With the rapid rise in its prevalence, hepatocellular carcinoma (HCC) has become the sixth most aggressive malignant tumor and the second-leading cause of cancer-related deaths worldwide (1). Among all therapeutic strategies, surgical resection remain the mainstay of the curative approach for HCC nowadays (2). Even though surgical technique and perioperative care have significantly improved over the past few years, post-hepatectomy liver failure (PHLF) is still the primary driver of morbidity and mortality after hepatectomy in HCC patients (3, 4). In patients with low liver regeneration capability and reduced function reservation of remnant liver tissue following hepatectomy, PHLF occurs most frequently. Therefore, it is critically important to predict the risk of PHLF, which is essential for surgeons to choose individualized treatment.

To accurately predict PHLF, several articles relating to PHLF have been published (5–8). Although these efforts on the preoperative prediction of PHLF have been made, an effective prediction model is still lacking (9). For many years, the clinical scoring systems, such as CP score (Child-Pugh) (10, 11) and MELD score (Model of End-Stage Liver Disease) (12), are widely used for preoperative assessment of liver function. The Child-Pugh score system has some drawbacks and limitations because of its two subjective clinical variables-ascites and hepatic encephalopathy (11). Similarly, the MELD score is not optimal for the prediction of PHLF. Recent research indicated that a new evidence-based model, called the albumin-bilirubin (ALBI) score (13). has been developed to assess liver function reserve. And it has been proven to be superior in estimating PHLF and survival of HCC patients undergoing liver resection (14, 15). Indocyanine green (ICG) (16), a nontoxic, infrared, and photosensitive dye, can be combined with albumin and beta lipoprotein. As a quantitative test to assess hepatic blood flow and liver function, the ICG clearance test at 15 minutes (ICG-R15) is now commonly used to evaluate reserved liver function in surgical patients. Moreover, ICG-R15 has been proven to be a reliable predictor of PHLF recently (17, 18).

Serum inflammatory indices are a reflection of the systemic inflammatory, which plays a significant role in the pathogenesis and progression of liver cirrhosis (19). Recently, a study has demonstrated that chronic inflammation can increase the operative risk of liver resection (20). However, the exact relationship between Serum inflammatory indices and PHLF is not very clear. And whether the combination of inflammatory indices and ICG-R15 could add more benefit in predicting PHLF is worth exploring.

Therefore, this study aims to investigate the possibility of inflammatory markers in predicting PHLF. Moreover, we develop a nomogram based on ICG-R15 and inflammatory markers, and compare its predictive value with traditional models, such as CP score, MELD score and ALBI score, in HCC patients undergoing hepatectomy.

Post-hepatectomy liver failure is one of the most feared complications after hepatectomy in HCC patients. There is a need to prospectively identify HCC patients at risk of PHLF. Therefore, establishing a prediction model of PHLF is necessary to improve clinical decisions.

Many models have been put forward to predict the occurrence of PHLF. But the predictive model is still evolving due to its multifactorial causative factors (25). Based on our clinical data, we performed this study to recognize the risk of PHLF in HCC patients in order to construct a nomogram for predicting PHLF.

In our analysis, we found that tumor size, blood loss(≥400ml), cirrhosis, PT, ICG-R15 and APRI were the independent risk factors for PHLF in HCC patients through the multivariable logistic regression analysis. Based on the risk factors, we developed the nomogram to predict the occurrence of PHLF. As for tumor size, we think the size of tumor influence the scope of resection of liver parenchyma, consequently affecting the volume of the remaining healthy liver. Several reports suggested that patients with a smaller liver remnant have a greater chance of developing PHLF (26, 27). Also, Heng Zou and his team found liver remnant is a good predictor of PHLF (28). As for blood loss, Osamu Aramaki (29) in his article demonstrated that intraoperative blood loss was the most crucial factor related to postoperative complications, including PHLF. Also back in 2007, Marieke T. de Boer found that there is a significant and clinically relevant association between blood loss and postoperative mortality and morbidity (30). Considering that the liver has abundant blood flow, excessive bleeding inevitably leads to impairment of liver cells, with the liver function decline. Liver cirrhosis has a great effect on liver regeneration after hepatectomy. In other words, Cirrhosis is a negative predictor of liver regeneration and liver function. It is extremely important to evaluate the degree of cirrhosis since it is the dominant risk factor for both PHLF and the prognosis of HCC patients (31). HCC patients with severe cirrhosis have higher morbidity and mortality rates after hepatectomy when compared with non-cirrhotic patients (32). Our study also confirms this point of view that cirrhosis is an independent risk factor of PHLF. Prothrombin time (PT) is an important reflection of coagulation status. It represents an essential parameter in many models that evaluate liver function, such as the Child-Pugh score system. Similarly, PT plays an important role in PHLF according to several studies (33, 34). For decades, ICG-R15 has been applied to test liver function prior to hepatectomy. Especially in Eastern countries, ICG-R15 was the most common approach to select suitable HCC patients for liver resection (35, 36). Admittedly, tumor size, blood loss, cirrhosis, PT and ICG-R15 have been demonstrated to predict PHLF in many studies which are consistent with our conclusion.

In so many serum inflammatory indices, only APRI is a unique independent factor in predicting PHLF in our analysis. APRI was used to predict the degree of liver fibrosis in patients since it is a non-invasive test (37). In 2015, the World Health Organization recommend APRI for non-invasive evaluation for liver cirrhosis in patients with chronic hepatitis B infection. APRI consists of two components, AST and Plt. The progression of liver cirrhosis in HCC patients is inevitably accompanied by sustained damage to liver cells, which results in the release of AST and the increase of its concentration in peripheral blood (38). The platelet count could be decreased because of sequestration and destruction of platelets in the enlarging spleen (portal hypertension) (39). And, Thrombopoietin (TPO) synthesis in the liver is reduced because of liver cirrhosis which could stimulate platelet formation (40). These may explain why APRI, not other inflammatory indices could be used to predict PHLF. In our study, APRI has an OR value of 4.652, which is higher than the other independent risk factors. That means it has a higher correlation with PHLF than others. APRI presented in nomogram confirmed the conclusion as well.

Compared to CP score, ALBI score and MELD score, our nomogram performed well in predicting PHLF and its prediction was supported by the C-index (0.845 and 0.854 for the training cohort and validation cohort, respectively). Traditionally, nomogram is assessed using metrics of diagnostic performances such as specificity, sensitivity and the C-index which fail to determine the clinical value. DCA is a wide-used tool for assessing the benefit of a diagnostic test across a variety of patient preferences for recognizing risks of undertreatment and overtreatment to facilitate decisions about test selection and use. In our study, the DCA indicated that our nomogram brought more benefits than other models in the training cohort and validation cohort. So our nomogram could be used uniformly in clinical practice.

Our nomogram is helpful in predicting PHLF, which can guide therapeutic decisions. By doing this, specific monitoring strategies can be established according to the specific risk categories. For example, if HCC patients are evaluated as a high-risk group of PHLF, we would recommend early use of hepatic protectant, close supervision and intensive care after surgery.

To our knowledge, this is the first nomogram based on inflammatory indices and ICG-R15 to predict PHLF. We emphasized the importance of APRI in the prediction model. However, there are several limitations in the present study. The main limitation is its retrospective nature. Although our data came from two academic centers, our study has a relatively small sample size. A future multicenter study including a larger number of HCC patients is needed to confirm our findings. Then, as mentioned above, there are many definitions of PHLF, resulting in a wide variation in the incidence of PHLF. We could compare different diagnostic criteria to determine which one is more beneficial to patients. Finally, the main etiology of HCC was chiefly HBV. Even though it didn’t play an important role in the prediction of PHLF, it is necessary to include more populations with different etiologies such as alcoholic liver disease or HCV.

In conclusion, we demonstrated that tumor size, blood loss, cirrhosis, PT, ICG-R15 and APRI are the independent risk factors of prediction of PHLF. We present a novel prediction nomogram of PHLF by combining the independent risk factors. The nomogram showed a good predictive performance and would be a convenient tool for us to facilitate clinical decisions.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The study was approved by the Ethics Committee of the Xiangya Hospital of Central South University and the Second Xiangya Hospital of Central South University in compliance with the Declaration of Helsinki. Written informed consent was obtained from all patients for use of their data in this study.

TF and LeZ contributed to conception and design of the sstudy. TF and LeZ organized the database. TF and LeZ performed the statistical analysis. TF wrote the first draft of the manuscript. TF wrote sections of the manuscript. All authors contributed to the article and approved the submitted version.

This study was funded by the National Nature Science Foundation of China (NO. 81771932).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.