CE: M.S.; ANNSURG-D-17-02099; Total nos of Pages: 3; ANNSURG-D-17-02099 RESPONSE Response: ‘‘Which is the True Role of Bridging Therapies for HCC Patients Waiting for Liver Transplantation?’’ Reply: e very much appreciate the comments of Drs Lai, Vitale, Rossi, Cillo, and Lerut on behalf of the European Hepatocellular Cancer Liver Transplantation (ErHeCaLT) Study Group regarding our manuscript,1 examining the effect of pretransplant bridging locoregional therapy (LRT) on recurrence of hepatocellular carcinoma (HCC) after liver transplantation (LT) in patients within Milan criteria. By virtue of examining outcomes in listed HCC patients who actually received LT, the dataset from our US Multicenter HCC Transplant Consortium (UMHTC) could only be expected to assess the utility endpoint (ie, risk of post-LT HCC recurrence). A thorough intention-totreat analysis, as Lai et al allude to, would have allowed for a robust analysis of urgency (ie, pre-LT outcome) and transplant benefit as well; unfortunately, the UMHTC dataset does not currently include the waitlist outcomes for HCC patients who dropped out due to tumor progression or otherwise did not receive LT. Nonetheless, the results from our large, multicenter study add significantly to the growing body of evidence that a static HCC Model for End-stage Disease (MELD) allocation policy based on tumor size and number alone fails to take into consideration dynamic evaluations of an individual’s tumor biology that have been shown to significantly impact waitlist dropout and post-LT HCC recurrence. Our findings that an increasing number of locoregional treatments (which undoubtedly serves as a surrogate for radiographic assessments demonstrating persistent, progressive, or de novo lesions) negatively affect recurrence outcomes is consistent with numerous studies identifying radiographic and pathologic response to LRT as important determinants of both waitlist dropout and post-LT recurrence.2– 4 To this end, we are in complete agreement with Dr Lai and colleagues that inclusion of dynamic measures of tumor response must be included in any HCC allocation policy. W Disclosure: The authors have no conflicts of interest to declare. Copyright ß 2017 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0003-4932/16/XXXX-0001 DOI: 10.1097/SLA.0000000000002577 TO THE LETTER However, several points need to be clarified. While our UMHTC data unequivocally demonstrate the lowest risk of posttransplant HCC recurrence in recipients achieving a complete pathologic response, we are certainly not advocating for these patients to be prioritized for liver transplantation. Rather, similar to the recent publication of Lai et al5 examining the intention-totreat survival benefit of LT, we are drawing the parallel conclusion that HCC patients requiring an increasing number of pretransplant LRT, which may be viewed as a surrogate for a poor radiographic response to LRT, may have a low transplant benefit. Along these lines, we disagree with the statement that the US liver transplant selection policy for HCC patients allocates organs to the ‘‘earliest first’’ (ie, patients with compensated cirrhosis and CR). In October of 2015, in an attempt to mitigate the overprioritization of HCC recipients compared with non-HCC patients with end-stage liver disease, the Organ Procurement and Transplantation Network (OPTN) instituted a 6month mandatory waiting period before awarding MELD exception points, with capping of the MELD score at 34 for HCC recipients.6 More recently, the OPTN Liver and Intestinal Organ Transplantation Committee has approved the newest iteration of the HCC allocation policy that will allow MELD exception prioritization of HCC recipients presenting with beyond OPTN T2 criteria tumors (ie, beyond Milan criteria) who demonstrate favorable response to LRT and are ‘‘down-staged’’ to T2 criteria. Similarly, candidates within T2 criteria who have serum alphafetoprotein (AFP) greater than 1000, who previously were automatically awarded MELD exception points, will now be required to demonstrate a reduction of AFP to less than 500 after LRT before receiving exception points.7 In perhaps the most liberal policy to maximize ‘‘transplant benefit’’ for patients with HCC, the UNOS Region 5 Board recently approved an HCC ‘‘allcomers’’ protocol, where patients who have tumors of any size, provided there is no vascular invasion and AFP <500, will be given an opportunity to receive MELD exception points if they are successfully down-staged by LRT to within Milan criteria and demonstrate stability for 6 months.8 Taken in aggregate, it is clear that US liver allocation policy is actively adopting strategies to maximize transplant benefit for HCC recipients that is not based on the ‘‘earliest first’’, while preserving ‘‘equity’’ for noncancer patients awaiting LT. Finally, we appreciate Lai et al’s concerns regarding the statistical validity of our finding that locoregional treatment itself was an independent predictor of post-transplant Annals of Surgery Volume XX, Number XX, Month 2017 HCC recurrence in the subset of patients not achieving a complete pathologic response. This finding was unexpected, and quite intriguing. We generally agree that the multivariable analysis performed does not necessarily guarantee a complete elimination of the selection bias inherent to observational study data. To alleviate this concern, we performed an inverse probability of treatment weighted (IPTW) multivariable survival analysis, which we agree is the best methodology to evaluate the true effect of the locoregional treatment short of the gold-standard randomized controlled trial. In brief, a logistic regression analysis was used to model the probability of receiving LRT among patients not achieving complete pathologic response as a function of numerous potential confounders including age, sex, liver disease etiology, MELD, neutrophil-lymphocyte ratio, pretransplant AFP, pathologic maximum diameter, pathologic grade, vascular invasion, and tumor stage. The propensity score was the probability of receiving LRT after conditioning on these covariates. Each patient was assigned a weight corresponding to the inverse of the propensity score under the logistic model. Comparison of the propensity scores in treated and untreated patients confirmed that there was sufficient overlap in the propensity score to allow proper adjustment. Comparison of all potential cofounders between HCC recipients with and without LRT in both the original (unweighted) and weighted sample confirmed that the propensity matching successfully balanced the 2 groups (Table 1). Utilizing this propensity matching in the weighted sample, our IPTW competing risk Cox regression model found that the adjusted hazard ratio for the effect of LRT on post-LT recurrence was highly statistically significant (hazard ratio 1.3089, 95% confidence interval 1.04–1.65, P ¼ 0.023), confirming our initial findings reported in the manuscript. Contrary to Lai et al’s assertion, our UMHTC finding does not represent the ‘‘only study’’ that implicates LRT as potentially negatively impacting HCC outcomes. In a recent study by Xu et al,9 examining 384 HCC LT recipients (268 having undergone pre-LT LRT), the authors found a significantly elevated incidence of post-transplant HCC recurrence in patients with treatmentinduced or spontaneous partial necrosis in the explant (22.6%) compared with recipients with no necrosis (5.2%) or complete necrosis (6.1%; P < 0.001). Interestingly, these increased recurrences, which were primarily lymphatic metastases, were associated with higher circulating plasma levels of the angiogenic factors VEGF-A and VEGF-C, increased expression of VEGFR-2 and VEGFR-3 in the peritumoral tissue, and www.annalsofsurgery.com | 1 Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. CE: M.S.; ANNSURG-D-17-02099; Total nos of Pages: 3; ANNSURG-D-17-02099 Annals of Surgery Volume XX, Number XX, Month 2017 Response to the Letter TABLE 1. Distribution of Covariates by LRT Before (Unweighted) and After (Weighted) Propensity Matching Unweighted Analysis Covariate Age, yrs Sex (male), % Diagnosis, % HBV HCV NASH Cryptogenic Alcohol Autoimmune PBC/PSC Noncirrhotic HCC Other Lab MELD NLR Maximum pre-LT AFP Immediate pre-LT AFP Pathologic max diam, cm Pathologic grade, % Well Moderate Poor Vascular invasion, % None Microvascular Macrovascular AJCC T stage, % T1 T2 T3a T3b LRT (n ¼ 2082) 59 (54.1–64.0) 76.2 13 2.8 23.6 13.1 2.5 10.0 65.4 5.8 4.1 9.3 1.0 1.7 0.5 2.2 (9–18) (1.8–4.7) (8.7–123) (5.5–44.2) (1.9–3.5) No LRT (n ¼ 711) 56 (52.0–62.0) 76.5 16 3.3 16.0 10.1 2.3 7.7 61.9 7.0 4.1 12.9 1.1 2.7 0.8 1.7 (12–23) (2.0–6.0) (6.3–65.0) (5.0–45.7) (1.8–3.1) 28.9 60.9 10.3 30.7 57.5 11.8 73.1 22.5 4.4 70.7 23.9 5.3 35.5 55.7 4.2 4.5 43.0 49.6 2.0 5.3 Weighted Analysis LRT (n ¼ 2082) P <0.001 0.877 0.039 <0.001 <0.001 <0.001 0.154 <0.001 0.255 No LRT (n ¼ 711) 58.5 (54.0–63.0) 76.2 58.0 (53.0–64.0) 75.9 9.3 64.5 6.1 4.1 10.2 1.0 2.0 0.7 2.0 13.3 2.9 20.7 12.7 2.5 9.1 65.4 6.3 4.4 9.6 1.0 2.0 0.8 1.5 14.0 3.0 22.0 12.2 2.5 (10–19) (1.8–4.9) (8.0–107) (5.3–42.2) (1.8–3.5) 29.1 60.2 10.7 28.6 59.9 11.5 72.4 22.9 4.7 72.3 23.3 4.5 37.7 53.9 3.7 4.8 38.7 50.9 5.9 4.5 (10–18.4) (1.9–5.0) (7.5–98.6) (5.2–53.0) (1.9–3.5) 0.395 P 0.815 0.924 0.999 0.919 0.989 0.582 0.847 0.479 0.891 0.956 <0.001 0.234 Continuous variables summarized as medians and interquartile ranges. AJCC indicates American Joint Committee on Cancer; HBV, hepatitis B virus; HCV, hepatitis C virus; NASH, nonalcoholic steatohepatitis; NLR, neutrophil lymphocyte ratio; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis. increased immunofluorescent staining of peritumoral lymphatics in patients with tumor necrosis compared with patients without necrosis. This constellation of findings supports the contention that LRT-related tissue injury may induce a local lymphangiogenic stimulus, which, in the absence of complete tumor destruction, allows lymphatic invasion of any remaining viable tumor cells. Ravaioli et al10 examined the impact of TACE in LT recipients with a single HCC, and found that partial necrosis of the tumor after treatment was an independent risk factor for post-LT HCC recurrence, and was associated with low E-cadherin expression, a protein whose down-regulation has been implicated in tumor invasion and metastasis.11 Thus, instead of attributing our UMHTC findings to statistical artifact, there is now enough collective evidence that locoregional treatment itself may unmask aggressive underlying tumor behavior in the subset of HCC patients who do not achieve the intended response of complete tumor destruction. Our scientific community should be compelled to further investigate the 2 | www.annalsofsurgery.com biological mechanisms by which locoregional treatment alters the peritumoral microenvironment and potentially facilitates the lymphovascular spread of HCC. Over the past 2 decades, our transplant community has indeed seen a ‘‘revolution’’ in the risk assessment and selection of HCC patients for liver transplantation, with a clear bias now in favor of important dynamic measures of tumor biology over the static features of size and number alone. Furthermore, significant advances in the molecular pathogenesis of HCC have provided an accurate landscape of the genetic alterations responsible for HCC tumor development and progression. As a result, we are on the cusp of finding the ‘‘holy grail’’ in LT for HCC, namely, a constellation of radiologic, serum, and molecular biomarkers that will identify the patients who stand to gain the most transplant benefit while preserving excellent outcomes with the utilization of a scarce donor resource. Vatche G. Agopian, MD Dumont-UCLA Liver Transplant and Cancer ß Centers, Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, CA email@example.com William C. Chapman, MD Section of Transplantation, Department of Surgery, Washington University in St. Louis, St. Louis, MO Ronald W. Busuttil, MD PhD Dumont-UCLA Liver Transplant and Cancer Centers, Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, CA On behalf of the United States Multicenter Hepatocellular Carcinoma Transplant Consortium (UMHTC) REFERENCES 1. Agopian VG, Harlander-Locke MP, Ruiz RM, et al. Impact of pretransplant bridging locoregional therapy for patients with hepatocellular carcinoma within Milan criteria undergoing liver transplantation: analysis of 3601 patients from the 2017 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. CE: M.S.; ANNSURG-D-17-02099; Total nos of Pages: 3; ANNSURG-D-17-02099 Annals of Surgery Volume XX, Number XX, Month 2017 US Multicenter HCC Transplant Consortium. Ann Surg. 2017;266:525–535. 2. Lai Q, Avolio AW, Graziadei I, et al. Alphafetoprotein and modified response evaluation criteria in solid tumors progression after locoregional therapy as predictors of hepatocellular cancer recurrence and death after transplantation. Liver Transpl. 2013;19:1108–1118. 3. Lai Q, Nicolini D, Inostroza Nunez M, et al. A novel prognostic index in patients with hepatocellular cancer waiting for liver transplantation: Time-Radiological-response-Alpha-fetoproteinINflammation (TRAIN) score. Ann Surg. 2016; 264:787–796. 4. Agopian VG, Morshedi MM, McWilliams J, et al. Complete pathologic response to pretransplant locoregional therapy for hepatocellular carcinoma defines cancer cure after liver ß transplantation: analysis of 501 consecutively treated patients. Ann Surg. 2015;262:536–545 [discussion 543–535]. 5. Lai Q, Vitale A, Iesari S, et al. Intention-to-treat survival benefit of liver transplantation in patients with hepatocellular cancer. Hepatology. 2017. doi: 10.1002/hep.29342 [Epub ahead of print]. 6. Liver and Intestinal Organ Transplantation Committee ‘‘Capping the HCC Exception Score at 34’’ 2014. Available at: http://optn.transplant.hrsa.gov/media/1140/policy_notice_12-2014. Accessed October 6, 2017. 7. OPTN Liver and Intestinal Transplantation Committee ‘‘HCC auto approval criteria changes’’; 2017. Available at: https://optn. transplant.hrsa.gov/governance/publiccomment/hcc-auto-approval-criteria-changes/. Accessed October 6, 2017. 2017 Wolters Kluwer Health, Inc. All rights reserved. Response to the Letter 8. UNOS Region 5 Fall Meeting Liver Breakout: HCC ‘‘All-Comers’’ Protocol; 2017. Available at: https://www.transplantpro.org/community/ regions/region-5/. Accessed October 6, 2017. 9. Xu M, Doyle MM, Banan B, et al. Neoadjuvant locoregional therapy and recurrent hepatocellular carcinoma after liver transplantation. J Am Coll Surg. 2017;225:28–40. 10. Ravaioli M, Grazi GL, Ercolani G, et al. Partial necrosis on hepatocellular carcinoma nodules facilitates tumor recurrence after liver transplantation. Transplantation. 2004;78:1780– 1786. 11. Wei Y, Van Nhieu JT, Prigent S, et al. Altered expression of E-cadherin in hepatocellular carcinoma: correlations with genetic alterations, beta-catenin expression, and clinical features. Hepatology. 2002;36:692–701. www.annalsofsurgery.com | 3 Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.