By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
Hepatocellular carcinoma (HCC) is 1 of the 5 most common malignancies worldwide, and its incidence is increasing in Western countries.1 For patients with HCC and cirrhosis, liver transplantation (LT) represents the treatment of choice and provides excellent oncological results and a cure for cirrhosis. However, not all patients with HCC and cirrhosis can undergo transplantation because of the scarcity of liver donors. HCC patients on the waiting list for transplantation can experience tumor growth beyond the accepted criteria for LT; for this reason, the practice of treating HCC patients with hepatic resection or locoregional therapies before they are placed on the waiting list or while they are waiting has gained favor and is now the standard of care in most transplant centers.2-5 Radiofrequency ablation (RFA), percutaneous ethanol injection (PEI), and transarterial chemoembolization (TACE) have considerably improved in the last decade, and they can have a positive impact on tumor growth control by reducing the risk of dropout or down-staging advanced disease.2-6 Similarly, the great improvements in diagnostic techniques and surveillance schedules have led to earlier diagnoses and better accuracy, and this has resulted in the increased curability of liver tumors.
In an effort to continue to balance justice and utility, a national conference on LT for HCC, gathering transplant physicians, surgeons, and policy makers, was held in 2009 in the United States. The discussed issues also included a proposal for performing procedures for select candidates with favorable biology markers (eg, responses to bridge therapy and slower tumor growth).7 Recent evidence has shown that an unsatisfactory response to bridge therapy could be a criterion for prioritizing patients with a high risk of dropout due to tumor progression8, 9 when the overall strategy is inspired by the survival benefit principle. On the other hand, if higher priority is advocated for these patients, we should investigate whether the priority for other patients is currently too high in light of their real dropout risk. In 2004, the allocation policy was changed to eliminate any increased priority for stage 1 HCC candidates; similarly, there is probably room for reducing the priority of some select stage 2 candidates according to their responses to bridge therapy. If this assumption is true, the result will be an increased probability of transplantation for the remaining HCC and non-HCC patients, and this will result in a reduced wait-list dropout rate. The main aim of this study was to investigate this issue.
Abbreviations: AASLD, American Association for the Study of Liver Diseases; AFP, alpha-fetoprotein; CEUS, contrast-enhanced ultrasound; CT, computed tomography; EASL, European Association for the Study of the Liver; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; MR, magnetic resonance; MVI, microscopic vascular invasion; PEI, percutaneous ethanol injection; RECIST, Response Evaluation Criteria in Solid Tumors; RFA, radiofrequency ablation; TACE, transarterial chemoembolization; TNM, tumor-node-metastasis; UNOS, United Network for Organ Sharing.
PATIENTS AND METHODS
All consecutive HCC patients referred to our institution and listed for LT between January 1, 2005 and December 1, 2010 were included in this analysis. For the study group, we considered patients known to have HCC and cirrhosis at the time of their listing and patients diagnosed with HCC after their listing for liver cirrhosis. The efficacy of bridge therapies was assessed 3 months after the treatment (either before the patient was listed or when the patient was already on the waiting list) with computed tomography (CT), magnetic resonance (MR) imaging, contrast-enhanced ultrasound (CEUS), or a combination of these techniques. In particular, MR was the preferred imaging technique for patients who underwent TACE.6 HCC patients who underwent transplantation or died without an efficacy assessment of the bridge therapy were excluded from this analysis.
The preoperative diagnosis of HCC was based on the latest guidelines of the European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Diseases (AASLD).10, 11 All patients were subjected to at least 2 dynamic imaging techniques; nodule biopsy was performed whenever it was technically feasible for patients with atypical or noncoincidental contrast-enhancement findings. Bridge therapies were proposed and tailored to each patient during multidisciplinary meetings, which always included transplant surgeons, hepatologists, and interventional radiologists, within the framework of the EASL and AASLD guidelines for the treatment of HCC. When a decision was being made for each patient, the aim was always a complete tumor cure; thus, the most radical treatments proposed by the guidelines were preferred, but a concurrent goal was to prevent a patient's exposure to excessive risks of treatment-related complications, including irreversible rapid liver failure. Consequently, consideration was always given to the technical feasibility of the various treatment modalities (including the number, size, and location of the nodules), the comorbidities, and the severity of the liver dysfunction. In accordance with a previously published policy,5, 6 liver resection was first proposed for patients with a single nodule or 2 nodules and preserved liver function; RFA was used for patients with deep and/or multiple nodules and preserved liver function in the absence of ascites; PEI was preferred when nodules were adjacent to a major vessel; and TACE was adopted for patients with multiple nodules and clinical signs of mild to moderate liver dysfunction.6 Patients with an advanced degree of liver dysfunction [ie, a Model for End-Stage Liver Disease (MELD) score > 20 or Child-Pugh class C (including refractory ascites or frank jaundice)] were not subjected to any bridge therapy because of the excessively high risk of liver failure with the treatment. In our transplant program, HCC patients were considered transplantable if they did not exceed the Milan criteria (a single nodule ≤5 cm or up to 3 nodules with each ≤3 cm and no major vascular invasion or distant metastases), and this included the possibility of a down-staging protocol. Since January 1, 2003, 53 patients initially not meeting the Milan criteria have in fact been included on the waiting list after a down-staging protocol, as previously reported.12 While they were on the waiting list, all HCC patients were followed with liver function tests, serum alpha-fetoprotein (AFP) measurements, and CT scans, MR imaging, and/or CEUS every 3 months as previously described. Tumors were staged on the basis of preoperative imaging according to the United Network for Organ Sharing (UNOS) tumor-node-metastasis (TNM) classification.13
The response to bridge therapy was recorded according to the modified Response Evaluation Criteria in Solid Tumors (RECIST) and EASL criteria.9, 10, 14 The modified RECIST criteria were chosen because the original RECIST criteria had been shown to be misleading with other anticancer drugs (eg, molecular-targeted therapies) and other therapeutic interventions.15 In the case of HCC, recent studies have shown a poor correlation between the clinical benefits provided by new agents (eg, sorafenib) and by locoregional interventional therapies and the conventional methods of response assessment.16, 17 Consequently, the local response to treatment was measured according to the radiologist's estimate of the reduction in the viable tumor volume, which was recognized from unenhanced areas on CT scans, MR images, or CEUS scans. Responses to bridge therapy were thus classified as follows: (1) a complete response (the disappearance of any intratumoral arterial enhancement in all target lesions), (2) a partial response [a decrease of at least 30% in the sum of the diameters of viable (enhanced in the arterial phase) target lesions, with the baseline sum of the diameters of target lesions used as a reference], or (3) a poor response [an increase of at least 20% in the sum of the diameters of viable (enhanced) target lesions, with the smallest sum of the diameters of viable target lesions recorded since the beginning of treatment used as a reference]. All patients underwent a retreatment of the residual vital tumor whenever this was possible. HCC candidates with a partial or poor response (or no response) also included patients with a diagnosis of HCC who were not subjected to bridge therapies (ie, any patients not qualifying for either a partial response or progressive disease).
Our allocation policy has already been published.18 In short, since March 1, 2003, patients were listed according to their MELD scores, and an extra score based on the HCC stage, the down-staging protocol, and every month on the list was added to the native MELD score. We also applied the January 2004 UNOS modification, which removed all increased priority for stage 1 tumors. The removal of HCC candidates for death/sickness severity or tumor progression were counted as dropout events. Criteria for exclusion from the waiting list due to tumor progression included evidence of gross neoplastic vascular invasion, tumor progression beyond the Milan criteria, and/or evidence of extrahepatic or lymph node metastases. Portal thrombosis was not an exclusion criterion when it could be proved to be nonneoplastic.6 Patients whose status was inactive but who were not removed from the list were considered at risk and were included in the analysis.19 The time to event (dropout or transplantation) was calculated from the day of listing with the HCC diagnosis; for living candidates, the survival calculation ended at the last follow-up visit (they were censored). The collection of follow-up data ended on February 1, 2011.
The histological examination of the explanted liver included an assessment of tumor necrosis,6 the tumor grade, and the presence of microscopic vascular invasion (MVI). The tumor grade was scored with the modified nuclear grading scheme outlined by Edmondson and Steiner.20, 21 In all cases, the tumor grade was defined by the poorest degree of differentiation identified within the tumor by a pathological analysis of the entire specimen. MVI was defined as the presence of tumor emboli within the central hepatic vein, the portal vein, or the large capsular vessels.22 The study protocol received a priori approval by the appropriate institutional review committee.
Continuous variables are expressed as medians and ranges, and differences between subgroups were calculated with the Mann-Whitney test or the Kruskal-Wallis test (whichever was appropriate). Differences between subgroups for categorical variables were calculated with Person's chi-square test, Fisher's exact test, or the Freeman-Halton extension of Fisher's test. The wait-list survival analysis was performed with a competing risk method.23 When data come from patients who experience a single event and censored individuals, a nonparametric estimate of the cumulative incidence can be obtained with the Kaplan-Meier method. With this approach, the survival time of an individual (or the time at which a subject experiences an event) is assumed to be independent of a mechanism that would cause the patient to be censored. On the waiting list for LT, a patient may experience an event other than the one of interest (a competing risk event such as transplantation), and this can alter the probability of experiencing the event of interest (dropout). In this setting, the calculation of the cumulative incidence of the specific event of interest (dropout) is more informative23, 24; thus, to calculate the cumulative incidence function for an event of interest, we must appropriately account for the presence of the competing risk event. The cumulative incidence function was calculated according to the method described by Gray24 with the cmprsk package of R version 2.12.0 (R Foundation for Statistical Computing). Specifically, the cuminc function provided statistical tests, estimates of the cumulative incidences, and variance estimates for different groups. An inference analysis was added with the crr function of R, which was used to model the hazard of the cumulative incidence function with respect to covariates. Candidates alive at the end of study were censored. Survival after LT was calculated with the Kaplan-Meier method, and differences between subgroups were compared with the log-rank test; a multivariate Cox regression model was finally applied to identify independent predictors of posttransplant tumor recurrence and patient survival. A P value < 0.05 was considered statistically significant for all analyses.
Three hundred thirty-seven patients with cirrhosis and a diagnosis of HCC were listed for LT during the study period; 22 were not included in the analysis because they underwent transplantation (11 patients), died (5 patients), or were still on the waiting list (6 patients) without an efficacy assessment of the bridge therapy (the time to event was less than 3 months). Thus, the final study group consisted of 315 patients. The baseline characteristics of the study population are reported in Table 1. During a median follow-up of 17.8 months (range = 3-53 months) after HCC treatment, 176 patients (55.9%) underwent LT (median time to transplantation = 10 months, range = 3-45 months), 33 (10.5%) died, 3 (0.95%) were removed from the waiting list because they were too sick (multiorgan failure), and 52 (16.5%) were removed because of tumor progression beyond the Milan criteria. Tumors were treated in most cases (293 patients or 93.0%). A complete response at the 3-month visit was observed in 144 of 293 treated patients (49.1%), whereas a partial response or no response was observed in the remaining 149 patients (50.9%). If we include the 22 patients not subjected to bridge therapies, 171 of the 315 patients (54.3%) had a partial response or no response.
Table 1. Baseline Characteristics of the Study Population
Candidates (n = 315)
NOTE: Continuous variables are reported as medians and ranges. HCC patients with T3-T4a tumors were subjected to a down-staging protocol.
Male sex [n (%)]
Diagnosis [n (%)]
HBV and HCV
Biochemistry at listing
International normalized ratio
AFP >400 ng/mL [n (%)]
MELD score at listing
6-10 [n (%)]
11-14 [n (%)]
15-20 [n (%)]
>20 [n (%)]
AFP at diagnosis (ng/dL)
AFP at diagnosis >400 ng/mL [n (%)]
Maximum size at diagnosis (cm)
Tumor number at diagnosis [n (%)]
UNOS TNM stage at diagnosis [n (%)]
Bridge therapy adopted after HCC diagnosis [n (%)]
TACE plus percutaneous therapies
Percutaneous therapies only
Factors Associated With Dropout
The dropout rates for the whole study population and the factors related to different dropout probabilities for patients on the waiting list are reported in Table 2. The dropout risk was related to the MELD score at listing (P = 0.032), the UNOS TNM stage at diagnosis (P = 0.041), the use of bridge therapy (P = 0.018), and the response to bridge therapy (P < 0.001), whereas the remaining variables showed no significant impact on dropout rates (P > 0.05 in all cases). A trend toward a higher risk of dropout was observed in patients with AFP levels > 400 ng/mL at listing (P = 0.086). The observation that the 22 patients who were not subjected to bridge therapy had an increased risk of dropout reflects their median MELD score of 20 (range = 9-29), which was significantly higher than the median MELD score of those subjected to bridge therapies (13, range = 7-32, P = 0.001). The cause of dropout for these patients was death or sickness severity in 6 of 8 cases and tumor progression in the remaining 2 cases. Moreover, the down-staging procedure was not significantly related to the risk of dropout for patients already within the Milan criteria at the time of their HCC diagnosis (P = 0.360). Table 3 reports the results from an inference analysis of variables significantly related to dropout in the univariate competing risk analysis: the response to bridge therapy was the sole variable independently related to the event (P = 0.025). The removal of this variable resulted in the adoption of bridge therapy as the second variable independently related to dropout, whereas the MELD score and the UNOS TNM stage played minor roles.
Table 2. Competing Risk Analysis of Removal for Death/Sickness Severity or Tumor Progression Beyond the Milan Criteria in the HCC Study Population
Dropout Probability [% (Standard Error)]
Serum AFP at listing
MELD score at listing
Serum AFP at diagnosis
Tumor diameter at diagnosis
Tumor number at diagnosis
UNOS TNM stage at diagnosis
Response to bridge therapies
Partial response or no response
Table 3. Multivariate Competing Risk Analysis of Removal for Death/Sickness Severity or Tumor Progression Beyond the Milan Criteria
95% Confidence Interval
The ranges for the MELD classes were 6 to 10, 11 to 14, 15 to 20, and >20.
The response to bridge therapies was removed from the model.
Relationships between clinical and tumor factors and responses to bridge therapy are reported in Table 4. The variables that were found to be significantly related to the response at the 3-month visit were the AFP level at the diagnosis of HCC (P = 0.026), the tumor number (P = 0.001), the use of a down-staging procedure (P = 0.005), and the type of bridge therapy (P = 0.001). The use of hepatic resection (31 cases) and percutaneous ablation (42 cases) with or without TACE (potentially curative treatments) was significantly related to a higher prevalence of complete tumor responses (73 of 121 patients treated with potentially curative therapies or 60.3%) in comparison with the prevalence observed for patients who received the other treatments (71 of 194 patients or 36.6%, P = 0.001). The remaining variables were not significantly related to the response to bridge therapy. Notably, the presence of a single tumor and the type of bridge therapy were strictly related. In fact, 75 of 169 patients (44.4%) with a solitary tumor were subjected to hepatic resection or percutaneous ablation with or without TACE; this proportion was significantly higher than the proportion observed for patients with multiple nodules (46 of 146 patients or 31.5%, P = 0.019). Forty-nine of 70 patients (70.0%) with a solitary tumor who were treated with potentially curative treatments showed a complete response at the 3-month visit; this dropped to 47 of 99 patients (47.5%) when other treatments were adopted (P = 0.004). Nineteen of 43 patients (44.2%) with multiple tumors who were treated with potentially curative treatments showed a complete response at the 3-month visit; this dropped to 29 of 103 patients (28.2%) when other treatments were adopted (P = 0.060). Notably, 4 patients who underwent hepatic resection developed postoperative liver failure and subsequently underwent transplantation, but no patient became too sick for transplantation because of the bridge therapy.
Table 4. Clinical and Tumor Characteristics and Responses to Bridge Therapy at the 3-Month Assessment of Treatment Efficacy
Complete Response (n = 144)
Partial Response/No Response (n = 171)
NOTE: Continuous variables are reported as medians and ranges.
The dropout risk of HCC patients with respect to the TNM stage, the response to bridge therapy, and the down-staging procedure is depicted in Fig. 1. The dropout risk for patients with T1 tumors was very similar to the risk observed for patients with T2 tumors and a complete response to bridge therapy (P = 0.964). Conversely, the dropout risk for patients with T2 tumors and a partial response or no response to bridge therapy was significantly higher than the risk for both T1 patients (P = 0.001) and T2 patients with a complete response (P = 0.001). The dropout risk for patients who were subjected to a down-staging procedure (T3-T4a patients) was significantly higher than the risk for all T1 patients and T2 patients with a complete response (P = 0.024) and was lower than the risk for T2 patients with a partial response or no response to bridge therapy (P = 0.037). The median MELD scores were 13 (range = 7-32) for T1 patients, 12 (range = 7-27) for T2 patients, and 12 (range = 7-29) for down-staged patients (P = 0.228). The median MELD scores were 12 (range = 7-26) for T2 patients with a complete response to bridge therapy and 12 (range = 7-29) for T2 patients with a partial response or no response (P = 0.721).
Response to Therapy and Survival After Transplantation
The baseline characteristics of the 176 patients who underwent LT are reported in Table 5. Fifty percent (88 patients) had a complete response to bridge therapies, and 25% (44 patients) had no viable tumor tissue according to the pathological examination of the explanted liver. Patients with a complete response showed a significantly higher prevalence of complete necrosis (29 of 88 cases or 33.0%) in comparison with patients with a partial response or no response (15 cases or 17.0%, P = 0.023). During the follow-up after transplantation (median = 24.3 months, range = 1-65 months), 18 of the 176 patients (10.2%) experienced tumor recurrence, and 34 (19.3%) died. Tumor recurrence was the cause of death in 13 patients (38.2%), whereas infections and hepatitis recurrence accounted for 20 cases (58.8%). One patient died from a cerebrovascular accident, and 5 patients were still alive with tumor recurrence. The 1-, 3-, and 5-year survival rates were 86.6%, 76.9% and 74.3%, respectively, and the 1-, 3-, and 5-year recurrence rates were 8.3%, 10.8%, and 12.8%, respectively. The response to bridge therapy was found to be significantly related to the recurrence rate (P = 0.017), as outlined in Table 6. Other clinical and histological variables that were found to be related to the recurrence rate were the histological presence of viable tumor tissue (P = 0.045) and the presence of MVI (P = 0.001). A multivariate Cox regression model confirmed MVI as the sole independent variable significantly associated with tumor recurrence (hazard ratio = 4.34, 95% confidence interval = 1.89-30.85, P = 0.010), whereas a partial response or no response to bridge therapy (hazard ratio = 2.45, 95% confidence interval = 0.80-7.55, P = 0.116) and the presence of tumor tissue (hazard ratio = 6.10, 95% confidence interval = 0.81-15.85, P = 0.206) lost their statistical significance.
Table 5. Baseline Characteristics of the Transplant Patients
Transplant Patients (n = 176)
NOTE: Continuous variables are reported as medians and ranges.
Histological grading and the presence of MVI were assessed only in nodules with viable tumor tissue.
Age at transplant (years)
Male sex [n (%)]
Diagnosis [n (%)]
HBV and HCV
Biochemistry at transplant
International normalized ratio
AFP > 400 ng/mL [n (%)]
MELD score at transplant
6-10 [n (%)]
11-14 [n (%)]
15-20 [n (%)]
>20 [n (%)]
Down-staging protocol: T3-T4a at diagnosis [n (%)]
Complete response to bridge therapy [n (%)]
Donor male sex [n (%)]
Donor age (years)
<18 years [n (%)]
18-39 years [n (%)]
40-49 years [n (%)]
50-59 years [n (%)]
≥60 years [n (%)]
Cerebrovascular accident as cause of donor death [n (%)]
Patients who had a complete response to bridge therapy showed a trend toward longer overall survival in comparison with patients who had a partial response or no response (P = 0.098), but the variables that most influenced posttransplant survival (see Supporting Table 1) were a hepatitis C virus (HCV)–positive status (P = 0.002) and the presence of MVI (P = 0.016); both were independently related to patient survival in a multivariate Cox analysis (HCV-positive status, hazard ratio = 2.24, 95% confidence interval = 1.12-6.86, P = 0.036; presence of MVI, hazard ratio = 5.19, 95% confidence interval = 2.01-38.89, P = 0.008). When the entire study population of 315 patients was reconsidered, the intention-to-treat survival proved to be significantly affected by the response to bridge therapy; in fact, patients with a complete response had 1-, 3-, and 5-year survival rates of 87.9%, 74.3%, and 66.4%, respectively, which were significantly higher than the rates of patients with a partial response or no response (68.3%, 47.6%, and 45.0%, respectively, P = 0.001).
The introduction of the MELD allocation system, which allows priority scores for patients with HCC, has resulted in consistent increases in the proportion of LT recipients with HCC; in the United States, more than 25% of donor livers are currently allocated to patients with HCC, and many of these patients have low MELD scores.7, 19 In our geographical region (Italy), the proportion of recipients with HCC is even higher (up to 45%) because HCC is much more common in this area.25 Consequently, a more equitable allocation policy that also considers possible alternative treatments for patients with HCC and their responses to therapy is highly endorsed.7 In an era of limited organ availability, better predictors of HCC prognosis are needed for the selection of appropriate transplant candidates and for the prioritization of HCC patients within and beyond the Milan criteria. Novel molecular data can change the approach to the diagnosis, staging, and prognostic assessment of HCC patients, but at present, these advances are still far from being introduced into clinical practice as daily tools for management and treatment selection.26, 27 The response to bridge therapy could represent a surrogate but extremely practical marker of tumor biology and could be helpful in the selection of candidates and prioritization for LT.
In 2004, the allocation policy was changed for patients with T1 HCC; similarly, this study suggests that there is probably room for reducing the priority of select T2 patients. The selection should be based on the response to bridge therapy, which has been proven (together with the MELD score) to be the main determinant of removal from the waiting list. This result is very interesting because it could lead to an increased probability of transplantation for the remaining HCC and non-HCC patients and, consequently, to reduced wait-list dropout rates. In particular, the finding that the dropout risk for T2 patients with a complete response to bridge therapy is similar to the risk observed for T1 patients suggests a reduction of the priority of these patients close to the priority currently adopted for T1 patients. This should not be considered a recommendation for removing any degree of priority from these patients; instead, we are simply suggesting a novel approach to be explored, perhaps through the application of simulation models, in order to investigate the benefits that could be obtained for the entire candidate population.28 The next step is the identification of patients who could probably be adequately treated with bridge therapy. The present results suggest that patients treated with potentially effective treatments for a single tumor (eg, hepatic resection or percutaneous ablation) are those most likely to experience a good response to bridge therapy and, consequently, a low dropout rate. However, the need to down-stage the tumor and the presence of an elevated AFP level at the diagnosis of HCC will reduce the probability of a complete response 2.5- and 6.0-fold, respectively. Consequently, the latter patients should probably not be considered candidates for whom a reduction in HCC priority could be sustained.
This study supports recent findings of different levels of dropout risk for HCC candidates according to their responses to bridge therapy,8, 9 but it also provides more useful and practical information. First, we correctly applied a competing risk analysis rather than the previously reported Kaplan-Meier survival estimate.8 In fact, it is well known that Kaplan-Meier analyses largely overestimate the risk of dropout because they cannot take into account the possibility of transplantation: patients who do not achieve the outcome of interest (dropout) during the observation period are censored, and this acknowledges that any of the outcomes are still possible. The Kaplan-Meier result for death is an estimate of what the death rate would be if all other endpoints were removed (eg, the discontinuation of the transplant programs); it may be appropriate for answering certain questions (eg, when a patient is being counseled about his or her prognosis without transplantation). Conversely, a competing risk analysis produces an estimate of death rates with the possibility of transplantation, and this results in a more appropriate metric.29 Previous conclusions from Kaplan-Meier analyses (ie, the observed tumor progression rate of T2 patients is lower than the expected rate with the current UNOS allocation policy and is strongly influenced by the response to bridge therapy) are not incorrect but have been further confirmed by this competing risk analysis. Second, in contrast to a previous competing risk analysis,9 we applied the Milan criteria as the acceptable criteria for transplantation. In fact, in the previous study, removal from the waiting list was adopted when tumor staging and nodule biopsy revealed macroscopic vascular invasion, extrahepatic spread, and poor differentiation at biopsy.9 Thus, the wider the accepted criteria for LT are, the lower the dropout rate is; in fact, the reported dropout rate was extremely low (5% 1 year after listing) and did not reflect the policies of most transplant centers worldwide. The results of our study support the evidence showing that the response to bridge therapy is strongly related to different dropout risks, but they set this finding in a clinical scenario in which the Milan criteria are the accepted criteria for LT.
Another important finding of this study is the relationship observed between the response to bridge therapy, tumor recurrence after LT, and intention-to-treat survival.8, 9 This finding could help with the refinement of the prognosis after listing and eventually with the identification of patients whose tumors exceed the Milan criteria but who have a good prognosis. One method for selecting this type of patient could be the incorporation of a time-from-diagnosis variable7; alternatively, a histological evaluation of specimens obtained by preoperative fine-needle biopsy could be used to identify tumors with good biology.9, 12 The adoption of the response to bridge therapy as a variable for the selection of candidates for transplantation seems to be reasonable because it appears to well reflect tumor aggressiveness and, consequently, posttransplant outcomes.2, 12, 30, 31 One additional point of interest of this study is that the results are in agreement with the recommendations of a national conference in the United States, which suggested an allocation system that incorporates the MELD score, tumor factors, and the time within the Milan criteria (our policy since 2003).18 This feature gives us a unique opportunity to analyze how the HCC dropout risk can be changed by the adoption of this policy versus the current policy in the United States. It should be noted that our dropout rate over time is much lower than the rate reported by the UNOS database.13, 19 Even when a Kaplan-Meier estimate was performed, the 3-, 6-, and 12-month unadjusted dropout fractions for our study population were 3.7%, 7.1%, and 25.5%, respectively (still below the values of 9.4%, 16.9%, and 31.8% reported by Freeman et al. in 2006,19 although not by much).
In conclusion, our results suggest that the priority of HCC candidates should be modulated with respect to the response to bridge therapy and that T2 patients who achieve a complete response to bridge therapy should have the same dropout risk as T1 patients. Consequently, there is probably room for reducing the priority of select T2 patients, and this could lead to a reduction of the wait-list removal rate for the remaining candidates awaiting transplantation.
The authors are indebted to the members of the Bologna Liver Oncology Group for their contributions to the clinical management of HCC patients at our institution: L. Bolondi, L. Gramantieri, P. Pini, F. Trevisani, P. Andreone, P. Caraceni, A. G. Gramenzi, M. Biselli, C. Sama, C. Serra, S. Berardi, Lenzi, G. Bianchi, G. Mazzella, A. Colecchia, M. R. Tamè, S. Brillanti, and F. Lodato (Department of Internal Medicine and Gastroenterology, Sant'Orsola-Malpighi Hospital, Bologna, Italy); E. Giampalma, A. Cappelli, M. Renzulli, and C. Mosconi (Radiology Unit, Sant'Orsola-Malpighi Hospital, Bologna, Italy); W. F. Grigioni and A. D. Grigioni (Pathology Unit, Addarii Institute of Oncology, Sant'Orsola-Malpighi Hospital, Bologna, Italy); G. Ballardini (Department of Internal Medicine, Rimini Hospital, Rimini, Italy); F. G. Foschi (Department of Internal Medicine, Faenza Hospital, Faenza, Italy); and M. Del Gaudio and M. Zanello (Liver and Multiorgan Transplant Unit, Department of General Surgery, Sant'Orsola-Malpighi Hospital, Bologna, Italy). The authors also thank Susan West for her assistance with the writing.