Liver Transplantation in Cirrhotic Patients with ... - Semantic Scholar

Original Paper Dig Surg 2004;21:152–160 DOI: 10.1159/000078741

Received: August 25, 2003 Accepted: December 11, 2003 Published online: May 27, 2004

Liver Transplantation in Cirrhotic Patients with Small Hepatocellular Carcinoma: An Analysis of Pre-Operative Imaging, Explant Histology and Prognostic Histologic Indicators Shantanu Bhattacharjyaa Tanushree Bhattacharjyab Alberto Quagliac Amar Pal Dhillonc Andrew K. Burroughsd David W. Patchd Jonathan M. Tibballsb Anthony F. Watkinsonb Keith Rollesa Brian R. Davidsona Departments of a Surgery, b Radiology, c Pathology and d Hepatology, Royal Free Hospital, Royal Free and University College School of Medicine, London, UK

Key Words Hepatocellular carcinoma W Iodized oil CT W Dual-phase CT W Microvascular invasion

Abstract Background: In recent years, liver transplantation in patients with hepatocellular cancers and cirrhosis has been restricted to those with small cancers (!5 cm for solitary and ! 3 cm for multifocal HCC with ! 3 nodules). The selection of patients for liver transplantation is based on pre-operative imaging. The accuracy of imaging correlated with explant histology and the effect of tumour stage has not been evaluated in this selected population. Methods: In this study, prospectively collected data for 30 patients who underwent orthotopic liver transplantation for cirrhosis complicated by small hepatocellular carcinoma (HCC) at a single centre have been reviewed with the aim of correlating radiological findings, explant histology and patient outcome. Patients who underwent orthotopic liver transplantation between 1995 and 1999 had plain and contrast-enhanced dual-phase spiral CT (DCT) scans of the liver. Patients suspected of having HCC on CT scan or due to elevated serum alpha-fetopro-

ABC

© 2004 S. Karger AG, Basel 0253–4886/04/0212–0152$21.00/0

Fax + 41 61 306 12 34 E-Mail [email protected] www.karger.com

Accessible online at: www.karger.com/dsu

tein underwent iodized oil CT (IOCT). Following transplantation, the explanted liver was serially sectioned at 10-mm intervals and examined by a pathologist blinded to the results of imaging. Data collected prospectively on imaging and histology were compared with outcome data. The median period of follow-up was 1,139 days (range 690–1,955 days) after transplantation. All patients were followed up by clinical assessment, assessment of serum alpha-protein levels and imaging when indicated. Results: All the patients transplanted fulfilled the selective criteria on the basis of imaging (solitary HCC !5 cm in diameter or multifocal HCC ! 3 cm in diameter with !3 nodules). Of the 30 patients transplanted, 46 HCCs were detected on explant histology with a median size of 24 mm (range 6–75 mm). Ten patients had multifocal disease (median number of lesions 2, range 2–4). No significant difference was observed between IOCT and DCT with regards to the sensitivity (67.4 vs. 68%) and specificity (78.97 vs. 88.6%) of detecting HCCs. IOCT had a positive predictive value of 78.9% as compared to 82.8% for DCT. IOCT had an overall sensitivity of 40% as compared to 30% for DCT in detecting multifocal disease (not significant). Histological assessment of the explanted livers showed that 8 patients had well-, 17 mod-

Brian R. Davidson Professor of Hepatobiliary Surgery and Liver Transplantation The Royal Free Hospital, Pond Street London NW3 2QG (UK) Tel. +44 207 794 0500, Fax +44 207 830 2688, E-Mail [email protected]

erate and 5 poorly differentiated HCCs. Tumour size and the presence of multifocal disease did not influence survival in this study. Microvascular invasion was more common with larger tumours (from 38% with lesions less than 40 mm in diameter to 60% with lesions 140 mm in diameter; p ! 0.01) and with moderately (29.4%) or poorly differentiated (60%) HCCs than welldifferentiated HCC (12.5%) (p ! 0.04 and 0.01 for well- vs. moderately and poorly differentiated HCC, respectively). Microvascular invasion on explant histology was associated with poor survival. Of the 17 transplant recipients without vascular invasion, 15 were alive at 1 and 2 years in comparison to 7 of 9 with microscopic vascular invasion (p ! 0.01). Four patients died in the post-transplant period due to recurrent HCC. Overall survival [after excluding early post-transplant sepsis-induced deaths (n = 4)] at 1 year was 83.3%. Conclusions: Selective criteria for transplantation of HCC in cirrhosis are associated with a 1-year and 3-year survival rate of 73.3% (including early post-transplant sepsis-induced deaths). IOCT and DCT are similar in their ability to detect unifocal or multifocal HCC. Tumour size and number are not predictive of recurrence with these selective criteria, but microscopic vascular invasion is a bad prognostic factor. Copyright © 2004 S. Karger AG, Basel

Introduction

Liver transplantation as a therapeutic option for cirrhosis complicated by hepatocellular carcinoma (HCC) lost popularity when outcome studies revealed that survival at 2 years was as low as 17–26% [1–3], with early recurrence rates as high as 50–80%. Bismuth et al. [4] were the first to highlight the fact that perhaps the selection criteria being used were inappropriate, as most of the patients who were transplanted in the early era of transplantation had advanced disease. Extrahepatic micrometastases, tumour cell migration during liver manipulation and immunosuppression [5–7] were possible factors contributing to early recurrence. Tumour size 15 cm, the presence of nodal metastasis and major vascular involvement [1] were identified as poor prognostic factors [8]. Mazzaferro et al. [9] suggested that the criteria for selection should include unifocal lesions !5 cm in diameter or multifocal HCC with no more than 3 nodules, none 13 cm in diameter, based on their retrospective study. Jonas et al. [10] recently questioned whether the proposed criteria were in fact surrogate markers for vascular infiltration that cannot be detected on conventional imaging.

Liver Transplantation in Cirrhotic Patients with Small HCC

Currently, most transplant centres are more selective in their approach to transplanting patients with cirrhosis complicated by HCC, and only transplant patients with small HCC. The difficulty is detecting HCC in cirrhosis. Most imaging modalities that have been studied have avoided the issue of the ability to detect small HCCs (diameter !20 mm), principally because of the difficulty in characterizing and obtaining histological confirmation of such lesions [11, 12]. Correlation with explant histology is therefore essential for evaluation of HCC imaging in cirrhosis. Improved outcome has been reported following transplantation using the criteria of Mazzaferro et al. [9]. However, to our knowledge, there are no reported studies that have looked at the ability of currently available imaging modalities to detect small and multifocal HCCs and at prognostic factors that may influence the outcome following transplantation of these carefully selected patients. Iodized oil CT (IOCT) and dual-phase spiral CT (DCT) have been used as imaging modalities to detect HCCs in cirrhosis. These imaging modalities have not been compared for their ability to detect small HCC with correlation with explant histology, nor have the imaging and histological findings been compared with outcome in patients selected for transplantation using the criteria of Mazzaferro et al. [9]. In this study, we have prospectively evaluated patients with cirrhosis and HCC who were transplanted using the selective criteria. The results of IOCT and DCT in detecting and characterizing HCCs were compared and correlated with the findings on histopathology after serial sectioning of the explanted liver. Finally, the radiological and histological findings were compared with survival.

Patients and Methods We performed a review of prospectively collected data for 30 patients transplanted for pre-operatively diagnosed cirrhosis complicated by HCC between 1995 and 1999. All the 30 patients included in this study underwent both a DCT and an IOCT. All patients had cirrhosis of Child’s grade B or C. The aetiology was hepatitis C infection in 17 patients, chronic hepatitis B infection in 7, alcoholic liver disease in 1, and coexistent hepatitis B and hepatitis C, coexistent hepatitis C and alcoholic liver disease, cryptogenic cirrhosis and primary biliary cirrhosis in 1 patient each. DCT was performed as an out-patient procedure, prior to IOCT, on a GE Hispeed spiral CT with 7- to 10-mm collimation with a table speed of 10–15 mm/s. Intravenous contrast [100 ml of Omnipaque 300 (Nycomed)] was infused at a rate of 4 ml/s through a peripheral vein. Images of the liver were acquired in two phases. An initial precontrast scan was performed. An arterial-phase scan was initiated 20 s after the start of contrast injection, followed by a portal venous-

Dig Surg 2004;21:152–160

153

phase scan 70 s after contrast injection. Scanning was performed with the patient in the supine position. The abdominal scans were performed from the level of the domes of the diaphragm to the symphysis pubis. Contiguous 10-mm-thick axial images were reconstructed from the volumetric data. The diagnosis of HCC was based on the presence of lesions with arterial-phase hypervascularity relative to adjacent liver parenchyma. In the portal venous phase, these lesions appeared iso- or hypointense as compared to the adjacent parenchyma. IOCT was performed in two stages. A selective hepatic angiogram was first performed using a 5-Fr sidewinder catheter inserted through a transfemoral route after correcting any clotting abnormalities (INR ! 1.5, platelet count 1 50,000/ml). 10 ml of lipiodol ultrafluid emulsified with 8 ml of urograffin 370 was injected into both the right and left hepatic arteries. The patient was admitted to hospital for a day when the angiogram was performed for post-procedure observation. A follow-up non-enhanced spiral CT was performed as an out-patient procedure after 10–14 days to assess the number and site of lesions. Dense focal retention of lipiodol was taken to represent HCC as previously described [13]. Following orthotopic liver transplantation, the explanted liver was serially sectioned in 10-mm slices according to a standard protocol [14], so that the morphological characteristics could be compared with the characteristics of the corresponding image on CT. A histopathologist who was blinded to the results of the scans examined the explanted liver. Tissue samples from all nodules and excised lymph nodes were examined by haematoxylin and eosin staining, and where equivocal, with reticulin staining and CD34 antigen staining. The entire circumference of the tumours was examined, and therefore, the number of sections from each nodule was determined by its size. The lesions were classified as HCC, regenerative nodules or dysplastic nodules using established criteria [15]. Results of the scans and histopathology were compared for the degree of concordance in terms of number, site and size of lesions, presence of lymph node metastasis and vascular invasion. Multifocal disease was considered to be present if there were 2 or more foci of HCC in anatomically separate segments of the liver. Tumour bulk was calculated by assuming that all the lesions represented small spheres and therefore had a volume of 4/3r3 (where r is the radius of the lesion). In multifocal tumours, the volume of individual lesions was added up to give an estimate of the total volume of intrahepatic disease. Of the 30 patients transplanted, there were 4 early postoperative deaths due to sepsis. The remaining 26 patients were followed up regularly as out-patients. The median period of follow-up was 1,139 days (range 690–1,955 days) after transplantation. All patients were followed up by clinical assessment (initially weekly after discharge then two weekly and monthly by 3 months following discharge and then at 3-monthly intervals for 1 year and biannually thereafter), assessment of serum alpha-protein levels and imaging when indicated. Kaplan-Meier survival curves were used to assess the effect of tumour size, multifocal disease, tumour volume, nodal status and microvascular invasion. ¯2 and Fisher’s exact tests were used where appropriate. Statistical significance was considered to be present if p ! 0.05.

154

Dig Surg 2004;21:152–160

Fig. 1. Incidence of microvascular invasion versus tumour size. Light bars represent total number of tumours and dark bars represent proportion of tumours with microvascular invasion.

Results

All 30 patients fulfilled the selective criteria for liver transplantation proposed by Mazzaferro et al. [9] on the basis of imaging. Cadaveric orthotopic liver transplant was carried out with caval replacement. At the time of surgery, there was no evidence of extrahepatic cancer. Portal lymphadenectomy was performed as part of the explant operation. The median waiting time between imaging and transplant was 36 days (range 1–370 days). Explant Histology On histopathological examination of the explanted liver, 19 patients had a solitary unilobar HCC, 2 had a solitary HCC with satellite nodules in the same anatomical segment and 1 had a bilobar solitary HCC. Five patients had 2 hepatocellular cancers, 3 patients had 3 and 2 had 4, yielding a total of 46 lesions. An additional 29 benign and 16 dysplastic lesions were also found in the 30 explanted livers. The median diameter of the cancers was 24 mm (range 6–75 mm). Based on segmental anatomy of the liver, 2 patients had lesions in segment 4, 7 in segment 5, 2 in segment 6, 6 in segment 7 and 4 in segment 8. Three patients had lesions in both segments 4 and 8, 3 in segments 4 and 7, 1 in segments 5 and 6 and 1 in segments 7 and 8. Eight patients had well-differentiated, 17 moderately differentiated and 5 poorly differentiated HCCs. Two patients had metastatic HCC in the portal lymph nodes, and 9 patients (1 with well-, 5 with moderately and 3 with poorly differentiated HCCs) showed evidence of microscopic vascular invasion surrounding the main tu-

Bhattacharjya et al.

Fig. 2. Kaplan-Meier curves comparing survival following transplan-

Fig. 3. Kaplan-Meier curves comparing survival following transplan-

tation of HCC patients stratified on the basis of tumour size. Deaths are marked by black squares, and the numbers on the figure represent the number of patients alive at the end of the follow-up period.

tation of patients with HCC in the presence and absence of vascular invasion. The black squares represent deaths, and the numbers on the figure represent those patients alive at the end of the follow-up period.

Table 1. Tumour detection rate on imaging versus lesion size

ter greater than 40 mm, and microscopic vascular invasion was detected in 5 (62.5%) (fig. 1). Microscopic vascular infiltration was found more frequently with large cancers (62.5% of tumours with diameter 140 mm) than with small HCCs (33% of tumours with diameter !40 mm) (p ! 0.01).

Tumour size

Number of tumours on histology

Number of tumours detected on IOCT

Number of tumours detected on DCT

! 9 mm 10–20 mm 21–30 mm 31–40 mm 1 40 mm

4 15 16 3 8

0 8 (53.3) 12 (75) 3 (100) 8 (100)

0 5 (33.3) 13 (81.2) 3 (100) 8 (100)

Figures in parentheses represent percentages.

mour. Out of the 46 lesions, 4 were less than 9 mm in diameter and were not found to have any microscopic vascular invasion. 15 lesions had a diameter between 10 and 20 mm, and 1 of these 15 lesions (6.6%) showed evidence of microscopic vascular invasion. 16 lesions had a diameter in the range of 21–30 mm, and 2 of these 16 (12.5%) had microscopic vascular invasion. 3 lesions had a diameter between 31 and 40 mm, and 1 of these (33.3%) had microscopic vascular invasion. 8 lesions had a diame-


Imaging versus Histology IOCT scans correctly identified 31 of the 46 tumour nodules (67.4%). The mean diameter of the lesions missed on IOCT was 13.1 B 3.2 mm. An additional 9 lesions were detected on IOCT but were shown on histology to be nodules with dysplasia, not HCC (false positives 22.5%). Hypervascular lesions seen during the arterial phase of DCT were considered to represent HCC. On comparison with histology, DCT correctly identified 29 of the 46 tumour nodules (63%). An additional 6 nodules were detected on DCT but were shown to be benign on histology (false positives 16.6%). The mean diameter of the lesions missed on DCT was 10.81 B 3.4 mm. Table 1 gives the number of HCCs detected on both modalities with respect to size. The size of the HCCs measured (to the closest mm) on IOCT as compared to histology was correct in 31% and

Dig Surg 2004;21:152–160

155

Table 2. Effect of multifocal HCC on outcome as compared to unifocal HCC (of similar size to the dominant nodule

in multifocal HCC) Size of largest HCC

Unifocal HCC

Multifocal HCC

number of patients

mean tumour volume, mm3

number of deaths due to recurrent HCC

number of patients

mean tumour volume, mm3

number of deaths due to recurrent HCC

20–29 mm 30–39 mm 40–49 mm

5 1 3

7,686.4 14,142.8 47,732.1

0 0 1

3 5 2

13,081.0 35,441.5 51,543.5

2 0 0

6 patients had unifocal HCCs ! 20 mm in diameter and are not included here. 1 patient who died had an HCC 1 50 mm in diameter.

underestimated in 69% of lesions (mean B SD 4.5 B 1.7 mm). DCT was accurate in measuring the size of the lesions in 46% of lesions, overestimated the size in 4% and underestimated the size in 50% (mean B SD 3.6 B 1.2 mm). No enlarged lymph nodes or extrahepatic disease were detected on imaging. Tumour Variables and Survival Analysis To assess the effect of tumour size on survival, the tumours were stratified into 3 groups based on the diameter of the largest or principal lesion. In group 1, maximum tumour diameter was !20 mm (n = 6); in group 2, maximum tumour diameter was 21–40 mm (n = 15), and in group 3, maximum tumour diameter was 140 mm (n = 5). Survival using Kaplan-Meier curves (fig. 2) was better in group 1 than in groups 2 and 3, but the difference did not reach statistical significance (p ! 0.07). 10 patients were found to have multifocal disease on pathology. Five patients had 2 lesions, 3 had 3 and 2 had 4 lesions. Pre-operatively, IOCT correctly identified the number of lesions in 2 of the 10 patients. IOCT failed to detect the presence of multifocal disease in 5 patients, overestimated the number of lesions in 1 patient and underestimated the number of lesions in 2 patients, yielding an overall sensitivity of 40% (i.e. ability to detect multifocal disease). IOCT accurately represented the true number of HCCs in 2 patients (20%). DCT correctly identified the presence and number of lesions in 1 patient, failed to detect lesions in 7 patients and underestimated the number of lesions in 2 patients, with an overall sensitivity of 30%. The significance of multifocal disease for outcome was assessed by grouping the patients according to the diameter of the HCC or largest tumour in the case of multifocal HCC. No difference was found in survival of patient

156

Dig Surg 2004;21:152–160

Table 3. Tumour characteristics of 4 patients with early HCC recur-

rence Patient No.

Tumour size mm

Lymph nodal status

Microscopic vascular invasion

Unifocal/ multifocal

1 2 3 4

25 25, 25, 12, 10 40 7, 25

negative negative positive positive

absent absent present present

unifocal multifocal unifocal multifocal

groups with unifocal or multifocal disease with HCC of similar diameter (table 2). Tumour volume in patients with multifocal HCC was greater than that in patients with unifocal HCC for a similar size of the largest tumour, but this did not influence survival (table 2). Pre-operative imaging in these 30 patients failed to identify lymph nodal metastasis or vascular invasion in any of the patients. Two patients had lymph node metastases. Vascular invasion was present in both the cases with lymph node metastasis. There was no procedure-related morbidity or mortality in these patients during IOCT or DCT. Four patients died post-transplant secondary to poor graft function and overwhelming sepsis. Following transplant, the remaining 26 patients were discharged from hospital. Of these 26 patients, 4 died after 94, 151, 264 and 335 days, respectively, due to recurrent HCC (tumour characteristics are shown in table 3). Survival at 1 year was 88.2% (15 of 17) in the absence of microscopic vascular invasion and 77.7% (7 of 9) in the presence of microscopic vascular invasion (p ! 0.01) after excluding early deaths related to sepsis (fig. 3).


Discussion

In this study, we analysed a group of patients who had undergone a liver transplant for HCC in cirrhosis applying the selective criteria proposed by Mazzaferro et al. [9]. The 1- and 3-year survival rate of 83.3% following transplantation in this cohort of patients is comparable to that for transplantation for non-malignant disease and is superior to reported results for liver resection for HCC in cirrhosis. Selective criteria were identified on the basis of imaging findings and anticipated survival. However, these were biased by the limitations of the imaging modalities that were used, and also by the waiting times from diagnosis to transplantation, which may have led to the selection of a cohort of patients with cancers having biologically less aggressive behaviour. Therefore, it is not clear whether the criteria of size, number of lesions and lymph node metastases are truly independent markers of HCC behaviour or are in fact surrogate markers for events occurring earlier in the course of the natural history of these tumours that are not detectable by conventional imaging. In this study, IOCT had a sensitivity of 67.4% in detecting HCCs, which is within the reported range (58– 97%) [12, 16, 17]. The wide range of sensitivities previously reported reflects differences in the method adopted to confirm the nature of the observed lesion on histopathology, which has included percutaneous biopsy [11], radiological criteria or examination of resected specimens [12, 16, 17] or explanted livers in the setting of liver transplantation. Another factor that may be contributory is observer variation in what is considered to represent an HCC on the IOCT scan. Diagnostic criteria are poorly defined. Four main patterns of uptake of lipiodol have been described in the liver on CT [13], but only wellcircumscribed lesions with dense homogenous uptake of lipiodol are highly suggestive of representing HCCs (92%). In light of these findings, only focal dense uptake of lipiodol was considered to represent HCC. In this series, there was no procedure-related morbidity or mortality. However, IOCT has a 1.4% reported incidence of intimal dissection during angiography and a similar incidence of exacerbating portal hypertension and causing variceal bleeds due to obstruction of the intrahepatic portal channels by opening a pre-sinusoidal anastomosis [11]. Retention of lipiodol by HCC forms the basis of performing IOCT. However, lipiodol can also be retained by dysplastic nodules, and in this study, 9 out of 16 dysplastic nodules (56%) retained lipiodol. In this study, IOCT had a high false-positive rate of 22.5%.


The sensitivity of DCT in detecting HCC in this study was 63%. The addition of arterial-phase imaging to portal venous-phase imaging has been reported to enhance the detection of HCCs from 68 to around 84% [18]. The results in this study are inferior to those previously reported. This may be due to verification bias, as in other studies histological confirmation was obtained by needle biopsy of the lesions that were observed on the scans or by demonstrating an increase in size on serial scans. Such a technique does not take into account lesions that are present but missed on imaging. Lee et al. [18] reported that hypervascularity during the arterial phase has a positive predictive value of 100% for HCC. However, a dysplastic nodule and a regenerative nodule may also appear hypervascular [19], and in these situations, the final decision as to the true nature of a lesion has to be based on histopathology. In the current study, all the lesions considered to represent HCC showed arterial-phase hypervascularity. DCT had a sensitivity of 63.2% in detecting HCCs. The advantage of serially sectioning the explanted liver using a slice thickness of the same order (10 mm) as that of the imaging technique is that it allows direct comparison of morphology with imaging. This allows a more accurate estimate of the true sensitivities and specificities of an imaging modality as compared to confirming diagnoses by percutaneous biopsies, as with the latter, there is no way to know how many tumours are actually present but have not been detected on imaging. In this study, DCT and IOCT were unreliable in the detection of lesions less than 11 and 13 mm, respectively, this difference not being significant. Both DCT and IOCT were able to detect all the lesions greater than 40 mm in this study. Reformatting the volumetric data on CT to finer sections, as is possible with the newer multidetector CT scans, may detect sub-centimetre lesions and thereby improve the sensitivity of imaging. In this study, the images were reformatted to 10-mm sections (on the basis of hardware and software available at the time), and this may have contributed to the lower sensitivities observed with IOCT and DCT. Tumour size did not significantly alter outcome in patients transplanted for HCC in cirrhosis in this study. This could be due to the small number of patients with tumours 140 mm in diameter in this group of patients, a consequence of the criteria used for selecting patients with cirrhosis complicated by HCC for transplantation. Mazzaferro et al. [9] suggest that in the presence of multifocal disease, the transplant recipient should have no more than 3 nodules, none greater than 30 mm in diameter. In this study, neither IOCT nor DCT reliably

Dig Surg 2004;21:152–160

157

detected multifocal disease. The incidence of multifocal disease was not related to the size of the principal dominant lesion and the outcome of multifocal disease was found to be no different from that of unifocal disease in this study. This is probably because of the limitations of imaging, and what may appear as a solitary HCC or !3 nodules of HCC in cirrhosis may actually have a much higher incidence of multifocality on serial sectioning. Only a third of the patients in this study had multifocal disease based on explant histology. This is similar to data published by Fasani et al. [20], in which a lower percentage of patients with a single aetiology underlying their chronic liver disease had multifocal HCC. The incidence of microvascular invasion was assessed in all 46 HCCs in the 30 explanted livers. The incidence was found to vary with tumour differentiation and tumour size. Jonas et al. [10] recently reported similar findings with HCCs 15 cm in diameter. To our knowledge, this has not been previously reported in patients transplanted for HCC using selective criteria. Microvascular invasion was noted in one lesion as small as 10 mm; the incidence of microvascular invasion rose rapidly once the lesion size was greater than 40 mm in diameter. The issue of whether microvascular invasion is indeed a prognostic marker or is in fact a surrogate marker for previously welldescribed adverse prognostic factors is difficult to resolve and is the basis of further ongoing studies in our unit. Similar results have been reported for patients who would have fulfilled the criteria for transplantation but had a resection [21, 22]. A larger patient cohort transplanted for small HCC in cirrhosis would be required to establish whether microvascular invasion is indeed an independent risk factor influencing outcome following transplantation. Lymph node metastases or major vascular invasion have not been reported in patients with small HCC. Our data would suggest that the incidence of microscopic vascular invasion, lymph node metastases and extrahepatic disease in this group of patients is small. The only factor found to alter outcome following transplantation was the presence of microvascular invasion. The true incidence of microvascular infiltration in this group of patients may be underreported. In the absence of microscopic vascular infiltration, a survival rate of 88.2% 1 year following liver transplantation was observed. A significantly increased incidence of microvascular invasion was found with moderately and poorly differentiated tumours in this study. Similar results have been reported in patients who had a resection for small HCC [22] but would have fulfilled the criteria for transplantation. This would suggest that microvascular invasion is

158

Dig Surg 2004;21:152–160

related to the degree of dedifferentiation of the HCC and occurs early in the course of the disease. The incidence of microvascular invasion rises rapidly once the tumour size is greater than 40 mm, and this would suggest that perhaps the size of the tumour is in fact a surrogate marker for microvascular invasion as has been suggested by Jonas et al. [10]. The results of IOCT and DCT in detecting and characterizing HCCs in this study were similar. Both had a high false-positive rate, i.e. 22% with IOCT and 17% with DCT. This difference was not statistically significant. The benefits of DCT over IOCT rest in the fact that it is less invasive, can be performed on an out-patient basis, has lower morbidity and is cheaper. We therefore suggest that DCT be used as the investigation of choice in the detection of HCC in cirrhosis in patients being considered for liver transplantation. Teefey et al. [23] prospectively compared CT, MRI and ultrasound in detection of small HCC in liver transplant candidates compared to explant histology and concluded that CT and MRI have similar efficacy. Untreated cirrhosis complicated by small HCC has a survival of less than 50% at 1 year [24]. Surgical resection for unifocal small HCC in cirrhosis has a reported survival of 50% at 5 years for well-differentiated HCC and 50% at 2 years for moderately and poorly differentiated HCC. In this study, the results of transplantation for small HCC were superior to resection or non-intervention. In addition, multifocal disease did not alter outcome, and therefore transplantation should be the preferred treatment in all patients with cirrhosis complicated by small HCC (!40 mm in diameter).


References 1 Koneru B, Cassavilla A, Bowman J, Iwatsuki S, Starzl TE: Liver transplantation for malignant tumors. Gastroenterol Clin North Am 1988;17: 177–193. 2 Ringe B, Wittekind C, Bechstein WO, Buzendahl H, Pichalmayr R: The role of liver transplantation in hepatobiliary malignancy. A retrospective analysis of 95 patients with particular regard to tumor stage and recurrence. Ann Surg 1989;209:88–98. 3 Olthoff KM, Millis JM, Rosove MH, Glodstein LI, Ramming KP, Busuttil R: Is liver transplantation justified for the treatment of hepatic malignancies? Arch Surg 1990;125:1261– 1266. 4 Bismuth LH, Chiche L, Adam R, Castaing D, Diamond T, Dennison A: Liver resection versus transplantation for hepatocellular carcinoma in cirrhotic patients. Ann Surg 1993;227: 145–151. 5 O’Grady JG, Polson RJ, Rolles K, Calne RY, Williams R: Liver transplantation for malignant disease. Results in 93 consecutive patients. Ann Surg 1988;207:373–379. 6 O’Grady JG, Johnson PJ, Zaman S, Calne RY, Williams R: Decreased rate of growth of hepatocellular carcinoma recurrence after liver transplantation in patients maintained on cyclosporine immunosuppression. Transplant Proc 1988;20:394–396. 7 Carr B, Kar S: An assay for hepatoma micrometastases: Albumin gene expression in peripheral blood from patients with advanced hepatocellular carcinoma (HCC) (abstract). Hepatology 1991;14:108A. 8 McPeake JR, O’Grady JG, Zaman S, et al: Liver transplantation for primary hepatocellular carcinoma: Tumor size and number determine outcome. J Hepatol 1993;18:226–234.

9 Mazzaferro V, Regalia E, Doci R, et al: Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693–699. 10 Jonas S, Bechstein WO, Steinmuller T, et al: Vascular invasion and histopathologic grading determine outcome after liver transplantation for hepatocellular carcinoma in cirrhosis. Hepatology 2001;33:1080–1086. 11 Bizollon T, Rode A, Bancel B, et al: Diagnostic value and tolerance of lipiodol-computed tomography for the detection of small hepatocellular carcinoma: Correlation with pathologic examination of explanted livers. J Hepatol 1998;28:491–496. 12 Taourel PG, Pageaux GP, Coste V, et al: Small hepatocellular carcinoma in patients undergoing liver transplantation: Detection with CT after injection of iodized oil. Radiology 1995; 197:377–380. 13 Ngan H: Lipiodol computerized tomography: How sensitive and specific is the technique in the diagnosis of hepatocellular carcinoma? Br J Radiol 1990;63:771–775. 14 Demetris AJ, Jaffe R, Starzl TE: A review of adult and pediatric post-transplant liver pathology. Pathol Annu 1987;22:347–386. 15 Terminology of nodular hepatocellular lesions. International Working Party. Hepatology 1995;22:983–993. 16 Spreafico C, Marchiano A, Mazzaferro V, et al: Hepatocellular carcinoma in patients who undergo liver transplantation: Sensitivity of CT with iodized oil. Radiology 1997;203:457– 460. 17 Saada J, Bhattacharjya S, Dhillon AP, et al: Detection of small hepatocellular carcinomas in cirrhotic livers using iodised oil computed tomography. Gut 1997;41:404–407.

18 Lee HM, Lu DSK, Krasny RM, Busuttil R, Kadell B, Lucas J: Hepatic lesion characterization in cirrhosis: Significance of arterial hypervascularity on dual-phase helical CT. AJR Am J Roentgenol 1997;169:125–130. 19 Freeny PC, Grossholz M, Kaakaji K, Schmiedl UP: Significance of hyperattenuating and contrast-enhancing hepatic nodules detected in the cirrhotic liver during arterial phase helical CT in pre-liver transplant patients: Radiologic-histopathologic correlation of explanted livers. Abdom Imaging 2003;28:333–346. 20 Fasani P, Sangiovanni A, Fazio C, et al: High prevalence of multinodular hepatocellular carcinoma in patients with cirrhosis attributable to multiple risk factors. Hepatology 1999;29: 1704–1707. 21 Esnaola NF, Lauwers GY, Mirza NQ, et al: Predictors of microvascular invasion in patients with hepatocellular carcinoma who are candidates for orthotopic liver transplantation. J Gastrointest Surg 2002;6:224–232. 22 Lauwers GY, Terris B, Balis UJ, et al: Prognostic histologic indicators of curatively resected hepatocellular carcinomas: A multi-institutional analysis of 425 patients with definition of a histologic prognostic index. Am J Surg Pathol 2002;26:25–34. 23 Teefey SA, Hildeboldt CC, Dehdashti F, et al: Detection of primary hepatic malignancy in liver transplant candidates: Prospective comparison of CT, MR imaging, US, and PET. Radiology 2003;226:533–542. 24 Llovet JM, Bustamante J, Castells A, et al: Natural history of untreated nonsurgical hepatocellular carcinoma: Rationale for the design and evaluation of therapeutic trials. Hepatology 1999;29:62–67.

Invited Commentary Preoperative Imaging versus Pathologic Findings after Transplantation for Hepatocellular Carcinoma: Are Current Transplant Criteria Being Met? Charles Cha a, Ronald P. DeMatteo b a Departments

of Gastrointestinal Surgery and Surgical Oncology, Yale University School of Medicine, New Haven, Conn., b Hepatobiliary Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, N.Y., USA

The precise evaluation of patients with hepatocellular carcinoma for transplantation is necessary to allocate the limited supply of available grafts to those who will derive the most benefit. Currently, radiologic assessment is a major determinant of transplant eligibility. Tumor size and number are used for predicting tumor vascular invasion, risk of recurrence and overall survival. However,

there are only scant data regarding the accuracy of current imaging modalities, and the study by Bhattacharjya et al. of the Royal Free Hospital helps to address this deficiency. The authors have reviewed their experience with liver transplantation in 30 patients with hepatocellular carcinoma who were selected according to the Milan criteria (solitary hepatocellular carcinoma ^5 cm or 2–3 lesions


Dig Surg 2004;21:152–160

159

^3 cm) [1]. The findings on triple-phase helical CT and iodized oil CT were compared to the pathologic examination of the explant. The investigators found that iodized oil CT did not provide additional information, confirming their previous report [2] and the conclusion of others [3]. Both techniques had a sensitivity of approximately 70% and a false-positive rate of about 20%. Strikingly, nearly 20% of tumors that measured between 2 and 3 cm were missed on radiologic evaluation. Furthermore, the sensitivity of CT in detecting multifocal disease was only 30%, which may have been even lower had the explant been transected at 5-mm and not 10-mm intervals. Another group has recently demonstrated that MR angiography is superior (sensitivity and specificity of 75%) to helical CT [4]. In particular, MR angiography was more sensitive (89 vs. 65%; p = 0.03) in detecting 1- to 2-cm nodules. Nevertheless, the overall accuracy in determining the extent of hepatocellular carcinoma by MR angiography or helical CT was similar at about 60%. Not unexpectedly, the detection of subcentimeter nodules by either technique was poor. With over 30% of lesions missed by standard imaging, the question that arises is whether current transplant criteria are truly being met. The simple answer is yes. The Milan criteria were established as a clinical means to super-select those patients whose survival after transplantation would mirror that of patients transplanted for benign disease. Although a certain percentage of patients will have more extensive disease on pathologic examination, the size and number criteria were created based upon preoperative imaging. Indeed, even in the original report of Mazzaferro et al. [1], 27% of patients pathologically exceeded the criteria. Multiple authors including the current ones have confirmed that these criteria are useful in selecting patients who fare well after transplantation. An issue not addressed in the current report, though, is how often the patients on the waiting list were scanned. The median time between imaging and transplantation was 36 days, but 1 patient had not been scanned for 1 year. Bhattacharjya et al. conclude that transplantation should be preferred in all patients with cirrhosis who have tumors !4 cm. Liver resection is infrequently performed in patients with Child’s grade B or C cirrhosis, which was the status of all the patients in the current study. Therefore, the 3-year survival of 73% after transplantation for

160

Dig Surg 2004;21:152–160

hepatocellular carcinoma is commendable, and few would dispute the role of transplantation for early hepatocellular carcinoma in these patients. The survival rate may be an overestimation since it does not account for dropouts from the waiting list, which may be as high as 20%, as emphasized by Llovet et al. [5]. The debate over resection versus transplantation actually concerns patients with preserved liver function (i.e. Child’s grade A). We and others have shown that patients with hepatocellular carcinoma with intact hepatic function who were eligible for transplantation by the Milan criteria but were instead treated with partial hepatectomy, had a 5-year survival of about 70%, which is equivalent to the results of transplantation in this patient subset [6, 7]. The assertion by Bhattacharjya et al. that transplantation is also indicated for multifocal disease requires more than an initial experience with 10 such patients. The future for preoperative evaluation of patients for transplantation for hepatocellular carcinoma will likely consist of more accurate diagnostic imaging in combination with some molecular-based analysis of tumors. Perhaps with improved preoperative risk assessment, even more refined criteria will be established to optimally identify patients with HCC who are suitable for transplantation. References 1 Mazzaferro V, Regalia E, Doci R, Andreola S, Pulvirenti A, Bozzetti F, et al: Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693–699. 2 Saada J, Bhattacharya S, Dhillon AP, Dick R, Burroughs AK, Rolles K, et al: Detection of small hepatocellular carcinomas in cirrhotic livers using iodised oil computed tomography. Gut 1997;41:404–407. 3 Bruix J, Sherman M, Llovet JM, Beaugrand M, Lencioni R, Burroughs AK, et al: Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona-2000 EASL conference. European Association for the Study of the Liver. J Hepatol 2001;35:421–430. 4 Burrel M, Llovet JM, Ayuso C, Iglesias C, Sala M, Miquel R, et al: MRI angiography is superior to helical CT for detection of HCC prior to liver transplantation: An explant correlation. Hepatology 2003;38:1034–1042. 5 Llovet JM, Fuster J, Bruix J: Intention-to-treat analysis of surgical treatment for early hepatocellular carcinoma: Resection versus transplantation. Hepatology 1999;30:1434–1440. 6 Poon RT, Fan ST, Lo CM, Liu CL, Wong J: Long-term survival and pattern of recurrence after resection of small hepatocellular carcinoma in patients with preserved liver function: Implications for a strategy of salvage transplantation. Ann Surg 2002;235:373–382. 7 Cha CH, Ruo L, Fong Y, Jarnagin WR, Shia J, Blumgart LH, et al: Resection of hepatocellular carcinoma in patients otherwise eligible for transplantation. Ann Surg 2003;238:315–321.
