https://orcid.org/0000-0002-2068-7335
Figures
Abstract
Background
Prophylactic transarterial chemoembolization (p-TACE) is frequently conducted for patients with hepatocellular carcinoma (HCC) in China, but the question of who could benefit from it remains controversial. Hence, we wanted to establish a nomogram model to identify patients eligible for p-TACE.
Methods
Data from HCC patients receiving R0 resection with or without p-TACE between January 2013 and December 2014 were identified, using primary liver cancer big data, to establish a nomogram model to predict overall survival (OS). Based on the model, Patients receiving R0 resection between January 2015 and December 2015 were divided into three subgroups, and survival curves were constructed using the Kaplan–Meier method and analyzed by the log-rank test among patients in each subgroup.
Results
A nomogram integrating the neutrophil to lymphocyte ratio, AFP, tumor diameter, and microvascular invasion was developed to predict the OS of patients with HCC receiving R0 resection, and significant differences were observed in the median OS of the subgroups of low-risk (≤20), intermediate-risk (20~120), and high-risk (>120) identified by the current model. This model showed good calibration and discriminatory power in the validation cohort and the external cohort (c-index of 0.669 and 0.676, respectively). In the external cohort, the Kaplan–Meier curves showed that p-TACE could only significantly prolong the median OS of high-risk patients (25.6 vs. 33.7 months, P<0.05), but no differences were observed in any subgroups stratified by the current staging systems (all P>0.05).
Citation: Lin N, Wang L, Huang Q, Zhou W, Liu X, Liu J (2022) A simplified model for prophylactic transarterial chemoembolization after resection for patients with hepatocellular carcinoma. PLoS ONE 17(10): e0276627. https://doi.org/10.1371/journal.pone.0276627
Editor: Pavel Strnad, Medizinische Fakultat der RWTH Aachen, GERMANY
Received: April 12, 2022; Accepted: October 10, 2022; Published: October 31, 2022
Copyright: © 2022 Lin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data are available from the the big database of hepatobiliary in fuzhou, researchers who meet the criteria for access to confidential data. (Due to the data contain potentially patients' information) Please contact the PLCBD with the telephone:0591-88116149 in requesting for data.
Funding: This study was supported by Fujian provincial medical center of hepatobiliary, Key Clinical Specialty Discipline Construction Program of Fuzhou (201912002),Science and Technology Project of Fuzhou (Grant number: 2019-SZ-47) and Startup Fund for scientific research, Fujian Medical University (2021QH1061). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Hepatocellular carcinoma is the sixth most common cancer globally, and approximately 850,000 patients are newly diagnosed with HCC every year [1]. Surgical resection is still the preferred strategy worldwide, although great progress has been made in the fields of transplantation, radiofrequency ablation (RFA), and stereotactic body radiotherapy (SBRT) [2]. Nonetheless, the long-term prognosis of patients receiving surgical resection remains far from satisfactory, mainly owing to the high incidence of intrahepatic recurrence. Transarterial chemoembolization (TACE) have been applied in recent decades to prevent recurrences and improve the prognosis [3], but unfortunately, neither has been widely confirmed nor recommended by current guidelines for that uses [4, 5].
Among the adjuvant strategies, prophylactic TACE (p-TACE) has become dominant in China [6], although the underlying mechanism of p-TACE remains controversial [6, 7]. p-TACE has been verified to benefit patients who are “high risk” by numerous studies and meta-analyses [8–10], but agreements on the definition of “high risk” are far from being reached. In addition, selecting patients to receive p-TACE according to the so-called “high-risk factors” is generally impractical [11]. Hence, an accurate and user-friendly model for patients receiving surgical resection to identify the candidates that will benefit the most from p-TACE is urgently needed.
In the current study, data of HCC patients receiving R0 resection were extracted from primary liver cancer big data (PLCBD) in China to establish and validate a selection model for patients who should receive p-TACE following R0 resection.
Materials and methods
Patients
This study was approved by our institution’s Ethics Committee (No. 2019_039_01) and performed according to the Declaration of Helsinki. Due to the retrospective nature of the study, informed consent was waived.
Patients receiving surgical resection but no p-TACE from January 2013 to December 2014 were identified to establish a prediction model, which was randomly divided into training and validation cohorts. Patients treated from January 2015 to December 2015 were used to verify whether the current model could identify the potential beneficiaries of p-TACE, which was used as the external cohort.
Patients who underwent R0 resection and were diagnosed with HCC by pathology were eligible for this study. Patients were excluded from this study if they met the following criteria: 1) resection for relapsed HCC, 2) distant metastasis, 3) macrovascular or bile duct tumor thrombus, 4) other malignant cancers, 5) death within four weeks following surgery, and 6) preoperative treatment or any adjuvant treatments, 7) tumor residue found by p-TACE angiography, or 8) incomplete data.
The data of patients receiving surgical resection were extracted from PLCBD by an IT engineer, and they were independently verified by the researchers: age, sex, HBV infection, serum AFP level, Child–Pugh grading, cirrhosis, intraoperative transfusion, tumor number, tumor diameter, differentiation grading, capsule, presence of tumor satellites, and outcomes.
Interventions and definitions
Before hepatectomy and TACE, all laboratory values, including albumin, bilirubin, neutrophil, and lymphocyte counts, AFP levels, and viral tests (HBsAg, HBeAg, and HBV-DNA); routine imaging, including abdominal ultrasound, dynamic contrast-enhanced computed tomography (CT) and/or magnetic resonance imaging (MRI); and dynamic changes in liver function evaluated by the Child–Pugh score or the indocyanine green clearance rate at 15 minutes (ICG-15); were conducted.
Hepatectomy was generally performed openly or laparoscopically via the anterior approach. Major hepatectomy was defined as the resection of Couinaud segments ≥3, applied only to those with Child–Pugh class A and ICG-15 ≤ 15%.
Briefly, p-TACE was conducted based on the consensus of Chinese experts [12], performed within four to eight weeks after hepatectomy. Under digital subtraction angiography (DSA), an 5-F catheter or microcatheter was selectively inserted into the predesigned hepatic artery, and then an emulsion of lipiodol (2–5 mL) was infused instantly after a slow injection of cisplatin (10–30 mg), doxorubicin hydrochloride (10 mg) and pharmorubicin (20–40 mg).
Adverse events (AEs) related to surgery and TACE were subjected to the Clavien–Dindo classification [13] and were extracted from the medical records, and grade III and above AEs were defined as severe AEs.
Follow-up and endpoints
Patients were followed up periodically based on the Chinese guidelines for treating primary liver cancer [11]. Generally, routine items of serum AFP and abdominal ultrasound were examined at three-month intervals after resection. Dynamic enhanced CT/MRI was performed when a recurrence was clinically suspected. Further intervention was immediately started when a recurrence was confirmed, including re-resection, RFA, TACE, and SBRT.
The primary endpoint in this study was overall survival (OS), which was calculated from the date of resection to either the date of death or the last follow-up (October 2018).
Clinical and pathological variables
Generally, all the variables potentially associated with the OS of HCC were reclassified based on previous studies [14, 15]. Specifically, the tumor number was categorized as single vs. multiple according to the National Comprehensive Cancer Network (NCCN) guidelines on HCC [4], and differentiation was categorized as I/II vs. III/IV according to the Edmondson-Steiner grading system [16]. Microvascular invasion (MVI) detected by pathologist after surgery. Of note, regardless of the status of HBsAg and HBV-DNA, HBV infection was defined as a history of HBV infection. Blood transfusion included intraoperative transfusion of red blood cells and plasma, which was extracted from the anesthesia records. Age was categorized as<50 years vs. ≥50 years as previously reported [17], and serum AFP levels were categorized as <400 ng/ml vs. ≥400 ng/ml using the preferred cutoff value [18, 19]. The neutrophil to lymphocyte ratio (NLR) was calculated initially and then dichotomized as low or high. The albumin-bilirubin score (ALBI) was calculated based on the following formula: (-0.085× (albumin g/L) + 0.66 × log (bilirubin μmol/L)) and graded as 1 (score ≤-2.60), 2 (score -2.6 to -1.39), or 3 (score>-1.39) based on a previous report [20].
Statistics
Variables were compared using the t-test, chi-square test, or Fisher’s exact test (two-tailed). The variable NLR was dichotomized for OS using the optimal cutoff values determined by the “surv_cutpoint” function of the “survminer” package [21]. Variables associated with OS were identified using the univariate Cox model, and those with P< 0.05 were then used to conduct multivariate Cox regression through backward selection to identify independent prognostic factors.
A nomogram was developed based on the Cox multivariate analysis of OS through the “rms” package in R. The performance of the new nomogram was evaluated using the concordant index (C-index) and calibration with 1000 bootstrap samples and then was compared with current staging systems of the Barcelona Clinic Liver Cancer (BCLC) [5], and China Liver Cancer (CNLC) [11] using the package “rcorrp.cens”.
In the external cohort, survival curves were depicted using the Kaplan–Meier method between the groups of p-TACE and non-TACE in each subgroup and compared using the log-rank test with 95% confidence intervals (CIs).
All analyses were conducted via R project version 3.6.3 (http://www.r-project.org/), and P <0.05 with two tails was considered statistically significant.
Results
Patients’ characteristics and prognosis
Initially, 744 patients receiving R0 resection from January 2013 to December 2014 were divided into training and validation cohorts. The basic characteristics of the training and validation cohorts are shown in Table 1. The optimal cutoff value for NLR was 2.46 for OS (S1 Fig); hence, an NLR score <2.46 was defined as low and ≥2.46 as high in this study.
Independent prognostics factors for OS
NLR, ALBI, AFP, tumor number, tumor diameter, capsule, presence of tumor satellites, tumor differentiation, and MVI were identified as prognostic factors for OS (all P<0.05, Table 2) using univariate analysis, but only NLR (HR = 1.80, 95%CI = 1.27–2.56), AFP (HR = 1.94, 95%CI = 1.38–2.75), tumor diameter (HR = 1.09, 95%CI = 1.04–1.13), and MVI (HR = 1.48, 95%CI = 1.01–2.17) remained independent risk factors for OS (all P<0.05, Table 2) in the multivariate Cox model.
Development and validation of the selection model
A nomogram consisting of all the independent risk factors for OS was established, as shown in Fig 1. The C-index for the OS prediction was 0.703 (95% CI, 0.659–0.747), and the calibration curves for the probabilities of OS at 3 and 5 years after hepatectomy exhibited optimal agreement between the prediction by the nomogram and the actual observation (Fig 2A and 2B).
Nomogram for predicting overall survival of patients receiving R0 resection.
The calibration curve for predicting overall survival by the nomogram at 3 year and 5 years in training cohort (A, B), validation cohort (C, D) and 1 year and 3 years in external cohort (E, F).
In the validation cohort, the C-index for OS prediction was 0.669 (95% CI, 0.628–0.710). The calibration plots showed satisfactory consistency for the probabilities of OS after hepatectomy between those predicted by the nomogram and the observed probabilities for 3- and 5-year OS (Fig 2C and 2D).
Comparison between the current nomogram model and other staging models in the training cohort
The discriminatory power of the current nomogram model and other widely used staging models, including the BCLC, and CNLC staging systems, were compared using the C-index in the training cohort. The current nomogram model has the highest discriminatory power in predicting the OS of patients receiving R0 resection, and its C-index was 0.703 (95% CI: 0.659–0.747), which was higher than the staging systems of, BCLC (C-index, 0.667, 95% CI: 0.620–0.714, P = 0.047), and CNLC (C-index, 0.668, 95% CI: 0.621–0.714, P = 0.046, Table 3).
Risk stratification based on the current nomogram model
The median OS in the training and validation cohorts was 34.8 and 34.4 months, respectively. Using 20 and 120 as the cutoff values of the current nomogram model, which corresponded to the 15th and 85th centiles of the score in the derivation cohort, patients in the derivation cohort were divided into three subgroups: low-risk, intermediate-risk, and high-risk, respectively. Patients in the three subgroups exhibited a well-stratified prognosis: low-risk (3-year OS, 80.3%), intermediate-risk (3-year OS, 59.8%), and high-risk (3-year OS, 33.3%) (P<0.001, Fig 3A). Based on the cutoff values selected in the training cohort, good prognostic stratification was observed among all the subgroups in terms of OS in the validation cohort (P<0.001, Fig 3B) and the external cohort (P<0.001, Fig 3C).
Kaplan-Meier curve of OS for subgroups stratified by the nomogram model in the training cohort (A), validation cohort (B) and external cohort (C).
Survival curves for patients receiving p-TACE or not stratified by subgroups according to the current nomogram model
Patients receiving R0 resection from January 2015 to December 2015 were enrolled as the external cohort, including 108 patients (108/599, 18.0%) who received p-TACE. Their basic characteristics are depicted in Table 4. Notably, the proportions of HBV infection, multiple tumors, presence of tumor satellites, and MVI were prominently higher in the TACE group (all P<0.05, Table 4).
The C-index for OS prediction was 0.676 (95% CI, 0.640–0.714), and the calibration curves for the probability of OS at 1 and 3 years after hepatectomy exhibited satisfactory agreement between the prediction by the nomogram and the actual observation (Fig 2E and 2F).
Based on the current nomogram model and the cutoff values established in the training cohort, patients in the external cohort were divided into three subgroups: 89 patients in the low-risk subgroup, 419 patients in the intermediate-risk subgroup, and 91 patients in the high-risk subgroup. The median OS of patients in the high-risk subgroup receiving p-TACE was significantly longer than that of patients receiving R0 resection only (25.6 vs. 33.7 months, P<0.05, Fig 4), but no difference was observed in any subgroup stratified by the staging systems of BCLC, or CNLC (all P>0.05, Figs 5 and 6). The basic characteristics of high-risk subgroup are depicted in Table 5.
Discussion
In the current study, we developed a nomogram model that had higher discriminatory power than that of the current staging systems. The Kaplan–Meier curves showed that p-TACE could only benefit patients scoring >120 as determined by the nomogram model but not any subgroup stratified by the current staging systems.
p-TACE is frequently conducted with the initial aim of preventing recurrences, but its efficacy remains controversial [22, 23]. The reasons might be as follows: 1) great heterogeneity exists among HCC patients undergoing resection [14], 2) the mechanisms by which p-TACE may prevent recurrence remain unknown [24], and 3) adverse effects, including hepatic dysfunction and immune function impairment, are hard to avoid [25]. Nonetheless, p-TACE has been identified repeatedly as an independent protective factor for OS, which is often considered to be a hard endpoint of studies [3, 24, 25]. Hence, we selected OS as the primary endpoint rather than recurrence or DFS.
The novelty of this study lies in its methodology. Patients with “high-risk” factors, including aggressive tumor characteristics, are recommended to receive p-TACE [11, 12], but the definition of “high-risk” is hard to standardize and normalize [3, 24]. A model incorporating “high-risk” factors is the solution to this difficult problem. In the current study, we selected patients receiving R0 resection without p-TACE to establish a prediction model and then selected an external cohort including patients with or without p-TACE to verify whether the model could guide the application of p-TACE. Different from previous models [26, 27], which selected patients receiving p-TACE to establish models directly, the methodology applied in this study had several advantages: 1) it decreased the interference of p-TACE because systematic reviews and meta-analyses failed to confirm a definite benefit of p-TACE [9, 28]; and 2) it complied with the clinical treatment process completely because the surgical margin and postoperative pathology were the determinant factors of TACE.
The major advantage of the current nomogram model is its ready availability. AFP is an important part of the diagnosis of HCC, and elevated preoperative AFP levels often indicate a potential recurrence and a worse prognosis [18, 19, 29]. Diameter is an essential element of the tumor that can be easily measured [30]. MVI has been well studied in the last decade and is often deemed the origin of intrahepatic recurrence [23, 31]. More importantly, the NLR is considered to be a new biomarker of recurrence and prognosis [32]; it is also readily available using preoperative routine blood examinations and represents the systematic immune status. In this study, the current model not only showed optimal performance for predicting OS among patients receiving R0 resection but it also exhibited good stratification among patients receiving p-TACE, indicating that the current model is robust.
To the best of our knowledge, p-TACE is rarely referred to in the current guidelines, especially in Western countries [4, 5]. p-TACE is recommended for patients with “high-risk” factors in the new CNLC [11], but the definition of “high risk” is not clear, which makes it unfeasible in the clinic. Based on the current model, patients were divided into three subgroups, and a survival benefit from p-TACE was observed only in the subgroup of patients with scores >120 according to the current model, which indicated that the current model could guide the application of p-TACE.
There are several limitations of this study. First, considering the differences in HCC between the east and the west [1], the conclusion needs further validation in the west. Second, variates of AFP is not always measured in Western countries according to current guidelines [2], which suggests that the conclusion may not be directly applicable to patients in Western countries. Third, recall bias and selection bias are hard to avoid in a retrospective study. Fourth, various treatments were immediately adopted once a recurrence was confirmed, which could influence the OS of the patients. Last but not least, with limited patients, the conclusion in low-risk group needed further validation, even the enrolled patients were higher in high-risk group, a further study is warrant to variated the result.
Conclusion
With the current data, we concluded that the newly constructed and readily available nomogram model was able to predict the OS of patients receiving R0 resection and that it could be used to guide the management of p-TACE. However, the model needs additional external validation.
Supporting information
S1 Fig. The cut-off value of the neutrophil to lymphocyte ratio.
https://doi.org/10.1371/journal.pone.0276627.s001
(TIFF)
S1 Checklist. STROBE statement—checklist of items that should be included in reports of observational studies.
https://doi.org/10.1371/journal.pone.0276627.s002
(DOC)
References
- 1. Bray F, Ferlay J, Soerjomataram I, et al(2018). Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin;68:394–424 pmid:30207593
- 2. Bruix J, Forner A, Reig M (2018). Hepatocellular carcinoma. LANCET;391:1301–1314 pmid:29307467
- 3. Huang J, Liu FC, Li L, et al(2020). Prognostic Nomogram for Hepatitis B Virus-related Hepatocellular Carcinoma With Adjuvant Transarterial Chemoembolization After Radical Resection. Am J Clin Oncol; 43:20–27 pmid:31633514
- 4. Abbott DE, Benson AB, D’Angelica MI, et al(2019). Guidelines Insights: Hepatobiliary Cancers, Version 2.2019. J Natl Compr Canc Netw; 17:302–310 pmid:30959462
- 5. Cervantes A, Chau I, Vogel A, et al(2018). Hepatocellular carcinoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up†. ANN ONCOL; 29:v238–v255
- 6. Guo L, Liu S, Li H, et al(2018). Postoperative Adjuvant Trans-Arterial Chemoembolization for Patients with Hepatocellular Carcinoma and Portal Vein Tumor Thrombus. ANN SURG ONCOL; 25:2098–2104 pmid:29728879
- 7. Wang H, Du PC, Wu MC, Cong WM(2018). Postoperative adjuvant transarterial chemoembolization for multinodular hepatocellular carcinoma within the Barcelona Clinic Liver Cancer early stage and microvascular invasion. Hepatobiliary Surg Nutr; 7:418–428 pmid:30652086
- 8. Cheng X, Hu QG, Sun P, et al(2014). Transarterial (chemo)embolization for curative resection of hepatocellular carcinoma: a systematic review and meta-analyses. J Cancer Res Clin Oncol; 140:1159–1170 pmid:24752339
- 9. Liao M, Zhu Z, Wang H, Huang J(2017). Adjuvant transarterial chemoembolization for patients after curative resection of hepatocellular carcinoma: a meta-analysis. Scand J Gastroenterol; 52:624–634 pmid:28276833
- 10. Chen Y, Ren Z, Wang Z, et al(2018). Adjuvant Transarterial Chemoembolization for HBV-Related Hepatocellular Carcinoma After Resection: A Randomized Controlled Study. CLIN CANCER RES; 24:2074–2081 pmid:29420221
- 11. Sun HC, Wang Z, Zhou J, et al(2018). Guidelines for Diagnosis and Treatment of Primary Liver Cancer in China (2017 Edition). Liver Cancer; 7:235–260 pmid:30319983
- 12. Chinese College of Interventionalists. Clinical guideline: Effect of transarterial chemoembolization in patient with hepatocellular carcinoma in China. Chinese medical journal. 2018, 98(47): 3811–3819.
- 13. Barkun J, Clavien PA, de Oliveira ML, et al(2009). The Clavien-Dindo classification of surgical complications: five-year experience. ANN SURG; 250:187–196 pmid:19638912
- 14. Villanueva A (2019). Hepatocellular Carcinoma. N Engl J Med;380:1450–1462 pmid:30970190
- 15. Bagante F, Sahara K, Tsilimigras DI, et al (2019). Prognosis After Resection of Barcelona Clinic Liver Cancer (BCLC) Stage 0, A, and B Hepatocellular Carcinoma: A Comprehensive Assessment of the Current BCLC Classification. ANN SURG ONCO; 26:3693–3700
- 16. Rui JA, Zhou L, Zhou WX, et al (2017). Edmondson-Steiner grade: A crucial predictor of recurrence and survival in hepatocellular carcinoma without microvascular invasio. PATHOL RES PRACT; 213:824–830 pmid:28554743
- 17. Choi JS, Yi SW, Yi JJ, Lee YH, Han KJ (2018). Risk factors for hepatocellular carcinoma by age, sex, and liver disorder status: A prospective cohort study in Korea. CANCER-AM CANCER SOC; 124:2748–2757
- 18. Foerster F, Galle PR, Kudo M, et al (2019). Biology and significance of alpha-fetoprotein in hepatocellular carcinoma. LIVER INT; 39:2214–2229 pmid:31436873
- 19. Andreu-Oller C, Bassaganyas L, Montal R, et al (2019). Molecular portrait of high alpha-fetoprotein in hepatocellular carcinoma: implications for biomarker-driven clinical trials. Br J Cancer; 121:340–343 pmid:31285588
- 20. Berhane S, Johnson PJ, Kagebayashi C, et al (2015). Assessment of liver function in patients with hepatocellular carcinoma: a new evidence-based approach-the ALBI grade. J CLIN ONCOL; 33:550–558 pmid:25512453
- 21. Li X, Yuan Y, Ren J, Shi Y, Tao X (2018). Incremental Prognostic Value of Apparent Diffusion Coefficient Histogram Analysis in Head and Neck Squamous Cell Carcinoma. ACAD RADIOL; 25:1433–1438 pmid:29599009
- 22. Jian PE, Li SH, Wei W, et al (2018). Adjuvant transcatheter arterial chemoembolization after curative resection for hepatocellular carcinoma patients with solitary tumor and microvascular invasion: a randomized clinical trial of efficacy and safety. Cancer Commun (Lond); 38:61 pmid:30305149
- 23. Liang ZY, Qi YP, Zhong JH, et al (2019). Adjuvant transarterial chemoembolization for patients with hepatocellular carcinoma involving microvascular invasion. AM J SURG; 217:739–744 pmid:30103903
- 24. Ke Q, Deng M, Wang L, et al (2019). Adjuvant transarterial chemoembolization for patients with hepatocellular carcinoma after radical hepatectomy: a real world study. Scand J Gastroenterol; 54:1403–1411 pmid:31686555
- 25. Chen ZH, Zhang XP, Zhou TF, et al (2019). Adjuvant transarterial chemoembolization improves survival outcomes in hepatocellular carcinoma with microvascular invasion: A systematic review and meta-analysis. Eur J Surg Oncol; 45:2188–2196 pmid:31256949
- 26. Huang J, Liu FC, Li L, et al (2020). Prognostic Nomogram for Hepatitis B Virus-related Hepatocellular Carcinoma With Adjuvant Transarterial Chemoembolization After Radical Resection. Am J Clin Oncol; 43:20–27 pmid:31633514
- 27. Jing CY, Fu YP, Zheng SS, et al (2017). Prognostic nomogram for patients with hepatocellular carcinoma underwent adjuvant transarterial chemoembolization following curative resection. Medicine (Baltimore); 96:e6140 pmid:28296727
- 28. Lin N, Ke Q, Wang L, Zeng Y, Liu J (2019). Does postoperative adjuvant transarterial chemoembolization benefit for all patients with hepatocellular carcinoma combined with microvascular invasion: a meta-analysis. Scand J Gastroenterol; 54:528–537 pmid:31081401
- 29. Lu Y, Li W, Zhu M, et al (2016). Hepatitis B Virus X Protein Driven Alpha Fetoprotein Expression to Promote Malignant Behaviors of Normal Liver Cells and Hepatoma Cells. J CANCER; 7:935–946 pmid:27313784
- 30. Jiang R, Wang X, Zhang W, et al (2015). Effect of Tumor Size on Cancer-Specific Survival in Small Hepatocellular Carcinoma. MAYO CLIN PROC; 90:1187–1195 pmid:26231292
- 31. Lee S, Kang TW, Song KD, et al (2019). Effect of Microvascular Invasion Risk on Early Recurrence of Hepatocellular Carcinoma After Surgery and Radiofrequency Ablation. ANN SURG
- 32. Motomura T, Mano Y, Shirabe K, et al (2013). Neutrophil-lymphocyte ratio reflects hepatocellular carcinoma recurrence after liver transplantation via inflammatory microenvironment. J HEPATOL; 58:58–64 pmid:22925812