bs_bs_banner
doi:10.1111/jgh.12438
H E PAT O L O G Y
Evaluation of alpha-fetoprotein in detecting hepatocellular carcinoma recurrence after radiofrequency ablation Surachate Siripongsakun,*,† Sindy H Wei,* Stephanie Lin,* Jeffrey Chen,* Steven S Raman,* James Sayre,*,‡ Myron J Tong§,¶ and David S Lu*,§ *Department of Radiology, ‡School of Public Health, §Pfleger Liver Institute and the Division of Digestive Diseases, David Geffen School of Medicine at UCLA, Los Angeles, and ¶The Liver Center, Huntington Medical Research Institutes, Pasadena, California, USA; and †Department of Radiology, Chalubhorn Hospital, Bangkok, Thailand
Key words AFP producing, hepatocellular carcinoma, radiofrequency ablation, recurrence. Accepted for publication 11 September 2013. Correspondence David S Lu, Department of Radiology, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095, USA. Email: [email protected] Conflict of interest: Nothing to report. Financial support: Nothing to report. Author contributions: Surachate Siripongsakun1,2,3,4,5, Sindy H. Wei2,4, Stephanie Lin2, Jeffrey M. Chen2, Steven Raman1,4, James Sayre3, Myron J. Tong1,3,5, David S. Lu1,3,4,5 Note: 1.Study concept and design 2.Acquisition of data 3.Analysis and interpretation of data 4.Drafting the manuscript 5.Critical revision of the manuscript.
Abstract Background and Aim: The performance of alpha-fetoprotein (AFP) in the detection of hepatocellular carcinoma (HCC) recurrence after radiofrequency ablation was analyzed. Methods: One hundred and forty-six solitary HCC lesions treated by radiofrequency ablation were evaluated. Using the AFP cutoff level at ≥ 20 ng/mL, tumors were categorized into AFP or non-AFP-producing HCC. Factors associated with true and false interpretations for cancer recurrence including analysis of elevated alanine aminotransferase (ALT) were evaluated. The performance of AFP using different cutoff levels adjusted for abnormal ALT was compared. Results: Of 146 HCCs, 103 demonstrated no HCC recurrence while 43 had local recurrence. In non-recurrence HCC cases, increased AFP levels (false positive) were associated with concomitant ALT elevations, while those with normal AFP (true negative) had correspondingly normal ALT values (P < 0.001). The AFP false positive rate in cases of elevated ALT was significantly higher than those with normal ALT levels (31.9% vs 5.4%, P = 0.001). Among all positive AFP tests, those with false positive values (non-recurrence) had a significantly lower AFP level than the true positive (recurrence) HCC cases (39.8 ng/mL vs 372 ng/mL, P < 0.001). At the 20 ng/mL cutoff level, the sensitivities of AFP for detecting recurrence in non-AFPproducing HCC and AFP-producing HCC were 12.0%, and 72.2%, respectively. Using a modified AFP criteria of ≥ 100 ng/mL for cases where ALT ≥ 40 U/L, the sensitivity and specificity in AFP-producing tumors increased from 72.2% and 56% to 100% and 85%, respectively. Conclusion: Serum AFP is a useful test in the detection of HCC recurrence in AFPproducing HCC. The performance in AFP-producing HCC was significantly improved after adjusting for elevation of serum ALT.
Introduction Alpha-fetoprotein (AFP), a 70 kD glycoprotein with a half-life of 5–7 days, has been implicated in the regulation of fatty acids in both fetal and proliferating adult liver cells.1 Historically, serum AFP level has been a valuable tool in the clinical management of hepatocellular carcinoma (HCC). As a tumor marker, AFP has been used as a diagnostic test, a surrogate marker for predicting tumor response, and for the detection of HCC recurrence.2–7 Despite its role in clinical practice, the value of AFP for the diagnosis and detection of HCC recurrence remains controversial.8 Since AFP lacks adequate sensitivity and specificity, the current American Association for the Study of Liver Disease (AASLD)
guidelines currently recommend that surveillance of HCC in highrisk patients should be based only on ultrasound examinations at 6-month intervals.8 Serum AFP has been excluded from the current HCC diagnostic criteria, which are now solely based on radiological and histological features.8 A rising serum AFP is not specific for HCC but may also be found in benign conditions commonly encountered in clinical practice, such as liver inflammation and cirrhosis.9–13 In a large AFP analytic study from the National Veterans’ Affair Clinical Case registry which involved 76 357 hepatitis C infected patients, a strong positive correlation was found between alanine aminotransferase (ALT) and AFP in both HCC and non-HCC patients.13 As a result, an increasing level of ALT is a major confounding
Journal of Gastroenterology and Hepatology 29 (2014) 157–164 © 2013 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
157
Evaluation of alpha-fetoprotein
S Siripongsakun et al.
RFA therapies session for HCC (n = 530) Excluded (n = 384) Mulple lesions at presentaon (n = 225) New lesion occurred within 3 months of RFA (n = 72) New lesions present at me of recurrence (n = 31) No AFP at diagnosis or follow up (n = 30) No CT/MRI to confirm recurrence (n = 26)
RFA therapies included in the study (n = 146)
No local recurrence (n = 103)
Local recurrence (n = 43)
AFP cut off at ≥ 20 ng/mL for posive
True negave* (n = 85)
False posive* (n = 18)
AFP cut off at ≥ 20 ng/mL for posive
False negave* (n = 27)
True posive* (n = 16)
Figure 1 Flow diagram of patients. *All of the four groups were subjected to analysis on Table 2. AFP, Alpha-fetoprotein; CT, computerized tomography; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; RFA, radiofrequency ablation.
factor which influences the diagnostic performance of AFP. As the majority of HCC arise in a background of liver cirrhosis or chronic viral hepatitis, a better understanding of factors that can cause elevation of serum AFP is necessary to avoid a false interpretation. Despite suboptimal accuracy in detecting HCC recurrence posttreatment, serum AFP is still being utilized for this purpose because it is a simple, inexpensive, and widely available test. A few studies reported that AFP may be helpful for detection of HCC recurrence in patients with high pretreatment serum AFP levels.3,4 Particularly, in patients with a high pretreatment serum AFP that normalized after treatment, the subsequent elevation of AFP may suggest tumor recurrence or progression.8 Therefore, the purpose of this study is to evaluate the diagnostic performance of serum AFP in detecting HCC recurrence after radiofrequency ablation (RFA), both in patients with high pretreatment AFP (AFPproducing HCC) levels and in patients with normal pretreatment AFP (non-AFP-producing HCC) levels. In addition, the false positive and true positive AFP results were analyzed to determine feasibility of improving the diagnostic performance of AFP after adjusting for significant confounding factors.
Methods Institutional Review Board approval was obtained for this retrospective study and the requirements for informed consent were 158
waived. The study was performed in compliance with the Health Insurance Portability and Accountability Act, United States 1996. Patient and ablation session selection. From a database of patients with HCC who underwent RFA from January 1999 to September 2012, we selected those with solitary HCC, who had available follow-up by contrast-enhanced computerized tomography (CT) or magnetic resonance imaging (MRI) at our institution (See below “imaging follow-up and end point determination”), and who had available and adequate AFP measurements (See below “AFP follow-up and abnormal cutoff”). The flow diagram of patients is shown in Figure 1. Diagnosis of HCC and HCC recurrence after tumor ablation. Diagnosis of HCC was made either by using the AASLD imaging criteria guidelines, or by pathological confirmation of HCC on biopsy or surgical resection specimens. A typical diagnostic feature on dynamic CT/MRI included arterial phase hyperenhancement followed by hypoenhancement on the portal venous phase. HCC recurrence on imaging was defined as new nodular enhancement around the ablation site at more than 1 month postablation with demonstration of arterial hyperenhancement and venous hypoenhancement, or as interval growth on subsequent follow-up.
Journal of Gastroenterology and Hepatology 29 (2014) 157–164 © 2013 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
S Siripongsakun et al.
Evaluation of alpha-fetoprotein
In cases with rising AFP but no imaging detection, the patients received either a short-term imaging follow-up within 1–2 month or a complimentary contrast study (e.g. if contrast-enhanced CT did not detect recurrence, then the patient was further evaluated with contrast-enhanced liver MRI).
20 ng/mL and were designated as false negative if there was tumor recurrence, and true negative if there was no tumor recurrence on imaging. Clinical parameters including tumor size, liver function test, and level of AFP among four groups were analyzed and compared.
Imaging follow-up and end-point determination. At our institution, the protocol for post-RFA follow-up includes CT/MRI within 1 month after the initial treatment, then at 3 months, then every 3 months up to 1 year, and at least every 3–6 months thereafter. The end points were considered in two circumstances. First, if tumor recurrence occurred, the date of imaging that first detected the recurrence was considered the date of recurrence. If no tumor recurrence occurred but a new suspicious HCC lesion emerged elsewhere in the liver, the last date of imaging anteceding the appearance of new lesions was used as the last imaging follow-up and considered as no recurrence. This would eliminate the question as to whether the AFP rise was due to the recurrence of HCC at the RFA site or from a newly appeared HCC. The follow-up time after tumor ablation was calculated from date of ablation to the date of recurrence, or to the last follow-up imaging date.
Subgroup analysis on normal and abnormal ALT cases. Abnormal ALT is a known factor associated with increasing AFP levels which may result in false AFP interpretation. Therefore, the level of AFP and false interpretation rate between normal ALT (< 40 U/L) and abnormal ALT (≥ 40 U/L) groups were compared. The underlying liver disease status in patients with normal or abnormal serum ALT levels was assessed in viral-related liver diseases and non-viral-related liver diseases, respectively. The viral-related liver disease included infection with hepatitis C virus (HCV), hepatitis B virus (HBV), or co-infection by both viruses. Active viral activity of HCV was considered if there was either detectable HCV RNA or grade 2 or higher HCV inflammation on corresponding histopathology.14,15 Active HBV activity was considered if there was presence of HBV DNA ≥ 104 copies/mL.16 In cases with co-infection by HCV and HBV, either active HCV or HBV was considered active viral activity. In non-viralrelated liver diseases, the Child-Pugh classification was used to assess liver status between abnormal ALT and normal ALT cases. The non-viral-related liver diseases in this study included alcoholic liver disease, non-alcoholic steatohepatitis, primary biliary cirrhosis, autoimmune hepatitis, and glycogen storage disease.
MRI and CT protocol. Over the span of the study period, scanner technology has improved considerably with an increased number of detector rows on CT scanners as well as with increased magnetic field strengths of MR scanners. However, the basic elements of CT and MR technology remained similar. For CT scans, only dual- or triple-phase contrast-enhanced protocols on multidetector scanners (4, 16, and 64 detector) were considered adequate. For MRI, the minimum requirements included T1 weighted dual-echo (in-phase and opposed-phase) and fat saturated gradient recalled echo, T2 weighted single-shot or multishot sequences, and dynamic contrast-enhanced multiphasic T1 weighted 2D or 3D acquisitions. Scans were mainly performed on 1.5 T scanners but a subset was performed on 3 T scanners with phased array body coils. AFP follow-up and abnormal cutoff. For inclusion in this study, AFP measurement follow-up was considered adequate if the AFP measurements were performed at the time of HCC diagnosis, at the initial RFA treatment, at 1–3 months post-RFA, and at the time of recurrence or last follow-up. At the time of diagnosis, we divided the HCCs into two subgroups which consisted of non-AFP-producing HCC if the patient’s initial serum AFP was < 20 ng/mL, and AFP-producing HCC if the patient’s initial AFP level was ≥ 20 ng/mL. The AFP value cutoff considered positive for HCC recurrence was ≥ 20 ng/mL at the time when tumor recurrence was detected by contrast-enhanced CT or MRI. An AFP < 20 ng/mL was considered negative for tumor recurrence. Analysis of AFP by test results. At the end point, AFP tests were considered positive if AFP ≥ 20 ng/mL and were designated as false positive if there was no tumor recurrence, and true positive if there was tumor recurrence on imaging. Alternatively, AFP tests were considered negative if AFP
doi:10.1111/jgh.12438
H E PAT O L O G Y
Evaluation of alpha-fetoprotein in detecting hepatocellular carcinoma recurrence after radiofrequency ablation Surachate Siripongsakun,*,† Sindy H Wei,* Stephanie Lin,* Jeffrey Chen,* Steven S Raman,* James Sayre,*,‡ Myron J Tong§,¶ and David S Lu*,§ *Department of Radiology, ‡School of Public Health, §Pfleger Liver Institute and the Division of Digestive Diseases, David Geffen School of Medicine at UCLA, Los Angeles, and ¶The Liver Center, Huntington Medical Research Institutes, Pasadena, California, USA; and †Department of Radiology, Chalubhorn Hospital, Bangkok, Thailand
Key words AFP producing, hepatocellular carcinoma, radiofrequency ablation, recurrence. Accepted for publication 11 September 2013. Correspondence David S Lu, Department of Radiology, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095, USA. Email: [email protected] Conflict of interest: Nothing to report. Financial support: Nothing to report. Author contributions: Surachate Siripongsakun1,2,3,4,5, Sindy H. Wei2,4, Stephanie Lin2, Jeffrey M. Chen2, Steven Raman1,4, James Sayre3, Myron J. Tong1,3,5, David S. Lu1,3,4,5 Note: 1.Study concept and design 2.Acquisition of data 3.Analysis and interpretation of data 4.Drafting the manuscript 5.Critical revision of the manuscript.
Abstract Background and Aim: The performance of alpha-fetoprotein (AFP) in the detection of hepatocellular carcinoma (HCC) recurrence after radiofrequency ablation was analyzed. Methods: One hundred and forty-six solitary HCC lesions treated by radiofrequency ablation were evaluated. Using the AFP cutoff level at ≥ 20 ng/mL, tumors were categorized into AFP or non-AFP-producing HCC. Factors associated with true and false interpretations for cancer recurrence including analysis of elevated alanine aminotransferase (ALT) were evaluated. The performance of AFP using different cutoff levels adjusted for abnormal ALT was compared. Results: Of 146 HCCs, 103 demonstrated no HCC recurrence while 43 had local recurrence. In non-recurrence HCC cases, increased AFP levels (false positive) were associated with concomitant ALT elevations, while those with normal AFP (true negative) had correspondingly normal ALT values (P < 0.001). The AFP false positive rate in cases of elevated ALT was significantly higher than those with normal ALT levels (31.9% vs 5.4%, P = 0.001). Among all positive AFP tests, those with false positive values (non-recurrence) had a significantly lower AFP level than the true positive (recurrence) HCC cases (39.8 ng/mL vs 372 ng/mL, P < 0.001). At the 20 ng/mL cutoff level, the sensitivities of AFP for detecting recurrence in non-AFPproducing HCC and AFP-producing HCC were 12.0%, and 72.2%, respectively. Using a modified AFP criteria of ≥ 100 ng/mL for cases where ALT ≥ 40 U/L, the sensitivity and specificity in AFP-producing tumors increased from 72.2% and 56% to 100% and 85%, respectively. Conclusion: Serum AFP is a useful test in the detection of HCC recurrence in AFPproducing HCC. The performance in AFP-producing HCC was significantly improved after adjusting for elevation of serum ALT.
Introduction Alpha-fetoprotein (AFP), a 70 kD glycoprotein with a half-life of 5–7 days, has been implicated in the regulation of fatty acids in both fetal and proliferating adult liver cells.1 Historically, serum AFP level has been a valuable tool in the clinical management of hepatocellular carcinoma (HCC). As a tumor marker, AFP has been used as a diagnostic test, a surrogate marker for predicting tumor response, and for the detection of HCC recurrence.2–7 Despite its role in clinical practice, the value of AFP for the diagnosis and detection of HCC recurrence remains controversial.8 Since AFP lacks adequate sensitivity and specificity, the current American Association for the Study of Liver Disease (AASLD)
guidelines currently recommend that surveillance of HCC in highrisk patients should be based only on ultrasound examinations at 6-month intervals.8 Serum AFP has been excluded from the current HCC diagnostic criteria, which are now solely based on radiological and histological features.8 A rising serum AFP is not specific for HCC but may also be found in benign conditions commonly encountered in clinical practice, such as liver inflammation and cirrhosis.9–13 In a large AFP analytic study from the National Veterans’ Affair Clinical Case registry which involved 76 357 hepatitis C infected patients, a strong positive correlation was found between alanine aminotransferase (ALT) and AFP in both HCC and non-HCC patients.13 As a result, an increasing level of ALT is a major confounding
Journal of Gastroenterology and Hepatology 29 (2014) 157–164 © 2013 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
157
Evaluation of alpha-fetoprotein
S Siripongsakun et al.
RFA therapies session for HCC (n = 530) Excluded (n = 384) Mulple lesions at presentaon (n = 225) New lesion occurred within 3 months of RFA (n = 72) New lesions present at me of recurrence (n = 31) No AFP at diagnosis or follow up (n = 30) No CT/MRI to confirm recurrence (n = 26)
RFA therapies included in the study (n = 146)
No local recurrence (n = 103)
Local recurrence (n = 43)
AFP cut off at ≥ 20 ng/mL for posive
True negave* (n = 85)
False posive* (n = 18)
AFP cut off at ≥ 20 ng/mL for posive
False negave* (n = 27)
True posive* (n = 16)
Figure 1 Flow diagram of patients. *All of the four groups were subjected to analysis on Table 2. AFP, Alpha-fetoprotein; CT, computerized tomography; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; RFA, radiofrequency ablation.
factor which influences the diagnostic performance of AFP. As the majority of HCC arise in a background of liver cirrhosis or chronic viral hepatitis, a better understanding of factors that can cause elevation of serum AFP is necessary to avoid a false interpretation. Despite suboptimal accuracy in detecting HCC recurrence posttreatment, serum AFP is still being utilized for this purpose because it is a simple, inexpensive, and widely available test. A few studies reported that AFP may be helpful for detection of HCC recurrence in patients with high pretreatment serum AFP levels.3,4 Particularly, in patients with a high pretreatment serum AFP that normalized after treatment, the subsequent elevation of AFP may suggest tumor recurrence or progression.8 Therefore, the purpose of this study is to evaluate the diagnostic performance of serum AFP in detecting HCC recurrence after radiofrequency ablation (RFA), both in patients with high pretreatment AFP (AFPproducing HCC) levels and in patients with normal pretreatment AFP (non-AFP-producing HCC) levels. In addition, the false positive and true positive AFP results were analyzed to determine feasibility of improving the diagnostic performance of AFP after adjusting for significant confounding factors.
Methods Institutional Review Board approval was obtained for this retrospective study and the requirements for informed consent were 158
waived. The study was performed in compliance with the Health Insurance Portability and Accountability Act, United States 1996. Patient and ablation session selection. From a database of patients with HCC who underwent RFA from January 1999 to September 2012, we selected those with solitary HCC, who had available follow-up by contrast-enhanced computerized tomography (CT) or magnetic resonance imaging (MRI) at our institution (See below “imaging follow-up and end point determination”), and who had available and adequate AFP measurements (See below “AFP follow-up and abnormal cutoff”). The flow diagram of patients is shown in Figure 1. Diagnosis of HCC and HCC recurrence after tumor ablation. Diagnosis of HCC was made either by using the AASLD imaging criteria guidelines, or by pathological confirmation of HCC on biopsy or surgical resection specimens. A typical diagnostic feature on dynamic CT/MRI included arterial phase hyperenhancement followed by hypoenhancement on the portal venous phase. HCC recurrence on imaging was defined as new nodular enhancement around the ablation site at more than 1 month postablation with demonstration of arterial hyperenhancement and venous hypoenhancement, or as interval growth on subsequent follow-up.
Journal of Gastroenterology and Hepatology 29 (2014) 157–164 © 2013 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
S Siripongsakun et al.
Evaluation of alpha-fetoprotein
In cases with rising AFP but no imaging detection, the patients received either a short-term imaging follow-up within 1–2 month or a complimentary contrast study (e.g. if contrast-enhanced CT did not detect recurrence, then the patient was further evaluated with contrast-enhanced liver MRI).
20 ng/mL and were designated as false negative if there was tumor recurrence, and true negative if there was no tumor recurrence on imaging. Clinical parameters including tumor size, liver function test, and level of AFP among four groups were analyzed and compared.
Imaging follow-up and end-point determination. At our institution, the protocol for post-RFA follow-up includes CT/MRI within 1 month after the initial treatment, then at 3 months, then every 3 months up to 1 year, and at least every 3–6 months thereafter. The end points were considered in two circumstances. First, if tumor recurrence occurred, the date of imaging that first detected the recurrence was considered the date of recurrence. If no tumor recurrence occurred but a new suspicious HCC lesion emerged elsewhere in the liver, the last date of imaging anteceding the appearance of new lesions was used as the last imaging follow-up and considered as no recurrence. This would eliminate the question as to whether the AFP rise was due to the recurrence of HCC at the RFA site or from a newly appeared HCC. The follow-up time after tumor ablation was calculated from date of ablation to the date of recurrence, or to the last follow-up imaging date.
Subgroup analysis on normal and abnormal ALT cases. Abnormal ALT is a known factor associated with increasing AFP levels which may result in false AFP interpretation. Therefore, the level of AFP and false interpretation rate between normal ALT (< 40 U/L) and abnormal ALT (≥ 40 U/L) groups were compared. The underlying liver disease status in patients with normal or abnormal serum ALT levels was assessed in viral-related liver diseases and non-viral-related liver diseases, respectively. The viral-related liver disease included infection with hepatitis C virus (HCV), hepatitis B virus (HBV), or co-infection by both viruses. Active viral activity of HCV was considered if there was either detectable HCV RNA or grade 2 or higher HCV inflammation on corresponding histopathology.14,15 Active HBV activity was considered if there was presence of HBV DNA ≥ 104 copies/mL.16 In cases with co-infection by HCV and HBV, either active HCV or HBV was considered active viral activity. In non-viralrelated liver diseases, the Child-Pugh classification was used to assess liver status between abnormal ALT and normal ALT cases. The non-viral-related liver diseases in this study included alcoholic liver disease, non-alcoholic steatohepatitis, primary biliary cirrhosis, autoimmune hepatitis, and glycogen storage disease.
MRI and CT protocol. Over the span of the study period, scanner technology has improved considerably with an increased number of detector rows on CT scanners as well as with increased magnetic field strengths of MR scanners. However, the basic elements of CT and MR technology remained similar. For CT scans, only dual- or triple-phase contrast-enhanced protocols on multidetector scanners (4, 16, and 64 detector) were considered adequate. For MRI, the minimum requirements included T1 weighted dual-echo (in-phase and opposed-phase) and fat saturated gradient recalled echo, T2 weighted single-shot or multishot sequences, and dynamic contrast-enhanced multiphasic T1 weighted 2D or 3D acquisitions. Scans were mainly performed on 1.5 T scanners but a subset was performed on 3 T scanners with phased array body coils. AFP follow-up and abnormal cutoff. For inclusion in this study, AFP measurement follow-up was considered adequate if the AFP measurements were performed at the time of HCC diagnosis, at the initial RFA treatment, at 1–3 months post-RFA, and at the time of recurrence or last follow-up. At the time of diagnosis, we divided the HCCs into two subgroups which consisted of non-AFP-producing HCC if the patient’s initial serum AFP was < 20 ng/mL, and AFP-producing HCC if the patient’s initial AFP level was ≥ 20 ng/mL. The AFP value cutoff considered positive for HCC recurrence was ≥ 20 ng/mL at the time when tumor recurrence was detected by contrast-enhanced CT or MRI. An AFP < 20 ng/mL was considered negative for tumor recurrence. Analysis of AFP by test results. At the end point, AFP tests were considered positive if AFP ≥ 20 ng/mL and were designated as false positive if there was no tumor recurrence, and true positive if there was tumor recurrence on imaging. Alternatively, AFP tests were considered negative if AFP
