A novel nomogram for prognosis stratification in salvage liver transplantation: a national-wide study with propensity score matching analysis in China

Introduction

Liver resection (LR) is considered as a curative option for patients with hepatocellular carcinoma (HCC) with 5-year overall survival (OS) of 30–50%, whereas 2-year recurrence rates was still up to 76% (1). Within the Milan criteria, primary liver transplantation (PLT) is the most effective treatment to eliminate both tumors and potential liver lesion (2,3). However, the drop-out rates of waiting list exceed 20% for these recipients due to organ shortage (4). Thus, many centers attempt to perform LR first and then salvage liver transplantation (SLT) in case of recurrence (5). When compared with PLT, different centers report controversial results regarding SLT. In the previous studies, propensity score matching (PSM) analysis has not been examined and thereby introduced uncontrolled bias, which affected the veracity of the results. It also should be noted that all the studies regarding SLT and PLT lacked a large sample size. Furthermore, previous reports did not present any new classification of the recipients in SLT group when they found that the survival of SLT was inferior to PLT.

In the present study, based on the data from the China Liver Transplant Registry (CLTR), OS and disease-free survival (DFS) of recipients fulfilling Milan criteria and Hangzhou criteria with PSM analysis were compared to verify if the criteria for candidate assessment in SLT were applicable or not. Additionally, landmark OS and DFS, which were measured from the relevant time point to the time of the event of interest (death or recurrence), were compared between the two groups. Finally, the nomograms plotted based on the results of multivariate analysis were used to classify SLT group into low- and high-risk groups. And OS and DFS of low-risk recipients were compared with PLT recipients fulfilling Hangzhou criteria. We present this article in accordance with the STROBE reporting checklist (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-22-304/rc).

Methods

Patients and informed consent

This study retrospectively analyzed 23,476 liver transplants performed from January 2015 to December 2019, and whose clinicopathological data were obtained from the CLTR. This was performed according to the Declaration of Helsinki (as revised in 2013) and approved by the CLTR (No. 20200039). Informed consent was obtained from all the patients for the use of their data for research purposes. The civilian organ donation is the sole source for organ transplant in China from January 2015 (6). All patients in this study did not receive liver grafts from prisoners. All patients with HCC scheduled for LT were evaluated preoperatively using ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET)-CT, bone scintigraphy, and colonoscopy to rule out extrahepatic disease which was unsuitable for LT. Preoperative loco-regional therapies, including radio frequency ablation (RFA), TACE and percutaneous ethanol injection, were performed to control or reduce the tumor lesion preoperatively. Moreover, the radiological information was acquired from the latest CT or MRI examination before LT. HCC recurrence was diagnosed based on the China Guidelines for Diagnosis and Treatment of Primary Liver Cancer in China.

Inclusion and exclusion criteria

The subject selection process is depicted in Figure 1. The excluded criteria were as follows: patients with concurrent other malignancies, or with pathologically-confirmed tumor types, including intrahepatic cholangiocarcinoma (ICC), combined HCC, fibrolamellar hepatocellular carcinoma (FLC) or secondary metastatic tumor; patients without histological confirmation of HCC for transplantation. Subsequently, a total of 8,340 patients were available for analysis. Additionally, patients who received a living donor LT, split LT, reduced-size LT or simultaneous transplantation were excluded, thereby obtaining 8,078 patients who underwent a full-size orthotopic LT. Next, pediatric patients (<18 years) or older patients (>65 years) and patients who died within 3 months after transplantation were excluded, thus, obtaining 6,573 patients suitable for analysis. Finally, patients with incomplete follow-up, having doubtful or contradictory data and missing essential data for analysis or re-transplantation during follow-up time were excluded. Therefore, 4,244 patients were included in the study.

Figure 1 Flow chart for recipients enrollment. LT, liver transplantation; ICC, intrahepatic cholangiocarcinoma; HCC, hepatocellular carcinoma; c HCC-CC, combined hepatocellular carcinoma-cholangiocarcinoma; FLC, fibrolamellar hepatocellular carcinoma; PLT, primary liver transplantation; SLT, salvage liver transplantation.

Baseline characteristics and postoperative follow-up

Patient’s baseline clinical characteristics were collected along with demographic characteristics, including age, sex, body mass index (BMI), preoperative alpha fetoprotein (AFP) level; ABO blood matched or mismatched, hepatitis B virus (HBV) infection, Child-Pugh grade, model for end-stage liver disease (MELD), preoperative loco-regional therapy, largest tumor size, tumor number, satellite lesions, creatinine level, hypertension, diabetes and chronic pulmonary. Cold ischemia time, operative time, intensive care unit (ICU) stay time, operative bleeding and the use of glucocorticoids were also collected.

All the patients received regular follow-ups after discharge from the hospital. Any patients who failed to attend a follow-up appointment was contacted by a research nurse via phone. Liver function tests, serum tumor marker assays, abdominal ultrasound and CT, chest CT and bone scans were routinely performed every 6 months for the first 2 or 3 years and annually thereafter.

Statistical analyses

Continuous variables with a normal distribution are expressed as the mean ± standard deviation or median (interquartile range). Categorical variables are expressed as numbers (n) or proportions (%). The Student’s t-test was used for comparison of continuous variables when applicable, otherwise, the Mann-Whitney U test was applied. Categorical variables were compared using χ² test or Fisher’s exact test, as appropriate. Using the Kaplan-Meier method generated by the log-rank test, the DFS and OS rates were compared between SLT and PLT groups.

Landmark OS and DFS were calculated. Time-to-event endpoints (OS and DFS) were measured among the recipients event-free at the time point post randomization: 1 and 2 years. They were measured from the relevant time point to the time of the event of interest (DFS event or death) occurred (7). Recipients who did not experience the event of interest were censored at the date that they were last known to be alive. Survival rates were also evaluated using the Kaplan-Meier method and were compared using log-rank tests. Univariate and multivariate Cox proportional hazards models were performed to identify independent prognostic factors of OS and DFS in the SLT population and whole population respectively. In multivariate analysis, factors with P≤0.2 in univariate analysis were finally tested, and P≤0.05 in the Cox model was considered statistically significant.

Then, the results of the multivariate analysis served as the basis for the construction of the nomograms. The accuracy of the final OS and DFS models were evaluated by estimating the models’ calibration, and the discrimination was measured using the concordance index (C-index) (8). The C-index is the probability that one event had a higher probability to happen according to the model in two selected group. A c-index of 0.5 indicated agreement by chance alone, and a c-index of 1 indicated perfect discrimination. Furthermore, SAS Version 9.4 software (SAS Institute Inc., Cary, NC, USA) was used to perform all statistical analyses.

To explain some differences in the baseline patients demographics and tumor characteristics in the PLT and SLT groups that could impact outcomes, an analysis using a PSM was performed to match the population (9). The baseline characteristics of recipients which were not balanced in the two groups before PSM and the potential prognostic factors of LT recipients were included in our study [such as: age, sex, Child-Pugh, pre-operative TACE, pre-operative RFA, HBV, international normalized ratio (INR), BMI, MELD, tumor numbers, tumor maximum size, blood loss, cold ischemia time and ICU stay time]. The two groups were paired on a 1:1 ratio.

Results

Patient demographics and survivals

Of the 4,244 LTs performed to treat HCC, 3,498 patients underwent PLT (PLT group), and 746 patients underwent SLT for recurrence after LR (SLT group). Overall and PSM-matched patient demographics and tumor characteristics are listed in detail in Table 1. An imbalance in some variables between the PLT and SLT groups was observed (Table 1). The median BMI was 23.60 kg/m² (range, 21.80–25.63 kg/m²) and 23.18 kg/m² (range, 21.26–24.80 kg/m²) in the PLT and SLT groups, respectively (P<0.001). The PLT group had an impaired liver functional reserve before LT when compared with the SLT group (Child-Pugh A 18.87% vs. 27.35%, Child-Pugh C 46.46% vs. 40.88%, P<0.001). In the preoperative condition, the median AFP was 39.13 ng/mL (range, 5.58–525.60 ng/mL) and 23.00 ng/mL (range, 4.04–242.00 ng/mL) in the PLT and SLT groups, respectively (P<0.001). Additionally, significant differences were observed between the two groups in terms of TACE, RFA and systemic chemotherapy (P<0.001). The HBV (+) rate was significantly higher in the SLT group (93.03%) than in the PLT group (90.19%) (P<0.001). Meanwhile, there was a difference between the two groups if the LT was blood type matched or not (P=0.026). In the operative condition, the median blood loss and operation time showed a significant difference between the two groups (P=0.005 and P=0.048, respectively). In the post-operative condition, the median ICU stay time was 72 hours (range, 39.0–144.0 hours) and 90 hours (range, 48.0–168.0 hours) in the PLT and SLT groups, respectively (P=0.001). Additionally, a difference was observed between the two groups whether glucocorticoid was used or not (P=0.016).

PSM analysis of the variants generated 728 matched recipients of the PLT group and SLT groups. No significant differences were found between the matched cases (Table 1). When compared the OS and DFS in the SLT and PLT groups after PSM, the 1-, 3- and 5-year OS between the two groups had no significant difference (92.17%, 74.10% and 68.81% vs. 92.42%, 78.74%, 69.80%; P=0.160; Figure S1A). No significant difference in DFS was also noted between the SLT group and PLT group (P=0.189; Figure S1B).

Landmark OS and DFS

The probabilities of surviving an additional 1-, 2- and 3-year given survival to 1 (N=630 in the PLT group, N=602 in the SLT group; after PSM) and 2 (N=381 in the PLT group, N=332 in the SLT group; after PSM) years were 92.06%, 85.20% and 82.08% in the PLT group vs. 88.56%, 80.39% and 76.31% in the SLT group, respectively (P=0.095, Figure S2A); 92.54%, 89.16% and 82.04% in the PLT group vs. 90.78%, 86.17% and 84.30% in the SLT group, respectively (P=0.611, Figure S2B). And the disease free survival an additional 1-, 2- and 3-year, given that DFS event was not experienced at 1 and 2 years were 93.12%, 86.55% and 81.06% in the PLT group vs. 87.21%, 81.81% and 78.83% in the SLT group, respectively (P=0.037, Figure S2C); 92.94%, 87.05% and 85.43% in the PLT group vs. 93.81%, 90.40% and 83.21% in the SLT group, respectively (P=0.623, Figure S2D). Indeed, the recipients who had given survival up to 1 year showed poor DFS in the SLT group compared with those in the PLT group.

Comparison of survival between the SLT and PLT groups with Milan criteria or Hangzhou criteria

For recipients fulfilling the Milan criteria between the SLT group and PLT group before PSM, significant differences in OS and DFS were noted (P=0.027, P=0.002; Figure S3). Meanwhile, significant differences in OS and DFS were also noted between the recipients fulfilling the Hangzhou criteria in SLT group and PLT group before PSM (P=0.003, P<0.001; Figure S4).

After PSM, the 1-, 3- and 5-year OS of recipients fulfilling the Milan criteria between the SLT group and PLT group had significant difference (96.70%, 86.38%, and 79.99% vs. 98.61%, 93.24%, and 90.05%; P=0.01; Figure 2A). The 1-, 3- and 5-year DFS of recipients fulfilling the Milan criteria between the SLT group and PLT group also had significant difference (92.34%, 82.44% and 79.01% vs. 97.57%, 91.58%, 90.65%; P<0.001; Figure 2B). Meanwhile, the 1-, 3- and 5-year OS of recipients fulfilling the Hangzhou criteria between the SLT group and PLT group had significant difference (94.78%, 81.87% and 75.24% vs. 96.28%, 87.50% and 82.43%; P=0.03; Figure 3A). The 1-, 3- and 5-year DFS of recipients fulfilling the Hangzhou criteria between the SLT group and PLT group also had significant difference (89.89%, 76.34% and 68.13% vs. 92.82%, 84.71% and 81.70%; P=0.003; Figure 3B). The OS and DFS of the recipients exceeding Milan criteria or Hangzhou criteria in SLT group were comparable to those in PLT group (Figure S5).

Figure 2 Survival curves for recipients fulfilling Milan criteria in SLT and PLT groups. OS (A) and DFS (B) in recipients fulfilling Milan criteria after propensity score matching between SLT group (n=351) and PLT group (n=292). PLT, primary liver transplantation; SLT, salvage liver transplantation; OS, overall survival; DFS, disease free survival.

Figure 3 Survival curves for recipients fulfilling Hangzhou criteria in SLT and PLT groups. OS (A) and DFS (B) in recipients fulfilling Hangzhou criteria after propensity score matching between SLT group (n=496) and PLT group (n=466). PLT, primary liver transplantation; SLT, salvage liver transplantation; OS, overall survival; DFS, disease free survival.

Univariate and multivariate analysis of OS and DFS in the SLT group

Univariate and multivariate analysis were applied to determine the risk factors affecting OS in the SLT group. According to the univariate analysis, factors in the preoperative condition associated with OS included AFP >400 ng/mL (P<0.0001), tumor maximum size >8 cm (P=0.031), tumor number >3 (P<0.0001), total tumor diameter >8 cm (P=0.002) and preoperative TACE (P=0.009). The complete univariate analyzes were shown in Table 2. According to cox regression model, factors in the preoperative condition associated with OS included AFP >400 ng/mL (P<0.0001), tumor maximum size >8 cm (P=0.031), total tumor diameter >8 cm (P=0.002) and preoperative TACE (P=0.036) (Table 2). Determinant DFS factors in univariate analysis are shown in Table S1. On including the variables in the preoperative condition, AFP >400 ng/mL (P<0.0001), tumor maximum size >8 cm (P=0.008), tumor number >3 (P=0.011), total tumor diameter >8 cm (P=0.012), preoperative TACE (P=0.009) and sorafenib use (P=0.002) were found to be poor prognostic factors for DFS (Table S1).

Nomograms to estimate the predicted probability of 3-year OS and DFS

The results of multivariable analysis were used to create a nomogram to estimate the predicted probability of 3-year OS (Figure 4A). The nomogram is a graphic depiction of the model wherein points are assigned based on the rank order of the effect estimated. Factors assigned the highest number of points included preoperative AFP and tumor total diameters. As shown in Figure 4B, the calibration plot for the OS model was plotted between the predicted probability of 3-year OS and the observed data. The modeled 3-year estimates of OS closely approximated those of observed estimates, but deviated relatively among individuals with poor survival. Model discrimination was evaluated using the C-index, which quantifies the level of concordance between the predicted and observed OS. The C-index for the final OS model was 0.70 [95% confidence interval (CI): 0.65–0.75].

Figure 4 Nomograms to classify recipients in SLT group. The nomogram for the prediction of 3-year OS (A); to use the nomogram, draw a line straight upward from the location on corresponding axis to the top line labeled “points”. Sum the points for all predictors then draw a line straight downward from the axis labeled “Total Points” to find the 3-year OS of recipients in SLT group. The calibration curves of nomogram for the prediction of 3-year OS (B) in the internal validation. The OS of low-risk and high-risk recipients was compared in SLT group (C). The OS of recipients in low-risk SLT group vs. recipients fulfilling Hangzhou criteria in PLT group (D). TACE, transarterial chemoembolization; AFP, alpha fetoprotein; OS, overall survival; PLT, primary liver transplantation; SLT, salvage liver transplantation.

In the SLT group, there was no difference in the OS between the recipients fulfilling the Hangzhou criteria and Milan criteria (P=0.091; Figure S6). Factors independently associated with DFS in multivariable analysis were applied to construct a nomogram to estimate the predicted probability of 3-year DFS (Figure S7A). Meanwhile, the nomogram was a graphic depiction of the model, in which points were assigned based on the rank order of the effect estimated. Figure S7B shows the calibration plot for the DFS model wherein the predicted probability of 3-year DFS was plotted against the observed data. The C-index for the final DFS model was 0.69 (95% CI: 0.65–0.73).

Furthermore, based on the nomograms and median risk score, we conducted SLT risk stratification system to further classify SLT recipients into low- (< median) and high- (> median) risk groups. The 1-, 3- and 5-year OS of recipients between low-risk SLT group and high-risk group had significant difference (95.34%, 84.26%, and 77.20% vs. 85.38%, 54.87%, and 51.23%; P<0.001; Figure 4C). The DFS of the two groups mentioned above had similar results (P<0.001; Figure S7C). Moreover, OS (P=0.107, Figure 4D) and DFS (P=0.055, Figure S7D) were not statistically significant between the recipients in the low-risk group who underwent SLT and the recipients in PLT fulfilling the Hangzhou criteria.

Discussion

This study indicated that no statistically significant difference was observed in OS between SLT group and PLT group before (P=0.143) or after (P=0.189) PSM. These results indicated that the two surgical modalities had similar prognoses in the non-screened setting. Additionally, a cohort study also reported that the vascular and biliary complications rates of in-hospital mortality and OS were similar in the recipients undergoing salvage living donor liver transplantation (LDLT) and primary LDLT (10). In addition, Guo et al. found that SLT was a viable treatment with similar outcomes for 5-year OS (P=0.345) and 5-year recurrence-free survival (P=0.263). Meanwhile, OS from the point of primary resection over 10 years period also showed no significant difference between the two groups (11). Thus, the outcomes of SLT were quite similar to that of PLT in the total population, except for technical difficulties in some SLT recipients from a surgical oncology standpoint (12). Notably, the median MELD score was 14/16 in SLT and PLT groups respectively, which was a little higher when compared with western population. But based on the Chinese population, Ling et al. (13) reported that mean MELD score was approximately 19 in a cohort of 125 LT recipients. Another recent study on extended criteria donor also reported mean MELD score in Chinese LT recipients up to 16 (14). LT indication was different in China, with a higher proportion of high-MELD recipients.

Meanwhile, the probabilities of surviving additional 1-, 2- and 3-year given survival to 1 or 2 years were investigated. The probabilities of disease free surviving additional 1-, 2- and 3-year, given that DFS event was not experienced at 1 year was better in the PLT group than in the SLT group. Therefore, the criteria that used survival time were not solely applicable to the selection of SLT recipients in China. There was certainly room for SLT recipients to research of selection criteria further. And our study was based on the huge data set from mainland China to compare the survival of SLT and PLT recipients after PSM. In the recent 10 years, such a large sample study with PSM was not reported.

The 1-, 3- and 5-year OS and DFS of recipients fulfilling the Milan criteria in SLT group were significantly reduced when compared with those in PLT group (OS, P=0.01; DFS, P<0.001). The same conclusion was also drawn for recipients fulfilling the Hangzhou criteria. Given this, these criteria were not appropriately applied to the selection of SLT recipients. Previously, similar conclusions have been reported, which indicated that the selection criteria for SLT have not been well-established and that the Milan criteria were not suitable for selecting recipients undergoing LT for recurrent HCC after LR (15). Lee et al. (16) reported that the extent and timing of HCC recurrence after LR played important roles in the outcomes of SLT. Another study suggested that the residual tumor burden after the treatment of recurrent HCC played a more vital role in post-transplant tumor recurrence than the timing of tumor recurrence after LR (17). Thus, these studies indicated that there is a lack of data on the selection of suitable SLT recipients. And we found that the OS and DFS of recipients was worse in SLT group than in PLT group. It is well-known that LR is widely used in HCC recipients at early stages and provides an opportunity to control tumor progression. However, it has been previously suggested that surgical stress will promote HCC cell entry into the circulation (18). Ren et al. (19) demonstrated that platelet TLR4-ERK5 axis implicates surgery-driven distant metastasis with a murine model of localized surgical stress. Recent clinical study has also reported that hepatectomy is associated with increased circulating tumor cells (CTCs) counts, which is an independent risk factor for recurrence after LT (20). Furthermore, a previous single cell transcriptome study indicated that the existence of active crosstalk between relapse HCC and immune cells in HCC, which will compromise antitumor function and promote tumor cell migration (21). In addition, the SLT recipients tend to have poor hepatic function and sarcopenia as a result of liver cirrhosis. And the sarcopenia has been reported to reduce recipients survival and be associated with high risk of postoperative complications (22,23). In addition, sarcopenia was showed to be associated with tumor recurrence and metastasis in a multicenter, large sample clinical study (24). Therefore, the above mechanisms may cause the relatively poor survival for LT recipients with a recurrent HCC after resection.

Consequently, a reliable model for the filtering of SLT recipients was needed. To these effect, multivariate analysis in the SLT group was performed based on the results of the univariate analysis. Preoperative TACE, AFP >400 ng/mL, tumor maximum size >8 cm and total tumor diameter >8 cm were found to be independent risk factors affecting OS in the SLT group. As for the factors affecting DFS, they were preoperative TACE, AFP >400 ng/mL, tumor maximum size >8 cm, tumor number >3, total tumor diameter >8 cm and preoperative sorafenib administration. Indeed, for the use of SLT in cirrhotic patients with HCC, an intention-to-treat analysis reported that the pre-resection predictors of a SLT strategy included the absence of TACE (25). Several studies reported that preoperative TACE up-regulated the expression of vascular endothelial growth factor protein (26), which is associated with metastasis and recurrence (27). Meanwhile, TACE causes hypoxia and HIF-1α expression in HCC cells, which are associated with poor outcomes (28). However, the adverse effects of preoperative TACE and sorafenib require further studies in the future. Moreover, it is well-known that tumor size and tumor biology are important factors associated with recipients OS and DFS (29,30). Notably, compared with Hangzhou criteria and Milan criteria, the results of our multivariate analysis take into account more pre-transplant treatments.

The nomograms were plotted based on the results of the multivariate analysis. In order to make the nomograms more accurate for SLT recipients selection, the variables (tumor maximum size and tumor number) used in nomograms were the post-operative pathological data of the tumor (31). However, the variables (tumor maximum size and tumor number) could be obtained from imaging date before LT. So that our model could be used to select recipients in SLT group before LT. Based on the median risk score, the SLT group was classified into the low-risk (< median) and high-risk (> median) groups. And in SLT group, no statistical difference was observed in the OS between the recipients fulfilling Hangzhou criteria and Milan criteria. Other studies also reported that the OS of recipients fulfilling the Hangzhou criteria was similar to those fulfilling the Milan criteria (29,32,33). To confirm the validity of models, we found OS and DFS were not statistically significant between the low-risk SLT group and PLT recipients fulfilling the Hangzhou criteria. So that, the internally validated nomograms in this study established the selection criteria for selecting appropriate recipients to undergo SLT.

Our results were pertinent and sensible from at least two perspectives. Firstly, this study based on a huge data set from mainland China compared post-transplant survival of recipients fulfilling Milan or Hangzhou criteria between SLT group and PLT group with PSM analysis. So that, these criteria were not appropriately applied to the selection of SLT recipients in China. Secondly, nomograms were plotted based on the results of the multivariate analysis to classify the SLT group into low-risk and high-risk groups. The low-risk group had a better outcome than the high-risk group, meanwhile, similar OS and DFS were observed in the recipients fulfilling Hangzhou criteria when compared with the recipients in the low-risk group. Thus, the nomogram can be used to improve the selection of SLT recipients.

Despite the advantages, this study has some limitations. This study is a retrospective analysis of observational data over a long period of time, which is mostly associated with the quality of data available in CLTR database. With lack of the survival data from initial LR to SLT, an intention-to-treat analysis could not be performed. In this regarding, prospective, multicenter, randomized clinical study remains to be conducted to provide a higher level of clinical evidence for SLT recipient selection and therapeutic strategy.

Conclusions

This study demonstrated that the OS and DFS of recipients fulfilling the Milan criteria or Hangzhou criteria showed a significant reduction in the SLT group compared to PLT group. A novel nomogram was plotted by using the independent risk factors from multivariate analysis. SLT risk stratification system based on this nomogram could further classify SLT recipients into low- and high-risk groups. And low-risk SLT group showed comparable OS and DFS to PLT group fulfilling Hangzhou criteria. In this regarding, the nomogram-based risk stratification system could assist for clinical decision guidance in SLT management, so as to achieve an improved long-term prognosis.

Acknowledgments

We thank the China Liver Transplant Registry (CLTR) for providing a qualified public database, and the local administrative staff who contributed to the quality control of the data in China.

Funding: This work was supported by Youth Program of National Natural Science Foundation of China(82003248); Key Program, National Natural Science Foundation of China (81930016); Key R&D Project of Zhejiang Province (2019C03050); National Key Research and Development Program of China (2021YFA1100500); The Major Research Plan of the National Natural Science Foundation of China (92159202).

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-22-304/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-22-304/coif). Q.X. serves as an unpaid editorial board member of Hepatobiliary Surgery and Nutrition. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was performed according to the Declaration of Helsinki (as revised in 2013) and approved by the CLTR (No. 20200039). Informed consent was obtained from all the patients for the use of their data for research purposes. No organs from executed prisoners were used.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Poon RT, Fan ST, Lo CM, et al. 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;235:373-82. [Crossref] [PubMed]
2. Yang JD, Larson JJ, Watt KD, et al. Hepatocellular Carcinoma Is the Most Common Indication for Liver Transplantation and Placement on the Waitlist in the United States. Clin Gastroenterol Hepatol 2017;15:767-775.e3. [Crossref] [PubMed]
3. Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. Hepatology 2011;53:1020-2. [Crossref] [PubMed]
4. Wertheim JA, Petrowsky H, Saab S, et al. Major challenges limiting liver transplantation in the United States. Am J Transplant 2011;11:1773-84. [Crossref] [PubMed]
5. Majno PE, Sarasin FP, Mentha G, et al. Primary liver resection and salvage transplantation or primary liver transplantation in patients with single, small hepatocellular carcinoma and preserved liver function: an outcome-oriented decision analysis. Hepatology 2000;31:899-906. [Crossref] [PubMed]
6. Huang J, Millis JM, Mao Y, et al. Voluntary organ donation system adapted to Chinese cultural values and social reality. Liver Transpl 2015;21:419-22. [Crossref] [PubMed]
7. McNamara MG, Lopes A, Wasan H, et al. Landmark survival analysis and impact of anatomic site of origin in prospective clinical trials of biliary tract cancer. J Hepatol 2020;73:1109-17. [Crossref] [PubMed]
8. Wang Y, Li J, Xia Y, et al. Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol 2013;31:1188-95. [Crossref] [PubMed]
9. Rubin DB. Propensity score methods. Am J Ophthalmol 2010;149:7-9. [Crossref] [PubMed]
10. Yong CC, Elsarawy AM, Wang SH, et al. The surgical challenges of salvage living donor liver transplantation for Hepatocellular carcinoma; The cumulative experience of 100 cases - A retrospective cohort study and a propensity score analysis. Int J Surg 2018;54:187-92. [Crossref] [PubMed]
11. Guo Y, Tan EK, Krishnamoorthy TL, et al. Outcomes of salvage liver transplant for recurrent hepatocellular carcinoma: A comparison with primary liver transplant. Ann Hepatobiliary Pancreat Surg 2019;23:1-7. [Crossref] [PubMed]
12. Hwang S, Lee SG, Moon DB, et al. Salvage living donor liver transplantation after prior liver resection for hepatocellular carcinoma. Liver Transpl 2007;13:741-6. [Crossref] [PubMed]
13. Ling Q, Xie H, Lu D, et al. Association between donor and recipient TCF7L2 gene polymorphisms and the risk of new-onset diabetes mellitus after liver transplantation in a Han Chinese population. J Hepatol 2013;58:271-7. [Crossref] [PubMed]
14. Tong L, Hu XG, Huang F, et al. Clinical Impacts and Outcomes With Possible Donor-Derived Infection in Infected Donor Liver Transplantation: A Single-Center Retrospective Study in China. J Infect Dis 2020;221:S164-73. [Crossref] [PubMed]
15. 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;334:693-9. [Crossref] [PubMed]
16. Lee S, Hyuck David Kwon C, Man Kim J, et al. Time of hepatocellular carcinoma recurrence after liver resection and alpha-fetoprotein are important prognostic factors for salvage liver transplantation. Liver Transpl 2014;20:1057-63. [Crossref] [PubMed]
17. Rhu J, Kim JM, Choi GS, et al. Continuing five or more locoregional therapies before living donor salvage liver transplantation for hepatocellular carcinoma is related to poor recurrence-free survival. Ann Surg Treat Res 2018;95:152-60. [Crossref] [PubMed]
18. Martin OA, Anderson RL, Narayan K, et al. Does the mobilization of circulating tumour cells during cancer therapy cause metastasis? Nat Rev Clin Oncol 2017;14:32-44. [Crossref] [PubMed]
19. Ren J, He J, Zhang H, et al. Platelet TLR4-ERK5 Axis Facilitates NET-Mediated Capturing of Circulating Tumor Cells and Distant Metastasis after Surgical Stress. Cancer Res 2021;81:2373-85. [Crossref] [PubMed]
20. Yu JJ, Xiao W, Dong SL, et al. Effect of surgical liver resection on circulating tumor cells in patients with hepatocellular carcinoma. BMC Cancer 2018;18:835. [Crossref] [PubMed]
21. Sun Y, Wu L, Zhong Y, et al. Single-cell landscape of the ecosystem in early-relapse hepatocellular carcinoma. Cell 2021;184:404-421.e16. [Crossref] [PubMed]
22. Meza-Junco J, Montano-Loza AJ, Baracos VE, et al. Sarcopenia as a prognostic index of nutritional status in concurrent cirrhosis and hepatocellular carcinoma. J Clin Gastroenterol 2013;47:861-70. [Crossref] [PubMed]
23. Valero V 3rd, Amini N, Spolverato G, et al. Sarcopenia adversely impacts postoperative complications following resection or transplantation in patients with primary liver tumors. J Gastrointest Surg 2015;19:272-81. [Crossref] [PubMed]
24. Fujihata S, Sakuramoto S, Morimoto Y, et al. Impact of loss of skeletal muscle mass within 6-12 months after gastrectomy and S1 adjuvant chemotherapy on the survival prognosis of elderly patients with gastric cancer. Surg Today 2022;52:1472-83. [Crossref] [PubMed]
25. de Haas RJ, Lim C, Bhangui P, et al. Curative salvage liver transplantation in patients with cirrhosis and hepatocellular carcinoma: An intention-to-treat analysis. Hepatology 2018;67:204-15. [Crossref] [PubMed]
26. Xiao EH, Guo D, Bian DJ. Effect of preoperative transcatheter arterial chemoembolization on angiogenesis of hepatocellular carcinoma cells. World J Gastroenterol 2009;15:4582-6. [Crossref] [PubMed]
27. Liang B, Zheng CS, Feng GS, et al. Correlation of hypoxia-inducible factor 1alpha with angiogenesis in liver tumors after transcatheter arterial embolization in an animal model. Cardiovasc Intervent Radiol 2010;33:806-12. [Crossref] [PubMed]
28. Herzog J, Ehrlich SM, Pfitzer L, et al. Cyclin-dependent kinase 5 stabilizes hypoxia-inducible factor-1α: a novel approach for inhibiting angiogenesis in hepatocellular carcinoma. Oncotarget 2016;7:27108-21. [Crossref] [PubMed]
29. Xu X, Lu D, Ling Q, et al. Liver transplantation for hepatocellular carcinoma beyond the Milan criteria. Gut 2016;65:1035-41. [Crossref] [PubMed]
30. Vidal J, Muinelo L, Dalmases A, et al. Plasma ctDNA RAS mutation analysis for the diagnosis and treatment monitoring of metastatic colorectal cancer patients. Ann Oncol 2017;28:1325-32. [Crossref] [PubMed]
31. Fakhry C, Zhang Q, Nguyen-Tân PF, et al. Development and Validation of Nomograms Predictive of Overall and Progression-Free Survival in Patients With Oropharyngeal Cancer. J Clin Oncol 2017;35:4057-65. [Crossref] [PubMed]
32. Li Y, Ruan DY, Jia CC, et al. Surgical resection versus liver transplantation for hepatocellular carcinoma within the Hangzhou criteria: a preoperative nomogram-guided treatment strategy. Hepatobiliary Pancreat Dis Int 2017;16:480-6. [Crossref] [PubMed]
33. Xiao GQ, Yang JY, Yan LN. Combined Hangzhou criteria with neutrophil-lymphocyte ratio is superior to other criteria in selecting liver transplantation candidates with HBV-related hepatocellular carcinoma. Hepatobiliary Pancreat Dis Int 2015;14:588-95. [Crossref] [PubMed]