Original Clinical ScienceçLiver
Follow-up Imaging After Liver Transplantation Should Take Into Consideration Primary Hepatocellular Carcinoma Characteristics Marco Dioguardi Burgio,1,7 Maxime Ronot,1,2,3 David Fuks,2,4 Federica Dondero,4 François Cauchy,4 Sébastien Gaujoux,2,4 Safi Dokmak,4 Valérie Paradis,2,3,5 François Durand,2,6 Jacques Belghiti,2,4 and Valérie Vilgrain1,2,3
Background. Tumor recurrence occurs in almost 10% after liver transplantation (LT) for early stage hepatocellular carcinoma (HCC). Evidence-based recommendation regarding these techniques and their schedule remains unclear. The aim was to analyze the imaging characteristics of HCC recurrence after LT according to the initial primary tumor characteristics to provide relevant evidence regarding their role. Methods. Among 336 cirrhotic patients undergoing LT between 2000 and 2011, 25 (7.4%) experienced HCC recurrence. Their pre-LT characteristics, long-term outcome and imaging characteristics of recurrences were retrospectively analyzed. Results. All patients were men and median age at the time of recurrence was 54 (41-64) years. Before LT, 13 (52%) patients had HCC outside the Milan criteria. Median time to HCC recurrence was 13.8 months (1-75) after LT and 8 patients (32%) experienced recurrence longer than 24 months after LT. The most frequently involved organs were the lungs (n = 13, 52%) and the bones (n = 9, 36%). Recurrent HCC involved more than 1 organ in 11 patients (44%). Recurrences were limited to the liver in 1 patient (4%), exclusively extrahepatic in 18 patients (72%), and were both intra and extrahepatic in 6 patients (24%). Delay of recurrence was shorter in the presence of microvascular invasion (17 vs 28 months), but this was not significant (P = 0.208). Poorly differentiated tumors showed a trend toward decreased recurrence-free survival (8 vs 28 months, P = 0.075). Conclusions. Both pattern and delay of HCC recurrence support performing regular whole-body imaging initially every 6 to 12 months and continuing beyond 2 years after LT for HCC, especially in patients with pejorative pathological features of the primary tumor that could influence HCC recurrence pattern.
(Transplantation 2015;99: 1613–1618)
L
iver transplantation (LT), which removes both the tumor and the underlying preneoplastic liver, is the best curative option in patients with early stage hepatocellular carcinoma (HCC) with 70% overall survival (OS) at 5 years.1 However, recurrence of HCC after LT has been reported to occur in up to 40% of patients2 and is associated with poor survival.3 Factors influencing recurrence include tumor features including size and number of tumors (Milan criteria [MC]), but also vascular invasion and differentiation
grade, as well as previous locoregional therapy, biomarkers, such as α-fetoprotein, and type of immunosuppressive regimen.4-11 For instance, patients within the MC have been clearly shown to have a better outcome and lower recurrence rates after LT than those outside the criteria.1 Because surgical resection of recurrent HCC is the only curative therapeutic option,3,12,13 early detection of recurrence would probably improve the prognosis in these patients. However, even though both repeated liver ultrasound (US) examinations,
Received 22 July 2014. Revision requested 24 October 2014.
design, in the writing of the paper, in the performance of the research, and in data analysis. D.F. participated in research design in the writing of the paper and in data analysis. F.D. participated in research design and in the writing of the paper. F.C. participated in the writing of the paper and contributed with new analytic tools. S.G. participated in the writing of the paper and contributed with new analytic tools. S.D. participated in the writing of the paper and contributed with new reagents or analytic tools. V.P. contributed for manuscript drafting or manuscript revision for important intellectual content. F.D. contributed for manuscript drafting or manuscript revision for important intellectual content. J.B. participated in research design, in the writing of the paper, and manuscript revision for important intellectual content. V.V. participated in research design, in the writing of the paper, in the performance of the research and contributed for manuscript revision for important intellectual content.
Accepted 15 December 2014. 1 Department of Radiology, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 2
University Paris Diderot, Sorbonne Paris Cité, Paris, France.
3
INSERM U1149, CRI, Paris, France. Department of Surgery, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France.
4 Department of Surgery, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 5 Department of Pathology, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 6 Department of Hepatology, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 7
Department of Radiology, University of Palermo, Palermo, Italy.
Correspondence: Maxime Ronot, MD, PhD, Department of Radiology, Beaujon Hospital, 100 bd du Général Leclerc, 92110 Clichy, France. (maxime.ronot@bjn. aphp.fr)
The authors declare no funding or conflicts of interest.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
M.D.B. participated in research design, in the writing of the paper, in the performance of the research, and in data analysis. M.R. participated in research
ISSN: 0041-1337/15/9908-1613
Transplantation
■
August 2015
■
Volume 99
■
Number 8
DOI: 10.1097/TP.0000000000000659
www.transplantjournal.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1613
1614
■
August 2015
■
Volume 99
■
Number 8
www.transplantjournal.com
and cross-sectional imaging are used during follow-up to detect tumor recurrence, there is currently no clear consensus regarding their use and schedule.14 In this setting, imaging patterns of recurrent HCC15,16 or the delay of occurrence after LT17-19 could be highly important to determine the follow-up strategy. Thus, the aim of this study was to analyze the imaging characteristics of HCC recurrence after LT to provide relevant evidence regarding both value and schedule of current imaging techniques, with a focus on the influence of primary tumor characteristics.
enhancement on the arterial phase and a washout on the portal or delayed phase acquisitions.
Transplantation
Confirmation and Management of Recurrence
Recurrence was confirmed from follow-up images showing tumor growth and/or increased serum AFP levels. Percutaneous guided biopsy was performed when the diagnosis of recurrence was uncertain to obtain histological confirmation. Details on the management of recurrent HCC (i.e., surgery, transarterial chemoembolization, percutaneous ablation, or systemic chemotherapy) were also collected. Statistical Analysis
METHODS Patient Selection and Baseline Characteristics
Between 2000 and 2011, 343 cirrhotic patients underwent LT for HCC at Beaujon University Hospital and follow-up was available for 336 (98%) patients. Of these, 25 patients (7.4%) with pathologically confirmed HCC experienced tumor recurrence and constituted the study population. Their characteristics were retrieved from a prospective database and included: (i) baseline characteristics, such as demographic data, etiology of the underlying liver disease, serum α-fetoprotein (AFP) levels, and (ii) type of pre-LT therapeutic management including details of locoregional bridging therapies, and (iii) pathological characteristics of the primary tumors (see below).
Values are expressed as medians (range) or percentages, as appropriate. The Fisher exact test was used to compare differences in discrete or categorical variables, and the Wilcoxon rank-sum test was used for continuous variables. Characteristics of recurrent HCC tumors were compared according to the pathological features of the primary tumor. A backward variable procedure with P less than 0.05 was used to identify independent predictive factors. Overall survival was defined as the time from surgery to death (all causes). Overall survival was estimated by the Kaplan-Meier method, and the log-rank test was used to compare survival curves. All statistical tests were 2-sided. Statistical significance was defined as P less than 0.05. All statistical analyses were performed using SPSS for Windows version 18.0 (SPSS Inc, Chicago, IL).
Primary Tumor and Recurrence Characteristics
The following features of the primary tumor were recorded: (1) number, (2) size (largest axial diameter in mm), (3) tumor differentiation, (4) microvascular invasion, and (5) satellite nodules. Characteristics of recurrent HCC were analyzed at diagnosis on post-LT examinations as follows: (1) site of recurrence defined as hepatic, extrahepatic, or both, (2) size of the largest lesions (measured in mm on axial sections at each site), and (3) the time to recurrence at each site, defined as the time between LT and when recurrence was first detected. Extrahepatic recurrence was defined as follows: lung, bone, lymph node, parietal nodule, and peritoneum. The number of lesions was noted in each site. In case of multiple lesions, the following upper limits were applied: n = 100 for the lung, n = 50 for the liver, n = 20 for lymph nodes and n = 10 for both bone and peritoneum. Initial Imaging Workup and Follow-up
Baseline pretransplantation imaging included body multidetector computed tomography and liver magnetic resonance imaging (MRI). After LT, follow-up included US every 3 months for the first year then every 6 months thereafter. Cross-sectional imaging exams (computed tomography [CT] and/or MRI) were performed when recurrence was suspected on follow-up US, in case of elevated serum AFP levels, or in the event of clinical signs, such as weight loss or fatigue. Pre- and post-LT CT scan and MR images were reviewed in consensus by 2 observers (MDB and MR with 4 and 10 years of experience in liver imaging, respectively). The delay between the last CT scan and/or MRI before LT, and between LT and tumor recurrence were recorded in months in each patient. Recurrence was considered to be delayed if it occurred more than 24 months after LT. Typical imaging features of HCC was defined as the association of a tumoral
RESULTS Patient Characteristics
Our study was performed with approval of the Institutional review board. The preoperative characteristics of patients with recurrent HCC are summarized in Table 1. All patients were men, and the median age was 54 years (4164). Thirteen patients (52%) were outside the MC. Before LT, 21 patients (84%) received local bridging therapy while on the waiting list with no significant difference between patient within MC (MC+) and those beyond (MC−) (83% vs 85%, respectively, P = 0.65). The median time from the last imaging test before LT to LT was 35 days (6-181) and was similar between MC+ and MC− patients (46 vs 28 days, respectively, P = 0.37). None of the patients had evidence of macroscopic vascular tumoral invasion in the last imaging test before LT. On final pathological examination, 5 patients (20%) had poorly differentiated HCC, 8 patients (32%) had associated satellite nodules, 14 patients (56%) displayed microvascular invasion, and 2 patients (8%) had complete tumor necrosis. Detection of Recurrences
Postoperative routine US examination allowed identifying recurrence in 7 patients (28%), based on the following anomalies: detection of liver lesions (n = 3), parietal nodule (n = 3), or lymphadenopathies in the hepatoduodenal ligament (n =1). Recurrences were then confirmed by CT scan or MR imaging. In the other 18 patients (72%), recurrences were identified using CT/MRI for various reasons. Five patients (20%) had abnormal blood tests, that is, increased AFP serum level (n = 2), abnormal liver tests (n = 2), or elevated inflammatory
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Burgio et al
© 2015 Wolters Kluwer
TABLE 1.
Characteristics of the 25 patients with recurrent HCC after liver transplantation Patients
Mean age (range) 54 (41-64) Sex (Male/female), n (%) 25:0 (100:0) Cirrhosis etiology, n (%) Hepatitis B Virus 7 (28%) Hepatitis C Virus 6 (24%) Chronic alcohol consumption 9 (36%) Other 3 (12%) Median delay on the waiting list in days (range) 182 (1-683) Tumors Median number of nodule (range) 3 (1-8) Median size of the largest nodule in mm (range) 35 (20-88) Median sum of the size of all nodules in mm (range) 65 (23-165) Bilobar tumor, n (%) 9 (36%) Patients with Locoregional therapy prior to the LT, n (%) 21 (84%) No. TACE 16 No. Percutaneous radiofrequency ablation 10 No. Percutaneous ethanol injection 3 No. Resection 6 LT within/beyond Milan criteria (n + %) 12/13 (48%/52%) Median serum AFP level (before LT in ng/mL (range) 13 (2-3572) Pathological analysis Poorly differentiated HCC, n (%) 5 (20%) Presence of associated satellite nodules 8 (32%) Presence of microvascular invasion 14 (56%) Complete tumor necrosis 2 (8%)
markers (n = 1). Two (8%) patients underwent crosssectional imaging for complications of LT: biliary stenosis (n = 1) and focal liver necrosis (n = 1). One (4%) patient was followed-up for a nonspecific liver nodule present on the graft. In the remaining 6 patients (24%), imaging was performed for clinical changes: cough (n = 1), dyspnea (n = 1), lumbar pain (n = 1), septic shock (n = 1), parietal tumefaction (n = 1), and incisional hernia (n = 1). Data regarding the indication for cross-sectional imaging that led to the diagnosis of recurrence were not available in 4 patients.
1615
4 (16%) occurred between 2 and 3 years, and 4 (16%) beyond 3 years after LT. Delay of recurrence was shorter in patients with microvascular invasion on the explanted liver (median, 17 vs 28 months), but this did not reach significance level (P = 0.208). It did not differ according to the presence of satellite nodules (20 vs 25, P = 0.602). Even though the existence of poorly differentiated component was associated with decreased recurrence-free survival, the difference did not reach significance level (8 vs 28 months, P = 0.075). However, all patients with poorly differentiated lesion recurred within the first 2 years. Characteristics of HCC recurrence according to the initial features of the primary tumor are detailed in Table 3. Recurrence was limited to the liver in 1 patient (4%), was extrahepatic in 18 patients (72%), and was both in 6 patients (24%). Recurrence involved only 1 site in 14 patients (56%), and 2 or more sites in the remaining 11 patients (44 %). Patients with pretransplant HCC larger than 3 cm mostly recurred in multiple sites (67% vs 31% for pretransplant HCC smaller than 3 cm, P = 0.11). The organs involved were the lungs (Figure 1) (n = 13, 52%), bones (Figure 2) (n = 9, 36%), liver (Figure 3), (n = 7, 28%), abdominal wall (Figure 4) (n = 4, 16%), peritoneum (n = 4, 16%), lymph node (n = 3, 12%), and spleen (n = 1, 4%). The number of involved recurrence sites was not significantly associated with presence of microvascular invasion (P = 0.514), tumor differentiation (P = 0.855), and presence of satellite nodules (P = 0.316). Management of Recurrence and Survival
Management of HCC recurrence is detailed in Table 4. The 1, 3-, and 5-year OSs of patients experiencing HCC recurrence after LT were 84%, 39%, and 28%, respectively, and were significantly better in patients with initial MC+ compared to those with initial MC− patients (91%, 64%, 51% vs. 69%, 17%, 8%, respectively, P = 0.02; Figure 5) and in patients with delayed recurrence compared to patients with early recurrences (100%, 85%, 57% vs. 76%, 14%, 0%, respectively, P = 0.006). The 1, 3-, and 5-year OSs after recurrence were 50%, 9%, and 0%, respectively. The 2 patients who underwent resection of parietal tumor survived 29 and 45 months after the diagnosis of recurrence.
Characteristics of Recurrent HCC
Details on the characteristics of recurrent HCC are summarized in Table 2. Median time from LT to HCC recurrence was 13.8 months (1-75), 17 (68%) occurred within the first 2 years, and 8 (32%) were delayed. Among the latter,
DISCUSSION The present study, which aimed at analyzing the imaging characteristics of recurrent HCC after LT, shows that recurrence may be influenced by primary tumor characteristics,
TABLE 2.
Patterns of HCC recurrence after liver transplantation Location
Liver, N = 7 (28%) Lung, N = 13 (52%) Bone,a N = 9 (36%) Abdominal wall, N = 4 (16%) Lymph node, N = 3 (12%) Peritoneum, N = 4 (16%) Spleen, N = 1 (25%) a
No. lesions median (range)
Largest tumor size (mm), median (range)
Exclusive location
Bilateral lesions (%)
Delay (mo) median (range)
4 (1-50) 6 (1-100) 2 (1-10) 1 (1-1) 10 (5-20) 1 (1-10) 5
27 (5-36) 10 (5-30) 37 (25-100) 45 (38-50) 40 (5-60) 20 (18-40) 23
1/7 (14%) 5/13 (38%) 3/9 (33%) 2/4 (50%) 1/3 (33%) 2/4 (50%) 0/1 (0%)
4/7 (57%) 11/13 (84%) — — — — —
13.8 (5-28) 15 (5-46) 9.5 (1-28) 42 (9-73) 11 (3-22) 40 (15-75) 9
The spine was the most frequent location (n = 5) of bone recurrence.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
August 2015
■
Volume 99
■
Number 8
www.transplantjournal.com
8 (57%) 6 (55%) 3 (75%) 7 (58%) 3 (60%) 1 (25%) 7 (50%) 7 (64%) 5 (62%) 9 (60%) 0 (0%) 11 (69%) 3 (33%)
6 (43%) 5 (45%) 1 (25%) 5 (42%) 2 (40%) 3 (75%) 7 (50%) 4 (36%) 3 (38%) 6 (40%) 2 (100%) 5 (31%) 6 (67%)
2 (14%) 5 (45%) 2 (50%) 2 (17%) 2 (40%) 1 (25%) 5 (36%) 2 (18%) 1 (12%) 5 (33%) 1 (50%) 3 (19%) 4 (44%)
8 (57%) 5 (45%) 1 (25%) 5 (41%) 3 (60%) 4 (100%) 5 (36%) 8 (73%) 3 (37%) 9 (60%) 1 (50%) 8 (50%) 5 (56%)
3 (21%) 6 (55%) 1 (25%) 4 (33%) 3 (60%) 1 (25%) 6 (43%) 3 (27%) 3 (37%) 5 (33%) 1 (50%) 3 (19%) 6 (67%)
3 (21%) 1 (9%) 1 (25%) 2 (17%) 0 (0%) 1 (25%) 2 (14%) 2 (18%) 1 (12%) 3 (20%) 0 (0%) 4 (25%) 0 (0%)
3 (21%) 1 (9%) 0 (0%) 4 (33%) 0 (0%) 0 (0%) 2 (14%) 2 (18%) 2 (25%) 2 (17%) 0 (0%) 4 (25%) 0 (0%)
1 (7%) 2 (18%) 0 (0%) 2 (17%) 0 (0%) 1 (25%) 1 (7%) 2 (18%) 1 (12%) 1 (7%) 1 (0%) 0 (0%) 3 (33%)
12 (86%) 6 (55%) 2 (50%) 10 (83%) 3 (60%) 3 (75%) 9 (64%) 9 (82%) 7 (87%) 10 (67%) 1 (50%) 13 (81%) 5 (56%)
10 (71%) 7 (64%) 2 (50%) 8 (67%) 5 (100%) 2 (50%) 11 (79%) 6 (55%) 8 (100%) 7 (47%) 2 (50%) 12 (75%) 5 (56%)
414 420 824 518 288 523 322 684 322 560 166 466 288
■
FIGURE 1. Bilateral lung recurrence 28 months after LT in a 60-yearold man. Coronal chest CT with maximum intensity projection reconstruction shows multiple bilateral lung nodules (arrows). Increased AFP (1218 mg/dL) confirmed the diagnosis of HCC recurrence after LT.
with earlier and more diffuse recurrence in patients with initial pejorative pathological features. Recurrences were mainly extrahepatic, involved several organs at diagnosis and may be observed after a long time interval after LT. Up to now, it had remained unclear whether prognostic factors of the primary tumor, such as histologic features influenced the type of recurrence. Despite the lack of significant difference in imaging patterns regarding these factors, our study supports that HCC recurrence tends to develop earlier, and more frequently with disseminated pattern in patients with more aggressive tumors.20 This is also suggested by the lower survival rates of patients recurring while initially transplanted beyond the MC. Hence, these results suggest that the protocol for follow-up after LT could be tailored according to the characteristics of the primary tumor, to favor closer follow-up for those with factors of poor prognosis. Our study has shown that the lungs, the liver, and the bones are involved in recurrent HCC in most patients. Moreover, recurrent HCC was frequently observed in several organs at diagnosis. This supports previous data by Davis et al who reported that in 218 patients with recurrent HCC after LT the liver, lungs, and bones were involved in 49%, 50%, and 26% of patients, respectively, whereas recurrence was located in a single site in only 54% of the patients.21 Indeed, the mechanism of recurrence can be explained by the
Size of the largest HCC
Satellite nodules
Tumor differentiation
≤3, N = 14 >3, N = 11 Well, N = 4 Moderate, N = 12 Poor, N = 5 Unknown, N = 4 Yes, N = 14 No, N = 11 Yes, N = 8 No, N = 15 Unknown, N = 2 ≤3 cm, N = 16 >3 cm, N c 9 Number of nodules
Baseline pathological features
Recurrence delay No. recurrence site
Site of recurrence TABLE 3.
HCC Recurrence pattern according to the initial pathological features of the primary tumor
Extrahepatic Only, Single, N = 14 Multiple, N = 9 Liver, N = 7 Lung,N = 13 Bone, N = 9 Wall, N = 4 Peritoneum, N = 4 Node, N = 4 N = 18
Follow-up Imaging After Liver Transplantation Should Take Into Consideration Primary Hepatocellular Carcinoma Characteristics Marco Dioguardi Burgio,1,7 Maxime Ronot,1,2,3 David Fuks,2,4 Federica Dondero,4 François Cauchy,4 Sébastien Gaujoux,2,4 Safi Dokmak,4 Valérie Paradis,2,3,5 François Durand,2,6 Jacques Belghiti,2,4 and Valérie Vilgrain1,2,3
Background. Tumor recurrence occurs in almost 10% after liver transplantation (LT) for early stage hepatocellular carcinoma (HCC). Evidence-based recommendation regarding these techniques and their schedule remains unclear. The aim was to analyze the imaging characteristics of HCC recurrence after LT according to the initial primary tumor characteristics to provide relevant evidence regarding their role. Methods. Among 336 cirrhotic patients undergoing LT between 2000 and 2011, 25 (7.4%) experienced HCC recurrence. Their pre-LT characteristics, long-term outcome and imaging characteristics of recurrences were retrospectively analyzed. Results. All patients were men and median age at the time of recurrence was 54 (41-64) years. Before LT, 13 (52%) patients had HCC outside the Milan criteria. Median time to HCC recurrence was 13.8 months (1-75) after LT and 8 patients (32%) experienced recurrence longer than 24 months after LT. The most frequently involved organs were the lungs (n = 13, 52%) and the bones (n = 9, 36%). Recurrent HCC involved more than 1 organ in 11 patients (44%). Recurrences were limited to the liver in 1 patient (4%), exclusively extrahepatic in 18 patients (72%), and were both intra and extrahepatic in 6 patients (24%). Delay of recurrence was shorter in the presence of microvascular invasion (17 vs 28 months), but this was not significant (P = 0.208). Poorly differentiated tumors showed a trend toward decreased recurrence-free survival (8 vs 28 months, P = 0.075). Conclusions. Both pattern and delay of HCC recurrence support performing regular whole-body imaging initially every 6 to 12 months and continuing beyond 2 years after LT for HCC, especially in patients with pejorative pathological features of the primary tumor that could influence HCC recurrence pattern.
(Transplantation 2015;99: 1613–1618)
L
iver transplantation (LT), which removes both the tumor and the underlying preneoplastic liver, is the best curative option in patients with early stage hepatocellular carcinoma (HCC) with 70% overall survival (OS) at 5 years.1 However, recurrence of HCC after LT has been reported to occur in up to 40% of patients2 and is associated with poor survival.3 Factors influencing recurrence include tumor features including size and number of tumors (Milan criteria [MC]), but also vascular invasion and differentiation
grade, as well as previous locoregional therapy, biomarkers, such as α-fetoprotein, and type of immunosuppressive regimen.4-11 For instance, patients within the MC have been clearly shown to have a better outcome and lower recurrence rates after LT than those outside the criteria.1 Because surgical resection of recurrent HCC is the only curative therapeutic option,3,12,13 early detection of recurrence would probably improve the prognosis in these patients. However, even though both repeated liver ultrasound (US) examinations,
Received 22 July 2014. Revision requested 24 October 2014.
design, in the writing of the paper, in the performance of the research, and in data analysis. D.F. participated in research design in the writing of the paper and in data analysis. F.D. participated in research design and in the writing of the paper. F.C. participated in the writing of the paper and contributed with new analytic tools. S.G. participated in the writing of the paper and contributed with new analytic tools. S.D. participated in the writing of the paper and contributed with new reagents or analytic tools. V.P. contributed for manuscript drafting or manuscript revision for important intellectual content. F.D. contributed for manuscript drafting or manuscript revision for important intellectual content. J.B. participated in research design, in the writing of the paper, and manuscript revision for important intellectual content. V.V. participated in research design, in the writing of the paper, in the performance of the research and contributed for manuscript revision for important intellectual content.
Accepted 15 December 2014. 1 Department of Radiology, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 2
University Paris Diderot, Sorbonne Paris Cité, Paris, France.
3
INSERM U1149, CRI, Paris, France. Department of Surgery, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France.
4 Department of Surgery, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 5 Department of Pathology, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 6 Department of Hepatology, APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France. 7
Department of Radiology, University of Palermo, Palermo, Italy.
Correspondence: Maxime Ronot, MD, PhD, Department of Radiology, Beaujon Hospital, 100 bd du Général Leclerc, 92110 Clichy, France. (maxime.ronot@bjn. aphp.fr)
The authors declare no funding or conflicts of interest.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
M.D.B. participated in research design, in the writing of the paper, in the performance of the research, and in data analysis. M.R. participated in research
ISSN: 0041-1337/15/9908-1613
Transplantation
■
August 2015
■
Volume 99
■
Number 8
DOI: 10.1097/TP.0000000000000659
www.transplantjournal.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
1613
1614
■
August 2015
■
Volume 99
■
Number 8
www.transplantjournal.com
and cross-sectional imaging are used during follow-up to detect tumor recurrence, there is currently no clear consensus regarding their use and schedule.14 In this setting, imaging patterns of recurrent HCC15,16 or the delay of occurrence after LT17-19 could be highly important to determine the follow-up strategy. Thus, the aim of this study was to analyze the imaging characteristics of HCC recurrence after LT to provide relevant evidence regarding both value and schedule of current imaging techniques, with a focus on the influence of primary tumor characteristics.
enhancement on the arterial phase and a washout on the portal or delayed phase acquisitions.
Transplantation
Confirmation and Management of Recurrence
Recurrence was confirmed from follow-up images showing tumor growth and/or increased serum AFP levels. Percutaneous guided biopsy was performed when the diagnosis of recurrence was uncertain to obtain histological confirmation. Details on the management of recurrent HCC (i.e., surgery, transarterial chemoembolization, percutaneous ablation, or systemic chemotherapy) were also collected. Statistical Analysis
METHODS Patient Selection and Baseline Characteristics
Between 2000 and 2011, 343 cirrhotic patients underwent LT for HCC at Beaujon University Hospital and follow-up was available for 336 (98%) patients. Of these, 25 patients (7.4%) with pathologically confirmed HCC experienced tumor recurrence and constituted the study population. Their characteristics were retrieved from a prospective database and included: (i) baseline characteristics, such as demographic data, etiology of the underlying liver disease, serum α-fetoprotein (AFP) levels, and (ii) type of pre-LT therapeutic management including details of locoregional bridging therapies, and (iii) pathological characteristics of the primary tumors (see below).
Values are expressed as medians (range) or percentages, as appropriate. The Fisher exact test was used to compare differences in discrete or categorical variables, and the Wilcoxon rank-sum test was used for continuous variables. Characteristics of recurrent HCC tumors were compared according to the pathological features of the primary tumor. A backward variable procedure with P less than 0.05 was used to identify independent predictive factors. Overall survival was defined as the time from surgery to death (all causes). Overall survival was estimated by the Kaplan-Meier method, and the log-rank test was used to compare survival curves. All statistical tests were 2-sided. Statistical significance was defined as P less than 0.05. All statistical analyses were performed using SPSS for Windows version 18.0 (SPSS Inc, Chicago, IL).
Primary Tumor and Recurrence Characteristics
The following features of the primary tumor were recorded: (1) number, (2) size (largest axial diameter in mm), (3) tumor differentiation, (4) microvascular invasion, and (5) satellite nodules. Characteristics of recurrent HCC were analyzed at diagnosis on post-LT examinations as follows: (1) site of recurrence defined as hepatic, extrahepatic, or both, (2) size of the largest lesions (measured in mm on axial sections at each site), and (3) the time to recurrence at each site, defined as the time between LT and when recurrence was first detected. Extrahepatic recurrence was defined as follows: lung, bone, lymph node, parietal nodule, and peritoneum. The number of lesions was noted in each site. In case of multiple lesions, the following upper limits were applied: n = 100 for the lung, n = 50 for the liver, n = 20 for lymph nodes and n = 10 for both bone and peritoneum. Initial Imaging Workup and Follow-up
Baseline pretransplantation imaging included body multidetector computed tomography and liver magnetic resonance imaging (MRI). After LT, follow-up included US every 3 months for the first year then every 6 months thereafter. Cross-sectional imaging exams (computed tomography [CT] and/or MRI) were performed when recurrence was suspected on follow-up US, in case of elevated serum AFP levels, or in the event of clinical signs, such as weight loss or fatigue. Pre- and post-LT CT scan and MR images were reviewed in consensus by 2 observers (MDB and MR with 4 and 10 years of experience in liver imaging, respectively). The delay between the last CT scan and/or MRI before LT, and between LT and tumor recurrence were recorded in months in each patient. Recurrence was considered to be delayed if it occurred more than 24 months after LT. Typical imaging features of HCC was defined as the association of a tumoral
RESULTS Patient Characteristics
Our study was performed with approval of the Institutional review board. The preoperative characteristics of patients with recurrent HCC are summarized in Table 1. All patients were men, and the median age was 54 years (4164). Thirteen patients (52%) were outside the MC. Before LT, 21 patients (84%) received local bridging therapy while on the waiting list with no significant difference between patient within MC (MC+) and those beyond (MC−) (83% vs 85%, respectively, P = 0.65). The median time from the last imaging test before LT to LT was 35 days (6-181) and was similar between MC+ and MC− patients (46 vs 28 days, respectively, P = 0.37). None of the patients had evidence of macroscopic vascular tumoral invasion in the last imaging test before LT. On final pathological examination, 5 patients (20%) had poorly differentiated HCC, 8 patients (32%) had associated satellite nodules, 14 patients (56%) displayed microvascular invasion, and 2 patients (8%) had complete tumor necrosis. Detection of Recurrences
Postoperative routine US examination allowed identifying recurrence in 7 patients (28%), based on the following anomalies: detection of liver lesions (n = 3), parietal nodule (n = 3), or lymphadenopathies in the hepatoduodenal ligament (n =1). Recurrences were then confirmed by CT scan or MR imaging. In the other 18 patients (72%), recurrences were identified using CT/MRI for various reasons. Five patients (20%) had abnormal blood tests, that is, increased AFP serum level (n = 2), abnormal liver tests (n = 2), or elevated inflammatory
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Burgio et al
© 2015 Wolters Kluwer
TABLE 1.
Characteristics of the 25 patients with recurrent HCC after liver transplantation Patients
Mean age (range) 54 (41-64) Sex (Male/female), n (%) 25:0 (100:0) Cirrhosis etiology, n (%) Hepatitis B Virus 7 (28%) Hepatitis C Virus 6 (24%) Chronic alcohol consumption 9 (36%) Other 3 (12%) Median delay on the waiting list in days (range) 182 (1-683) Tumors Median number of nodule (range) 3 (1-8) Median size of the largest nodule in mm (range) 35 (20-88) Median sum of the size of all nodules in mm (range) 65 (23-165) Bilobar tumor, n (%) 9 (36%) Patients with Locoregional therapy prior to the LT, n (%) 21 (84%) No. TACE 16 No. Percutaneous radiofrequency ablation 10 No. Percutaneous ethanol injection 3 No. Resection 6 LT within/beyond Milan criteria (n + %) 12/13 (48%/52%) Median serum AFP level (before LT in ng/mL (range) 13 (2-3572) Pathological analysis Poorly differentiated HCC, n (%) 5 (20%) Presence of associated satellite nodules 8 (32%) Presence of microvascular invasion 14 (56%) Complete tumor necrosis 2 (8%)
markers (n = 1). Two (8%) patients underwent crosssectional imaging for complications of LT: biliary stenosis (n = 1) and focal liver necrosis (n = 1). One (4%) patient was followed-up for a nonspecific liver nodule present on the graft. In the remaining 6 patients (24%), imaging was performed for clinical changes: cough (n = 1), dyspnea (n = 1), lumbar pain (n = 1), septic shock (n = 1), parietal tumefaction (n = 1), and incisional hernia (n = 1). Data regarding the indication for cross-sectional imaging that led to the diagnosis of recurrence were not available in 4 patients.
1615
4 (16%) occurred between 2 and 3 years, and 4 (16%) beyond 3 years after LT. Delay of recurrence was shorter in patients with microvascular invasion on the explanted liver (median, 17 vs 28 months), but this did not reach significance level (P = 0.208). It did not differ according to the presence of satellite nodules (20 vs 25, P = 0.602). Even though the existence of poorly differentiated component was associated with decreased recurrence-free survival, the difference did not reach significance level (8 vs 28 months, P = 0.075). However, all patients with poorly differentiated lesion recurred within the first 2 years. Characteristics of HCC recurrence according to the initial features of the primary tumor are detailed in Table 3. Recurrence was limited to the liver in 1 patient (4%), was extrahepatic in 18 patients (72%), and was both in 6 patients (24%). Recurrence involved only 1 site in 14 patients (56%), and 2 or more sites in the remaining 11 patients (44 %). Patients with pretransplant HCC larger than 3 cm mostly recurred in multiple sites (67% vs 31% for pretransplant HCC smaller than 3 cm, P = 0.11). The organs involved were the lungs (Figure 1) (n = 13, 52%), bones (Figure 2) (n = 9, 36%), liver (Figure 3), (n = 7, 28%), abdominal wall (Figure 4) (n = 4, 16%), peritoneum (n = 4, 16%), lymph node (n = 3, 12%), and spleen (n = 1, 4%). The number of involved recurrence sites was not significantly associated with presence of microvascular invasion (P = 0.514), tumor differentiation (P = 0.855), and presence of satellite nodules (P = 0.316). Management of Recurrence and Survival
Management of HCC recurrence is detailed in Table 4. The 1, 3-, and 5-year OSs of patients experiencing HCC recurrence after LT were 84%, 39%, and 28%, respectively, and were significantly better in patients with initial MC+ compared to those with initial MC− patients (91%, 64%, 51% vs. 69%, 17%, 8%, respectively, P = 0.02; Figure 5) and in patients with delayed recurrence compared to patients with early recurrences (100%, 85%, 57% vs. 76%, 14%, 0%, respectively, P = 0.006). The 1, 3-, and 5-year OSs after recurrence were 50%, 9%, and 0%, respectively. The 2 patients who underwent resection of parietal tumor survived 29 and 45 months after the diagnosis of recurrence.
Characteristics of Recurrent HCC
Details on the characteristics of recurrent HCC are summarized in Table 2. Median time from LT to HCC recurrence was 13.8 months (1-75), 17 (68%) occurred within the first 2 years, and 8 (32%) were delayed. Among the latter,
DISCUSSION The present study, which aimed at analyzing the imaging characteristics of recurrent HCC after LT, shows that recurrence may be influenced by primary tumor characteristics,
TABLE 2.
Patterns of HCC recurrence after liver transplantation Location
Liver, N = 7 (28%) Lung, N = 13 (52%) Bone,a N = 9 (36%) Abdominal wall, N = 4 (16%) Lymph node, N = 3 (12%) Peritoneum, N = 4 (16%) Spleen, N = 1 (25%) a
No. lesions median (range)
Largest tumor size (mm), median (range)
Exclusive location
Bilateral lesions (%)
Delay (mo) median (range)
4 (1-50) 6 (1-100) 2 (1-10) 1 (1-1) 10 (5-20) 1 (1-10) 5
27 (5-36) 10 (5-30) 37 (25-100) 45 (38-50) 40 (5-60) 20 (18-40) 23
1/7 (14%) 5/13 (38%) 3/9 (33%) 2/4 (50%) 1/3 (33%) 2/4 (50%) 0/1 (0%)
4/7 (57%) 11/13 (84%) — — — — —
13.8 (5-28) 15 (5-46) 9.5 (1-28) 42 (9-73) 11 (3-22) 40 (15-75) 9
The spine was the most frequent location (n = 5) of bone recurrence.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
August 2015
■
Volume 99
■
Number 8
www.transplantjournal.com
8 (57%) 6 (55%) 3 (75%) 7 (58%) 3 (60%) 1 (25%) 7 (50%) 7 (64%) 5 (62%) 9 (60%) 0 (0%) 11 (69%) 3 (33%)
6 (43%) 5 (45%) 1 (25%) 5 (42%) 2 (40%) 3 (75%) 7 (50%) 4 (36%) 3 (38%) 6 (40%) 2 (100%) 5 (31%) 6 (67%)
2 (14%) 5 (45%) 2 (50%) 2 (17%) 2 (40%) 1 (25%) 5 (36%) 2 (18%) 1 (12%) 5 (33%) 1 (50%) 3 (19%) 4 (44%)
8 (57%) 5 (45%) 1 (25%) 5 (41%) 3 (60%) 4 (100%) 5 (36%) 8 (73%) 3 (37%) 9 (60%) 1 (50%) 8 (50%) 5 (56%)
3 (21%) 6 (55%) 1 (25%) 4 (33%) 3 (60%) 1 (25%) 6 (43%) 3 (27%) 3 (37%) 5 (33%) 1 (50%) 3 (19%) 6 (67%)
3 (21%) 1 (9%) 1 (25%) 2 (17%) 0 (0%) 1 (25%) 2 (14%) 2 (18%) 1 (12%) 3 (20%) 0 (0%) 4 (25%) 0 (0%)
3 (21%) 1 (9%) 0 (0%) 4 (33%) 0 (0%) 0 (0%) 2 (14%) 2 (18%) 2 (25%) 2 (17%) 0 (0%) 4 (25%) 0 (0%)
1 (7%) 2 (18%) 0 (0%) 2 (17%) 0 (0%) 1 (25%) 1 (7%) 2 (18%) 1 (12%) 1 (7%) 1 (0%) 0 (0%) 3 (33%)
12 (86%) 6 (55%) 2 (50%) 10 (83%) 3 (60%) 3 (75%) 9 (64%) 9 (82%) 7 (87%) 10 (67%) 1 (50%) 13 (81%) 5 (56%)
10 (71%) 7 (64%) 2 (50%) 8 (67%) 5 (100%) 2 (50%) 11 (79%) 6 (55%) 8 (100%) 7 (47%) 2 (50%) 12 (75%) 5 (56%)
414 420 824 518 288 523 322 684 322 560 166 466 288
■
FIGURE 1. Bilateral lung recurrence 28 months after LT in a 60-yearold man. Coronal chest CT with maximum intensity projection reconstruction shows multiple bilateral lung nodules (arrows). Increased AFP (1218 mg/dL) confirmed the diagnosis of HCC recurrence after LT.
with earlier and more diffuse recurrence in patients with initial pejorative pathological features. Recurrences were mainly extrahepatic, involved several organs at diagnosis and may be observed after a long time interval after LT. Up to now, it had remained unclear whether prognostic factors of the primary tumor, such as histologic features influenced the type of recurrence. Despite the lack of significant difference in imaging patterns regarding these factors, our study supports that HCC recurrence tends to develop earlier, and more frequently with disseminated pattern in patients with more aggressive tumors.20 This is also suggested by the lower survival rates of patients recurring while initially transplanted beyond the MC. Hence, these results suggest that the protocol for follow-up after LT could be tailored according to the characteristics of the primary tumor, to favor closer follow-up for those with factors of poor prognosis. Our study has shown that the lungs, the liver, and the bones are involved in recurrent HCC in most patients. Moreover, recurrent HCC was frequently observed in several organs at diagnosis. This supports previous data by Davis et al who reported that in 218 patients with recurrent HCC after LT the liver, lungs, and bones were involved in 49%, 50%, and 26% of patients, respectively, whereas recurrence was located in a single site in only 54% of the patients.21 Indeed, the mechanism of recurrence can be explained by the
Size of the largest HCC
Satellite nodules
Tumor differentiation
≤3, N = 14 >3, N = 11 Well, N = 4 Moderate, N = 12 Poor, N = 5 Unknown, N = 4 Yes, N = 14 No, N = 11 Yes, N = 8 No, N = 15 Unknown, N = 2 ≤3 cm, N = 16 >3 cm, N c 9 Number of nodules
Baseline pathological features
Recurrence delay No. recurrence site
Site of recurrence TABLE 3.
HCC Recurrence pattern according to the initial pathological features of the primary tumor
Extrahepatic Only, Single, N = 14 Multiple, N = 9 Liver, N = 7 Lung,N = 13 Bone, N = 9 Wall, N = 4 Peritoneum, N = 4 Node, N = 4 N = 18
