Association between preoperative aspartate aminotransferase-to-platelet ratio index and prognosis in colorectal liver metastases: a systematic review and meta-analysis

Introduction

Colorectal cancer (CRC), a common malignancy, stands third in incidence and second in mortality among all cancers worldwide (1,2). Studies have shown that the liver is the predominant site of distant metastasis in CRC (3). It is reported that 30% of individuals with CRC have synchronous liver metastases at diagnosis (4), whereas another 15–25% are diagnosed with postoperative liver metastases (5). Notably, 80–90% of these liver metastases cannot be radically resected (6). Therefore, liver metastasis is also the main cause of death among individuals with CRC (1). Surgical resection is still the primary potentially curative approach for individuals with resectable liver metastases. For unresectable liver metastases, a comprehensive strategy including systemic chemotherapy, targeted therapy, and local ablation is mainly used (7). However, therapeutic efficacy is limited by factors such as tumor biology, overall conditions of patients, and the accuracy of preoperative risk assessment (8). Therefore, identifying cost-effective, readily available, and reproducible indicators to accurately predict prognosis and thus improve preoperative risk stratification and perioperative counseling is still a major clinical challenge and a research focus.

The aspartate aminotransferase (AST)-to-platelet ratio index (APRI) is a simple indicator of liver fibrosis on the basis of AST level and platelet count (PLT). Investigations have shown that APRI is significantly linked to chemotherapy-induced adverse events and perioperative complications (9,10). Moreover, APRI specifically predicts colorectal liver metastases (CRLM) (11). This finding indicates its potential value for prognostic assessment in individuals with CRLM. However, the majority of the existing studies are single-center, small-sample, and retrospective studies (9,12,13) Furthermore, evidence regarding the utilization of preoperative APRI for predicting prognosis in individuals with CRLM remains limited and inconclusive. For example, research by Ashouri et al. (14) identifies APRI as an independent forecast factor of postoperative complications in CRLM, while the work of Hubert et al. (12) suggests its limited value for such predictions. On the one hand, significant discrepancies are found in selected cut-off values of APRI, patient stratification, and follow-up durations across studies. For instance, Chen et al. (15) utilize cut-off values of 0.33 and 0.25, respectively, to predict different endpoints, whereas Ashouri et al. (14) determine 0.365 as the optimal cut-off value through receiver operating characteristic (ROC) curve analysis. The enrolled patients across studies receive diverse treatment regimens, including neoadjuvant chemotherapy (NACT), local therapy, and systemic therapy. Furthermore, no systematic reviews or meta-analyses synthesize the existing findings, restricting the promotion and application of APRI in clinical practice.

Therefore, this study aimed to systematically collect and synthesize the available evidence regarding the association between preoperative APRI and outcomes in patients with CRLM. We performed a meta-analysis to evaluate the relationship of APRI with overall survival (OS), recurrence-related outcomes, and postoperative adverse events. The goal was to clarify the potential role of APRI as a supplementary marker for preoperative prognostic and perioperative risk stratification in CRLM, rather than as a direct determinant of treatment selection. We present this article in accordance with the PRISMA reporting checklist (16) (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1069/rc).

Methods

The research protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (registration number: CRD420251058027).

Data sources

Two investigators independently retrieved PubMed, EMBASE, Web of Science (WOS), and Cochrane from creation until April 23, 2025. The search was conducted without language restrictions, while only articles published in English were considered for final inclusion. The retrieval strategy integrated medical subject headings (MeSH) with free-text terms (FT) for key concepts such as “colorectal cancer”, “aspartate aminotransferases”, and “blood platelets”, along with all their potential synonyms and related expressions. The retrieval strategy is detailed in Appendix 1.

Inclusion and exclusion criteria

The inclusion and exclusion criteria for the current study were developed strictly as per the population, intervention, comparison, outcome, and study design (PICOS) principle. The inclusion criteria were defined below: (I) adults diagnosed with CRC presenting with liver metastases; (II) studies that calculated and reported the APRI values based on AST and PLTs measured preoperatively; (III) studies that provided data on the correlation of APRI with the prognosis of CRLM or diagnostic data [e.g., false positive (FP), false negative (FN), sensitivity, specificity, and ROC-area under the curve (AUC)]; (IV) observational study designs, including cohort studies and case-control studies. The exclusion criteria were outlined as follows: (I) publications with ineligible study types, such as meta-analyses, reviews, animal experiments, case reports, letters, or guidelines; (II) articles with full texts unavailable; (III) studies with data that could not be extracted; (IV) articles published not in English.

Literature screening

Two investigators screened the retrieved studies independently as per the pre-established inclusion and exclusion criteria. All records were imported into Endnote 20 to remove duplicates, and the remaining studies were screened against titles and abstracts. The entire content of the initially included literature was searched and thoroughly reviewed to determine eligible studies. Any disagreement between the two investigators was addressed through consultation with a third investigator.

Data extraction

Two investigators collected data from the eligible articles independently, involving author, publication year, country, study design, source of population, cancer type, sample size, proportion of male participants, preoperative treatments, APRI values, and outcome indicators. All collected data were cross-checked by the two investigators, and any disagreement was addressed through consultation with a third investigator.

Quality assessment

Two independent investigators appraised the quality of the included literature leveraging the Newcastle-Ottawa Scale (NOS) (17). The NOS involved 8 specific items across three fields. The comparability domain was awarded a maximum of 2 points, while a maximum of one point was assigned to each of the other seven items. Studies with total scores of 7–9 were considered high in quality, while those with 4–6 points were deemed moderate in quality. After completing the assessments, the two investigators cross-checked their evaluations. Any disagreement was resolved by consulting a third investigator. The itemized scores for each study across the three domains (selection, comparability, and outcome) are detailed in Table S1.

Statistical analysis

A meta-analysis of correlation data was implemented utilizing the metan module. All diagnosis-related data were analyzed leveraging the bivariate mixed-effects models via the MIDAS module. All analyses were conducted in Meta-Disc1.4 and STATA15.1. Heterogeneity was quantified employing the Higgins I² statistic. I²>50% revealed considerable heterogeneity among the studies. Thus, a random-effects model was applied. Otherwise, a fixed-effects model was leveraged. When substantial heterogeneity was present, sensitivity and subgroup analyses were implemented to examine the potential origins of heterogeneity. For survival outcomes (OS and recurrence-related outcomes), we extracted and pooled only hazard ratios (HRs). Odds ratios (ORs) were strictly excluded from all survival analyses to ensure statistical rigor and avoid bias inherent in cross-sectional measures. Pooled ORs were reserved solely for categorical outcomes, such as clinical postoperative complications and pathological liver injuries. Publication bias was initially assessed visually utilizing funnel plots and then statistically evaluated leveraging Egger’s test. If publication bias was ascertained, the trim-and-fill method was employed to analyze its influence on the results of the meta-analysis. A P<0.05 was deemed to be statistically significant for the pooled statistics. Regarding diagnosis-related data, forest plots were generated to compute pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR). A greater DOR value illustrated better diagnosis performance. AUC was determined based on the summary ROC (SROC) curve. Diagnostic efficiency was considered low, moderate, or high for AUC values of 0.5–0.7, 0.7–0.9, and 0.9–1.0, respectively. The presence of a threshold effect was examined utilizing Spearman’s correlation coefficient. A P<0.05 suggested no heterogeneity related to a threshold effect across studies. Finally, Deeks’ funnel plot asymmetry test was leveraged to appraise publication bias. A P<0.05 was deemed to be statistically significant. For survival outcomes, we extracted HRs from original studies, preferentially using the most fully adjusted estimates when available. Separate pooled analyses of adjusted versus unadjusted estimates were not performed due to the limited number of studies.

Results

Results of literature retrieval

In total, 432 studies were searched from PubMed, EMBASE, WOS, and Cochrane. After 68 duplicates were removed, 131 studies were excluded during the preliminary screening. Furthermore, additional studies with ineligible populations, unextractable data, or ineligible outcome indicators were eliminated. Ultimately, 10 studies were incorporated in this study, as depicted in the flow diagram (Figure 1).

Figure 1 Flow diagram of study selection. APRI, aspartate aminotransferase-to-platelet ratio index; CRLM, colorectal liver metastases.

Basic characteristics of included studies

A total of 10 studies (9,12-15,18-22) were included, published between 2010 and 2024. These studies were implemented across various regions: two in North America, six in Europe, and two in East Asia (China and Japan). Among these studies, nine were cohort studies, while the remaining one was a case-control study. Four were multi-center studies, and six were single-center studies. The sample sizes varied from 32 to 2,374 individuals. The study populations included a higher proportion of men in the vast majority of included studies, and only one study by Hubert et al. [2013] reported a higher proportion of women. The mean or median age of participants across the included studies ranged from 57 to 66 years, while the reported overall age span was 17 to 92 years. All 10 included studies involved the administration of NACT. Among them, six studies incorporated targeted therapy, and four employed a combination of NACT and targeted therapy. The cut-off values of APRI generally ranged from 0.17 to 0.40 in studies using the standard formula. However, two studies (by Amptoulach et al. and Shimagaki et al.) employed a scaled index [(AST/ULN)/PLT × 100], resulting in higher thresholds of 3.63 and 5.64, respectively. OS was quantitatively synthesized from three studies, whereas recurrence-related outcomes and complications were reported in the remaining cohorts. Study quality was evaluated utilizing NOS. Their NOS scores ranged between 7 and 8, indicating high overall quality (Table 1).

Correlation analysis

OS

Three studies involving surgical cohorts and reporting HRs were incorporated into the quantitative synthesis of OS. The pooled analysis using a random-effects model demonstrated that an elevated preoperative APRI was significantly associated with a shorter OS [HR =1.14, 95% confidence interval (CI): 1.01–1.28, P<0.05]. Moderate statistical heterogeneity was observed (I²=66.2%, P=0.052). The leave-one-out method was adopted for sensitivity analysis to explore the source of heterogeneity. The multicenter study by Chen et al. [2024] was identified as the primary contributor to the observed heterogeneity. After the exclusion of this study, the I² was reduced to 0.0%, while the pooled effect size remained statistically significant (HR =1.08, 95% CI: 1.02–1.15, P<0.05). These findings suggested that although the association between APRI and OS was directionally consistent, the magnitude of the effect might be influenced by study-specific factors such as sample size and statistical adjustment methods [e.g., inverse probability of treatment weighting (IPTW)] (Figure 2).

Figure 2 Forest plot for overall survival. (A) General forest plot for overall survival; (B) sensitivity analysis for overall survival. CI, confidence interval; HR, hazard ratio.

Progression-free survival (PFS)

Non-surgical cohorts were excluded to ensure clinical homogeneity. Consequently, only the multicenter study by Chen et al. [2024], involving 1,323 individuals, provided a quantitative HR for PFS in a surgical population. Their results indicated that a high preoperative APRI was significantly associated with a shorter PFS (HR =1.24, 95% CI: 1.04–1.48, P=0.02). Two other surgical studies, by Shimagaki et al. [2022] and Amptoulach et al. [2017], also investigated recurrence-related outcomes (disease-free survival, DFS). However, both studies reported no significant association between preoperative APRI and recurrence (all P>0.05) and did not provide specific HR values. Due to the lack of additional quantitative data from homogeneous surgical cohorts, a pooled meta-analysis for PFS was not performed. Therefore, no forest plot for PFS is presented, as combining clinically distinct populations (e.g., surgical vs. radioembolization) would introduce potential bias.

Complications and pathological liver injury

To address clinical heterogeneity, we categorized the findings into two groups: clinical postoperative complications and pathological liver injuries.

Clinical postoperative complications

Four studies (by Ashouri et al., Ratti et al., Yamashita et al., and Amptoulach et al.) reported clinical postoperative outcomes, including postoperative liver failure, hepatic insufficiency, liver-related morbidity, and bleeding. Given the variability in outcome definitions across studies, these events were summarized under the broader category of clinical postoperative complications. The pooled analysis using a random-effects model showed that an elevated preoperative APRI was associated with a higher risk of clinical postoperative complications (OR =1.76, 95% CI: 1.00–3.07, P=0.05). Moderate statistical heterogeneity was observed (I²=63.4%, P=0.042). Sensitivity analysis identified the study by Amptoulach et al. [2017] as the main contributor to heterogeneity, likely owing to its relatively large statistical weight (46.77%) and more conservative effect estimates compared with the other cohorts. Most included studies did not report complications according to the Clavien-Dindo classification, and data were insufficient to distinguish major (grade ≥III) from minor complications. Therefore, separate meta-analyses by complication severity were not feasible. These findings suggested that a high APRI might be associated with increased perioperative risk, although they should be interpreted with caution clinically due to heterogeneity in endpoint definitions (Figure 3).

Figure 3 Forest plot for clinical postoperative complications. (A) General forest plot for clinical postoperative complications; (B) sensitivity analysis for clinical postoperative complications. CI, confidence interval; OR, odds ratio.

Pathological liver injury

Three studies evaluated the relationship between preoperative APRI and chemotherapy-associated liver injury (CALI) confirmed by pathological examination. High APRI was identified as a strong predictor of severe sinusoidal obstruction syndrome (SOS) (OR =14.67, 95% CI: 5.03–42.79; Soubrane et al., 2010) and nodular regenerative hyperplasia (NRH) (OR =3.01, 95% CI: 1.26–7.20; Viganò et al., 2015). Additionally, Hubert et al. [2013] reported a significant univariate correlation between APRI and SOS severity (P<0.001). Given the biological differences between SOS and NRH and the high statistical heterogeneity (I²=80.2%), the pooled estimate (OR =6.44, 95% CI: 1.37–30.36) should be interpreted as exploratory. These results suggested that APRI primarily reflected underlying pathological damage to the liver parenchyma induced by preoperative chemotherapy (Figure 4).

Figure 4 Forest plot for pathological liver injury. (A) General forest plot for postoperative complications; (B) forest plot by type of pathological liver injury. CI, confidence interval; NRH, nodular regenerative hyperplasia; OR, odds ratio; SOS, sinusoidal obstruction syndrome.

Exploratory analysis of APRI’s predictive power for complications

Among the seven studies reporting the association, six provided sufficient data to construct 2×2 diagnostic tables and were thus included in the exploratory analysis. To further explore the potential clinical utility of APRI in identifying high-risk individuals, we carried out an exploratory diagnostic accuracy meta-analysis based on data from six studies. Notably, APRI functioned as a continuous prognostic biomarker rather than a binary diagnostic test, and the optimal thresholds for complications varied significantly across the included studies (ranging from 0.17 to 0.365).

The exploratory pooled values for sensitivity and specificity were 75% (95% CI: 67–81%) and 64% (95% CI: 55–72%), respectively. The pooled PLR was 2.08 (95% CI: 1.57–2.75), NLR was 0.39 (95% CI: 0.27–0.56), and DOR was 5.30 (95% CI: 2.86–9.83). The SROC curve showed an AUC of 0.76 (95% CI: 0.72–0.80).

While these mathematical estimates suggested a certain level of discriminative capacity for postoperative complications in CRLM patients, they should be interpreted with caution. Due to the variability in cut-off values and the inherent nature of APRI as a continuous index, APRI could not be used as a standalone diagnostic tool. This analysis provided a preliminary quantitative summary of the strength of the association rather than a standardized clinical threshold (Figures 5,6, and Figures S1-S3).

Figure 5 Pooled sensitivity and specificity. CI, confidence interval.

Figure 6 ROC-AUC. AUC, area under the curve; ROC, receiver operating characteristic; SENS, sensitivity; SPEC, specificity; SROC, summary receiver operating characteristic.

Sensitivity analysis

Sensitivity analysis was performed for OS and clinical postoperative complications using the leave-one-out approach to evaluate the stability of the pooled estimates and explore sources of heterogeneity.

For the OS analysis, the pooled HR remained consistently above 1.0 and statistically significant across all iterations, indicating robust results. Notably, the exclusion of the study by Chen et al. [2024] resulted in the most significant reduction in statistical heterogeneity, with I² dropping from 66.2% to 0.0%. This finding confirmed that, compared with other single-center surgical studies, the multicenter cohort study by Chen et al., which utilized IPTW adjustment, was the primary source of heterogeneity (Figure S4).

In the analysis of clinical postoperative complications, the leave-one-out plot identified the study by Amptoulach et al. [2017] as the major outlier. When that study was omitted, the pooled OR indicated a stronger association (increasing from 1.76 to approximately 2.53), accompanied by a decrease in heterogeneity. The reason was that the study by Amptoulach et al. carried the largest weight (46.77%) and reported a more conservative effect size than other cohorts. Despite these variations, the direction of the association between high APRI and clinical complications remained unchanged.

Due to the limited number of studies (N=2) investigating pathological liver injury, a formal leave-one-out sensitivity analysis was not performed for this specific endpoint. Overall, the sensitivity analyses demonstrated that the core findings of this meta-analysis were robust and not unduly affected by any single study, although specific effect sizes were influenced by methodological differences across studies (Figure S5).

Publication bias

Formal assessment of publication bias using funnel plots or quantitative tests (e.g., Egger’s test) was not performed. According to established methodological guidelines, such tests are not recommended when the number of included studies for a specific endpoint is less than 10, as they lack sufficient power to distinguish chance from real asymmetry. However, we acknowledged that the limited number of studies and the predominance of positive findings in the surgical cohorts might indicate a potential risk of publication bias. This has been addressed in the Limitations section.

Discussion

This study synthesized evidence from 10 clinical studies evaluating the relationship between preoperative APRI and prognosis in patients with CRLM, involving OS, recurrence-related outcomes, postoperative complications, and pathological liver injury. Preliminary results suggested that a higher preoperative APRI was significantly linked to reduced OS and increased risks of postoperative clinical complications and pathological liver injury. Six studies were included in an exploratory analysis of APRI for predicting postoperative outcomes, which yielded a pooled sensitivity of 75%, specificity of 64%, and AUC of 0.76. These results suggest that APRI may have moderate value as a supplementary tool for perioperative risk stratification, although its predictive performance should be interpreted with caution due to heterogeneity across studies and the lack of standardized cut-off values.

Regarding OS, the pooled analysis restricted to surgical cohorts showed that elevated preoperative APRI was associated with poorer OS (HR =1.14, 95% CI: 1.01–1.28, P<0.05). This finding demonstrated statistical significance. However, the effect size was small, the significance was borderline, and moderate heterogeneity was observed (I²=66.2%). In sensitivity analysis, after the exclusion of the multicenter study by Chen et al., heterogeneity was reduced to 0%, while the pooled association remained statistically significant (HR =1.08, 95% CI: 1.02–1.15). These findings uncover that the association between APRI and OS is directionally stable, although its magnitude may be influenced by differences in sample size, statistical adjustment methods, and study design. Therefore, the current evidence supports the potential value of APRI as a supplementary preoperative prognostic marker in patients with CRLM. However, it remains insufficient to regard APRI as an independent and strong determinant of prognosis. This association is also biologically plausible. First, APRI is a simple marker reflecting liver function reserve and the degree of fibrosis. An elevated APRI indicates hepatocyte injury (elevated AST) and thrombocytopenia (decreased PLT), which are commonly observed in individuals with hepatic fibrosis, cirrhosis, and portal hypertension (23). These pre-existing liver conditions can compromise the tolerance of patients for major hepatic resection and systemic therapies (e.g., chemotherapy or targeted therapy) (24). Second, liver fibrosis can accelerate tumor progression by remodeling the hepatic microenvironment. Fibrosis-associated collagen deposition, sinusoidal capillarization (25,26), and the activation of Kupffer cells (KCs) and hepatic stellate cells (HSCs) promote the sustained release of pro-inflammatory cytokines [e.g., interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α)] and pro-fibrogenic signals [e.g., transforming growth factor beta (TGF-β)]. This process fosters an immunosuppressive microenvironment, which facilitates tumor cell survival, invasion, and immune escape (27,28). Elevated AST not only serves as a marker of hepatocyte injury but is also associated with metabolic dysfunction and oxidative stress. Thrombocytopenia is not only an indirect manifestation of portal hypertension but may also promote the formation of metastatic lesions by impairing platelet-mediated immune protection (15). Collectively, these mechanisms may drive disease progression and worsen the prognosis in CRLM patients, thereby explaining the association between APRI and OS.

Beyond long-term survival, the current evidence also suggests that elevated preoperative APRI may be associated with an increased risk of postoperative complications. A pooled analysis of four studies showed that high APRI was related to a higher risk of clinical postoperative complications (OR =1.76, 95% CI: 1.00–3.07, P=0.05). However, this result was also of borderline statistical significance and was accompanied by moderate heterogeneity (I²=63.4%). Sensitivity analysis further showed that, after the exclusion of the heavily weighted study by Amptoulach et al., the effect estimate increased to approximately 2.53 and heterogeneity decreased. This finding suggests that the overall direction of the association was relatively consistent, although its magnitude was affected by individual studies. Thus, a higher APRI may indicate increased perioperative risk. Nevertheless, this finding should be interpreted carefully given that the definitions of complications varied across studies, and most did not classify complication severity according to the Clavien-Dindo system.

From a mechanism perspective, the potential link between APRI and postoperative complications is also biologically plausible. As an integrated marker reflecting hepatocellular injury, thrombocytopenia, and underlying hepatic fibrosis, an elevated APRI may indicate the presence of chemotherapy-related liver injury, sinusoidal endothelial damage, or reduced hepatic parenchymal reserve. This state of hepatic vulnerability may be further amplified after liver resection, thereby increasing the risk of postoperative liver dysfunction, liver failure, bleeding, and other liver-related complications. In parallel, an elevated APRI is closely associated with pathological liver injuries, including SOS and NRH. These pathological changes may impair liver regenerative capacity, disrupt intrahepatic microcirculation, and reduce tolerance to surgical stress, thus providing a pathological basis for adverse postoperative outcomes. In addition, fibrosis-related inflammation, oxidative stress, and microenvironmental imbalance may not only influence tumor biological behavior but also delay postoperative recovery. Collectively, these findings suggest that an elevated APRI may serve not only as an indirect marker of poor long-term oncologic prognosis but also as a clinical signal of increased perioperative risk.

Moreover, the exploratory diagnostic accuracy analysis suggested a moderate discriminative capacity of APRI for postoperative outcomes. Nevertheless, this finding should be interpreted cautiously because APRI is a continuous prognostic biomarker rather than a standardized binary diagnostic test, and the applied cut-off values varied considerably across studies.

Conclusions

In conclusion, a higher preoperative APRI may be associated with worse survival and a greater risk of postoperative adverse outcomes in patients with surgically treated CRLM. Given its low cost and easy availability, APRI may serve as a supplementary biomarker for preoperative prognostic and perioperative risk stratification. However, the current evidence is limited by the small number of studies for some endpoints, heterogeneity in patient populations and outcome definitions, and variability in APRI cut-off values. Therefore, APRI should not be used as a standalone basis for treatment selection. Its incremental clinical value beyond established prognostic factors requires further validation in large-scale, multicenter, prospective studies.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1069/rc

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1069/prf

Funding: This work was supported by the National Natural Science Foundation of China (NSFC) (No. 82374462).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1069/coif). All authors declared that this study was supported by the National Natural Science Foundation of China. The authors have no other 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.

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. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Siegel RL, Wagle NS, Cercek A, et al. Colorectal cancer statistics, 2023. CA Cancer J Clin 2023;73:233-54. [Crossref] [PubMed]
3. Tsilimigras DI, Brodt P, Clavien PA, et al. Liver metastases. Nat Rev Dis Primers 2021;7:27. [Crossref] [PubMed]
4. Zhou H, Liu Z, Wang Y, et al. Colorectal liver metastasis: molecular mechanism and interventional therapy. Signal Transduct Target Ther 2022;7:70. [Crossref] [PubMed]
5. Cervantes A, Adam R, Roselló S, et al. Metastatic colorectal cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol 2023;34:10-32. [Crossref] [PubMed]
6. Adam R, De Gramont A, Figueras J, et al. The oncosurgery approach to managing liver metastases from colorectal cancer: a multidisciplinary international consensus. Oncologist 2012;17:1225-39. [Crossref] [PubMed]
7. Wang Y, Zhong X, He X, et al. Liver metastasis from colorectal cancer: pathogenetic development, immune landscape of the tumour microenvironment and therapeutic approaches. J Exp Clin Cancer Res 2023;42:177. [Crossref] [PubMed]
8. Kawashima J, Endo Y, Woldesenbet S, et al. Preoperative identification of early extrahepatic recurrence after hepatectomy for colorectal liver metastases: A machine learning approach. World J Surg 2024;48:2760-71. [Crossref] [PubMed]
9. Shimagaki T, Sugimachi K, Mano Y, et al. Simple systemic index associated with oxaliplatin-induced liver damage can be a novel biomarker to predict prognosis after resection of colorectal liver metastasis. Ann Gastroenterol Surg 2022;6:813-22. [Crossref] [PubMed]
10. Desjardin M, Bonhomme B, Le Bail B, et al. Hepatotoxicities Induced by Neoadjuvant Chemotherapy in Colorectal Cancer Liver Metastases: Distinguishing the True From the False. Clin Med Insights Oncol 2019;13:1179554918825450. [Crossref] [PubMed]
11. Hu X, Marietta A, Dai WX, et al. Prediction of hepatic metastasis and relapse in colorectal cancers based on concordance analyses with liver fibrosis scores. Clin Transl Med 2020;9:13. [Crossref] [PubMed]
12. Hubert C, Sempoux C, Humblet Y, et al. Sinusoidal obstruction syndrome (SOS) related to chemotherapy for colorectal liver metastases: factors predictive of severe SOS lesions and protective effect of bevacizumab. HPB (Oxford) 2013;15:858-64. [Crossref] [PubMed]
13. Soubrane O, Brouquet A, Zalinski S, et al. Predicting high grade lesions of sinusoidal obstruction syndrome related to oxaliplatin-based chemotherapy for colorectal liver metastases: correlation with post-hepatectomy outcome. Ann Surg 2010;251:454-60. [Crossref] [PubMed]
14. Ashouri Y, Hsu CH, Riall TS, et al. Aspartate Aminotransferase-to-Platelet Ratio Index Predicts Liver Failure After Resection of Colorectal Liver Metastases. Dig Dis Sci 2022;67:4950-8. [Crossref] [PubMed]
15. Chen Q, Deng Y, Li Y, et al. Association of preoperative aspartate aminotransferase to platelet ratio index with outcomes and tumour microenvironment among colorectal cancer with liver metastases. Cancer Lett 2024;588:216778. [Crossref] [PubMed]
16. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [Crossref] [PubMed]
17. Wells G, Shea B, O'Connell J. The Newcastle-Ottawa Scale (NOS) for Assessing The Quality of Nonrandomised Studies in Meta-analyses. Ottawa Health Research Institute Web site. 2014;7. Available online
18. Amptoulach S, Gross G, Sturesson C, et al. Preoperative Aspartate Aminotransferase-to-Platelet Ratio Index Predicts Perioperative Liver-Related Complications Following Liver Resection for Colorectal Cancer Metastases. Scand J Surg 2017;106:311-7. [Crossref] [PubMed]
19. Pereyra D, Rumpf B, Ammann M, et al. The Combination of APRI and ALBI Facilitates Preoperative Risk Stratification for Patients Undergoing Liver Surgery After Neoadjuvant Chemotherapy. Ann Surg Oncol 2019;26:791-9. [Crossref] [PubMed]
20. Ratti F, Cipriani F, Catena M, et al. Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis. Eur J Surg Oncol 2014;40:1550-6.
21. Viganò L, Rubbia-Brandt L, De Rosa G, et al. Nodular Regenerative Hyperplasia in Patients Undergoing Liver Resection for Colorectal Metastases After Chemotherapy: Risk Factors, Preoperative Assessment and Clinical Impact. Ann Surg Oncol 2015;22:4149-57. [Crossref] [PubMed]
22. Yamashita S, Shindoh J, Mizuno T, et al. Hepatic atrophy following preoperative chemotherapy predicts hepatic insufficiency after resection of colorectal liver metastases. J Hepatol 2017;67:56-64. [Crossref] [PubMed]
23. Ji SS, Jiang HD, Jiang JC, et al. Applicability of liver stiffness measurement based nomograms to the assessments of hepatitis B related significant fibrosis and cirrhosis. Clin Chim Acta 2019;489:75-82. [Crossref] [PubMed]
24. Affo S, Yu LX, Schwabe RF. The Role of Cancer-Associated Fibroblasts and Fibrosis in Liver Cancer. Annu Rev Pathol 2017;12:153-86. [Crossref] [PubMed]
25. Shimizu S, Yamada N, Sawada T, et al. In vivo and in vitro interactions between human colon carcinoma cells and hepatic stellate cells. Jpn J Cancer Res 2000;91:1285-95. [Crossref] [PubMed]
26. Shasha T, Gruijs M, van Egmond M. Mechanisms of colorectal liver metastasis development. Cell Mol Life Sci 2022;79:607. [Crossref] [PubMed]
27. Monvoisin A, Bisson C, Si-Tayeb K, et al. Involvement of matrix metalloproteinase type-3 in hepatocyte growth factor-induced invasion of human hepatocellular carcinoma cells. Int J Cancer 2002;97:157-62. [Crossref] [PubMed]
28. Kang N, Gores GJ, Shah VH. Hepatic stellate cells: partners in crime for liver metastases? Hepatology 2011;54:707-13. [Crossref] [PubMed]