Development and validation of a prognostic nomogram... : Medicine (original) (raw)

1. Introduction

Hepatocellular carcinoma (HCC) constitutes the predominant histological subtype of primary liver malignancies, representing roughly 85% to 90% of instances.[1] Globally, it ranks as the third leading cause of cancer-related deaths, contributing significantly to mortality and morbidity rates in China.[2,3] Surgical resection represents a curative treatment option for early-stage HCC, leading to significant improvement in survival outcomes for affected patients.[4,5] However, the majority of patients with HCC are diagnosed at advanced stages with metastatic tumors, rendering surgical intervention infeasible and necessitating systemic therapies. The progression from single-agent targeted therapies (sorafenib and lenvatinib) to combinations of checkpoint inhibitors and targeted therapies (atezolizumab plus bevacizumab therapy) has been observed.[6,7] Despite significant advancements, only a fraction of patients experiences lasting clinical benefit due to immune-related adverse events, posing considerable therapeutic challenges.[8,9] Consequently, these patients typically face a grim prognosis.[10,11] For patients with unresectable HCC, loco-regional treatments are considered a viable option.[12] These treatments encompass local ablation therapy, hepatic arterial intervention therapy, radiation therapy, and radioisotope immunotherapy, and others.[13] Alessandro et al discussed available evidence and current and future research on combined trans-arterial chemoembolization (TACE) and systemic treatments aimed at enhancing antitumor efficacy and producing synergistic effects.[14] This includes the use of antiangiogenic agents, immune checkpoint inhibitors, and immune-based combinations for the benefit of HCC patients.

Cancer staging is essential as it assists in determining prognosis and guiding treatment recommendations.[15] The Barcelona Clinic Liver Cancer staging system and tumor-node-metastasis (TNM) staging system are commonly used for assessing the prognosis of HCC patients, but these staging systems primarily consider the impact of tumor burden and other physical indicators on patient prognosis. However, even among patients with the same stage, there often exists significant variability in prognosis. Therefore, it is worthwhile to further explore more accurate and personalized prognostic assessment indicators. The prognosis of HCC patients is influenced by a range of factors, such as tumor grading and staging, liver function status, treatments received, and the presence of tumor local invasions. For instance, portal vein tumor thrombus (PVTT) is a common complication that signifies a very poor prognosis in HCC patients, with around 40% of these patients presenting with PVTT at the time of diagnosis.[16] Additional factors impacting prognosis encompass the overall health condition of the individual, nutritional status, and inflammatory response, indicating limitations in the ability of the TNM staging system to offer personalized prognostic predictions. Since anatomical features alone are far from sufficient to predict the prognosis of HCC,[17,18] it is imperative to develop a precise prognostic model to accurately discern individuals with varying survival risks, potentially aiding in informed clinical decision-making.

Nomograms are widely acknowledged as reliable and effective instruments for assessing cancer outcomes, as they integrate diverse pathological and clinical features. It is widely acknowledged that nomograms outperform traditional TNM staging systems in prognosticating cancer.[19–21] Nevertheless, currently, there is a lack of nomograms available to predict overall survival (OS) specifically for HCC patients who have undergone surgery or loco-regional therapy as their primary form of treatment.

In the past, prognosis assessment of HCC patients in clinical practice mainly relied on their physiological status (such as age, gender, etc.), pathological characteristics (including histological subtypes, tumor size and number, microvascular invasion, distant metastasis, etc.), and tumor microenvironment (presence of viral infections or liver cirrhosis, etc.). With the deepening research on tumors, multiple studies have shown the significant role of inflammation and nutritional status in tumor progression. Biological indicators such as neutrophils, lymphocytes, and fibrinogen not only represent inflammatory responses but also can predict tumor prognosis.[22–24] Due to the complexity and heterogeneity of tumors, the value of using single indicators is limited in assessing the overall inflammatory status. Therefore, some scholars have proposed a series of composite indicators to better predict the prognosis of HCC patients, including the neutrophil times γ-glutamyl transpeptidase to lymphocyte ratio (NRLR),[25] the platelet to lymphocyte ratio (PLR),[26] the lymphocyte to monocyte ratio (LMR),[27] the prognostic nutritional index (PNI),[28] the systemic immune-inflammation index (SII),[29] and the Royal Marsden Hospital (RMH) Score,[30] among others.

Given this context, our study focuses primarily on inflammatory indicators related to liver function and the objective was to construct a nomogram for HCC patients treated primarily with surgery or loco-regional therapy, utilizing data from cases at the First Affiliated Hospital of Anhui Medical University, and subsequently validate its predictive performance. According to the cutoff value of the overall score, the nomogram could differentiate high-risk from low-risk HCC patients and could elucidate the survival benefits of primary therapy.

2. Patients and methods

2.1. Patients and variables

Clinical data and medical radiological features for patients with HCC from the First Affiliated Hospital of Anhui Medical University between January 2017 and June 2023 were retrospectively analyzed. Inclusion criteria were HCC confirmed by pathology, patients treated primarily with surgery or loco-regional therapy, and no severe failure of the heart, lungs, kidneys, and other organs. Exclusion criteria were age < 18 years, intrahepatic cholangiocarcinoma or other malignancies confirmed by pathology, incomplete clinical and follow-up data, and peri-treatment death (survival time < 1 month). The selection of the study population is represented in Figure 1. Overall survival (OS) was chosen as the primary endpoint of the study. OS was defined as the time from date of diagnosis of HCC to the date of death from any cause or the last date of follow-up.

Figure 1.:

Flowchart of the patient screening. HCC = hepatocellular carcinoma.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This retrospective study protocol was approved by the Institutional Review Boards of the First Affiliated Hospital of Anhui Medical University and the requirement for informed consent was waived due to the retrospective nature of this study.

We applied electronic medical record system to collect the following data: demographic and laboratory characteristics. All enrolled patients had venous blood samples taken within 24 hours of admission and performed complete blood count analysis. About variables and biomarkers, body mass index (BMI) equaled weight in kilograms divided by height in meters squared, platelet to lymphocyte ratio (PLR) was the ratio of platelet count to lymphocyte count, neutrophil to lymphocyte ratio (NLR) was the ratio of neutrophil count to lymphocyte count, γ-glutamyl transpeptidase to platelet ratio (GPR) was the ratio of γ-glutamyl transpeptidase to platelet count, neutrophil times γ-glutamyl transpeptidase to lymphocyte ratio (NRLR) was defined as neutrophil count * γ-glutamyl transpeptidase/ lymphocyte count, neutrophil to (leukocyte subtract neutrophil) ratio (DNLR) was defined as neutrophil count/ (leukocyte count - neutrophil count), monocyte to lymphocyte ratio (MLR) was the ratio of monocyte count to lymphocyte count, systemic inflammation response index (SIRI) was defined as monocyte count * neutrophil count/ lymphocyte count, SII was defined as platelet count * neutrophil count/lymphocyte count, PNI was defined as albumin + 5* lymphocyte count, and aspartate aminotransferase to alanine aminotransferase ratio (AAR) was the ratio of aspartate aminotransferase to alanine aminotransferase. In all, 30 demographic and clinical variables were collected for risk factor analysis of OS, as shown in the Table 1 and a Kaplan–Meier survival curve was employed to identify the optimal cutoff value for converting continuous variables into categorical variables.

Table 1 - Demographics and clinical characteristics of 143 patients

AAR = aspartate aminotransferase to alanine aminotransferase ratio, AFP = α-fetal protein, BMI = body mass index, CA19-9 = cancer antigen 19-9, DNLR = neutrophil to (leukocyte subtract neutrophil) ratio, ECOG PS = Eastern Cooperative Oncology Group Performance Status, GPR = γ-glutamyl transpeptidase to platelet ratio, LDH = lactate dehydrogenase, MLR = monocyte to lymphocyte ratio, NLR = neutrophil to lymphocyte ratio, NRLR = neutrophil times γ-glutamyl transpeptidase to lymphocyte ratio, PLR = platelet to lymphocyte ratio, PNI = prognostic nutritional index, PT = prothrombin time, PVTT = portal vein tumor thrombus, SII = systemic immune-inflammation index, SIRI = systemic inflammation response index, TBIL = total bilirubin, UREA = serum urea nitrogen.

*n (%).

†Pearson’s Chi-squared test; Fisher’s exact test.

2.2. Development and validation of nomograms

All patients were allocated randomly into a training cohort and a validation cohort at a ratio of 7 to 3. Subsequently, univariate and multivariate Cox regression analyses were performed on the training cohort to identify risk factors affecting patients with HCC (P < .05). Utilizing the results from multivariate analysis, an OS nomogram was developed using R software to forecast the 1-, 2-, and 3-year survival rates in HCC patients. The accuracy of the nomogram’s predictions was assessed using Harrell’s C-index. Sensitivity and specificity were evaluated using receiver operating characteristic (ROC) curves. The calibration plot was utilized to gauge the agreement between predicted and observed outcomes. Additionally, decision curve analysis (DCA) was conducted to assess the clinical utility of the nomogram. Using the risk scores derived from the nomogram, patients were categorized into low-risk and high-risk groups. Kaplan–Meier curves were then utilized to assess prognostic disparities between the 2 groups, further validating the discriminatory capability of the nomogram model.

2.3. Statistical analysis

All statistical analyses were conducted using R software (version 4.2.1) and the associated R package. The X-tile software (version 3.6.1) was employed to ascertain the optimal cutoff values for the nomogram. The chi-square test was used to compare differences between the 2 groups. Univariate and multivariate Cox regression analyses were conducted to determine hazard ratios (HRs) and their corresponding 95% confidence intervals (CIs) for variables, identifying independent prognostic factors. Kaplan–Meier survival curve was used to determine the optimal cutoff value for converting continuous variables into categorical variables. Categorical variables were expressed as numbers (percentage). A two-tailed P value of less than .05 was considered statistically significant.

3. Results

3.1. Patient characteristics

In this study, we analyzed the baseline demographic and clinical characteristics of a total of 143 patients, randomly divided into a training cohort of 100 individuals and a validation cohort of 43 individuals. Among the patients, gender distribution showed no significant differences between cohorts (P = .185). The majority of patients in both cohorts were of male (87.4% overall, 85.0% in the training cohort, and 93.0% in the validation cohort), with a smaller representation of female (12.6% overall, 15.0% in the training cohort, and 7.0% in the validation cohort). Regarding age, the distribution was similar across cohorts (P = .316), with age ≤ 56 years representing 51.7% of the total cohort, and age > 56 years representing 48.3%. For BMI, no significant differences were observed between cohorts (P = .491), with distribution across 18.5 to 24.0 kg/m2, <18.5 kg/m2, 24.0 to 28.0 kg/m2, and ≥ 28 kg/m2. The demographic and clinical characteristics of all the patients are exhibited in Table 1.

3.2. Analysis of prognostic factors

In a univariate Cox regression analysis of the training cohort, gender, Child-Pugh scale, TNM stage, PVTT, PLR, NLR, GPR, NRLR, DNLR, MLR, SIRI, SII, PNI, and AAR were identified as risk factors for OS (P < .05). Multivariate analysis showed that hepatitis B (P = .005), hepatitis C (P < .001), tumor number (P = .018), TNM stage (P = .045), PVTT (P < .001), DNLR (P = .005), PNI (P = .038), AAR (P < .001), total bilirubin (P = .013), lactate dehydrogenase (P < .001), UREA (P = .047), and prothrombin time (P = .013) were independent predictors of OS. The results of univariate and multivariate analyses are shown in Table 2.

Table 2 - Univariate and multivariate analysis of OS in the training cohort

AAR = aspartate aminotransferase to alanine aminotransferase ratio, AFP = α-fetal protein, BMI = body mass index, CA19-9 = cancer antigen 19-9, CI = confidence interval, DNLR = neutrophil to (leukocyte subtract neutrophil) ratio, ECOG PS = Eastern Cooperative Oncology Group Performance Status, GPR = γ-glutamyl transpeptidase to platelet ratio, HR = hazard ratio, LDH = lactate dehydrogenase, MLR = monocyte to lymphocyte ratio, NLR = neutrophil to lymphocyte ratio, NRLR = neutrophil times γ-glutamyl transpeptidase to lymphocyte ratio, PLR = platelet to lymphocyte ratio, PNI = prognostic nutritional index, PT = prothrombin time, PVTT = portal vein tumor thrombus, SII = systemic immune-inflammation index, SIRI = systemic inflammation response index, TBIL = total bilirubin, UREA = serum urea nitrogen.

*P < .05,

**P < .01,

***P < .001.

3.3. Development and validation of a nomogram

To simplify the model, based on the results of multivariate Cox regression analysis, observe the HR and P values on the training cohort and validation cohort, artificially reselect variables once again to include independent predictive factors with smaller P values and statistical significance, while also including variables with clinical significance, to establish a nomogram to predict OS at 1-, 2-, and 3-year in HCC patients after surgery or loco-regional therapy (Fig. 2). Variables that were ultimately included in the model included BMI, PVTT, leukocyte, GPR, MLR, PNI, and therapy. It was evident from the nomogram that PVTT was the most important factor affecting patient survival. In the present study, C-index, ROC curves, calibration plots, and DCA curves were used to identify the superiority of the nomogram in predicting the prognosis of HCC patients after surgery or loco-regional therapy. The C-indexes of the nomogram were 0.745 (95% CI: 0.669–0.820) in the training cohort and 0.650 (95% CI: 0.507–0.792) in the validation cohort, showing satisfactory results. In the ROC curves, a high area under the ROC (AUC) was observed both in the training cohort and validation cohort. AUC values for 1-, 2-, and 3-year OS of the training cohort were 0.725, 0.854, and 0.820, as for the values of the validation cohort were 0.830, 0.655, and 0.585 (Fig. 3). Indicating that the model has good discriminatory ability. In addition, the calibration plots of the 1-, 2-, and 3-year OS showed that the actual observations of the 2 cohorts were in high agreement with the predictions of the nomogram (Figure S1, Supplemental Digital Content, https://links.lww.com/MD/O157). Furthermore, based on the DCA curve results, the nomogram demonstrated good clinical utility and positive net benefit in both cohorts, which can play a practical role in decision-making (Figure S2, Supplemental Digital Content, https://links.lww.com/MD/O158).

Figure 2.:

Nomogram for 1-, 2-, and 3-year OS. BMI = body mass index, GPR = γ-glutamyl transpeptidase to platelet ratio, MLR = monocyte to lymphocyte ratio, OS = overall survival, PNI = prognostic nutritional index, PVTT = portal vein tumor thrombus.

Figure 3.:

The ROC curves of the nomogram to predict OS at (A) 1-, 2-, and 3-year OS in training cohort, (B) 1-, 2-, and 3-year OS in validation cohort. OS = overall survival, ROC = receiver operating characteristic.

3.4. Performance of the nomogram in stratification

Based on the constructed nomogram model, all patients were categorized into 2 subgroups according to the cutoff values of the nomogram for OS: low-risk group (total score ≤ 152.04) and high-risk group (total score > 152.04). Kaplan–Meier survival curve analysis indicated that, both in the training cohort and the validation cohort, patients in the high-risk group suffered a poorer prognosis than those in the low-risk group (Fig. 4).

Figure 4.:

Kaplan–Meier curves of OS for risk classification based on the nomogram scores. (A) in training cohort; (B) in validation cohort. OS = overall survival.

4. Discussion

HCC ranks among the most prevalent malignancies globally,[1] imposing a significant burden on both individuals and society as a whole. To date, the prognosis for HCC patients treated primarily with surgery or loco-regional therapy is largely determined by the Barcelona Clinic Liver Cancer staging or TNM staging, and the follow-up patterns of all patients post-treatment are generally similar. Unfortunately, it is often observed that some patients experience disease recurrence and metastasis shortly after surgery or loco-regional therapy, leading to a shorter overall survival. Additionally, some low-risk HCC patients develop irreversible complications due to over-treatment, which may significantly impact their quality of life. Our novel nomogram, incorporating inflammatory indicators, is the first to predict OS in this context, effectively stratifying patients into risk groups. High-risk patients may benefit from intensified therapy and follow-up, while low-risk patients should avoid overtreatment. This nomogram offers precise survival predictions and highlights the need for personalized follow-up and treatment plans.

In this study, we analyzed the baseline demographic characteristics and clinical information of all patients. We observed a majority of male patients in both cohorts, which aligns with previous findings.[31,32] This gender disparity may be linked to the presence of androgen receptors (AR) in men. Research has shown that AR signaling and its interaction with endoplasmic reticulum stress can influence the development of HCC.[33] Furthermore, a significant portion of HCC patients tested positive for alpha fetoprotein. This underscores the practicality of utilizing alpha fetoprotein as a serum biomarker for diagnosing HCC.[34] Univariate analysis revealed significant associations between T stage and TNM stage with the prognosis of HCC patients. Liu et al[35] confirmed that TNM staging is an independent factor affecting the survival of liver cancer patients. However, T stage did not emerge as an independent prognostic factor in the multivariate analysis, suggesting that it may not be critical to OS. This could have varied depending on the HCC patients included in the screening process. Additionally, the impact of treatment on patients’ OS should not be overlooked. Our study found that treated primarily with loco-regional therapy, low BMI, present with PVTT, decreased leukocyte level, high GPR, high MLR and low PNI were the independent risk factors for poor prognosis. This study combined all the aforementioned independent prognostic factors to develop a nomogram for predicting OS in HCC patients primarily treated with surgery or loco-regional therapy. A series of validation tests confirmed the model’s accuracy and reliability.

This study demonstrates that patients with HCC who undergo surgery have a longer survival period compared to those receiving loco-regional therapy. For individuals with early-stage HCC, surgical resection can extend survival or potentially result in a cure.[36] Those who did not receive surgical treatment exhibited markedly lower survival rates.[37] Research has reported that there is no statistically significant difference in efficacy and prognosis between percutaneous radiofrequency ablation and laparoscopic liver resection for primary small liver cancers, and complications after radiofrequency ablation are fewer.[38] However, recent studies show that the rate of local recurrence after surgical resection is significantly lower than that after radiofrequency ablation, and the reason for no difference in long-term survival between the 2 treatments may be attributed to patients receiving more salvage therapy after recurrence.[39] Therefore, we recommend that patients with early HCC should undergo surgery as early as possible when the indication for surgery is available.

In a multivariate analysis of this study, BMI were significantly associated with the prognosis of HCC patients. Previous studies have found that lean patients with HCC are more likely to have a history of hepatitis C and alcohol abuse at the time of diagnosis, and they tend to present with larger and more aggressive tumors, leading to poorer prognosis,[40] which is consistent with the findings of this study.

It was evident from the nomogram that PVTT was the most important factor affecting patient survival. PVTT is considered the most common form of vascular invasion in HCC, with an incidence rate reaching 44.0% to 62.2% among HCC patients. This type of vascular invasion is associated with poor prognosis and poses a significant challenge in the treatment of HCC due to its high recurrence rate and the aggressive nature of the disease when PVTT is involved.[41]

Inflammatory signaling, as one of the critical hallmarks of malignant tumors, promotes the initiation and progression of cancer.[23] HCC is a prototypical inflammation-associated malignancy, where tumor-associated macrophages, neutrophils, and tumor-infiltrating lymphocytes are major components of the liver cancer microenvironment.[42] In recent years, typical inflammatory indicators such as the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) have been confirmed to be associated with the prognosis of HCC.[26,43–45] When patients have good liver function, clinicians can employ more flexible treatment approaches, and a well-functioning liver reserve is beneficial for the prognosis of HCC patients. Therefore, this study will primarily focus on inflammatory markers related to liver function, including leukocyte, GPR, MLR, and PNI.

Previous research on the relationship between leukocyte and cancer mortality has produced inconsistent findings.[46] Studies have shown a robust association between leukocyte levels and unfavorable outcomes in individuals with HCC. Elevated levels of neutrophils and monocytes, both types of leukocytes, are associated with advanced tumor stages, vascular invasion, and extrahepatic metastasis, all indicative of a poorer prognosis.[47,48] However, this study identified a correlation between decreased leukocyte level and poor prognosis in HCC patients. One aspect of the analysis is that patients with advanced HCC often have concurrent severe liver cirrhosis and splenomegaly, leading to increased destruction of leukocyte.[49] On the other hand, decreased leukocytes indicate compromised immune function in the body, making such patients prone to serious complications such as severe infections, which can affect prognosis.[50] Indeed, the relationship between leukocyte and the prognosis of HCC warrants further investigation to elucidate its underlying mechanisms.

Gamma-glutamyl transferase (GGT) is another important indicator reflecting liver function. HCC occurrence models indicate that GGT expression is associated with tumor formation and progression.[51] Reports suggest that serum GGT levels serve as prognostic markers for HCC patients undergoing radiofrequency ablation therapy.[52] Platelets, as markers associated with cancer, can be activated by cancer cells, leading to cancer-related inflammation, metastasis, and progression.[53] Lemoine et al[54] first reported that GPR could predict liver fibrosis and cirrhosis in patients with chronic hepatitis B in West African countries, thereby drawing increasing attention to the relationship between GPR and cancer. Preoperative GPR serves as an effective prognostic indicator for patients with hepatitis B-related HCC.[55,56]

As a systemic inflammatory marker in peripheral blood, the MLR is cost-effective to examine, easy to calculate using monocyte and lymphocyte counts, and can be measured repeatedly. Activated circulating monocytes can secrete multiple proinflammatory cytokines involved in tumor development and progression.[57] Conversely, a low lymphocyte count is associated with systemic inflammatory responses and promotes cancer progression through effects on cell-mediated immunity.[58] A meta-analysis showed that a low MLR significantly correlated with increased OS in HCC patients, regardless of sample size, publication type, or MLR cutoff value.[59] Wu et al[60] reported that combining the MLR with clinical risk factors helps clinicians identify high-risk HCC patients with early recurrence and improves their prognosis.

The PNI score is calculated as serum albumin plus 5 times the total peripheral lymphocyte count. A lower score indicates low serum albumin or low lymphocyte count. Relevant studies indicate that albumin not only reflects the body’s nutritional status but also indicates liver function reserve, inflammation, and other metabolic conditions. Lower levels of albumin are associated with elevated inflammatory factors, leading to poorer prognosis for patients and increased cancer-related mortality.[61] The PNI score was initially proposed by Onodera to assess the surgical risk and nutritional reserve of gastrointestinal surgery patients.[62] Subsequent studies have demonstrated that preoperative PNI scores can be used to assess the prognosis of HCC patients. Relevant research indicates that patients with low PNI scores have a poorer prognosis and are associated with tumor, lymph node, and metastasis staging.[63] Additionally, the PNI score can reflect patient nutritional status. Therefore, some scholars have combined the PNI score with another nutritional status indicator (BMI) to evaluate the outcomes of HCC patients. It has been confirmed that PNI and BMI are independent predictors of HCC patient outcomes,[64] and a low PNI score combined with BMI can accurately predict poorer outcomes, with a predictive range more sensitive than when used alone.

A nomogram is a visual and personalized tool for predicting prognosis. By incorporating a greater number of clinicopathological variables, nomograms can offer a more accurate prognosis than the TNM staging system.[19–21] Currently, several nomograms for HCC have been reported.[15,38,65,66] Li et al[67] developed a nomogram for predicting survival in lymph node-negative HCC patients using data from the SEER database and Zhongnan Hospital of Wuhan University. The model demonstrated good predictive ability with C-indexes greater than 0.70. However, this nomogram does not differentiate between surgery and loco-regional therapy, nor does it include inflammatory indicators. To our knowledge, the nomogram we developed is the first to predict OS in HCC patients treated primarily with surgery or loco-regional therapy that includes inflammatory indicators. The nomogram demonstrated strong predictive ability, with high C-index values (0.745 for the training cohort and 0.650 for the validation cohort). Additionally, ROC curves, calibration plots, and DCA curves all indicated satisfactory performance of the nomogram. It effectively identified patient subgroups with different risk levels, suggesting that high-risk patients may require intensive therapy and follow-up, while low-risk patients should avoid overtreatment.

Despite its widespread application, the nomogram, characterized as a unimodal modal, relies exclusively on a singular data modality – text data – for its training and predictive capabilities. Consequently, the volume of information accessible to this model is comparatively constrained, rendering it vulnerable to external perturbations. Conversely, large multimodal models possess the capacity to amalgamate data from diverse modalities, encompassing text, images, audio, and beyond, thereby furnishing a more exhaustive and enriched informational tapestry that elevates predictive precision. As technological advancements persist and data accumulate, large multimodal models tailored for HCC are poised to wield an increasingly significant influence. By relentlessly refining model architectures and algorithms, and integrating clinical data, imaging modalities, and genomic information, the accuracy and generalizability of these models can be augmented. Furthermore, fostering interdisciplinary collaborations will propel the extensive utilization of multimodal data in the anticipation, diagnosis, and management of HCC. Ultimately, this endeavor will facilitate earlier identification and intervention in HCC, ultimately enhancing patient survival rates and quality of life.

However, there are some limitations to our study. First, being retrospective, it cannot establish causation. Additionally, factors potentially affecting survival outcomes, such as genetic mutation status or specific treatment drugs, were not available in our data, which was limited to medical records. Prospective studies including more potential factors are needed to establish causation and validate the associations observed in our retrospective analysis. Moreover, while our model was internally developed and validated in a controlled environment for optimization, its generalizability is limited, and additional external validation data would strengthen our findings. Finally, since this study involved patients from a single institution, it is difficult to completely eliminate selection and information biases. Therefore, multicenter studies are essential for a more comprehensive investigation.

5. Conclusions

This study demonstrated that therapy, BMI, PVTT, leukocyte, GPR, MLR, and PNI were independent prognostic factors for OS in HCC patients primarily treated with surgery or loco-regional therapy. We developed a nomogram model incorporating inflammatory indicators to predict patients’ OS at 1-, 2- and 3-year. This nomogram provides precise survival predictions for HCC patients and helps identify individuals with varying prognostic risks, emphasizing the need for individualized follow-up and treatment plans.

Author contributions

Conceptualization: Guoping Sun.

Data curation: Xin Wang, Jing Xu.

Formal analysis: Xin Wang.

Funding acquisition: Guoping Sun.

Investigation: Xin Wang, Zhenya Jia, Guoping Sun.

Methodology: Xin Wang, Jing Xu.

Project administration: Xin Wang, Guoping Sun.

Resources: Xin Wang.

Software: Xin Wang, Jing Xu.

Supervision: Zhenya Jia, Guoping Sun.

Validation: Xin Wang, Jing Xu.

Visualization: Xin Wang.

Writing – original draft: Xin Wang, Jing Xu.

Writing – review & editing: Xin Wang, Zhenya Jia, Guoping Sun

Abbreviations:

AAR

aspartate aminotransferase to alanine aminotransferase ratio

AUC

area under the curve

BMI

body mass index

DCA

decision curve analysis

GGT

gamma-glutamyl transferase

GPR

γ-glutamyl transpeptidase to platelet ratio

HCC

hepatocellular carcinoma

MLR

monocyte to lymphocyte ratio

NLR

neutrophil to lymphocyte ratio

NRLR

neutrophil times γ-glutamyl transpeptidase to lymphocyte ratio

OS

overall survival

PLR

the platelet to lymphocyte ratio

PNI

prognostic nutritional index

PVTT

portal vein tumor thrombus

ROC

receiver operating characteristic

SII

systemic immune-inflammation index

SIRI

systemic inflammation response index

TNM

tumor-node-metastasis

References

[1]. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–49.

[2]. Kim BH, Lee D, Jung KW, Won YJ, Cho H. Cause of death and cause-specific mortality for primary liver cancer in South Korea: a nationwide population-based study in hepatitis B virus-endemic area. Clin Mol Hepatol. 2022;28:242–53.

[3]. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. 2020;70:7–30.

[4]. Zhang CY, Liu S, Yang M. Regulatory T cells and their associated factors in hepatocellular carcinoma development and therapy. World J Gastroenterol. 2022;28:3346–58.

[5]. Eilard MS, Naredi P, Helmersson M, et al. Survival and prognostic factors after transplantation, resection and ablation in a national cohort of early hepatocellular carcinoma. HPB (Oxford). 2021;23:394–403.

[6]. Rizzo A, Ricci AD. Challenges and future trends of hepatocellular carcinoma immunotherapy. Int J Mol Sci. 2022;23:11363.

[7]. Yang C, Zhang H, Zhang L, et al. Evolving therapeutic landscape of advanced hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol. 2023;20:203–22.

[8]. Guven DC, Erul E, Kaygusuz Y, et al. Immune checkpoint inhibitor-related hearing loss: a systematic review and analysis of individual patient data. Support Care Cancer. 2023;31:624.

[9]. Rizzo A, Santoni M, Mollica V, et al. Peripheral neuropathy and headache in cancer patients treated with immunotherapy and immuno-oncology combinations: the MOUSEION-02 study. Expert Opin Drug Metab Toxicol. 2021;17:1455–66.

[10]. Wong KM, King GG, Harris WP. The treatment landscape of advanced hepatocellular carcinoma. Curr Oncol Rep. 2022;24:917–27.

[11]. Bangaru S, Marrero JA, Singal AG. Review article: new therapeutic interventions for advanced hepatocellular carcinoma. Aliment Pharmacol Ther. 2020;51:78–89.

[12]. Raoul JL. Loco-regional treatments of unresectable hepatocellular carcinoma: safety first. Hepatobiliary Surg Nutr. 2022;11:324–6.

[13]. Kim N, Cheng J, Jung I, et al. Stereotactic body radiation therapy vs. radiofrequency ablation in Asian patients with hepatocellular carcinoma. J Hepatol. 2020;73:121–9.

[14]. Rizzo A, Ricci AD, Brandi G. Trans-arterial chemoembolization plus systemic treatments for hepatocellular carcinoma: an update. J Pers Med. 2022;12:1788.

[15]. Han R, Gan L, Lang M, et al. A retrospective study on predicting recurrence of intermediate-stage hepatocellular carcinoma after radical therapy. J Hepatocell Carcinoma. 2024;11:51–64.

[16]. Wang L, Li Q, Zhang J, Lu J. A novel prognostic scoring model based on albumin and γ-glutamyltransferase for hepatocellular carcinoma prognosis. Cancer Manag Res. 2019;11:10685–94.

[17]. Mei J, Lin WP, Shi F, et al. Prognostic nomogram predicting survival of patients with unresectable hepatocellular carcinoma after hepatic arterial infusion chemotherapy. Eur J Radiol. 2021;142:109890.

[18]. Lee SH, Jang HJ. Deep learning-based prediction of molecular cancer biomarkers from tissue slides: a new tool for precision oncology. Clin Mol Hepatol. 2022;28:754–72.

[19]. Yu C, Zhang Y. Establishment of prognostic nomogram for elderly colorectal cancer patients: a SEER database analysis. BMC Gastroenterol. 2020;20:347.

[20]. Fu R, Yang J, Wang H, et al. A nomogram for determining the disease-specific survival in invasive lobular carcinoma of the breast: a population study. Medicine (Baltimore). 2020;99:e22807.

[21]. Wu C, Wang N, Zhou H, Wang T, Zhao D. Development and validation of a nomogram to individually predict survival of young patients with nonmetastatic gastric cancer: a retrospective cohort study. Saudi J Gastroenterol. 2019;25:236–44.

[22]. Swierczak A, Mouchemore KA, Hamilton JA, Anderson RL. Neutrophils: important contributors to tumor progression and metastasis. Cancer Metastasis Rev. 2015;34:735–51.

[23]. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–74.

[24]. Diakos CI, Charles KA, McMillan DC, Clarke SJ. Cancer-related inflammation and treatment effectiveness. Lancet Oncol. 2014;15:e493–503.

[25]. Zeng J, Zeng J, Wu Q, et al. Novel inflammation-based prognostic nomograms for individualized prediction in hepatocellular carcinoma after radical resection. Ann Transl Med. 2020;8:1061.

[26]. Ma W, Zhang P, Qi J, et al. Prognostic value of platelet to lymphocyte ratio in hepatocellular carcinoma: a meta-analysis. Sci Rep. 2016;6:35378.

[27]. Lin ZX, Ruan DY, Li Y, et al. Lymphocyte-to-monocyte ratio predicts survival of patients with hepatocellular carcinoma after curative resection. World J Gastroenterol. 2015;21:10898–906.

[28]. Chan AW, Chan SL, Wong GL, et al. Prognostic Nutritional Index (PNI) predicts tumor recurrence of very early/early stage hepatocellular carcinoma after surgical resection. Ann Surg Oncol. 2015;22:4138–48.

[29]. Hu B, Yang XR, Xu Y, et al. Systemic immune-inflammation index predicts prognosis of patients after curative resection for hepatocellular carcinoma. Clin Cancer Res. 2014;20:6212–22.

[30]. Sahin TK, Rizzo A, Aksoy S, Guven DC. Prognostic significance of the Royal Marsden Hospital (RMH) score in patients with cancer: a systematic review and meta-analysis. Cancers (Basel). 2024;16:1835.

[31]. Zhang CH, Cheng Y, Zhang S, Fan J, Gao Q. Changing epidemiology of hepatocellular carcinoma in Asia. Liver Int. 2022;42:2029–41.

[32]. Zheng B, Zhu YJ, Wang HY, Chen L. Gender disparity in hepatocellular carcinoma (HCC): multiple underlying mechanisms. Sci China Life Sci. 2017;60:575–84.

[33]. Kanda T, Yokosuka O. The androgen receptor as an emerging target in hepatocellular carcinoma. J Hepatocell Carcinoma. 2015;2:91–9.

[34]. Hu X, Chen R, Wei Q, Xu X. The landscape of alpha fetoprotein in hepatocellular carcinoma: where are we? Int J Biol Sci. 2022;18:536–51.

[35]. Liu K, Huang G, Chang P, et al. Construction and validation of a nomogram for predicting cancer-specific survival in hepatocellular carcinoma patients. Sci Rep. 2020;10:21376.

[36]. Zhou H, Song T. Conversion therapy and maintenance therapy for primary hepatocellular carcinoma. Biosci Trends. 2021;15:155–60.

[37]. Zhu XD, Sun HC. Emerging agents and regimens for hepatocellular carcinoma. J Hematol Oncol. 2019;12:110.

[38]. Hsu WF, Hsu SC, Chen TH, et al. Modified albumin-bilirubin model for stratifying survival in patients with hepatocellular carcinoma receiving anticancer therapy. Cancers (Basel). 2022;14:5083.

[39]. Mohkam K, Dumont PN, Manichon AF, et al. No-touch multibipolar radiofrequency ablation vs. surgical resection for solitary hepatocellular carcinoma ranging from 2 to 5 cm. J Hepatol. 2018;68:1172–80.

[40]. deLemos AS, Zhao J, Patel M, et al. Lean body mass index is a marker of advanced tumor features in patients with hepatocellular carcinoma. World J Hepatol. 2024;16:393–404.

[41]. Xiao Y, Li K, Zhao Y, et al. Efficacy of radiotherapy in combined treatment of hepatocellular carcinoma patients with portal vein tumor thrombus: a real-world study. BMC Surg. 2024;24:54.

[42]. Ramon-Gil E, Geh D, Leslie J. Harnessing neutrophil plasticity for HCC immunotherapy. Essays Biochem. 2023;67:941–55.

[43]. Suner A, Carr BI, Akkiz H, et al. Inflammatory markers C-reactive protein and PLR in relation to HCC characteristics. J Transl Sci. 2019;5:3.

[44]. Uchinaka E, Amisaki M, Morimoto M, et al. Utility and limitation of preoperative neutrophil lymphocyte ratio as a prognostic factor in hepatocellular carcinoma. Yonago Acta Med. 2018;61:197–203.

[45]. Lou W, Liu J, Ding B, et al. Identification of potential miRNA-mRNA regulatory network contributing to pathogenesis of HBV-related HCC. J Transl Med. 2019;17:7.

[46]. Kabat GC, Kim MY, Manson JE, et al. White blood cell count and total and cause-specific mortality in the women’s health initiative. Am J Epidemiol. 2017;186:63–72.

[47]. Pan GQ, Yang CC, Shang XL, Dong ZR, Li T. The causal relationship between white blood cell counts and hepatocellular carcinoma: a Mendelian randomization study. Eur J Med Res. 2022;27:278.

[48]. Hong YM, Yoon KT, Hwang TH, Cho M. Pretreatment peripheral neutrophils, lymphocytes and monocytes predict long-term survival in hepatocellular carcinoma. BMC Cancer. 2020;20:937.

[49]. Zhang EL, Zhang ZY, Huang ZY. Splenectomy may facilitate systemic therapy for advanced hepatocellular carcinoma with hypersplenism. Hepatobiliary Surg Nutr. 2023;12:137–9.

[50]. Wang M, Xie X, Du Y, et al. Protective effects of biscoclaurine alkaloids on leukopenia induced by (60)Co-γ radiation. Evid Based Complement Alternat Med. 2020;2020:2162915.

[51]. Hanigan MH. Gamma-glutamyl transpeptidase: redox regulation and drug resistance. Adv Cancer Res. 2014;122:103–41.

[52]. Ma H, Zhang L, Tang B, et al. γ-Glutamyltranspeptidase is a prognostic marker of survival and recurrence in radiofrequency-ablation treatment of hepatocellular carcinoma. Ann Surg Oncol. 2014;21:3084–9.

[53]. Palacios-Acedo AL, Mège D, Crescence L, Dignat-George F, Dubois C, Panicot-Dubois L. Platelets, thrombo-inflammation, and cancer: collaborating with the enemy. Front Immunol. 2019;10:1805.

[54]. Lemoine M, Shimakawa Y, Nayagam S, et al. The gamma-glutamyl transpeptidase to platelet ratio (GPR) predicts significant liver fibrosis and cirrhosis in patients with chronic HBV infection in West Africa. Gut. 2016;65:1369–76.

[55]. Dai T, Deng M, Ye L, et al. Prognostic value of combined preoperative gamma-glutamyl transpeptidase to platelet ratio and fibrinogen in patients with HBV-related hepatocellular carcinoma after hepatectomy. Am J Transl Res. 2020;12:2984–97.

[56]. Huang L, Mo Z, Hu Z, et al. Diagnostic value of fibrinogen to prealbumin ratio and gamma-glutamyl transpeptidase to platelet ratio in the progression of AFP-negative hepatocellular carcinoma. Cancer Cell Int. 2020;20:77.

[57]. Lippitz BE, Harris RA. Cytokine patterns in cancer patients: a review of the correlation between interleukin 6 and prognosis. Oncoimmunology. 2016;5:e1093722.

[58]. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell. 2010;140:883–99.

[59]. Song W, Tian C, Wang K, Zhang RJ, Zou SB. The pretreatment lymphocyte to monocyte ratio predicts clinical outcome for patients with hepatocellular carcinoma: a meta-analysis. Sci Rep. 2017;7:46601.

[60]. Wu Y, Tu C, Shao C. Inflammatory indexes in preoperative blood routine to predict early recurrence of hepatocellular carcinoma after curative hepatectomy. BMC Surg. 2021;21:178.

[61]. Kosovali BD, Kucuk B, Balkiz Soyal O, Mehmet Mutlu N. Can prognostic nutritional index predict mortality in intensive care patients with COVID-19? Int J Clin Pract. 2021;75:e14800.

[62]. Onodera T, Goseki N, Kosaki G. Prognostic nutritional index in gastrointestinal surgery of malnourished cancer patients. Nihon Geka Gakkai Zasshi. 1984;85:1001–5.

[63]. Fan X, Chen G, Li Y, et al. The preoperative prognostic nutritional index in hepatocellular carcinoma after curative hepatectomy: a retrospective cohort study and meta-analysis. J Invest Surg. 2021;34:826–33.

[64]. Ji F, Liang Y, Fu S, et al. Prognostic value of combined preoperative prognostic nutritional index and body mass index in HCC after hepatectomy. HPB (Oxford). 2017;19:695–705.

[65]. Li H, Kang W, Rong P. Development and validation of a clinical factors and body fat distribution-based nomogram to predict refractoriness of transarterial chemoembolization in hepatocellular carcinoma. Quant Imaging Med Surg. 2024;14:447–61.

[66]. Ding DY, Liu L, Li HL, et al. Development of preoperative prognostic models including radiological features for survival of singular nodular HCC patients. Hepatobiliary Pancreat Dis Int. 2023;22:72–80.

[67]. Li Z, Hong Q, Li K. Nomogram predicting survival in patients with lymph node-negative hepatocellular carcinoma based on the SEER database and external validation. Eur J Gastroenterol Hepatol. 2024;36:904–15.

Keywords:

hepatocellular carcinoma; inflammatory indicators; nomogram; overall survival; primary therapies

Supplemental Digital Content

Copyright © 2024 the Author(s). Published by Wolters Kluwer Health, Inc.