Response to preoperative chemoembolization correlates with outcome after liver transplantation in patients with hepatocellular carcinoma

Authors


Abstract

Patients with small hepatocellular carcinoma (HCC) can be cured by liver transplantation (LT). However, many patients drop out during the waiting time as a result of tumor progression. We prospectively investigated the effect of transarterial chemoembolization on long-term survival of 116 patients with HCC listed for LT. Intention-to-treat analysis revealed that patients with either complete or partial response to therapy (no vital tumor or devascularization of ≥30%, respectively) as assessed by computed tomographic scan before LT had far better 1-, 2-, and 5-year survival rates (100, 93.2, and 85.7%; and 93.8, 83.6, and 66.2%, respectively) compared with those with no response or with tumor progression (82.4, 50.7, and 19.3%). Posttransplant survival analysis showed a marked survival benefit according to transarterial chemoembolization response: patients with complete or partial response had 1-, 2-, and 5-year survival rates of 89.1, 85.1, and 85.1%, and 88.6, 77.4, and 63.9%, respectively, compared with 68.6, 51.4, and 51.4% for patients whose disease did not respond to therapy. Subgroup analysis, however, showed that these benefits were only seen in patients whose disease met the Milan criteria, but not in disease exceeding the Milan criteria but fitting the expanded University of California at San Francisco criteria. These patients were also more likely to drop out as a result of tumor progression while waiting for LT (dropout rate 12.1 vs. 2.9%) and to develop recurrent HCC (21.6 vs. 7.6%). Downstaged patients did even worse, with a dropout rate of 26.7% and a 5-year survival rate of only 25%. In conclusion, the response to preoperative chemoembolization may predict long-term outcome after LT. Liver Transpl 13:272–279, 2007. © 2007 AASLD.

The last 3 decades have seen an increasing incidence of hepatocellular carcinoma (HCC) in western countries.1 HCC is one of the most common malignant tumors worldwide, with an estimated annual incidence of about 1 million cases.2–4 The main risk factor for the development of HCC is cirrhosis, regardless of cause.5

Standard therapy for HCC depends on the size of tumor and the stage of underlying liver disease. Patients with advanced tumor disease have no chance of curative treatment, whereas those with small HCC have the option of orthotopic liver transplantation (LT). LT has been considered the only curative treatment because it has been claimed to cure the malignant disease by replacing the premalignant cirrhotic liver. Survival and recurrence rates after LT for HCC have markedly improved over the years, in particular by appropriate selection criteria.6–11 Despite stringent criteria for patient selection, liver transplant waiting lists are getting longer, and both Europe and the United States are facing a continually worsening discrepancy between supply and demand for donor livers. For some patients, increasing waiting times lead to tumor progression. A recent intention-to-treat analysis showed that a continually growing number of patients must be removed from the waiting list because of tumor progression exceeding accepted transplant criteria.12 Dropout rates of >20% have recently been reported.12–14

Transplant centers have tried a variety of treatment approaches to prevent progression of tumor disease in patients on the waiting list, the most frequent among them being ablative therapies such as transarterial chemoembolization (TACE), percutaneous ethanol injection, and radiofrequency ablation. Several studies, including our own, have shown the efficacy of pre-LT locoablative therapies, in particular TACE, in preventing tumor progression in patients on the waiting list.15–20 The effect of therapy on post-LT survival and tumor recurrence, however, remains unclear. Although small studies have reported excellent tumor-free survival when TACE was performed before LT,15, 16 especially in patients with advanced-stage tumors,17 most studies failed to demonstrate any benefit of pre-LT therapy on intermediate or long-term outcome and tumor recurrence.6, 21–23

The present study was conducted to address the question of a possible beneficial effect of TACE before LT on long-term patient outcome and rate of recurrence.

Abbreviations

LT, liver transplantation; HCC, hepatocellular carcinoma; TACE, transarterial chemoembolization; UCSF-expanded, expanded University of California at San Francisco criteria; AFP, alfa-fetoprotein; CR, complete response; PR, partial response; CT, computed tomography.

PATIENTS AND METHODS

Between January 1994 and December 2004, a total of 495 patients with HCC were referred to our clinic for treatment. Of these, 175 patients (35.6%; 149 men and 26 women; mean age 58 ± 7 years) initially met the inclusion criteria and were listed for LT.

In 116 (66.3%) of 175 patients eligible for LT, imaging studies revealed that HCC was hypervascularized, and TACE was performed during the waiting time for LT (embolization group). Mean follow-up after tumor diagnosis was 3.8 ± 2.5 years; mean follow-up after LT was 3.1 ± 2.6 years. In 10 of 116 patients, marked tumor progression occurred during the waiting time, and the patients had to be removed from the waiting list (overall dropout rate 8.6%). Two of these patients showed tumor infiltration of the portal vein, and 1 developed pulmonary metastases. In the remaining 7 patients, the HCC was growing beyond the expanded University of California at San Francisco (UCSF-expanded) criteria, which was considered to be the limit of tumor growth during the waiting time. When patients were suspected to have progressive disease but imaging studies had ambiguous results, a tissue-based diagnosis was obtained to support the decision for removing the patient from the waiting list.

The remaining 106 patients successfully underwent transplantation. Median waiting time for LT was 9 months (range 1.2–34 months).

Liver Disease in Patients Treated With TACE

The underlying liver disease was viral hepatitis in 67 patients (57.9%), alcoholic liver cirrhosis in 29 patients (25.0%), and other diseases in 20 patients (17.1%). These 116 patients comprised the embolization group. Detailed patient characteristics are listed in Table 1. According to Child-Pugh classification, 61 patients (52.6%) presented with stage A, 52 (44.8%) with stage B, and 3 (2.6%) with stage C at the time of tumor diagnosis.

Table 1. Patient Characteristics (Embolization Group, n = 116)
CharacteristicValue
  1. Abbreviations: HCV, hepatitis C virus; HBV, hepatitis B virus; NAFLD, non-alcoholic fatty liver disease.

Gender (M/F)100/16
Mean age (yr)58 ± 7
Cause of underlying liver disease, n (%) 
 HCV50 (43.2%)
 HBV14 (12.1%)
 HCV/HBV3 (2.6%)
 Alcohol29 (20.0%)
 Cryptogenic10 (8.6%)
 Hemochromatosis4 (3.4%)
 NAFLD4 (3.4%)
 Alpha-1-AT-deficiency2 (1.7%)
Child-Pugh stage, n (%) 
 A61 (52.6%)
 B52 (44.8%)
 C3 (2.6%)

Diagnosis and Staging of HCC

HCC was diagnosed according to EASL guidelines (2 separate conclusive imaging studies and/or alfa-fetoprotein [AFP] levels >400 ng/mL and/or histology).24 On the basis of pre-LT radiologic assessment, 6 patients (5.2%) were diagnosed with stage I, 62 (53.5%) with stage II, 33 (28.4%) with stage III, and 15 (12.9%) with stage IV disease according to the modified UICC (International Union Against Cancer) criteria. Thus, the tumors of 68 patients (58.6%) were within the Milan criteria,6 and an additional 33 patients (28.5%) had tumors exceeding the Milan criteria but fitting the UCSF-expanded criteria.11 Fifteen patients (12.9%) did not meet current selection criteria but were listed for LT because of an initial response to TACE, and they were enrolled onto the downstaging program, as outlined elsewhere.20

Tumor Marker

AFP (normal value <11 ng/mL) was measured in the serum at the time of diagnosis and every 3 months thereafter during follow-up (Roche Diagnostics, Switzerland).

Treatment of HCC While Waiting for LT

TACE was performed in 116 patients on the waiting list for LT. Patients underwent multiple sessions of TACE (mean 2.7 ± 1.7 sessions, range 1–8 TACE sessions per patient). TACE was performed as described elsewhere in detail.20 Briefly, the feeding arteries to the lesion or lesions were catheterized as selectively as possible, and a lipiodol mixture (70 mg epirubicin, 10 mL lipiodol) was injected under fluoroscopic control. Computed tomographic (CT) scanning was performed the next day to determine uptake of lipiodol by the tumor tissue. Embolization was repeated every 6–8 weeks until complete devascularization was achieved.

Assessment of Response

Response before LT

Patients were followed up after complete devascularization by 4-phasic CT scans every 3 months. Response was graded into 4 groups according to RECIST (Response Evaluation Criteria in Solid Tumors) criteria,25 including modifications as published in the EASL guidelines24: complete response (CR; no vital tumor tissue visible), partial response (PR; regression of viable tumor of at least 30%), stable disease (neither PR nor progressive disease), and progressive disease (>20% increase in tumor size). Embolization was resumed only if hypervascularized vital tumor tissue was again visible on CT scan.

Response after LT

In the pathology laboratory, liver explants were sectioned into 1-cm slices for detecting tumor or nodules. All tumors and suspicious nodules were excised, microscopically sectioned into 4-μm sections, stained with hematoxylin and eosin, and examined under the microscope for malignancy.

Immunosuppression

Immunosuppressive therapy for patients after LT (n = 106) consisted of a triple-drug regimen of cyclosporine A or tacrolimus in combination with corticosteroids and either azathioprine or mycophenolate mofetil. Because of poor renal function, some patients received an interleukin-2 receptor antagonist instead of a calcineurin inhibitor in the early postoperative period. Corticosteroids were gradually tapered and discontinued within 3 months. Azathioprine or mycophenolate mofetil was continued for 1 year after LT unless contraindicated.

Patient Follow-up

All patients were followed up weekly in the outpatient clinic for the first month after discharge from the hospital. The frequency of outpatient clinic visits thereafter varied according to the patients' condition and types of complications. Screening for tumor recurrence was done by measurements of AFP as well as sonography every 3 months. A routine CT scan of the abdomen and chest was performed every year, and additional imaging techniques (bone scan, magnetic resonance imaging) were used if HCC recurrence was suspected. No adjuvant chemotherapy after LT was administered to any patient.

Statistical Analysis

Baseline characteristics of the patients are expressed as mean ± SD or median and range. Comparisons between groups were performed by χ2 test. Survival probabilities were analyzed by the Kaplan-Meier method combined with the log rank test. The intention-to-treat analysis included all patients treated with TACE while waiting for LT (embolization group, n = 116). Additionally, an analysis of patients who underwent LT (transplant group, n = 106) was performed.

All calculations were performed by SPSS version 11.0 for Windows (SPSS, Chicago, IL).

RESULTS

Response to TACE of Patients on the Waiting List for LT

Between 1994 and 2004, the 116 patients in the embolization group were treated with TACE while waiting for LT. Of these, 10 were removed from the waiting list because of tumor progression, and 106 underwent transplantation. According to the last radiologic assessment before LT or dropout, 33 patients (28.4%) showed no sign of vital tumor (CR), 66 patients (56.9%) had tumor necrosis of >30% (PR), 3 patients (2.6%) were stable, and 13 patients (11.2%) had progressive disease despite TACE.

Among the 10 patients who dropped off the waiting list, 2 had disease that met the Milan criteria, an additional 4 had disease that was outside the Milan criteria but that met the UCSF-expanded criteria, and 4 had disease that failed to meet either of these criteria. Thus, the dropout rate was 2.9, 12.1, and 26.7% for the Milan criteria group, the UCSF-expanded group, and the downstaged group, respectively (P = 0.08 Milan vs. UCSF-expanded; P = 0.009 Milan vs. downstaged group). AFP levels at the time of diagnosis were normal in 2 patients, mildly increased in 4 patients, and >400 ng/mL in 4 patients.

Patients whose illness responded to TACE who had no signs of vital tumor on CT scan (CR) and those with PR had actuarial survival rates at 1, 2, and 5 years after tumor diagnosis of 100, 93.2, and 85.7% and 93.8, 83.6, and 66.2%, respectively. Compared with patients with PR, there was a trend toward slightly longer survival for patients with CR, but this did not reach statistical significance (P = 0.08). Patients with stable or progressive disease did significantly worse than those with CR or PR (P < 0.01), with actuarial survival rates of 82.4, 50.7, and 19.3% at 1, 2, and 5 years after tumor diagnosis (Fig. 1, left).

Figure 1.

Intention-to-treat analysis of embolization group (n = 116).

The intention-to-treat analysis (irrespective of TACE response) showed that patients whose disease was within the Milan criteria had actuarial survival rates at 1, 2, and 5 years of 97.0, 88.4, and 70.3%, compared with 94.0, 75.8, and 65.9% for the UCSF-expanded group (P = 0.08) and 80, 65.5, and 24.5% for the downstaged group (P < 0.001 vs. Milan; P = 0.10 vs. UCSF-expanded).

Milan criteria group

Similar results were found in patients whose disease fulfilled the Milan criteria. Patients with CR demonstrated slightly longer survival rates at 1, 2, and 5 years (100, 94.7, and 89.5%) compared with PR (97.1, 91.3, and 73.3%); however, this difference did not reach statistical significance. Patients with stable or progressive disease had a 1-year survival of 100%, but this dropped to 40% after 2 years, and no patient was alive after 5 years. These patients did significantly worse than those with CR or PR (P < 0.01) (Fig. 1, middle).

UCSF-expanded group

There was no significant difference in survival between patients who responded to therapy and those who did not in patients whose disease exceeded the Milan criteria but fitted into the UCSF-expanded criteria. CR had a 1-, 2-, and 5-year survival of 83.3, 66.6, and 66.6%, compared with 100, 80.9, and 63.7% in PR and 66.7, 50, and 25% in patients with stable or progressive disease (Fig. 1, right).

Downstaged group

One-, 2- and 5-year survival rates among patients with advanced-stage tumor who underwent TACE and who were in the initial list for LT were 80, 66.7, and 25%, respectively. The number of patients was too small to analyze them in separate groups according to TACE response. However, compared with patients whose disease met the Milan criteria, the overall patient survival was shorter, but there was no difference compared with those within the UCSF-expanded group (data not shown).

Response to TACE of Patients Who Underwent LT

In addition to the intention-to-treat analysis, posttransplant survival was calculated in relation to pre-LT response to TACE. The response of 106 patients comprising the transplantation group was analyzed. Patients with CR and PR to TACE demonstrated 1-, 2-, and 5-year survival rates after LT of 89.1, 85.1, and 85.1%, and 88.6, 77.4, and 63.9%, respectively. The difference between both groups showed a trend toward slightly longer survival for patients with CR but did not reach statistical significance (P = 0.1). Patients whose disease did not respond to therapy and those with progressive disease had survival rates of 68.6, 51.4 and 51.4% at 1, 2, and 5 years after LT. They did significantly worse than those with CR or PR (P = 0.02) (Fig. 2, left).

Figure 2.

Survival after liver transplantation (n = 106).

Milan criteria group

Patients whose disease met the Milan criteria showed similar differences. The 1-, 2-, and 5-year survival rates of CR according to the last CT scan before LT were 89.2, 89.2, and 89.2% and for PR were 94.4, 91.1, and 73.4%, respectively. Patients without response or those with progression had a 1-year survival of 37.5%. This group did significantly worse compared with either CR (P = 0.04) or PR (P < 0.01) (Fig. 2, middle).

UCSF-expanded group

Survival rates at 1 and 2 years among patients with CR or PR were 83.3 and 66.6%, and 85.2 and 70.2%, respectively. All patients with CR underwent transplantation <5 years ago; therefore, data are not yet available, but in patients with PR, the 5-year survival after LT was 64.2%. There were only 2 patients in the group of stable or progressive disease, a fact that prevented statistical analysis (Fig. 2, right).

Histological Assessment of Explanted Livers

Explanted livers after LT were histologically examined for viable tumor tissue and extent of necrosis induced by TACE. Comparison between initial stage of tumor and explant histology revealed that the disease of 10 of 63 patients believed to have met Milan criteria had, in fact, failed to meet them (understaging rate 15.9%). Two of these 10 patients experienced CR and 4 experienced PR; the tumors of 4 did not respond to TACE.

In the group of those initially listed for LT, 29 patients (27.4%) treated with TACE experienced complete tumor necrosis, whereas 77 patients (72.6%) showed remaining tumor tissue on explant histology. Survival rates at 1, 2, and 5 years after LT were 85.7, 85.7, and 79.9% for patients with complete necrosis, and for those with remaining tumor tissue, the survival rates were 88.2, 74.5, and 63.5%. Although the survival rates diverged with increasing time from LT, this did not reach statistical significance (P = 0.2).

Recurrent HCC After LT

Fifteen (14.2%) of 106 patients who underwent LT experienced a recurrence of HCC. All of them were men. Five of these patients had disease that met the Milan criteria, 7 patients were in the UCSF-expanded group, and 3 had advanced-stage tumor. Therefore, the recurrence rate was 7.6% in the Milan criteria group, 21.2% in the UCSF-expanded group, and 26.7% in the downstaged group. This indicates that patients within the UCSF-expanded and downstaged groups are at a significantly higher risk of experiencing recurrent HCC than those in the Milan criteria group (P = 0.04 for both UCSF-expanded vs. Milan and downstaged group vs. Milan). Survival rate dropped significantly in patients with recurrent HCC (P < 0.01).

Patients whose disease met the Milan criteria whose tumor showed complete necrosis in the explant histology had a longer recurrence-free survival compared with those showing vital tumor tissue in histological examination, but this did not reach statistical significance (P = 0.06).

Side Effects of TACE

TACE was well tolerated by most patients. The most common complaints after TACE were pain, transient fever, nausea, and inguinal hematoma. Two patients developed a hepatic abscess after TACE, which resolved after antibiotic treatment alone in 1 patient and after drainage with a pigtail catheter and antibiotic treatment in the other.

AFP

AFP levels were determined at the time of diagnosis and at all follow-up visits. Only 8.6% of the patients showed substantially increased AFP levels (>400 ng/mL), whereas 45.2% had moderately increased levels (11–400 ng/mL) and 46.2% had AFP levels within normal ranges (<11 ng/mL). The median AFP level was 15 ng/mL (range 1.5–4,566 ng/mL).

Patients with normal AFP at the time of diagnosis had a far better outcome than those with moderately or substantially increased levels. This was found to be true in the intention-to-treat embolization group (n = 116, P < 0.05) and the LT group (n = 106, P < 0.05), regardless of whether disease met the Milan criteria (n = 68, P < 0.05) or the UCSF-expanded criteria (n = 33, P < 0.05).

Table 2 lists the follow-up values of AFP in patients with AFP values >100 ng/mL. In patients with AFP >100 ng/mL at the time of diagnosis, patients with CR had a median reduction in AFP from 119 to 16 ng/mL and patients with PR had a median reduction from 842 to 75 ng/mL, but patients with progressive disease had a median increase from 932 to 13,092 ng/mL.

Table 2. Follow-up of Increased Alfa-Fetoprotein (AFP) Values
PatientAFP at diagnosis (ng/mL)AFP at LT or dropout (ng/mL)No. of TACE sessionsResponse to TACEUnderlying diseaseTransplant criteriaRecurrence
  1. Abbreviations: LT, liver transplantation; TACE, transarterial chemoembolization; PR, partial response; HCV, hepatitis C virus; CR, complete response; HBV, hepatitis B virus; UCSF-expanded, expanded University of California at San Francisco criteria.

1109443PRHCVMilanNo
2110163CRHCVMilanNo
31191272CRHCVMilanNo
4122461PRHBVUCSF-expandedYes
5144232PRHCVMilanNo
61771043PRAlcoholMilanYes
7353263PRCryptogenicMilanNo
84264252PRHCVDownstagedYes
95389276PRHBVUCSF-expandedYes
105716263PDHCVUCSF-expandedDropout
116997704PRHCVMilanNo
1272454CRHemochromatosisMilanNo
1393214,8246PDHCVMilanDropout
149606481PRHBVMilanNo
151,282102PRHCVMilanNo
161,91113,0926PDHCVUCSF-expandedDropout
173,23746PRHCVDownstagedNo
184,56616,7005PRAlcoholMilanNo

Response of Patients Depending on Date of Diagnosis

On the basis of our previous report regarding pre-LT TACE in patients with HCC,20 we specifically addressed the question of whether there was any difference in patients within the previously published observation period and those outside of it. There was no significant difference in response either to TACE or survival (data not shown).

DISCUSSION

LT has emerged as the treatment of choice for carefully selected patients with HCC and liver cirrhosis. The outcome for patients with HCC meeting specific selection criteria is excellent and comparable to that in recipients with nonmalignant liver diseases.6, 11 However, tumor progression exceeding acceptable selection criteria while waiting for LT frequently results in patients being removed from the list.12, 14 Commonly used options for slowing down tumor progression before LT include TACE and local ablation with radiofrequency or alcohol. Their effectiveness in inducing tumor necrosis, thus providing temporary control of tumor spread, makes them ideal bridging procedures to LT. Several studies, including our own, have shown the favorable results of TACE before LT in terms of tumor control.15–20 Reports on its effect on post-LT outcome and rate of recurrence, however, are conflicting. The aim of the present study was to examine the effect of TACE before LT on tumor recurrence and patient survival.

In this study, the overall survival of LT patients who had been treated with TACE preoperatively was comparable to the recently reported 2004 United Network for Organ Sharing data.26 The overall recurrence rate in our transplant cohort was 14.2%, but patients whose disease met the Milan criteria did markedly better, with a recurrence rate of only 7.6%, whereas those whose disease exceeded Milan criteria but that fitted the UCSF-expanded criteria had a far higher recurrence rate of 21.6%, almost comparable to that seen in the downstaged group (26.7%). Compared with the findings of Yao et al.,11 our cohort was followed up for a longer period (mean follow-up in our cohort 3.1 years after LT vs. 2 years in the UCSF-expanded study). Several patients developed recurrence of disease only after 2 years and would have been missed in a shorter follow-up period.

We detected a lower incidence of recurrent HCC in patients with complete tumor necrosis either on last CT scan or on explant histology. However, these differences did not reach statistical significance, probably as a result of the relatively small number of patients with recurrence.

In our study, 2 patients (2.9%) whose disease met the Milan criteria and 4 (12.1%) whose disease met the UCSF-expanded criteria had to be removed from the waiting list because of tumor progression during a median waiting time of 9 months. Our numbers were far lower than those in the report of Yao et al.,27 who showed cumulative dropout probabilities of 7.2% at 6 months and 37.8% at 12 months of waiting time without any adjuvant therapy in patients who met their proposed expanded criteria.

Increased waiting time was associated with a higher dropout rate that resulted from tumor progression. Between 1994 and 1999, the median waiting time was 6.8 months (range 0–24 months), and no patient had to be removed from the waiting list because of tumor progression, whereas the waiting time increased to 9.8 months (range 1.23–25 months) between 2000 and 2004. During this period, 8 patients (12.5%) were removed from the list (4 from the UCSF-expanded group and 4 from the downstaged group, none of whom had disease that met the Milan criteria) because of tumor progression exceeding the selection criteria of our institution. Because increasing time from listing to LT demonstrates the major risk factor and predictor for dropout,12 TACE or other locoablative therapies have been considered as a means to prevent tumor progression, particularly if the waiting time is expected to exceed 6 months. However, the efficacy of pre-LT locoablative therapies to impede tumor progression while on the waiting list has not yet been proven in randomized, controlled trials.

Patients whose disease met the Milan criteria who had a CR or PR to TACE showed a far better outcome after LT than those with stable disease or progression. This does not apply to disease exceeding the Milan criteria, regardless of whether it fit into the UCSF-expanded criteria or not. However, the lack of effect of response to TACE within the UCSF-expanded group may be the result of the small sample size. Patients presenting with advanced tumors at the time of listing for LT but who then respond to TACE have no marked survival benefit in the long-term outcome after LT. These findings lead to a discussion about the effect of tumor biology and whether response to TACE reflects tumor biology rather than therapeutic efficacy. This issue has been discussed before, most recently by Otto et al.28 In their cohort, response to TACE seemed to be more effective in terms of long-term outcome rather than initial tumor size. This contradicts our own data; we found that patients with disease exceeding the Milan criteria showed no survival benefit, although tumor regression had been achieved by TACE. The main problem remains the lack of criteria incorporating both tumor size and tumor biology, such as tumor grading and microvascular invasion.

The progression that can be tolerated during the waiting time has not been defined yet. In our study, patients whose tumor exceeded the UCSF-expanded criteria were removed from the waiting list. But this is an arbitrary limit. For instance, a patient with a tumor diagnosed at a diameter of 2 cm but growing to 5 cm during the waiting time probably has a worse prognosis than a patient with a tumor diagnosed at 4.5 cm but with a small increase to 5 cm. Improved imaging methods, including new serum markers, would be helpful in sorting out these issues. As generally known, AFP is a poor tumor marker in HCC. In our cohort, only 8.6% of patients had diagnostically increased AFP (>400 ng/mL), and only 45.2% showed moderately increased levels (11–400 ng/mL). Up to now, there have been no guidelines in how to incorporate AFP values in the assessment of tumor progression. In our cohort, most cases of tumor progression leading to exclusion from LT showed a concomitant AFP increase. In patients with increased AFP at the time of diagnosis, none with CR on CT scan had a further increase in AFP. However, an increase in AFP should prompt additional investigations for tumor progression, even if imaging studies indicate a stable tumor disease.

In conclusion, our study suggests that tumor stage and waiting time are the most important factors influencing the outcome of patients with HCC after LT. Patients whose disease fulfills the Milan criteria seem to benefit from TACE if the tumor is completely devascularized or shows at least a 30% decrease in size before LT. This was not true for patients whose disease did not meet this criteria. To answer the question of whether or not to perform bridging therapies such as TACE while waiting for LT, randomized, controlled trials are needed.

Acknowledgements

We thank Rajam Csordas for critical reading and editorial assistance.

Ancillary