Next Article in Journal
MicroRNA Profiling of Red Blood Cells for Lung Cancer Diagnosis
Previous Article in Journal
Genome-Wide Extrachromosomal Circular DNA Profiling of Paired Hepatocellular Carcinoma and Adjacent Liver Tissues
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Cancers
Volume 15
Issue 22
10.3390/cancers15225310
cancers-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Construction of a Nomogram to Predict Overall Survival in Patients with Early-Onset Hepatocellular Carcinoma: A Retrospective Cohort Study

by
Tianrui Kuang
1,2,†,
Wangbin Ma
1,2,†,
Jiacheng Zhang
1,2,
Jia Yu
1,2,
Wenhong Deng
1,2,
Keshuai Dong
1,2,* and
Weixing Wang
1,2,*
1
Department of General Surgery, Renmin Hospital of Wuhan University, Wuhan 430060, China
2
Central Laboratory, Renmin Hospital of Wuhan University, Wuhan 430060, China
*
Authors to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Cancers 2023, 15(22), 5310; https://doi.org/10.3390/cancers15225310
Submission received: 18 October 2023 / Revised: 31 October 2023 / Accepted: 3 November 2023 / Published: 7 November 2023
(This article belongs to the Topic Public Health and Healthcare in the Context of Big Data)
Browse Figures
Versions Notes

Abstract

:

Simple Summary

Hepatocellular carcinoma (HCC) is a severe global health concern, and it is increasingly jeopardizing younger individuals. Despite this, there is a lack of available tools for the prognosis estimation of early-onset HCC. In our study, we conducted a retrospective analysis of early-onset HCC (EO-LIHC) using data of the period from 2004 to 2018. We identified independent risk factors using a Cox regression analysis, including age, sex, AFP level, the grading and staging of the tumor, the size of the tumor, and whether the patient was receiving therapy like surgery and chemotherapy. We developed a predictive nomogram to estimate 1-, 3-, and 5-year survival rates of EO-LIHC patients and a user-friendly web-based survival prediction model tailored for these patients. These findings provide valuable insights for personalized care and treatment decisions for individuals with EO-LIHC.

Abstract

Hepatocellular carcinoma (HCC) is a widespread and impactful cancer which has pertinent implications worldwide. Although most cases of HCC are typically diagnosed in individuals aged ≥60 years, there has been a notable rise in the occurrence of HCC among younger patients. However, there is a scarcity of precise prognostic models available for predicting outcomes in these younger patients. A retrospective analysis was conducted to investigate early-onset hepatocellular carcinoma (EO-LIHC) using data from the Surveillance, Epidemiology, and End Results (SEER) database from 2004 to 2018. The analysis included 1392 patients from the SEER database and our hospital. Among them, 1287 patients from the SEER database were assigned to the training cohort (n = 899) and validation cohort 1 (n = 388), while 105 patients from our hospital were assigned to validation cohort 2. A Cox regression analysis showed that age, sex, AFP, grade, stage, tumor size, surgery, and chemotherapy were independent risk factors. The nomogram developed in this study demonstrated its discriminatory ability to predict the 1-, 3-, and 5-year overall survival (OS) rates in EO-LIHC patients based on individual characteristics. Additionally, a web-based OS prediction model specifically tailored for EO-LIHC patients was created and validated. Overall, these advancements contribute to improved decision-making and personalized care for individuals with EO-LIHC.

1. Introduction

Cancer is a complex disease influenced by multiple factors, primarily impacting individuals aged ≥50 years. Early-onset malignancies, which are those detected in persons aged <50 years, have become more common in several nations in recent decades. This upward trend is particularly evident in colorectal, gallbladder, liver, pancreatic, and other gastrointestinal tumors [1,2,3,4]. Although the overall incidence and mortality rates of all types of cancer are significantly lower in younger adults compared to older adults, cancers in younger men and women have substantial economic and social implications. Moreover, they can lead to a higher loss of person-years of life compared to cancers diagnosed later in life [5,6,7].
Hepatocellular carcinoma (HCC) is recognized as one of the most common cancers worldwide, presenting a significant healthcare challenge on a global scale [8]. Although most liver cancer cases are diagnosed in individuals aged ≥60 years, a consistent increase in the incidence of liver cancer has been reported among young patients recently. Several risk factors for HCC are now clearly defined, such as chronic hepatitis B and C virus infections, chronic hepatitis, cirrhosis, and alcohol consumption [9]. However, limited information is available regarding risk factors specifically associated with early-onset hepatocellular carcinoma (EO-LIHC), which warrants further investigation and understanding. It has been shown that patients with EO-LIHC not only have fewer complications but also have a lower incidence of underlying cirrhosis compared to patients with advanced HCC, suggesting that there may be an underlying etiologic difference between the two groups of patients [10]. Distinctions in epidemiology, clinical manifestations, pathological characteristics, and molecular features between early-stage and late-stage cancers have been observed. However, it is unlikely to precisely demarcate these disparities at the age of 50 [11].
Current clinical practice recommendations heavily rely on the AJCC TNM staging system to predict prognosis and guide therapeutic decisions for patients with EO-LIHC [12]. However, the TNM staging system has some limitations due to various factors that can influence patient prognosis, including age, gender, the degree of tumor differentiation, serum biomarkers, and treatment-related variables. Nomograms have shown precise predictive capabilities for various tumor types and are widely used in clinical settings, surpassing the traditional TNM staging system or alternative staging systems [13,14,15]. However, currently, there is a lack of a nomogram model specifically developed to predict postoperative survival in patients with EO-LIHC.
In this context, we constructed and validated a novel nomogram for the first time, utilizing data from the SEER database and a Chinese cohort. With the use of this nomogram, patients with EO-LIHC can have their overall survival (OS) predicted. Additionally, recognizing the significance of improving medical care for patients, we developed a web-based model. This web-based tool enables clinicians to make more informed clinical decisions by accurately assessing the prognosis of patients with EO-LIHC.

2. Methods

2.1. Study Design and Cohort Selection

The precise definition of EO-LIHC remains controversial, particularly regarding the age cutoff. For this study, EO-LIHC was defined as HCC diagnosis in patients aged ≤50 years. The study cohorts were established by screening patients diagnosed with HCC between 2004 and 2018 using the SEER database. From the SEER database, 1287 EO-LIHC patients were randomly assigned to the training cohort (n = 899) and validation cohort 1 (n = 388) in a 7:3 ratio. Furthermore, 105 EO-LIHC patients from our hospital comprised validation cohort 2.
Specific inclusion and exclusion criteria were applied to all the cohorts. The inclusion criteria were as follows: (a) age < 50 years; (b) confirmed diagnosis of HCC; and (c) known cause of death. The exclusion criteria were as follows: (a) lack of clinicopathological information; (b) lack of follow-up information; and (c) missing treatment options. Figure 1 shows the detailed information regarding the inclusion and exclusion of the patients. This retrospective study has been reported in line with the Strengthening the Reporting of Cohort Studies in Surgery guidelines [16].

2.2. Variables Management

We retrieved 11 clinically relevant variables for EO-LIHC from the SEER database, which included data regarding age, sex, race, pathology, tumor size, AJCC TNM stage, alpha-fetoprotein (AFP) levels, chemotherapy, and survival. The primary outcome of the research was OS, which was calculated as the period until death. The tumors were staged according to the AJCC TNM staging criteria, the 8th edition.

2.3. Establishment and Validation of the Nomogram Model

A prognostic model was constructed using a nomogram with the training cohort. The stability of the model was assessed using validation cohorts 1 and 2. To determine the factors that substantially affect OS in patients with EO-LIHC, univariate and multivariate Cox regression analyses were conducted on all variables included in the research.
The models were verified based on the training and validation cohorts using the C-index, receiver-operating characteristic (ROC) curve, calibration curves, and decision curve analysis (DCA). The C-index was used to examine the nomogram’s performance and prediction accuracy, while the ROC curve was used to evaluate its sensitivity and specificity. Calibration curves were generated for 1-, 3-, and 5-year periods to assess the agreement between the model predictions and actual data.

2.4. Comparison of Nomogram-Based Risk Classification and the AJCC TNM Staging System

To evaluate the comparative net benefit and risk stratification of the nomogram model against the AJCC TNM staging system, several statistical measures were employed, including the net reclassification index (NRI), C-index, integrated discrimination improvement (IDI), and decision curve analysis (DCA). The DCA was specifically utilized to gauge the clinical utility of the nomogram. To facilitate risk stratification, all eligible patients were categorized into low-risk, middle-risk, and high-risk groups. The optimal threshold for determining the overall score was determined using an X-Tile analysis. The OS of patients in the various risk categories was then compared using Kaplan–Meier curves and log-rank testing.

2.5. Statistical Methods

Descriptive statistics for all study variables are presented as numbers and percentages. To explore the associations between variables and survival outcomes, both univariate and multivariable Cox regression analyses were performed using R, version 3.6.3, and relevant packages for the correlation analysis. Various evaluation metrics, such as the C-index and ROC curve, were used to assess the performance of the model. OS was analyzed using Kaplan–Meier and log-rank tests. The chi-square test was used to determine whether the distributions of the cohorts used for training and validation differed. Two-tailed p-values < 0.05 were considered statistically significant.

3. Results

3.1. Patient Characteristics

Patients confirmed with HCC between 2004 and 2018 were identified using the SEER database. In total, 1287 patients with EO-LIHC from the database were randomly divided into a training cohort (n = 899) and validation cohort 1 (n = 388) in a 7:3 ratio. Furthermore, validation cohort 2 comprised 105 EO-LIHC patients from our hospital. In the study sample (n = 1392), 985 patients were male, 452 had tumors measuring <5 cm, 690 tested positive for AFP, and 485 received chemotherapy. No significant differences were observed between the training cohort and validation cohort 1 in the patient characteristics (Table 1). Supplementary Table S1 shows the details of the patients in validation cohort 2.

3.2. Univariate and Multivariate Analyses

In our endeavor to comprehensively assess the influence of each prognostic factor on the ultimate outcome, we meticulously conducted both univariate and multivariate Cox regression analyses within the training cohort. The results of the univariate analysis notably unveiled statistically significant differences in the training cohort across several key variables, including age, gender, tumor stage, alpha-fetoprotein (AFP) level, tumor grade, tumor size, surgical procedures, the extent of lymph node removal, and even marital status, as thoughtfully summarized in Table 2.
Subsequently, employing a rigorous multiple regression analysis approach, we further distilled our findings to identify the truly independent prognostic factors for overall survival (OS). This analysis underscored the pivotal role of age, gender, AFP levels, tumor grade, tumor stage, tumor size, surgical interventions, and the judicious use of chemotherapy as determinants of OS. These critical factors were seamlessly incorporated into our comprehensive nomogram, as visually depicted in Figure 2. Our work serves as a testament to the vital role these factors play in the precise prediction of OS within our studied cohort.

3.3. Construction and Verification of the OS Prognostic Nomogram

Figure 2 presents the prediction results of the nomogram constructed to estimate the 1-, 3-, and 5-year OS rates of the EO-LIHC patients. The nomogram integrated all independent prognostic markers determined by the multivariate Cox regression model. The nomogram offers a personalized approach for prognostication by considering various clinicopathological characteristics of individual patients.
The ROC, DCA, and calibration curves were generated to complement this analysis (Figure 3, Figure 4 and Figure 5). The internal validation was performed using the C-index to evaluate the model’s correctness and showed the following values: training cohort, 0.784 (95% confidence interval [CI]: 0.777–0.797); validation cohort 1, 0.781 (95% CI: 0.763–0.804); and validation cohort 2, 0.766 (95% CI: 0.731–0.782) (Figure 6). The area under the ROC curve values were 0.878, 0.886, and 0.871 for the training cohort at the 1-, 3-, and 5-year time points, respectively. This indicates the robust predictive performance of the nomogram across these intervals. Additionally, the calibration curves exhibited strong concordance between the predicted and observed probabilities of the 1-, 3-, and 5-year OS rates. Furthermore, the DCA curves revealed favorable clinical net benefits at 1-, 3-, and 5-year intervals for the training cohort as well as the validation cohorts 1 and 2, underscoring the utility of the nomogram in guiding clinical decision-making.

3.4. Comparing the Clinical Applicability of the New Nomogram with the AJCC TNM Staging System

We used a thorough study based on the NRI, IDI, and C-index to compare the relative benefits and drawbacks of the nomogram with the conventional AJCC TNM staging system. Our investigation yielded noteworthy findings, illustrating that the nomogram outperformed the AJCC staging system, as evidenced by the higher C-index values in the former than in the latter (Figure 6). The IDI values for 1-, 3-, and 5-year OS in the training cohort were 0.197, 0.204, and 0.189, respectively, while those in validation cohort 1 were 0.228, 0.306, and 0.323, respectively (p < 0.05). These findings conclusively prove that the established nomogram is better than the AJCC TNM staging system. Furthermore, the NRI values in the training cohort at 1 year, 3 years, and 5 years were 0.794 (95% CI: 0.636–0.929), 0.699 (95% CI: 0.211–0.814), and 0.204 (95% CI: 0.099–0.328), respectively (Table 3). Additionally, we compared the net benefit of the nomogram and the AJCC TNM staging system using the DCA to assess their practical benefits. Notably, the DCA curves consistently demonstrated that the nomogram provided superior predictions for the 1-, 3-, and 5-year OS rates across all cohorts, delivering greater net benefits when compared to the AJCC TNM staging system.

3.5. Establishment of a Risk Stratification System According to the Nomogram

Using the nomogram and the findings from the X-tile software (Version 3.6.1), we classified the patients into the following three categories based on their total score: low risk (total score < 461), middle risk (461 ≤ total score < 519), and high risk (total score ≥ 519) (Figure 7). The prognostic capability of this novel classification system was significantly better than that of the traditional AJCC TNM staging system, as demonstrated by the Kaplan–Meier curves (Figure 8).

3.6. Constructing a Web-Based Survival Calculator

We have developed a publicly accessible web-based survival calculator (URL: https://doctoryyds.shinyapps.io/DynNomapp/ accessed on 2 November 2023) that utilizes the nomogram model to predict OS in EO-LIHC patients based on their personalized characteristics (Figure 9). The survival calculator collects variables, such as age, sex, AFP level, tumor grade, tumor stage, tumor size, surgery, and chemotherapy. For instance, a male patient aged 20–35 years with EO-LIHC, who had a Grade I or II tumor and underwent chemotherapy and surgery, would have a predicted 5-year OS rate of 60% (95% CI: 50–72%). The website is user-friendly and convenient and aims to assist in the individualized prediction of outcomes for EO-LIHC patients.

4. Discussion

Hepatocellular carcinoma has traditionally been more common among the elderly than among young patients [17]. However, recently, a noticeable increase in the incidence of EO-LIHC has been observed. The rise in early-onset cancers may be partly attributed to increased screening efforts and improved early-detection methods before the age of 50. The accurate prediction of a survival prognosis is crucial for healthcare professionals to facilitate individualized treatment and follow-up decisions. Although the AJCC TNM staging method is currently the most popular prognostic evaluation technique, focusing exclusively on the anatomical infiltration and metastasis of the tumor may restrict the accuracy of the survival prognosis [18,19,20]. Recently, several clinical prediction models that offer superior predictive power compared to the AJCC TNM staging system have emerged, expanding the options for the accurate prediction of tumor prognosis.
There are relatively few clinical prediction models for EO-LIHC patients. This study included 1392 EO-LIHC patients. Univariate and multivariate Cox regression analyses conducted to identify independent prognostic factors revealed nine such factors, including age, sex, tumor grade, surgery, chemotherapy, tumor size, and AFP level. Prognostic line graph models were constructed based on these factors. The performance of these models was evaluated using various assessment methods, including the C-index for a discriminant analysis, calibration curves for a calibration assessment, and the DCA for evaluating clinical usefulness. Additionally, a web-based OS prediction model specifically tailored for EO-LIHC patients was created and validated. Overall, these advancements contribute to improved decision-making and personalized care for individuals with EO-LIHC.
The AFP was initially identified as a tumor-associated antigen and a target for immunotherapy in HCC [21]. The AFP has been widely used for HCC surveillance and evaluating treatment response in HCC patients [22,23]. The results of this study corroborate the conclusions of previous studies that EO-LIHC patients may have increased AFP levels as a distinct risk factor. These results are also consistent with clinical observations and support the clinical significance of AFP as a prognostic marker in EO-LIHC.
Tumor size is a significant factor that influences the prognosis of hepatocellular carcinoma (HCC), and it is commonly used as a crucial criterion for HCC staging and treatment guidelines [24]. Several multicenter studies focusing on surgically-resected HCC have consistently demonstrated that tumors measuring >5 cm are associated with a poorer prognosis than those measuring ≤5 cm [25,26,27]. The nomogram clearly illustrates the relationship between the tumor size and prognosis, indicating that as the tumor size increases, the corresponding score also increases, indicating a worse prognosis. This observation underscores the importance of considering tumor size when assessing the prognosis of HCC patients.
The findings of this study also show that both chemotherapy and surgery function as distinct protective factors. Surgical resection remains the treatment of choice for HCC [28]. However, due to late-stage diagnosis and the lack of symptoms in early-stage HCC, many patients are not eligible for surgery [29]. Current guidelines may not endorse surgery for all cases of HCC [30,31]. However, a retrospective analysis by Mao et al. showed that patients who received viable surgery had better results than those who did not, even in the presence of distant metastases [32]. Additionally, a combination of surgery and chemotherapy has been shown to be beneficial for patients with HCC [33,34]. Systemic chemotherapy, utilizing agents such as gemcitabine, doxorubicin, or combination regimens, has shown improved survival rates in patients with HCC [33]. Consistent with these findings, the nomogram scores indicate that patients who received chemotherapy had a more favorable prognosis compared to those who did not. Beyond these mentioned factors, research has elucidated that cancer patients grappling with obesity exhibit a less favorable prognosis, potentially linked to the burdensome load of excessive body composition [35]. Conversely, individuals contending with underlying conditions like diabetes tend to experience a more promising prognosis [36]. These additional dimensions contribute to the intricate landscape of prognostication within the realm of cancer care, necessitating a holistic approach to comprehending and improving patient outcomes.
The nomogram is a simple-to-use statistical tool that considers various risk variables to offer patients personalized evaluations [37,38]. In this study, a new risk model was developed based on the risk points derived from the column line plot, allowing for the identification of high and low-risk patients with EO-LIHC [39]. The clinical relevance of this risk model was confirmed through an independent analysis, validating its predictive effectiveness across diverse patient populations. As the culmination of this research, we have developed the first online survival calculator for patients with EO-LIHC. This tool considers personalized clinicopathological features and can help predict postoperative outcomes on an individual basis.
This study has several limitations. First, it was a retrospective analysis, which introduces inherent limitations in data collection and potential bias. Second, data regarding important prognostic factors, such as etiology, hepatitis B surface antigen status, aspartate transferase levels, and vascular infiltration, were not available in the SEER database, which may have affected the comprehensive evaluation of prognostic factors. Third, the SEER database predominantly represents the US population; therefore, the findings may not be generalizable to other regions or populations. Fourth, although this study included a case-cohort from the Chinese population, the sample size was relatively small, which may limit the generalizability and overall robustness of the findings.

5. Conclusions

This study aimed to enhance the prediction of OS in EO-LIHC patients by combining data from the SEER database and a cohort from our hospital. Thus, a precise and specific nomogram was developed to predict OS in EO-LIHC patients. Moreover, a novel risk model and web-based survival calculator were developed for clinical application. These efforts intend to enhance the prognostic evaluation and individualized treatment decision-making in EO-LIHC patients.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cancers15225310/s1, Table S1. Characteristics of patients in external validation cohort.

Author Contributions

T.K. and J.Z. designed the research. W.M., J.Z. and T.K. performed the research and analyzed results. J.Y. and W.D. edited the manuscript. W.W. and K.D. provided critical comments and revised the manuscript. T.K. and W.D. collected and organized data. W.W. wrote the revised manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by grants from the National Natural Science Foundation of China (No. 82003063) and the Natural Science Foundation of Hubei Province, China (No. 2020CFB213).

Institutional Review Board Statement

Ethical approval for this study was obtained from the Ethics Committee of the Renmin Hospital of Wuhan University on 7 December 2022 (registration number: WDRY2022-K258).

Informed Consent Statement

This study did not involve personal privacy and commercial interests and collected patient diagnosis and treatment information retrospectively from the medical system of our hospital, which was almost at no risk to patients. After reviewed by the Ethical Committee of Renmin Hospital of Wuhan University, our study was approved as exempt from needing signed informed consent.

Data Availability Statement

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest

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.

References

  1. Ugai, T.; Sasamoto, N.; Lee, H.-Y.; Ando, M.; Song, M.; Tamimi, R.M.; Kawachi, I.; Campbell, P.T.; Giovannucci, E.L.; Weiderpass, E.; et al. Is early-onset cancer an emerging global epidemic? Current evidence and future implications. Nat. Rev. Clin. Oncol. 2022, 19, 656–673. [Google Scholar] [CrossRef]
  2. Islami, F.; Ward, E.M.; Sung, H.; Cronin, K.A.; Tangka, F.K.L.; Sherman, R.L.; Zhao, J.; Anderson, R.N.; Henley, S.J.; Yabroff, K.R.; et al. Annual Report to the Nation on the Status of Cancer, Part 1: National Cancer Statistics. J. Natl. Cancer Inst. 2021, 113, 1648–1669. [Google Scholar] [CrossRef]
  3. Gupta, S.; Harper, A.; Ruan, Y.; Barr, R.; Frazier, A.L.; Ferlay, J.; Steliarova-Foucher, E.; Fidler-Benaoudia, M.M. International Trends in the Incidence of Cancer Among Adolescents and Young Adults. J. Natl. Cancer Inst. 2020, 112, 1105–1117. [Google Scholar] [CrossRef]
  4. Sung, H.; Siegel, R.L.; Rosenberg, P.S.; Jemal, A. Emerging cancer trends among young adults in the USA: Analysis of a population-based cancer registry. Lancet Public. Health 2019, 4, e137–e147. [Google Scholar] [CrossRef]
  5. Fidler, M.M.; Gupta, S.; Soerjomataram, I.; Ferlay, J.; Steliarova-Foucher, E.; Bray, F. Cancer incidence and mortality among young adults aged 20–39 years worldwide in 2012: A population-based study. Lancet Oncol. 2017, 18, 1579–1589. [Google Scholar] [CrossRef] [PubMed]
  6. Lortet-Tieulent, J.; Soerjomataram, I.; Lin, C.C.; Coebergh, J.W.W.; Jemal, A.U.S. Burden of Cancer by Race and Ethnicity According to Disability-Adjusted Life Years. Am. J. Prev. Med. 2016, 51, 673–681. [Google Scholar] [CrossRef] [PubMed]
  7. Bradley, C.J.; Yabroff, K.R.; Dahman, B.; Feuer, E.J.; Mariotto, A.; Brown, M.L. Productivity costs of cancer mortality in the United States: 2000–2020. J. Natl. Cancer Inst. 2008, 100, 1763–1770. [Google Scholar] [CrossRef]
  8. Forner, A.; Reig, M.; Bruix, J. Hepatocellular carcinoma. Lancet 2018, 391, 1301–1314. [Google Scholar] [CrossRef] [PubMed]
  9. Kulik, L.; El-Serag, H.B. Epidemiology and Management of Hepatocellular Carcinoma. Gastroenterology 2019, 156, 477–491.e471. [Google Scholar] [CrossRef]
  10. Lam, C.M.; Chan, A.O.; Ho, P.; Ng, I.O.; Lo, C.M.; Liu, C.L.; Poon, R.T.; Fan, S.T. Different presentation of hepatitis B-related hepatocellular carcinoma in a cohort of 1863 young and old patients—Implications for screening. Aliment. Pharmacol. Ther. 2004, 19, 771–777. [Google Scholar] [CrossRef]
  11. Akimoto, N.; Ugai, T.; Zhong, R.; Hamada, T.; Fujiyoshi, K.; Giannakis, M.; Wu, K.; Cao, Y.; Ng, K.; Ogino, S. Rising incidence of early-onset colorectal cancer—A call to action. Nat. Rev. Clin. Oncol. 2021, 18, 230–243. [Google Scholar] [CrossRef]
  12. Tang, Y.Y.; Zhao, Y.N.; Zhang, T.; Chen, Z.Y.; Ma, X.L. Comprehensive radiomics nomogram for predicting survival of patients with combined hepatocellular carcinoma and cholangiocarcinoma. World J. Gastroenterol. 2021, 27, 7173–7189. [Google Scholar] [CrossRef]
  13. Balachandran, V.P.; Gonen, M.; Smith, J.J.; DeMatteo, R.P. Nomograms in oncology: More than meets the eye. Lancet Oncol. 2015, 16, e173–e180. [Google Scholar] [CrossRef]
  14. Iasonos, A.; Schrag, D.; Raj, G.V.; Panageas, K.S. How to build and interpret a nomogram for cancer prognosis. J. Clin. Oncol. 2008, 26, 1364–1370. [Google Scholar] [CrossRef] [PubMed]
  15. Liu, J.; Liu, Z.; Li, J.; Tian, S.; Dong, W. Personalizing prognostic prediction in early-onset Colorectal Cancer. J. Cancer 2020, 11, 6727–6736. [Google Scholar] [CrossRef] [PubMed]
  16. Mathew, G.; Agha, R.; Albrecht, J.; Goel, P.; Mukherjee, I.; Pai, P.; D’Cruz, A.K.; Nixon, I.J.; Roberto, K.; Enam, S.A.; et al. STROCSS 2021: Strengthening the reporting of cohort, cross-sectional and case-control studies in surgery. Int. J. Surg. 2021, 96, 106165. [Google Scholar] [CrossRef]
  17. Goodrich, J.A.; Walker, D.; Lin, X.; Wang, H.; Lim, T.; McConnell, R.; Conti, D.V.; Chatzi, L.; Setiawan, V.W. Exposure to perfluoroalkyl substances and risk of hepatocellular carcinoma in a multiethnic cohort. JHEP Rep. 2022, 4, 100550. [Google Scholar] [CrossRef]
  18. Yan, B.; Su, B.B.; Bai, D.S.; Qian, J.J.; Zhang, C.; Jin, S.J.; Jiang, G.Q. A practical nomogram and risk stratification system predicting the cancer-specific survival for patients with early hepatocellular carcinoma. Cancer Med. 2021, 10, 496–506. [Google Scholar] [CrossRef]
  19. Wu, J.; Zhang, H.; Li, L.; Hu, M.; Chen, L.; Xu, B.; Song, Q. A nomogram for predicting overall survival in patients with low-grade endometrial stromal sarcoma: A population-based analysis. Cancer Commun. 2020, 40, 301–312. [Google Scholar] [CrossRef] [PubMed]
  20. Yang, D.; Zhu, M.; Xiong, X.; Su, Y.; Zhao, F.; Hu, Y.; Zhang, G.; Pei, J.; Ding, Y. Clinical features and prognostic factors in patients with microvascular infiltration of hepatocellular carcinoma: Development and validation of a nomogram and risk stratification based on the SEER database. Front. Oncol. 2022, 12, 987603. [Google Scholar] [CrossRef]
  21. Di Bisceglie, A.M.; Hoofnagle, J.H. Elevations in serum alpha-fetoprotein levels in patients with chronic hepatitis B. Cancer 1989, 64, 2117–2120. [Google Scholar] [CrossRef]
  22. Yamada, R.; Hiramatsu, N.; Oze, T.; Morishita, N.; Harada, N.; Yakushijin, T.; Iio, S.; Doi, Y.; Yamada, A.; Kaneko, A.; et al. Impact of alpha-fetoprotein on hepatocellular carcinoma development during entecavir treatment of chronic hepatitis B virus infection. J. Gastroenterol. 2015, 50, 785–794. [Google Scholar] [CrossRef] [PubMed]
  23. Wang, X.; Mao, M.; He, Z.; Zhang, L.; Li, H.; Lin, J.; He, Y.; Dai, S.; Hu, W.; Liu, W. Development and Validation of a Prognostic Nomogram in AFP-negative hepatocellular carcinoma. Int. J. Biol. Sci. 2019, 15, 221–228. [Google Scholar] [CrossRef] [PubMed]
  24. Lee, Y.T.; Wang, J.J.; Luu, M.; Noureddin, M.; Nissen, N.N.; Patel, T.C.; Roberts, L.R.; Singal, A.G.; Gores, G.J.; Yang, J.D. Comparison of Clinical Features and Outcomes Between Intrahepatic Cholangiocarcinoma and Hepatocellular Carcinoma in the United States. Hepatology 2021, 74, 2622–2632. [Google Scholar] [CrossRef] [PubMed]
  25. He, T.; Chen, T.; Liu, X.; Zhang, B.; Yue, S.; Cao, J.; Zhang, G. A Web-Based Prediction Model for Cancer-Specific Survival of Elderly Patients With Early Hepatocellular Carcinoma: A Study Based on SEER Database. Front. Public. Health 2021, 9, 789026. [Google Scholar] [CrossRef] [PubMed]
  26. Qiu, X.; Cai, J.; Chen, H.; Yao, J.; Xiao, C.; Li, R.; Xiao, J.; Zhang, J.; Sui, X.; Lu, T.; et al. Chemotherapy combined with radiotherapy can benefit more unresectable HCC patients with portal and/or hepatic vein invasion: A retrospective analysis of the SEER database. Front. Oncol. 2023, 13, 1098686. [Google Scholar] [CrossRef]
  27. Xie, Q.; Yang, Y.; Qu, B.; Xiao, P.; Tang, F.; Shen, H. Comparison of surgical resection and radiofrequency ablation for stages I and II elderly hepatocellular carcinoma patients (≥65 years): A SEER population-based propensity score matching’s study. Front. Oncol. 2022, 12, 903231. [Google Scholar] [CrossRef]
  28. VoPham, T.; Bertrand, K.A.; Fisher, J.A.; Ward, M.H.; Laden, F.; Jones, R.R. Emissions of dioxins and dioxin-like compounds and incidence of hepatocellular carcinoma in the United States. Environ. Res. 2022, 204, 112386. [Google Scholar] [CrossRef]
  29. Zhan, G.; Peng, H.; Zhou, L.; Jin, L.; Xie, X.; He, Y.; Wang, X.; Du, Z.; Cao, P. A web-based nomogram model for predicting the overall survival of hepatocellular carcinoma patients with external beam radiation therapy: A population study based on SEER database and a Chinese cohort. Front. Endocrinol. 2023, 14, 1070396. [Google Scholar] [CrossRef]
  30. European Association for the Study of the Liver. EASL Clinical Practice Guidelines: Management of hepatocellular carcinoma. J. Hepatol. 2018, 69, 182–236. [Google Scholar] [CrossRef]
  31. Kong, C.; Zhao, Z.; Chen, W.; Lv, X.; Shu, G.; Ye, M.; Song, J.; Ying, X.; Weng, Q.; Weng, W.; et al. Prediction of tumor response via a pretreatment MRI radiomics-based nomogram in HCC treated with TACE. Eur. Radiol. 2021, 31, 7500–7511. [Google Scholar] [CrossRef] [PubMed]
  32. Mao, K.; Yan, Y.; Zhang, J.; Wang, J.; Wang, R.; Ling, X.; Liu, Y.; Lau, W.Y.; Jiang, S.; Liu, J.; et al. The impact of liver resection on survival outcomes of hepatocellular carcinoma patients with extrahepatic metastases: A propensity score matching study. Cancer Med. 2018, 7, 4475–4484. [Google Scholar] [CrossRef] [PubMed]
  33. Lin, S.; Hoffmann, K.; Schemmer, P. Treatment of hepatocellular carcinoma: A systematic review. Liver Cancer 2012, 1, 144–158. [Google Scholar] [CrossRef]
  34. Liang, L.J.; Hu, W.J.; Yin, X.Y.; Zhou, Q.; Peng, B.G.; Li, D.M.; Lu, M.D. Adjuvant intraportal venous chemotherapy for patients with hepatocellular carcinoma and portal vein tumor thrombi following hepatectomy plus portal thrombectomy. World J. Surg. 2008, 32, 627–631. [Google Scholar] [CrossRef] [PubMed]
  35. Ramai, D.; Singh, J.; Lester, J.; Khan, S.R.; Chandan, S.; Tartaglia, N.; Ambrosi, A.; Serviddio, G.; Facciorusso, A. Systematic review with meta-analysis: Bariatric surgery reduces the incidence of hepatocellular carcinoma. Aliment. Pharmacol. Ther. 2021, 53, 977–984. [Google Scholar] [CrossRef]
  36. Facciorusso, A. The influence of diabetes in the pathogenesis and the clinical course of hepatocellular carcinoma: Recent findings and new perspectives. Curr. Diabetes Rev. 2013, 9, 382–386. [Google Scholar] [CrossRef]
  37. Yuan, G.; Song, Y.; Li, Q.; Hu, X.; Zang, M.; Dai, W.; Cheng, X.; Huang, W.; Yu, W.; Chen, M.; et al. Development and Validation of a Contrast-Enhanced CT-Based Radiomics Nomogram for Prediction of Therapeutic Efficacy of Anti-PD-1 Antibodies in Advanced HCC Patients. Front. Immunol. 2020, 11, 613946. [Google Scholar] [CrossRef]
  38. Yang, Z.; Zi, Q.; Xu, K.; Wang, C.; Chi, Q. Development of a macrophages-related 4-gene signature and nomogram for the overall survival prediction of hepatocellular carcinoma based on WGCNA and LASSO algorithm. Int. Immunopharmacol. 2021, 90, 107238. [Google Scholar] [CrossRef]
  39. Liu, H.; Li, Z.; Zhang, Q.; Li, Q.; Zhong, H.; Wang, Y.; Yang, H.; Li, H.; Wang, X.; Li, K.; et al. Multi-institutional development and validation of a nomogram to predict prognosis of early-onset gastric cancer patients. Front. Immunol. 2022, 13, 1007176. [Google Scholar] [CrossRef]
Cancers 15 05310 g001
Figure 1. Flow diagram of selecting patients.
Figure 1. Flow diagram of selecting patients.
Cancers 15 05310 g001
Cancers 15 05310 g002
Figure 2. The nomogram predicts the 1-, 3-, and 5- Years OS rates in HCC patients with EO-LIHC. Significant difference: * p  <  0.05, *** p  <  0.001.
Figure 2. The nomogram predicts the 1-, 3-, and 5- Years OS rates in HCC patients with EO-LIHC. Significant difference: * p  <  0.05, *** p  <  0.001.
Cancers 15 05310 g002
Cancers 15 05310 g003
Figure 3. The ROC curves of the nomogram to predict 1-, 3-, and 5-year OS: (A) the training cohort; (B) the validation cohort 1; and (C) the validation cohort 2.
Figure 3. The ROC curves of the nomogram to predict 1-, 3-, and 5-year OS: (A) the training cohort; (B) the validation cohort 1; and (C) the validation cohort 2.
Cancers 15 05310 g003
Cancers 15 05310 g004
Figure 4. The calibration curves of the nomograms using three cohorts show how survival predictions from the model compare to the actual observed survival; the calibration curve of 1-, 3-, and 5-year OS: (A) the training cohort; (B) the validation cohort 1; and (C) the validation cohort 2.
Figure 4. The calibration curves of the nomograms using three cohorts show how survival predictions from the model compare to the actual observed survival; the calibration curve of 1-, 3-, and 5-year OS: (A) the training cohort; (B) the validation cohort 1; and (C) the validation cohort 2.
Cancers 15 05310 g004
Cancers 15 05310 g005
Figure 5. Decision curve analysis of the nomogram and AJCC tumor staging for the overall survival prediction of HCC patients with EO-LIHC. (A,D,G) 1-, 3- and 5-year overall survival benefits based on the training cohorts. (B,E,H) 1-, 3- and 5-year overall survival benefits based on the training cohort 1. (C,F,I) 1-, 3- and 5-year overall survival benefits based on the training cohort 2.
Figure 5. Decision curve analysis of the nomogram and AJCC tumor staging for the overall survival prediction of HCC patients with EO-LIHC. (A,D,G) 1-, 3- and 5-year overall survival benefits based on the training cohorts. (B,E,H) 1-, 3- and 5-year overall survival benefits based on the training cohort 1. (C,F,I) 1-, 3- and 5-year overall survival benefits based on the training cohort 2.
Cancers 15 05310 g005
Cancers 15 05310 g006
Figure 6. The comparison chart of C-index between nomogram and AJCC stage: (A) the training cohort; (B) the validation cohort 1; and (C) the validation cohort 2.
Figure 6. The comparison chart of C-index between nomogram and AJCC stage: (A) the training cohort; (B) the validation cohort 1; and (C) the validation cohort 2.
Cancers 15 05310 g006
Cancers 15 05310 g007
Figure 7. The basis for the grouping new risk stratification (cut-off point selected using the X-tile). blue part (low risk: total score < 461), gray part (middle risk: 461 ≤ total score < 519), and purple part (high risk: total score ≥ 519).
Figure 7. The basis for the grouping new risk stratification (cut-off point selected using the X-tile). blue part (low risk: total score < 461), gray part (middle risk: 461 ≤ total score < 519), and purple part (high risk: total score ≥ 519).
Cancers 15 05310 g007
Cancers 15 05310 g008
Figure 8. Kaplan–Meier curves of overall survival for the new risk classification and the AJCC tumor staging: (A) the AJCC tumor staging in the training cohort; (B) the AJCC tumor staging in the validation cohort 1; (C) the AJCC tumor staging in the validation cohort 2; (D) the new risk classification in the training cohort; (E) the new risk classification in the validation cohort 1; and (F) the new risk classification in the validation cohort 2.
Figure 8. Kaplan–Meier curves of overall survival for the new risk classification and the AJCC tumor staging: (A) the AJCC tumor staging in the training cohort; (B) the AJCC tumor staging in the validation cohort 1; (C) the AJCC tumor staging in the validation cohort 2; (D) the new risk classification in the training cohort; (E) the new risk classification in the validation cohort 1; and (F) the new risk classification in the validation cohort 2.
Cancers 15 05310 g008
Cancers 15 05310 g009
Figure 9. A web-based nomogram for predicting overall survival for HCC patients with EO-LIHC.
Figure 9. A web-based nomogram for predicting overall survival for HCC patients with EO-LIHC.
Cancers 15 05310 g009
Table 1. Patients’ characteristics.
Table 1. Patients’ characteristics.
CharacteristicsTraining CohortValidation Cohort 1p Value
n899388
Age, n (%) 0.662
≥35685 (53.2%)300 (23.3%)
<35214 (16.6%)88 (6.8%)
Sex, n (%) 0.418
Male693 (53.8%)291 (22.6%)
Female206 (16%)97 (7.5%)
Race, n (%) 0.859
White536 (41.6%)225 (17.5%)
Other246 (19.1%)111 (8.6%)
Black117 (9.1%)52 (4%)
Grade, n (%) 0.966
Well/Moderate366 (28.4%)158 (12.3%)
Poor/Undifferentiated158 (12.3%)66 (5.1%)
Unknown375 (29.1%)164 (12.7%)
AJCC_Stage, n (%) 0.699
Unknown166 (12.9%)64 (5%)
stage I–II393 (30.5%)174 (13.5%)
stage III–IV340 (26.4%)150 (11.7%)
Stage_T, n (%) 0.218
Unknown179 (13.9%)65 (5.1%)
T1/T2474 (36.8%)201 (15.6%)
T3/T4246 (19.1%)122 (9.5%)
Stage_N, n (%) 0.163
Unknown180 (14%)63 (4.9%)
N0648 (50.3%)286 (22.2%)
N171 (5.5%)39 (3%)
Stage_M, n (%) 0.633
Unknown73 (5.7%)29 (2.3%)
M0656 (51%)293 (22.8%)
M1170 (13.2%)66 (5.1%)
Dissected lymph nodes, n (%) 0.176
0820 (63.7%)344 (26.7%)
1–369 (5.4%)35 (2.7%)
≥410 (0.8%)9 (0.7%)
Surgery, n (%) 0.765
No564 (43.8%)240 (18.6%)
Yes335 (26%)148 (11.5%)
Chemotherapy, n (%) 0.248
No551 (42.8%)251 (19.5%)
Yes348 (27%)137 (10.6%)
AFP, n (%) 0.101
Negative178 (13.8%)90 (7%)
Positive477 (37.1%)213 (16.6%)
Unknown244 (19%)85 (6.6%)
Tumor_size, n (%) 0.358
≤5 cm317 (24.6%)135 (10.5%)
5–10 cm209 (16.2%)94 (7.3%)
>10 cm173 (13.4%)87 (6.8%)
Unknown200 (15.5%)72 (5.6%)
Tumor_Number, n (%) 0.133
1876 (68.1%)372 (28.9%)
223 (1.8%)16 (1.2%)
Table 2. Univariate and multivariate analyses for OS.
Table 2. Univariate and multivariate analyses for OS.
VariableUnivariatepMultivariatep
HR95% CIHR95% CI
Age
<35Reference
≥351.451.19–1.77<0.0011.681.36–2.09<0.001
Sex
Female0.630.51–0.76<0.0010.670.54–0.82<0.001
MaleReference Reference
AJCC Stages
I and IIReference Reference
III and IV4.113.41–4.95<0.0012.591.59–2.52<0.001
Unknown2.481.97–3.11<0.0011.810.99–1.49<0.001
Grade
WellReference Reference
Bad2.682.15–3.34<0.00121.59–2.52<0.001
Unknown2.191.82–2.64<0.0011.210.99–1.490.067
Surgery
NoReference Reference
Yes0.230.19–0.28<0.0010.30.24–0.38<0.001
Chemotherapy
NoReference Reference
Yes1.030.88–1.210.720.520.44–0.62<0.001
AFP
NegativeReference Reference
Positive2.712.14–3.43<0.0011.941.51–2.48<0.001
Unknown1.541.18–2.01<0.010.910.67–1.220.52
Tumor size
≤5 cmReference Reference
5–10 cm2.221.79–2.76<0.0011.361.07–1.73<0.05
>10 cm2.321.84–2.92<0.0011.331.01–1.76<0.05
Unknown2.952.37–3.67<0.0011.270.96–1.690.099
Dissected lymph nodes
0Reference Reference
1–30.570.41–0.79<0.0010.890.63–1.250.499
≥40.260.08–0.81<0.050.470.15–1.520.208
Table 3. NRI and IDI to evaluate the predictive power of the model.
Table 3. NRI and IDI to evaluate the predictive power of the model.
IndexTraining CohortpValidation Cohortp
Value95% CIValue95% CI
NRI
1-year OS0.7940.636–0.929 0.7570.326–0.927
3-year OS0.6990.211–0.814 0.5210.292–0.783
5-year OS0.2040.099–0.328 0.2230.114–0.558
IDI
1-year OS0.1970.161–0.243<0.0010.2280.161–0.300<0.001
3-year OS0.2040.156–0.251<0.0010.3060.224–0.398<0.001
5-year OS0.1890.145–0.237<0.0010.3230.234–0.406<0.001
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Kuang, T.; Ma, W.; Zhang, J.; Yu, J.; Deng, W.; Dong, K.; Wang, W. Construction of a Nomogram to Predict Overall Survival in Patients with Early-Onset Hepatocellular Carcinoma: A Retrospective Cohort Study. Cancers 2023, 15, 5310. https://doi.org/10.3390/cancers15225310

AMA Style

Kuang T, Ma W, Zhang J, Yu J, Deng W, Dong K, Wang W. Construction of a Nomogram to Predict Overall Survival in Patients with Early-Onset Hepatocellular Carcinoma: A Retrospective Cohort Study. Cancers. 2023; 15(22):5310. https://doi.org/10.3390/cancers15225310

Chicago/Turabian Style

Kuang, Tianrui, Wangbin Ma, Jiacheng Zhang, Jia Yu, Wenhong Deng, Keshuai Dong, and Weixing Wang. 2023. "Construction of a Nomogram to Predict Overall Survival in Patients with Early-Onset Hepatocellular Carcinoma: A Retrospective Cohort Study" Cancers 15, no. 22: 5310. https://doi.org/10.3390/cancers15225310

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop