evidence of cirrhosis by the responsible clinician within the clinic letters. It was permissible for patients to be included without histologic confirmation if the responsible clinician believed that there was enough clinical evidence for cirrhosis based on Fibroscan results or evidence of decompensation. Patients were excluded if their medical records could not confirm or exclude cirrhosis. Duplications were also excluded as a small number of patients would have been seen at more than 1 of the 3 surveillance centers during the study period. Patients that deceased before October 1, 2013 were excluded.
Of the 898 patients eligible for inclusion, 532 (59%) had an abdominal USS for HCC surveillance within the 6 months study period. When only including investigations where the intention was explicitly for surveillance purposes alone, an additional 55 patients were found to have received surveillance MRI or CT scans (25 MRI and 30 CT). Hence, 65% of patients overall underwent HCC screening appropriate to guidelines. Twelve patients were identified within the cohort to have clinical comorbidities that precluded them from undergoing surveillance. Allowing for the fact that some decompensated patients with cirrhosis within the cohort may have been too unwell to undergo surveillance, only 444 of the 587 (76%) Child-Pugh A patients with cirrhosis had HCC surveillance as per guidelines. Of the 587 patients who received appropriate routine HCC surveillance, 467 (80%) also had a combined serum AFP measurement taken at the time of liver scan. Three hundred eleven patients were classified as overdue for HCC surveillance. Reasons for failure to complete surveillance within the recommended 6-month period are summarized in Figure 1. The majority of overdue cases (42%, n = 131) were found to be due to patient-related factors with the most common being unexplained patient nonattendance for scheduled scans (n = 119). The remaining 12 patients were not surveyed due to documented clinical reasons by the responsible clinician that the patient was not for on-going screening due to medical comorbidities. Hospital-related factors accounted for a further 40% (n = 125) of patients being overdue, with 116 due to scans not being booked by physicians and 9 due to cancellations by the radiology department. It is possible that some of the 116 scans not booked by clinicians may have due to appropriate medical reasons that were not clearly stated in the notes. In 8 (3%) cases an appropriate surveillance scan was booked but later than guideline recommendations. Three overdue patients were found to have died during the study period. Across the whole cohort we identified 638 (71%) patients who had been under HCC surveillance when the interval was increased to 1 year.
HCC Detection We reviewed information on clinic letters and investigation request forms for each patient to establish the intention of each scan to ensure that all scans were correctly attributed to screening or for other indications. Computed tomography (CT) and magnetic resonance imaging (MRI) scans that were explicitly requested for screening were therefore also included for analysis. Where cases of HCC were detected, each was noted as being either “whilst on surveillance,” as an “incidental” finding from an unrelated investigation or as “unknown.” All patients with suspected tumors were assessed by the Imperial College Healthcare Liver Cancer Multidisciplinary Team. HCC diagnosis was based on EASL criteria using a 4-phase multidetector CT scan or dynamic contrastenhanced MRI with typical hallmark of HCC (hypervascular in the arterial phase with venous washout) or histology in indeterminate cases.8 We defined radiofrequency ablation, surgical resection, and liver transplant as curative treatments in line with Barcelona Clinic Liver Cancer classifications (BCLC).12 Childs-Pugh classification and BCLC staging at the time of diagnosis were ascertained retrospectively using information gathered from clinical letters and hospital reporting software.
Statistical Analysis Univariate analysis was performed using Wald w2 tests to assess the significance of potential predictor categorical variables. We used independent samples t test to compare the mean ages between “routine” and “overdue” groups and a Mann-Whitney U test to compare distances travelled from home to hospital. Comparison between HCC treatments between those found on “surveillance” and those with “incidental” findings was made using 2-tailed Fisher exact testing. Multivariate regression analysis to investigate the influence of these variables was considered in all factors with a P < 0.1 on univariate analysis.
RESULTS Patient Characteristics A total of 1721 patients were identified using the search term to identify potential cirrhotic patients. Postexclusion, demonstrated in Table 1, we identified 898 patients eligible for HCC surveillance in the study period. Copyright
r
Predictors of Surveillance Potential predictors of surveillance were investigated from a range of sociodemographic data retrieved from patient clinic letters. We found no statistically significant relationship between age (P = 0.521) or sex (P = 0.947) and the likelihood of a patient being overdue HCC surveillance. Although there were significantly more patients with Child-Pugh B cirrhosis in the routine screening group (P = 0.0002), the overall hepatic decompensation did not reveal any significant difference (P = 0.767). The distance travelled to surveillance center (P = 0.647) and etiology of underlying liver disease [including alcoholic (P = 0.509) viral (P = 0.166), autoimmune (P = 0.624), metabolic (P = 0.150), nonalcoholic fatty liver (P = 0.736), and mixed etiology cirrhosis (P = 0.120)] were not significantly associated with rates of HCC surveillance. Multivariate analysis was not performed due to the lack of significance on univariate analysis.
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
e9
J Clin Gastroenterol
Selvapatt et al
Volume 50, Number 1, January 2016
TABLE 1. Study Population Derivation and Comparison of Patient Characteristics and Demographics Between “Routine” and “Overdue” Screening Groups Screened Excluded by study criteria Inclusion for analysis Reasons for exclusion No evidence of cirrhosis Incomplete clinical documentation Duplicated cases (between 2 hospitals) Death (before study period) Demographic Variables Age (y)* < 40 [n (%)] 40-55 [n (%)] 56-70 [n (%)] > 71 [n (%)] Male sex [n (%)] Distance to center (miles)w < 10.0 [n (%)] 10.1-50.0 [n (%)] > 51.0 [n (%)] Etiology of liver disease [n (%)] ALD Viral hepatitis Autoimmune Metabolic Non alcoholic fatty liver disease Other Mixed Hepatic compensation [n (%)] Child-Pugh A Child-Pugh B Child-Pugh C Indeterminate
1721 823 898 655 66 57 45 All Patients (n = 898)
Routine (n = 587)
Overdue (n = 311)
P
59.7 (20.6-91.6) 44 (4.9) 287 (32.0) 395 (44.0) 172 (19.1) 599 (66.7) 3.6 (0.9-232.0) 773 (86.1) 105 (11.7) 19 (2.2)
59.5 (23.5-91.6) 29 (4.9) 198 (33.7) 248 (42.3) 112 (19.1) 392 (66.7) 3.6 (0.9-232.0) 506 (86.2) 69 (11.8) 12 (2.0)
60.0 (20.6-91.1) 15 (4.8) 89 (28.6) 147 (47.3) 60 (19.3) 207 (66.6) 3.5 (0.9-140.0) 267 (85.9) 36 (11.6) 8 (2.5)
0.52
0.95 0.65
305 338 42 20 52 32 109
(34.0) (37.6) (4.7) (2.2) (5.8) (3.6) (12.1)
201 235 28 9 31 15 68
(34.2) (40.0) (4.8) (1.5) (5.3) (2.6) (11.6)
104 103 14 11 21 17 41
(33.4) (33.1) (4.5) (3.5) (6.8) (5.5) (13.2)
0.51 0.17 0.62 0.15 0.73
682 124 63 29
(75.9) (13.8) (7.0) (3.3)
444 99 40 4
(75.6) (16.9) (6.8) (0.7)
238 25 23 25
(76.5) (8.0) (7.4) (8.0)
0.81 0.0002 0.78 < 0.0001
0.12
*Expressed as mean (range). wExpressed as median (range). Values in bold are statistically significant. ALD indicates alcoholic liver disease.
HCC Detection and Outcomes
FIGURE 1. Diagrammatic representation for reasons for nonadherence to 6-monthly hepatocellular carcinoma surveillance within the study cohort.
Sixty-one (6.8%) patients were diagnosed with HCC during the study period. Of these, we noted that 38 cases had been picked up as a direct result of a surveillance scan compared with 9 cases detected incidentally. Method of detection could not be identified from the notes for the remaining 14 cases. Liver biopsy was required for histologic confirmation of HCC in 2 cases. Figure 2A demonstrates treatment outcomes between screening and incidentally picked HCCs. Despite a numerical trend toward more curative outcomes in the screening group, statistical significance was not achieved. Out of the 38 cases picked up on surveillance 21 had surgical resection, radiofrequency ablation, or transplant; 12 transarterial chemoembolization (TACE)—of which 10 were for palliation and 2 were neoadjuvant; 5 patients were treated with palliative therapy of which 2 received Sorafenib. In the incidentally diagnosed group, 2 had curative surgery or ablation, 4 had TACE with palliative intent, and 2 further patients were treated with palliative therapy of which 1 received Sorafenib. In this group 1 patient’s treatment outcome could not be accurately ascertained from the data available. Figure 2B demonstrates the BCLC staging and ChildPugh scores at the time of diagnosis. In the incidental
e10 | www.jcge.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Clin Gastroenterol
Volume 50, Number 1, January 2016
FIGURE 2. A, Hepatocellular carcinoma (HCC) outcomes comparison between those diagnosed after screening and those diagnosed incidentally outside of the screening programme. B, Comparison of BCLC and Child-Pugh scores between HCCs identified through surveillance and those identified incidentally.
diagnosed group the clinical notes were insufficient to ascertain the BCLC stage of 1 patient. No statistical difference was noted in BCLC nor Child-Pugh Scores between those diagnosed with HCC on surveillance or incidental diagnosis.
DISCUSSION The overall surveillance utilization within this cirrhotic cohort is disappointing, when considering the intention to screen was for all but 12 patients (1.3%). Although literature in this area is limited, these results seem to support the reported experiences of many countries that HCC surveillance may not be practiced as efficiently as expected. In a previous UK audit performed in 184 cirrhotic patients, 6monthly surveillance was reported in 52% of the patients.13 However, in the United Kingdom and Europe little information exists on HCC surveillance utilization. In the United States, differences in the health care system structure mean that nongastroenterologists may be responsible for HCC surveillance. In a cohort of 1873 patients identified with HCC with prior diagnosis of cirrhosis, only 29% of patients were identified to have undergone regular surveillance. Being seen by a gastroenterologist/hepatologist ensured a 4.5-fold increased chance of receiving regular surveillance and 2.8-fold if seen by a physician with academic affiliation.14 In our cohort, all patients with cirrhosis were under the care of gastroenterologists or hepatologists in university teaching hospitals, which may in part explain the superior utilization rates in comparison. In a single-center study at the University of Michigan in 2010 out of 160 cirrhotic patients routine Copyright
r
Hepatocellular Carcinoma Surveillance
surveillance was performed in 75%. However, it should be noted that the authors deemed that routine ultrasound surveillance be performed yearly as opposed to 6 monthly as used in our study.15 When using these parameters 71% of our study patients were surveyed. Indeed, a worrying population-based systematic review of HCC screening literature in 2012 suggested that fewer than 20% of patients at high risk of developing HCC undergo any form of regular surveillance.11 Where possible we obtained reasons for cases failing to adhere to surveillance guidelines. As discovered in previous studies, our analysis suggested that potential points of failure in the service appeared to fall into 2 main groups: patient-related factors and hospital-related factors.16,17 We found that nonattendance at scheduled USS appointments was the most common patient-related reason for individuals being overdue surveillance (42% of cases). This may be because patients are wilfully choosing not to attend but could also occur if patients are forgetting appointments. Comorbities such as encephalopathy, cognitive impairment, and depression, which are all associated with cirrhosis, may in part contribute to this.18 However, continued patient reinforcement and education can only further improve chances of improved patient participation that has been shown to have a positive impact on HCC surveillance adherence.15 The main system-related factor appeared to be failure to schedule USS appointments. A 3-year audit cycle looking into HCC surveillance in 22 viral hepatitis patients in Adelaide, Australia identified doctor education as 1 of the 3 key areas for improvement and demonstrated a 100% improvement in adherence to routine surveillance at completion compared with baseline.19 Improving physician awareness, especially of nonhepatologists, of international guidelines may therefore reduce the number of missed scans. Given 25% of UK gastroenterologists surveyed were unaware of HCC surveillance guidelines this demonstrates the on-going need to reinforce the guidelines among those required to be responsible for surveillance.6 Similar issues of limitations in health care provider attitudes impeding adequate HCC surveillance have also been recognized in the United States. A web-based survey of 131 primary care providers in Dallas demonstrated only half believed that HCC screening reduced mortality. Furthermore, incorrect perceptions of appropriate surveillance were demonstrated. In total, 45% believed clinical examination was effective, 59% liver enzyme testing, and 89% believed AFP alone were effective tests.20 It must be appreciated as well that unbooked scans by physicians may also be reflective of time pressures or clinical decisions that may not have been captured within the remit of this study. Our study found no association between any sociodemographic factors and the likelihood of receiving routine guideline-recommended surveillance. Whereas previous studies have demonstrated possible associations between etiologies such as viral hepatitis and increased likelihood for utilization of HCC screening no such association was noted between the various etiologies in this study.15 To date cost-effective analyses has suggested that HCC surveillance is cost-effective when pick up incidence is at least 1.5% to 6% per year.21 Within this study cohort 61 (6.8%) new HCC cases were picked up overall within the 6-month surveillance period studied. Thirty-eight out of 587 (6.5%) patients that underwent routine screening were diagnosed with HCC. At least 62% of HCCs within this study period
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcge.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
e11
J Clin Gastroenterol
Selvapatt et al
were identified as a direct result of screening which compares favourably to previously published literature.7 HCCs picked up during surveillance were associated with a 61% chance of curative therapy compared with 25% in those picked up incidentally which supports the findings elsewhere.22 Although this study did not demonstrate a statistically improved chance of curative therapy with HCCs picked up through screening it must be emphasized that this was not the primary endpoint of the study and the total number of HCCs were too small to make any categorical conclusions. To our knowledge this is the largest study to date on HCC surveillance utilization in cirrhotic patients within the United Kingdom. However, limitations of the study are related to the retrospective nature of data collection. Missing or incomplete data capture from clinical letters introduces the risk of measurement bias. Furthermore, data were not collected on comorbidities or performance status as this could not be consistently measured in a standardized way between individual clinic letters. This may explain why clinicians, who may have failed to make their intentions clear in the clinic letters, did not survey those patients. We may therefore have underrepresented the surveillance utilization from a true intention to screen perspective. This study only looked at 1 cycle of HCC screening, it is possible that consistent surveillance rates (ie, over multiple 6monthly cycles) may be lower. The study population, being inner city London teaching hospitals, may not truly reflect both national or European practice and patient demographics per se. Further collaborative work with multicenter registry studies may better answer the question of how effectively HCC screening is being performed at an international level. In terms of assessment of outcomes no randomized clinical trials have been performed evaluating whether patients with cirrhosis that undergo surveillance versus no surveillance have a superior outcome, but this study certainly supports the notion that it does. HCC surveillance in patients with cirrhosis is suboptimally adhered to despite evidence that HCC detection during surveillance is more likely to be at an early enough stage to be treated curatively. Although patient nonattendance is a significant factor, physician-related and hospital-related factors are equally important to be addressed. To date the literature has demonstrated that we pick up less than half of HCC on surveillance. This study suggests that a major reason is that the guidelines are not being adhered to and that probably more formalized processes and infrastructure are required to ensure that surveillance is performed more adequately. REFERENCES 1. El-Serag HB. Hepatocellular carcinoma. N Engl J Med. 2011; 365:1118–1127. 2. El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemiology and molecular carcinogenesis. Gastroenterology. 2007;132:2557–2576. 3. El-Serag HB. Hepatocellular carcinoma: an epidemiologic view. J Clin Gastroenterol. 2002;35(suppl 2):S72–S78.
Volume 50, Number 1, January 2016
4. Ryder SD. British Society of G. Guidelines for the diagnosis and treatment of hepatocellular carcinoma (HCC) in adults. Gut. 2003;52(suppl 3):iii1–iii8. 5. Yuen MF, Cheng CC, Lauder IJ, et al. Early detection of hepatocellular carcinoma increases the chance of treatment: Hong Kong experience. Hepatology. 2000;31:330–335. 6. Coon JT, Rogers G, Hewson P, et al. Surveillence of cirrhosis for hepatocellular carcinoma: systematic review and economic analysis. Health Technol Assess. 2007;11:34. 7. Singal AG, Pillai A, Tiro J. Early detection, curative treatment, and survival rates for hepatocellular carcinoma surveillance in patients with cirrhosis: a meta-analysis. PLoS Med. 2014;11: e1001624. 8. Pelletier SF, Thyagarajan V, Romeo-Marrero C, et al. An intention-to-treat analysis of liver transplantation for hepatocellular carcinoma using organ procurement transplant network data. Liver Transpl. 2009;15:859–868. 9. European Association For The Study Of The L, European Organisation For R, Treatment Of C. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol. 2012;56:908–943. 10. Bruix J, Sherman M. American Association for the Study of Liver D. Management of hepatocellular carcinoma: an update. Hepatology. 2011;53:1020–1022. 11. Singal AG, Yopp A, Skinner C, et al. Utilization of hepatocellular carcinoma surveillance among American patients: a systematic review. J Gen Intern Med. 2012;27:861–867. 12. Pons F, Varela M, Llovet JM. Staging systems in hepatocellular carcinoma. HPB. 2005;7:35–41. 13. Townsend SA, Ch’ng CL. PWE-283 Surveillance for hepatocellular carcinoma: a clinical audit. Gut. 2012;61(suppl 2): A413. 14. Davila JA, Morgan RO, Richardson PA, et al. Use of surveillance for hepatocellular carcinoma among patients with cirrhosis in the United States. Hepatology. 2010;52:131–141. 15. Singal AG, Volk ML, Rakoski MO, et al. Patient involvement in healthcare is associated with higher rates of surveillance for hepatocellular carcinoma. J Clin Gastroenterol. 2011;45: 727–732. 16. Singal AG, Yopp AC, Gupta S, et al. Failure rates in the hepatocellular carcinoma surveillance process. Cancer Prev Res. 2012;5:1124–1130. 17. Davila JA, Henderson L, Kramer JR, et al. Utilization of surveillance for hepatocellular carcinoma among hepatitis C virus-infected veterans in the United States. Ann Intern Med. 2011;154:85–93. 18. Bajaj JS, Wade JB, Sanyal AJ. Spectrum of neurocognitive impairment in cirrhosis: implications for the assessment of hepatic encephalopathy. Hepatology. 2009;50:2014–2021. 19. Kennedy NA, Rodgers A, Altus R, et al. Optimisation of hepatocellular carcinoma surveillance in patients with viral hepatitis: a quality improvement study. Intern Med J. 2013; 43:772–777. 20. Dalton-Fitzgerald E, Tiro J, Kandunoori P, et al. Practice patterns and attitudes of primary care providers and barriers to surveillance of hepatocellular carcinoma in patients with cirrhosis. Clin Gastroenterol Hepatol. 2014;12:1958–1963. 21. Sarasin FP, Giostra E, Hadengue A. Cost-effectiveness of screening for detection of small hepatocellular carcinoma in western patients with Child-Pugh class A cirrhosis. Am J Med. 1996;101;422–434. 22. Van Meer S, de Man RA, Siersema OD, et al. Surveillence for hepatocellular carcinoma in chronic liver disease: evidence and controversies. World J Gastroenterol. 2013;19:6744–6756.
e12 | www.jcge.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
