The prognostic value of protein induced by vitamin K absence or antagonist-II in hepatocellular carcinoma patients undergoing surgical resection.

Introduction
Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide, with a high incidence and mortality rate [1]. In recent years, significant advancements have been achieved in the diagnosis and treatment of HCC. Surgical resection is currently one of the most effective curative therapies for HCC. However, because of the high recurrence rate of HCC, the prognosis of patients remains unsatisfactory [2–4]. Effectively predicting tumor recurrence and taking timely intervention measures for postoperative patients are crucial for improving patient prognosis and extending the overall survival (OS) of patients.
Alpha-fetoprotein (AFP) is the most widely used biomarker for the diagnosis and prognosis of HCC in clinical practice [5,6]. However, its sensitivity remains suboptimal, and a subset of HCC patients are AFP-negative, limiting its utility in both diagnostic and prognostic settings [7]. Therefore, there is an urgent need to search for new prognostic biomarkers to improve the predictive efficiency of survival prognosis in patients with HCC.
Prothrombin is a vitamin K-dependent plasma coagulation factor synthesized in the liver. Vitamin K deficiency impairs the activity of vitamin K-dependent carboxylase, resulting in defective posttranslational carboxylation of glutamic acid residues on prothrombin precursors. This leads to the accumulation of an abnormal prothrombin variant, decarboxylated prothrombin, also known as protein induced by vitamin K absence/antagonist-II (PIVKA-II) [8].
The association between serum PIVKA-II levels and HCC was first reported by Liebman et al. [9] in 1984. Since then, accumulating evidence has revealed that PIVKA-II may serve as a valuable diagnostic and prognostic biomarker for HCC. Several studies have indicated that PIVKA-II offers superior diagnostic performance compared with AFP [10–12]. However, research on the prognostic significance of PIVKA-II in HCC patients undergoing surgical resection remains limited, and no consistent conclusion has been reached. Some studies have shown that high expression of PIVKA-II is associated with poor prognosis, while others have found no significant correlation between PIVKA-II and recurrence or survival outcomes [13,14].
To address these discrepancies, this study investigates the relationship between PIVKA-II levels and the clinicopathological characteristics of HCC, as well as the prognostic value of preoperative serum PIVKA-II levels in predicting tumor recurrence and survival in patients undergoing surgical resection. By identifying reliable prognostic indicators, we aim to improve postoperative risk assessment and guide clinical management strategies for HCC patients.

Patients and methods

Patients and samples
This study analyzed a total of 467 serum samples obtained from 164 HCC patients undergoing surgical resection, 150 patients diagnosed with benign liver disease (BLD), and 153 healthy controls (HC). Samples of HCC were obtained 1 week before surgery at Tianjin Medical University Cancer Institute and Hospital (Tianjin, China) between December 2018 and January 2020.
Patients with hemorrhagic or thrombotic diseases (prothrombin time activity < 40% or prothrombin time-international normalized ratio  ≥ 1.5) and patients who have taken anticoagulant agents such as warfarin and similar drugs or vitamin K and similar preparations within 6 months before enrollment should be excluded. Patients with other malignant tumors, cardiac insufficiency, end-stage renal disease, or other severe systemic diseases were excluded, and patients who were pregnant or lactating at the time of recruitment were also excluded from the study.
All procedures performed in this study that involved human participants were approved by the Research Ethics Committee of Tianjin Medical University Cancer Institute and Hospital and conformed to the ethical standards of the 1964 Helsinki Declaration.
Clinicopathological feature data of HCC including gender (male/female), age (≥65/<55 years), tumor size (≥5/<5 cm), tumor number (single/multiple), hepatitis B surface antigen status (positive/negative), anti-hepatitis C virus status (positive/negative), and Child–Pugh classification (A–C) were collected (Table 1). According to the American Joint Committee on Cancer tumor node metastasis (TNM) classification, HCC patients were divided into those with early-stage (stage I + II) and advanced-stage (stage III + IV).
HCC patients undergoing surgical resection received regular follow-up every three months, with recurrence and survival status recorded accordingly. The final database was locked in January 2025, and the median follow-up duration for the entire cohort was 28 [95% confidence interval (CI): 8–60] months. After excluding cases with incomplete medical records, loss to follow-up, non-cancer-related mortality, and patients who failed to meet resection criteria after 6 months of conversion therapy, 164 patients were ultimately included in the final analysis. Recurrence was diagnosed following the ‘Chinese Expert Consensus on the Prevention and Management of Recurrent Hepatocellular Carcinoma after Hepatic Resection (2020 Edition)’.
To validate the primary findings of this study, an independent validation cohort was established, comprising 81 patients initially diagnosed with HCC who were undergoing surgical resection at Tianjin Medical University Cancer Institute and Hospital (Tianjin, China) between January and July 2020. All exclusion criteria and follow-up protocols were maintained identically to those implemented in the training cohort.

Detection of indicators
PIVKA-II serum levels were measured using an ELISA (Eitest, Tokyo, Japan) according to the operation instructions provided by the manufacturer. AFP, aspartate aminotransferase (AST), alanine aminotransferase (ALT), albumin, and total bilirubin serum levels were measured using a Roche Cobas E601 electrochemiluminescence analyzer (Roche, Mannheim, Germany) equipped with Roche dedicated reagents following the manufacturer’s instructions.

Statistical analysis
PIVKA-II, AFP, AST, ALT, albumin, and total bilirubin values were expressed as medians with interquartile range. Comparisons between groups were performed using the Wilcoxon rank-sum test. Clinical pathological feature data were expressed as percentage and analyzed using the χ² test. OS and recurrence-free survival (RFS) were assessed using the Kaplan–Meier method, and survival differences were evaluated using the log-rank test. The predictive performance of each indicator for survival and recurrence was assessed using the receiver operating characteristic (ROC) curve analysis. Comparisons of the predictive efficacy between different indicators were performed using DeLong’s test. To identify independent prognostic factors associated with survival and recurrence, various likely predictors associated with univariate analysis were subjected to multivariate analysis using Cox’s stepwise proportional hazard model. A P value less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 26.0 (SPSS, Chicago, Illinois, USA).

Results

Serum PIVKA-II levels are elevated in HCC patients
Serum PIVKA-II levels were measured in 164 HCC patients, 150 patients with BLD, and 153 HC. As shown in Fig. 1a, the median values for serum PIVKA-II in HCC patients, BLD patients, and HC were 181.50 (61.22–1314.00), 28.60 (21.80–36.39), and 21.82 (17.65–26.53) mAU/ml, respectively. Serum PIVKA-II levels in HCC patients were significantly higher than those of BLD patients and HC (all P < 0.01; Fig. 1a). These findings indicate that PIVKA-II may serve as a potential biomarker for HCC.
Further analysis revealed that serum PIVKA-II levels in HCC patients with advanced-stage (III + IV) (897.60, 130.10–5356.00) mAU/ml were significantly higher than those of HCC patients with early-stage (I + II) (125.80, 39.95–734.80) mAU/ml (P < 0.01; Fig. 1b). These results demonstrate a correlation between PIVKA-II levels and the disease progression of HCC.

Clinicopathological characteristics of HCC patients undergoing surgical resection with different serum PIVKA-II levels
Patients were classified into high and low PIVKA-II groups using a cut-off value of 100 mAU/ml. There were no significant differences between the two groups in terms of gender, age, tumor number, or gross tumor morphology. However, compared with the low PIVKA-II group, patients in the high PIVKA-II group had significantly larger tumor diameters, more advanced TNM staging, higher rates of lymph node infiltration and distant metastasis, poorer tumor differentiation, and a higher incidence of complications (all P < 0.05) (Table 2). These results indicate that the level of PIVKA-II is closely related to tumor aggressiveness and prognosis in HCC patients. Patients with high PIVKA-II levels exhibit more aggressive tumor characteristics, suggesting that PIVKA-II can serve as an important indicator for prognostic evaluation in HCC and provide a reference for clinical decision-making.

Overall survival and recurrence-free survival in HCC patients undergoing surgical resection
After applying exclusion criteria for incomplete data, loss to follow-up, failure to undergo surgical therapy, as well as deaths attributable to non-tumor-related causes, a final cohort of 164 resected patients was included in the analysis. The median overall survival (mOS) for the HCC patient cohort was 58.0 (95% CI: 13.0–60.0) months; the OS rates for all patients were 95.73, 75.61, and 49.39% at 1, 3, and 5 years, respectively (Fig. 2a).
During the follow-up period, the recurrence of HCC was observed in 139 of 164 patients (84.76%) who underwent surgical resection. The median recurrence-free survival (mRFS) of the HCC patient cohort was 27.0 (95% CI: 6.0–49.0) months; the 1-, 3-, and 5-year RFS rates were 81.71, 35.37, and 15.24%, respectively (Fig. 3a).

Comparison of overall survival and recurrence-free survival between HCC patients with and without PIVKA-II elevation who undergo surgical resection
In patients with PIVKA-II levels greater than or equal to 100 mAU/ml, the OS rates were 95.28, 72.64, and 41.51% at 1, 3, and 5 years, respectively. Whereas in patients with PIVKA-II levels less than 100 mAU/ml, the OS rates were 96.55, 81.03, and 63.79% at 1, 3, and 5 years, respectively. Patients in the high PIVKA-II group showed shorter mOS than those in the low PIVKA-II group (53.0 vs. 60.0 months, hazards ratio = 2.053, 95% CI: 1.291–3.262, P = 0.006; Fig. 2b).
Similarly, in patients with PIVKA-II levels greater than or equal to 100 mAU/ml, the RFS rates were 75.47, 27.26, and 12.26% at 1, 3, and 5 years, respectively. Whereas in patients with PIVKA-II levels less than 100 mAU/ml, the RFS rates were 91.67, 39.58, and 20.69% at 1, 3, and 5 years, respectively. Patients in the high PIVKA-II group showed significantly shorter mRFS than patients in the low PIVKA-II group (22.0 vs. 36.5 months, hazards ratio = 1.592, 95% CI: 1.139–2.226, P = 0.008; Fig. 3b).

The significant correlations between PIVKA-II levels and OS rate as well as RFS rate, indicate that PIVKA-II can serve as an important prognostic biomarker in HCC patients who are undergoing surgical resection. A higher PIVKA-II level is associated with an elevated recurrence rate and a shortened survival time. In clinical applications, serum PIVKA-II can be used as a prognostic biomarker for the risk-stratification management of patients and prognosis monitoring.
To compare the predictive efficacy of PIVKA-II and AFP on the 5-year OS rate of HCC patients who are undergoing surgical resection, we plotted the ROC curves and calculated the areas under the curve (AUCs). The analysis showed that the AUC of PIVKA-II was 0.817 (95% CI: 0.749–0.884), significantly higher than that of AFP, which was 0.707 (95% CI: 0.625–0.789). The direct comparison of the AUC values was statistically significant (DeLong test, P = 0.049; Fig. 4a). At the optimal cutoff value, the sensitivity and specificity of PIVKA-II reached 77.78% (95% CI: 67.58–85.46) and 81.93% (95% CI: 72.30–88.73), respectively, while the sensitivity and specificity of AFP were 60.49% (95% CI: 49.61–70.43) and 72.29% (95% CI: 61.84–80.77), respectively. These data indicate that PIVKA-II has superior discriminative ability and predictive accuracy compared with AFP in predicting long-term survival outcomes in HCC patients who are undergoing surgical resection.
Moreover, the predictive efficacy of combined PIVKA-II and AFP testing was significantly higher than that of a single marker, with an AUC of 0.905 (95% CI: 0.857–0.954). The sensitivity was further increased to 83.95% (95% CI: 74.45–90.37), and the specificity was also enhanced to 87.95% (95% CI: 79.22–93.32). The difference was statistically significant compared with using PIVKA-II alone (DeLong test, P = 0.001; Fig. 4a).
In the prediction of the 5-year RFS rate, the trends of all indicators were consistent with those of the 5-year OS rate prediction. The AUC of AFP was 0.789 (95% CI: 0.715–0.863), at the optimal cutoff value, the sensitivity was 61.15% (95% CI: 52.85–68.85), and the specificity was 92.00% (95% CI: 75.03–98.58). The predictive ability of PIVKA-II was still superior to that of AFP, with an AUC of 0.886 (95% CI: 0.835–0.937), at the optimal cutoff value, the sensitivity was 80.58% (95% CI: 73.21–86.29), and the specificity was 96.00% (95% CI: 80.46–99.79). The difference between PIVKA-II and AFP was statistically significant (DeLong test, P = 0.024; Fig. 4b).
The combined detection of PIVKA-II and AFP also showed the best predictive efficacy, with an AUC of 0.950 (95% CI: 0.916–0.984). The sensitivity was significantly increased to 89.93% (95% CI: 83.80–93.91), and the specificity remained at 96.00% (95% CI: 80.46–99.79). The difference was statistically significant compared with using PIVKA-II alone (DeLong test, P = 0.002; Fig. 4b).
Overall, in the prediction of both the 5-year OS rate and the 5-year RFS rate, the predictive efficacy of PIVKA-II was superior to that of AFP. The combined detection of PIVKA-II and AFP had significantly higher efficacy than a single marker, which could more accurately assess the survival and recurrence risks of HCC patients who are undergoing surgical resection and provide a more reliable reference for clinical prognosis.

Univariate and multivariate analysis of prognostic factors for overall survival in HCC patients undergoing surgical resection
Univariate analysis identified Child–Pugh classification (P = 0.013), PIVKA-II ≥ 100 mAU/ml (P = 0.006), albumin level (P = 0.049), tumor size (P = 0.026), and tumor number (P = 0.023) as significant prognostic factors for OS in patients undergoing surgical resection (Fig. 5).
Multivariate analysis using Cox’s stepwise proportional hazards model further revealed that Child–Pugh classification (P = 0.028) and PIVKA-II ≥ 100 mAU/ml (P = 0.036) were independent prognostic factors for OS in patients undergoing surgical resection (Fig. 6).
In conclusion, these findings indicate that Child–Pugh classification and PIVKA-II ≥ 100 mAU/ml were independent prognostic factors for OS in patients undergoing surgical resection. These findings may assist clinicians in better evaluating patient prognosis and developing individualized treatment strategies.

Univariate and multivariate analysis of the factors predicting recurrence-free survival in patients with HCC undergoing surgical resection
The results of univariate analysis showed that Child–Pugh classification (P = 0.004), albumin level (P = 0.032), tumor size (P = 0.027), number of tumors (P = 0.045), PIVKA-II ≥ 100 mAU/ml (P = 0.008), and AFP ≥ 20 ng/ml (P < 0.001) were important prognostic factors for RFS in HCC patients undergoing surgical resection (Fig. 7).
Multivariate analysis revealed that Child–Pugh classification (P = 0.047), PIVKA-II ≥ 100 mAU/ml (P = 0.001), and AFP ≥ 20 ng/ml (P < 0.001) were significant independent prognostic factors for RFS in HCC patients undergoing surgical resection (Fig. 8).
In summary, Child–Pugh classification, PIVKA-II ≥ 100 mAU/ml, and AFP ≥ 20 ng/ml are independent prognostic factors for RFS in HCC patients after surgical resection. These findings provide clinicians with important tools to assess patients’ recurrence risk and develop individualized treatment and follow-up strategies.

Validate the performance of the multivariate prognostic model using an independent validation cohort
In the validation cohort, multivariate analysis reaffirmed Child–Pugh classification (hazards ratio = 0.428, 95% CI: 0.220–0.830, P = 0.012) and PIVKA-II ≥ 100 mAU/ml (hazards ratio = 1.888, 95% CI: 1.058–3.368, P = 0.031) as independent prognostic factors of OS, and this analysis thereby provides external validation for the original findings from the training cohort (Fig. 9).
Multivariate analysis of the validation cohort also identified Child–Pugh classification, AFP, and PIVKA-II as independent predictors of postoperative recurrence in HCC patients. Specifically, Child–Pugh classification was associated with a lower risk of recurrence (hazards ratio = 0.622, 95% CI: 0.407–0.951, P = 0.028), whereas AFP ≥ 20 ng/ml (hazards ratio = 1.964, 95% CI: 1.229–3.138, P = 0.005) and PIVKA-II ≥ 100 mAU/ml (hazards ratio = 1.589, 95% CI: 1.029–2.454, P = 0.037) were significantly associated with increased recurrence risk. This analysis provides external validation for the original findings regarding recurrence prediction derived from the training cohort (Fig. 10).

Discussion
In summary, primary liver cancer is the fourth most common malignant tumor and the second leading cause of cancer-related mortality in China. It primarily comprises three pathological subtypes: HCC, intrahepatic cholangiocarcinoma, and combined hepatocellular–cholangiocarcinoma [1]. Among these, HCC is the most prevalent, accounting for ~75–85% of cases. It originates from the malignant transformation of hepatocytes and is closely associated with liver cirrhosis and viral hepatitis [15]. The present study focuses on patients with HCC.
Surgical resection remains the most effective curative treatment for early-stage HCC. In recent years, advancements in systemic antitumor therapy and multimodal treatment strategies have expanded the possibilities for radical resection in patients with advanced-stage HCC [2–4]. However, postoperative recurrence and metastasis remain significant challenges, underscoring the need for reliable prognostic biomarkers to guide clinical management and improve patient outcomes.
AFP is the most widely used serum biomarker for HCC diagnosis, treatment response assessment, and prognosis evaluation. However, its prognostic sensitivity and survival prediction accuracy are limited, particularly in AFP-negative HCC cases, thereby restricting its clinical utility [5–7]. This limitation highlights the urgent need to identify novel prognostic markers that can provide more precise risk stratification and facilitate timely therapeutic interventions. Liebman et al. [9] first reported the correlation between PIVKA-II levels and HCC in 1984. Since then, subsequent studies have provided increasing evidence supporting its role as a diagnostic and prognostic marker for HCC [10–12]. Our previous research demonstrated that PIVKA-II is a superior diagnostic biomarker for HCC compared with AFP, particularly in AFP-negative cases [16]. However, the prognostic value of PIVKA-II in HCC patients undergoing surgical treatment remains controversial. While some studies have reported that elevated PIVKA-II levels correlate with poor prognosis, others have found no significant association with clinical outcomes.
To address these inconsistencies, we investigated the relationship between PIVKA-II levels and OS, RFS in HCC patients undergoing surgical resection. In addition, we explored independent prognostic factors influencing recurrence and survival in this patient population. Our findings confirm that PIVKA-II serves as a significant predictive biomarker for recurrence and survival in HCC patients undergoing surgical treatment, providing valuable insights for clinical decision-making and patient management.
Our research results show that serum PIVKA-II levels in HCC patients show a significant upward trend with the progression of clinical staging, and serum PIVKA-II levels in advanced (III + IV) stage patients were significantly higher than that in early (I + II) stage patients (Fig. 1). This is consistent with the findings of several previous studies. Advanced HCC is often accompanied by severe impairment of liver function and vitamin K metabolic disorders, resulting in insufficient prothrombin carboxylation and the accumulation of PIVKA-II. Combining PIVKA-II with other HCC markers or imaging methods can improve the accuracy of staging, especially in distinguishing early-stage and advanced-stage HCC.
The serum levels of PIVKA-II in patients with HCC were related to tumor diameter, differentiation degrees, TNM staging, local lymph node infiltration, distant metastasis, and the occurrence of complications (Table 2). These results demonstrate that the level of PIVKA-II is closely correlated with tumor burden, malignancy degree, and clinical progression in HCC, serving as a critical tool for evaluating disease staging, predicting metastasis, and anticipating complications.
Using 100 mAU/ml as the optimal prognostic cutoff for PIVKA-II, patients were subgrouped into high PIVKA-II (≥100 mAU/ml) and low PIVKA-II (<100 mAU/ml) groups. Our results suggested that patients in the high PIVKA-II group had shorter mOS and mRFS compared with those in the low PIVKA-II group (both P < 0.05). Consistently, the OS rates and RFS rates at 1, 3, and 5 years after surgery in the high PIVKA-II group were lower than those in the low PIVKA-II group, respectively, as shown in Figs. 2 and 3. Implicating that PIVKA-II ≥ 100 mAU/ml is a poor prognostic factor for both OS and RFS in HCC patients undergoing surgical resection. PIVKA-II may serve as a risk-stratification tool for HCC patients after surgery: using a threshold of 100 mAU/ml, it can effectively distinguish between high-risk (poor prognosis) and low-risk (favorable prognosis) patients, and assist in formulating individualized follow-up plans (e.g. shortening the interval between imaging reviews for the high-risk group).
We established the prognostic value of PIVKA-II in postoperative survival and recurrence among HCC patients using Kaplan–Meier analysis. To further quantitatively evaluate its predictive capacity, we performed a comparative analysis of PIVKA-II and AFP using ROC curves. The results demonstrated that PIVKA-II had significantly higher AUC than AFP in predicting 5-year OS and RFS, highlighting its superior discriminative ability in long-term prognosis (Fig. 4). Furthermore, a combined model incorporating both PIVKA-II and AFP showed significantly enhanced predictive performance over either biomarker alone (Fig. 4). Therefore, combined detection of these two markers can provide a more comprehensive assessment of tumor behavior, leading to more accurate assessment of postoperative prognosis in HCC patients.
The sequential analytical approach, from univariate screening to multivariate adjustment, provides a robust framework for identifying true independent prognostic factors. While univariate analysis identified several clinically relevant variables, including Child–Pugh classification, PIVKA-II, albumin, tumor size, and tumor number as significant predictors of OS, only Child–Pugh classification and PIVKA-II ≥ 100 mAU/ml retained independent significance in the multivariate analysis (Fig. 6). The persistence of Child–Pugh classification through both analytical stages reaffirms the well established impact of liver function on long-term survival after surgery. The emergence of PIVKA-II as an independent prognostic factor highlights its value in predicting patients’ OS after surgery.
In the subsequent RFS studies, while multiple clinical and tumor-related parameters demonstrated significance in univariate screening, only three factors retained independent prognostic value in the multivariate model: Child–Pugh classification, PIVKA-II ≥ 100 mAU/ml and AFP ≥ 20 ng/ml (Fig. 8). These findings suggest that integrating specific serum biomarker with measures of hepatic functional reserve may create a more comprehensive prognostic framework, thereby enabling more personalized postoperative management strategies.
To validate the robustness and generalizability of the aforementioned multivariate analysis results, we replicated the analysis in an independent validation cohort (n = 81). Multivariate Cox regression analysis in the validation set confirmed that PIVKA-II ≥ 100 mAU/ml remained an independent negative prognostic factor for both OS (Fig. 9) and RFS (Fig. 10). The successful replication of this finding enhances the reliability of our initial conclusions.
Notably, compared with invasive diagnostic methods, the serological testing of PIVKA-II has significant advantages in clinical practice, including noninvasiveness and convenience, making it a practical tool for risk stratification and therapeutic decision-making. Existing evidence indicates that PIVKA-II can bind and activate the hepatocyte growth factor receptor MET, triggering the downstream rat sarcoma virus/mitogen-activated protein k inase and phosphoinositide 3-kinase/protein kinase B/mammalian target of rapamycin signaling pathways and promoting tumor angiogenesis [17–20]. PIVKA-II can promote epithelial–mesenchymal transition by upregulating transcription factors such as Snail and Twist, thereby increasing the migration and invasion abilities of tumor cells [21,22]. In addition, PIVKA-II can promote the secretion of CXCL1/CXCL5, recruit myeloid-derived suppressor cells to the tumor microenvironment, inhibit the antitumor immune response, and increase the risk of postoperative recurrence [23–25]. The clinical findings of this study provide clinical support for the molecular mechanisms reported in the literature. Future research should prioritize the direct validation of these mechanisms in clinical samples. Elevating PIVKA-II from a diagnostic and prognostic biomarker to a mechanism-based stratification tool holds the potential to ultimately improve postoperative outcomes in HCC.
In conclusion, PIVKA-II is a promising, convenient, and effective biomarker that can be used to evaluate the survival and recurrence of HCC patients undergoing surgical resection, identify those who may benefit from surgical resection treatment, and stratify the risk of patients after surgery; nevertheless, our findings must be viewed as preliminary. Despite the advantages of our study, there are still some limitations, and the retrospective, single-center design is susceptible to selection bias. Despite our efforts to control for established prognostic factors through multivariate analysis, the impact of unmeasured confounders cannot be discounted. To translate these findings into clinical practice, future prospective, multicenter, large-scale studies are essential to validate the prognostic utility of PIVKA-II in a wider Chinese demographic. Our subsequent research should validate the predictive performance of PIVKA-II in different treatment modalities, including systemic therapy and local intervention, to comprehensively delineate its prognostic role in HCC. In addition, studying the mechanism of PIVKA-II in HCC progression is crucial for determining its clinical utility in prognostic stratification. The regional predominance of HBV-related HCC in our cohort limits the generalizability of our findings to populations with different etiology patterns; further validation in multiethnic cohorts is required.

Acknowledgements
This work was supported by the Pharmaceutical Research Capacity Building Program of Bethune Charitable Foundation (Grant No. Z04J2023E095) and the Tianjin Key Medical Discipline Construction Project (Grant No. TJYXZDXK-3-003A).

Conflicts of interest
There are no conflicts of interest.