Development and validation of prognostic nomograms for patients with cervical cancer and liver metastasis: a SEER-based study

Introduction

Cervical cancer remains a leading malignancy of the female reproductive system globally. According to Global Cancer Statistics 2022, there were 662,301 new cases of cervical cancer worldwide and 348,874 deaths, with both incidence and mortality ranking fourth among malignancies affecting women (1). While human papillomavirus (HPV) screening has mitigated incidence rates, patients with distant metastasis, particularly to the liver, continue to face high mortality. While studies report a 5-year overall survival (OS) rate exceeding 90% for patients with localized cervical cancer, those who develop metastatic disease face a dismal prognosis, with a 5-year OS rate of 16.5% and a median OS of 8–13 months, as observed in multicenter cohorts (2,3). The prevalence of cervical cancer with liver metastases (CCLMs) varies between 1.2% and 2.2%, ranking third after lung and bone metastases. Despite its lower incidence compared to bone and lung metastasis, the 5-year survival rate for patients with CCLM is significantly lower than that for patients with bone or lung metastases (4,5). Nonetheless, investigations concerning patients diagnosed with CCLM, both nationally and globally, are predominantly confined to small sample studies or singular case reports (4,6,7). To date, no comprehensive large-scale studies have been conducted to predict the prognosis for these patients, likely due to the rarity of this condition. The prognosis of CCLM is correlated with various factors, including clinicopathological characteristics and treatment regimens. The conventional staging systems lack the granularity required to differentiate outcomes among patients already diagnosed with distant metastasis. Consequently, a precise, individualized prognostic tool is urgently required.

We specifically selected patients with cervical cancer harboring liver metastasis—with or without other distant metastases—as the modeling target for three reasons: (I) liver involvement confers a distinctly poor prognosis compared with other metastatic sites, with the lowest 5-year survival rate among all distant metastatic patterns; (II) restricting the cohort to liver-only metastasis would substantially reduce the sample size and compromise model stability, whereas modeling all metastatic cervical cancer as a single group would obscure the heterogeneity across different metastatic patterns; and (III) including concomitant metastases allows the model to explicitly account for the added prognostic burden of multi-organ dissemination, reflecting real-world clinical heterogeneity where liver metastasis rarely occurs in isolation.

A nomogram is a widely used tool in medical research for evaluating the incidence or prognosis of a particular disease. By converting complex regression equations into visual graphs, nomograms enhance the readability and accessibility of predictive model results. As such, nomograms have become an essential tool for clinicians in treatment planning and prognosis assessment across various cancer types (8-11).

In this study, we aim to develop and validate two accurate and effective prognostic nomograms for predicting OS and cancer-specific survival (CSS) in patients with CCLM. We obtained a significant number of cases from the Surveillance, Epidemiology, and End Results (SEER) database, which encompasses cancer patient data from many states and regions around the United States, so ensuring extensive geographic distribution and representation. We discovered independent risk factors influencing the prognosis of CCLM, developed two clinical prediction models, and constructed nomograms based on these results. These nomograms are reliable statistical tools that can predict long-term survival and guide treatment decisions. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0271/rc).

Methods

Data sources

Data were retrieved from the SEER database (www.seer.cancer.gov) using SEER*Stat 8.4.3 (17 registries, November 2022 submission). SEER captures roughly 50% of the U.S. population and provides comprehensive demographic, tumor, treatment, and survival information. Cases of CCLM diagnosed between January 1, 2010 and December 31, 2020 were identified (metastatic status was not recorded before 2010). The study was exempt from institutional review board approval due to the data being de-identified and publicly accessible (12). The design and analytic processes were performed in accordance with the principles outlined in the Declaration of Helsinki and its subsequent amendments. Inclusion criteria for the study were as follows: (I) a histopathological diagnosis of cervical cancer [site recode International Classification of Diseases-Third Edition (ICD-O-3)/World Health Organization (WHO) 2008 = Cervix Uteri]; (II) cervical cancer identified as the first primary malignancy; (III) behavior recorded for analysis labeled ‘Malignant’; (IV) the extent of disease SEER combined sets at DX-liver labeled ‘Yes’; and (V) initial diagnosis occurring between 2010 and 2020. Exclusion criteria included: (I) unclear race; (II) missing information on lung, bone, or brain metastasis; and (III) incomplete survival time data. After a thorough screening process, 704 patients remained available for analysis. The selection criterion and the specific process are shown in Figure 1.

Figure 1 Flowchart for selection patients with CCLM. CCLM, cervical cancer with liver metastasis; SEER, Surveillance, Epidemiology, and End Results.

Variable extraction

The variables were chosen based on data from the SEER database, clinical knowledge of liver metastases from cervical cancer, and previous research. The following variables were extracted: age, race, marital status, T stage, N stage, tumor grade, histologic type, primary site, surgery, radiotherapy, chemotherapy, lung/brain/bone metastasis, vital status, survival time, and cause‑specific death. The histological type was categorized as squamous cell carcinoma, adenocarcinoma, and ‘Other’. The ‘Other’ category encompasses carcinoma not otherwise specified (NOS), neuroendocrine carcinomas (including small cell and large cell neuroendocrine carcinoma), adenosquamous carcinoma, and other rare variants (e.g., clear cell carcinoma, serous carcinoma). The OS has been defined as the duration from the diagnosis of CCLM to the last follow-up or death from any cause. The CSS refers to the time span from the initial diagnosis to death attributed to cervical cancer.

Statistical analysis

We determined the optimal cutoff value for age using X-tile (version 3.6.1). This cutoff value served as a threshold to divide patients into groups that showed the most significant differences in survival outcomes (Figure 2). Before modeling, the samples were randomly allocated into training and validation cohorts in a 7:3 ratio via R’s random sampling function. The training cohort was employed for model construction, whilst the validation cohort was applied to assess the model’s performance and generalizability. Statistical analyses of the training and validation cohorts were performed utilizing chi-square and rank-sum tests. To minimize the risk of overfitting and address multicollinearity, the least absolute shrinkage and selection operator (LASSO) regression analysis was initially employed. Variables with non-zero coefficients were subsequently entered into a backward stepwise multivariate Cox regression model to ascertain independent predictive factors. This study identified variables with a P value below 0.05 as independent prognostic factors, which were utilized to develop a nomogram. The concordance index (C‑index) and the area under the curve (AUC) were employed to evaluate the discriminative efficacy of the nomograms. The precision of the nomograms was evaluated by calibration curves, which juxtaposed the anticipated OS rates with the actual OS rates. To mitigate overfitting bias, bootstrap resampling (B=100) was applied within the calibration procedure. Decision curve analysis (DCA) was employed to examine the clinical utility of the nomograms, assessing whether decisions made with the aid of the nomogram enhanced patient prognosis evaluation. All statistical analysis and data visualizations were conducted using R software (version 4.4.0) and PCPM_V5.48.1 (Jingding Medical Technology Co., Ltd.). LASSO regression was carried out using the “glmnet” package. The “rms” package was used to construct a prognostic nomogram and calibration curve, while the “survivalROC” package was employed for generating the receiver operating characteristic (ROC) curve. Furthermore, the “dcurves” packages facilitated the DCA. Survival analysis was performed separately for high-, medium-, and low-risk groups using Kaplan-Meier curves, with log-rank tests applied to assess differences across validation and training cohorts. A two-tailed P value <0.05 was considered statistically significant.

Figure 2 Estimation of the appropriate cutoff value for age at diagnose by X-tile analysis.

Results

Patient characteristics

From the SEER database, 748 patients diagnosed with CCLM between 2010 and 2020 were initially identified, with 44 patients excluded based on specific criteria. The remaining 704 patients were randomly allocated into a training cohort (n=492) and a validation cohort (n=212). Detailed demographic, clinicopathological characteristics, and therapy data for all patients are presented in Table 1. The training cohort is used to identify key prognostic factors and develop the nomogram. The validation cohort, on the other hand, provides an independent dataset to assess the model’s performance. This two-stage approach helps create an accurate and generalizable model for clinical application.

As a whole, the majority of patients were aged 51–74 years (59.52%), white (68.61%), and married (36.22%). The cervix uteri was the most common site for primary cervical cancer (86.08%), followed by the endocervix (10.94%). Histologically, squamous cell carcinoma accounted for 50.85%, adenocarcinoma 17.90%, and other subtypes 31.25%. Poorly differentiated tumors comprised 43.04% of the sample. Lung, bone, and brain metastases were present in 47.0%, 32.1%, and 3.3% of patients, respectively. Only 7.67% underwent surgery, whereas 64.63% received chemotherapy and 43.89% received radiotherapy. No significant baseline differences existed between training and validation cohorts (P>0.05). The 0.5-, 1-, and 2-year survival rates for CCLM patients were 48.3%, 23.0% and 8.0%, respectively.

Variable selection and nomogram construction

Based on the training cohort, initial screening of the variables was conducted using LASSO-Cox regression from 14 variables that may affect the prognostic risk of CCLM. This analysis allowed us to identify variables with non-zero coefficients. The LASSO regression identified five key prognostic factors for OS: age, grade, lung metastasis, radiation treatment, and chemotherapy. For CSS, six key prognostic factors emerged: age, tumor grade, histological type, lung metastasis, radiation treatment, and chemotherapy. The cross-validation error plot of the LASSO regression model demonstrated the model’s regularization effect, showing that the ideal model achieved its minimum cross-validation error (see Figure 3). Following the results of the LASSO regression, we incorporated the identified variables into a multivariate Cox regression analysis. The results indicated that lung metastasis [hazard ratio (HR): 1.493, 95% confidence interval (CI): 1.229–1.813, P<0.001], radiotherapy (HR: 0.747, 95% CI: 0.613–0.910, P=0.004), and chemotherapy (HR: 0.297, 95% CI: 0.240–0.367, P<0.001) were independent prognostic factors for OS (as shown in Table 2). Additionally, lung metastasis (HR: 1.543, 95% CI: 1.261–1.890, P<0.001), radiation (HR: 0.760, 95% CI: 0.619–0.932, P=0.008), chemotherapy (HR: 0.296, 95% CI: 0.238–0.369, P<0.001), and histological type (HR: 1.334, 95% CI: 1.069–1.664, P=0.01) were identified as independent prognostic factors for CSS (as shown in Table 3). To visually predict the 6-month, 1-year, and 2-year OS and CSS rates for CCLM patients, we plotted nomograms using R (see Figure 4A,4B). According to the OS nomogram, chemotherapy had the most significant impact on prognosis, followed by lung metastasis and radiation. Similarly, the CSS nomogram indicated that chemotherapy was the most influential factor, followed by lung metastasis, radiation, and histological type. Each variable was assigned a score, and the total score for patients with CCLM was obtained by summing the scores for all variables. A vertical line drawn at the total score on the horizontal axis indicates the 0.5-, 1-, and 2-year OS or CSS.

Figure 3 LASSO regression results for OS (A) and CSS (B). CSS, cancer-specific survival; LASSO, least absolute shrinkage and selection operator; OS, overall survival.

Figure 4 Nomogram predicting 0.5-, 1-, and 2-year OS (A) and CSS (B). CSS, cancer-specific survival; OS, overall survival.

Validation of the nomogram performance

The accuracy and effectiveness of our nomogram were verified using various methods. In the training and validation cohorts, the C-index of the OS nomogram was 0.720 (95% CI: 0.695–0.745) and 0.727 (95% CI: 0.686–0.768), respectively. The AUC metrics delineating the performance within the training cohort (0.788, 0.764, and 0.711 for the 0.5-, 1-, and 2-year OS, respectively) and the validation cohort (0.786, 0.738, and 0.754) demonstrate the commendable discriminatory capacity of the model (Figure 5A,5B). In the training cohort, the C-index of the CSS nomogram was 0.729 (95% CI: 0.702–0.756), while in the validation cohort, it was 0.724 (95% CI: 0.681–0.767). The model demonstrated a good discriminatory ability for CSS nomograms, as evidenced by the AUC values for the training cohort (0.795, 0.76, and 0.705 for the 0.5-, 1-, and 2-year CSS, respectively) and validation cohort (0.783, 0.725, and 0.737) (Figure 5C,5D). These findings indicate the consistent and robust discriminatory ability of our nomogram. The calibration plots (Figure 6) show the bootstrap-corrected predicted versus observed survival probabilities, demonstrating good agreement for both OS and CSS at 6, 12, and 24 months. Furthermore, both models produced net gains in the training and validation cohorts, as shown by the DCA curves at 0.5, 1, and 2 years (Figures 7,8). This signifies the commendable clinical utility of the models, affirming their good clinical efficacy and application value.

Figure 5 Receiver-operating characteristic curves. The ability of the model to be measured by the AUC. ROC curves (A) from the training set of OS, (B) from the validation set of OS, (C) from the training set of CSS, and (D) from the validation set of CSS. AUC, area under the curve; CI, confidence interval; CSS, cancer-specific survival; OS, overall survival; ROC, receiver operating characteristic.

Figure 6 Calibration curves of the nomogram for OS and CSS at 6, 12, and 24 months. (A) Training cohort OS; (B) validation cohort OS; (C) training cohort CSS; (D) validation cohort CSS. CSS, cancer-specific survival; OS, overall survival.

Figure 7 DCA of OS (A-C) from the training set and (D-F) from the validation set. DCA, decision curve analysis; OS, overall survival.

Figure 8 DCA of CSS (A-C) from the training set and (D-F) from the validation set. CSS, cancer-specific survival; DCA, decision curve analysis.

Risk stratification

A risk classification system for OS and CSS was developed based on the total scores of each patient produced by the nomogram. The OS nomogram categorizes patients into three risk groups: low risk (total score <27.06), intermediate risk (27.06≤ total score <100), and high risk (total score ≥100). Similarly, the CSS nomogram defines three risk groups: low risk (total score <36.53), intermediate risk (36.53≤ total score <124.07), and high risk (total score ≥124.07). There was a median OS time of 11 months in the low-risk group, while just 1 month was observed in the high-risk group. The Kaplan-Meier survival analysis with the log-rank test demonstrated significant differences in OS and CSS among the three risk groups of patients with CCLM (Figure 9A,9C). Both nomograms demonstrated that the high-risk group in the training cohort had worse prognoses than the low-risk group (P<0.001). Similar trends were observed within the validation cohort (Figure 9B,9D). Consequently, the nomogram was able to stratify risks more effectively. Utilizing a nomogram for risk stratification is instrumental in facilitating clinicians in accurately evaluating prognosis and devising personalized treatment strategies.

Figure 9 Survival curves showed the survival status classified by the OS nomogram of the training set (A) and the validation set (B). Survival curves showed the survival status classified by the CSS nomogram of the training set (C) and the validation set (D). CSS, cancer-specific survival; OS, overall survival.

Discussion

CCLM is characteristically associated with dismal clinical outcomes, highlighting the exigency for robust prognostic models to facilitate personalized management and optimize patient survival. Nomograms have emerged as indispensable graphical tools for clinical decision-making, offering intuitive and individualized risk assessments across various malignancies (13-18). While prognostic models for cervical cancer have been increasingly refined to address specific pathological subtypes, nodal status, and distinct patient populations (2,19-23), a dedicated tool for CCLM has been notably absent. This study addresses this gap by developing and validating the first comprehensive nomograms specifically designed to predict CSS and OS in patients with CCLM. By integrating heterogeneous clinicopathological variables into a user-friendly interface, our models allow clinicians to provide individualized survival estimates and optimize therapeutic decision-making.

Following stringent inclusion and exclusion criteria, 704 CCLM patients were identified from the SEER database and stratified into training and validation cohorts. The SEER database’s extensive sample size and longitudinal follow-up provide a robust foundation for high-fidelity cancer research (12). A key methodological strength of our study was the application of LASSO-Cox regression, which effectively mitigated overfitting—a common pitfall in high-dimensional clinical data analysis. Both models demonstrated superior discriminative ability, with C-index and AUC values exceeding 0.70. The DCA confirmed that our nomograms provide superior net clinical benefit compared to conventional staging, underscoring their practical utility in oncology clinics. Notably, our risk stratification system outperformed the conventional American Joint Committee on Cancer (AJCC) and International Federation of Gynecology and Obstetrics (FIGO) staging systems in terms of prognostic differentiation, underscoring its potential for clinical application.

While several studies have identified adenocarcinoma as an independent poor prognostic factor for cervical cancer patients with lung metastases, the survival disparities between adenocarcinoma and squamous cell carcinoma remain a subject of ongoing debate (24,25). In the present study, adenocarcinoma was identified as a poor prognostic factor for CSS in patients with CCLM, but not for OS. This discrepancy potentially arises from the “dilution effect” of competing non-cancer-related mortality risks in all-cause survival analyses; as patients with metastatic disease are susceptible to varied systemic complications, CSS provides a more precise reflection of tumor-specific aggressiveness. Whether prognostic outcomes differ between squamous cell carcinoma and adenocarcinoma specifically in the context of hepatic metastasis warrants further investigation. The proportion of ‘Unknown’ grade (42%) and ‘Other’ histology (31%) in our cohort merits comment. The high rate of unknown grading is characteristic of population-based registry data for metastatic disease, where histological re-biopsy at the metastatic site is not routine. The ‘Other’ histology subcategory captures a heterogeneous group of rare cervical carcinoma subtypes including adenosquamous, small cell neuroendocrine, clear cell, and carcinoma NOS variants—entities with biologically distinct behavior from the two dominant histotypes. These categories were retained rather than excluded to avoid sample-size reduction and selection bias; importantly, neither grade nor the histology “Other” subcategory reached independent significance in the OS model; histological type was only retained in the CSS model. These findings are robust because LASSO pre-selected them based on their predictive contribution. Numerous studies globally have demonstrated that lymph node metastasis is a key prognostic indicator of cervical cancer (19,26). However, our results indicate that once liver metastasis is established, the prognostic weight of regional lymph node status diminishes. This suggests that in the context of visceral dissemination, the disease enters a “systemic phase” where the distant metastatic burden becomes the primary driver of attrition, rendering regional nodal status a secondary concern. Conversely, lung metastasis emerged as a significant predictor of diminished OS and CSS in CCLM patients. This finding aligns with the “seed and soil” hypothesis, suggesting that multi-organ involvement reflects a high systemic metastatic burden and aggressive tumor biology. Consistently, and in agreement with observations in other metastatic malignancies, patients with multi-site metastases exhibited poorer outcomes compared to those with single-site involvement (27,28).

Regarding therapeutic interventions, our results underscore chemotherapy and radiation as the cornerstones of CCLM management. Notably, the nomogram indicated that chemotherapy provided the most substantial survival benefit, which is consistent with current guidelines advocating for systemic therapy as the primary approach for stage IVB disease. However, it should be noted that chemotherapy is the dominant predictor in both nomograms, contributing the largest point allocation. The model may therefore, to a substantial extent, reflect differences between chemotherapy-eligible and non-eligible patients. Because the SEER database lacks information on performance status [Eastern Cooperative Oncology Group (ECOG)/Karnofsky Performance Status (KPS)] and comorbidities-key determinants of treatment eligibility—the chemotherapy variable likely captures a mixture of treatment selection and treatment effect. We acknowledge this as an inherent limitation. Nevertheless, the remaining variables (lung metastasis, radiation, and histological type) provide additional prognostic discrimination that enables risk stratification within both chemotherapy-treated and untreated subgroups, supporting the model’s practical utility. The rapid advancement of targeted therapy and immunotherapy has transformed the treatment landscape for cervical cancer (29,30). To enhance the survival rate of patients with CCLM, a multimodal therapeutic approach must be pursued and tailored to each patient’s needs (31).

Despite the strengths of this study, several limitations must be acknowledged. First, its retrospective design inherently introduces selection bias. Second, the SEER database does not capture several potentially important pre-treatment variables, including performance status (ECOG/KPS), body mass index, serological biomarkers [squamous cell carcinoma antigen (SCC-Ag), cancer antigen 125 (CA-125), lactate dehydrogenase (LDH), albumin, hemoglobin], number/size of hepatic lesions, lymphovascular invasion, HPV status and genotype, and programmed death-ligand 1 (PD-L1) expression. Future studies utilizing institutional or prospective cohorts should comprehensively evaluate all available baseline variables to identify potentially overlooked prognostic factors. Finally, although our internal validation was successful, external validation in geographically diverse cohorts (particularly Eastern populations) is essential to ensure the global generalizability of the nomograms.

Conclusions

In conclusion, this study successfully developed and validated the first specialized nomograms for predicting 0.5-, 1-, and 2-year OS and CSS in patients with CCLM. By identifying chemotherapy, radiation, and lung metastasis as key prognostic determinants, we established a risk-stratification system that effectively differentiates patients into low-, intermediate-, and high-risk groups. These user-friendly tools provide accurate, individualized prognostic assessments, facilitating the optimization of treatment strategies and personalized follow-up in clinical practice. Future prospective, multicenter trials are warranted to further refine these models and validate their clinical impact. From a translational perspective, several prognostic markers identified in our study are already actionable. Clinicians may use the nomogram to identify high-risk patients who might benefit from more aggressive systemic regimens. In the current immunotherapy era, PD-L1 expression and microsatellite instability (MSI)/mismatch repair deficiency (dMMR) status represent actionable biomarkers that can guide the use of immune checkpoint inhibitors, as demonstrated by the KEYNOTE-826 trial. Future studies should integrate these molecular markers alongside treatment-response data to transition from purely prognostic to predictive modeling, enabling identification of patients most likely to benefit from specific therapeutic strategies including immunotherapy and targeted agents such as tisotumab vedotin.

Acknowledgments

We gratefully acknowledge the patients, investigators, and institutions contributing data to the SEER database.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0271/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0271/prf

Funding: This study was supported by the Academic Projects of Xi’an International Medical Center Hospital (Nos. 2022QN03 and 2022QN08).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0271/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Chang L, Zhao K. Construction and validation of an innovative prognostic nomogram for overall survival in cervical cancer patients with lung metastasis: an analysis utilizing the SEER database. Front Oncol 2024;14:1397454. [Crossref] [PubMed]
3. Xu M, Cao C, Wu P, et al. Advances in cervical cancer: current insights and future directions. Cancer Commun (Lond) 2025;45:77-109. [Crossref] [PubMed]
4. Yin Z, Tang H, Li L, et al. Impact of sites versus number of metastases on survival of patients with organ metastasis from newly diagnosed cervical cancer. Cancer Manag Res 2019;11:7759-66. [Crossref] [PubMed]
5. Shan Y, Ding Z, Cui Z, et al. Incidence, prognostic factors and a nomogram of cervical cancer with distant organ metastasis: a SEER-based study. J Obstet Gynaecol 2023;43:2181690. [Crossref] [PubMed]
6. Narayanan S, Biju P, Sundaramoorthy S, et al. A Unique Finding of Hepatogastric Fistula in Cervical Cancer Liver Metastasis. Cureus 2021;13:e20761. [Crossref] [PubMed]
7. Kundumadam S, Laharwani H, Azhikkalayil S, et al. Metastatic Mucinous Adenocarcinoma of the Cervix Presenting as Acute Hepatitis. Cureus 2020;12:e6620. [Crossref] [PubMed]
8. Yan Q, Zhang Z, Li R, et al. Clinical and nutritional-inflammatory biomarkers-based nomogram predicts survival in recurrent or metastatic cervical cancer treated with immune checkpoint inhibitors. Ann Med 2026;58:2603780. [Crossref] [PubMed]
9. Wu Y, Zhang W, Liang X, et al. Habitat radiomics analysis for progression free survival and immune-related adverse reaction prediction in non-small cell lung cancer treated by immunotherapy. J Transl Med 2025;23:393. [Crossref] [PubMed]
10. Ge X, Yang G, Wu H, et al. Prognostic Factors and Nomogram-Based Prediction Models for Colorectal Cancer Patients With Synchronous Peritoneal Metastasis Undergoing Cytoreductive Surgery: A Retrospective Cohort Study. Cancer Med 2026;15:e71464. [Crossref] [PubMed]
11. Tan L, Su X, Ma Y, et al. Cholesterol Metabolism-related Characteristics Predict Therapeutic Response and Survival in Esophageal Cancer. J Adv Res 2026;81:837-48. [Crossref] [PubMed]
12. Che WQ, Li YJ, Tsang CK, et al. How to use the Surveillance, Epidemiology, and End Results (SEER) data: research design and methodology. Mil Med Res 2023;10:50. [Crossref] [PubMed]
13. Lu Y, Sun J, Huang J, et al. Establishment and validation of a prognostic nomogram for overall survival in type II endometrial carcinoma patients. Sci Rep 2025;15:7801. [Crossref] [PubMed]
14. Nie GL, Geng L, Zhang H, et al. Nomograms of risk prediction and prognosis for the T1-T2 stage gastric cancer with lymph node metastasis: a population-based study. Front Med (Lausanne) 2025;12:1492041. [Crossref] [PubMed]
15. Liu X, Xue J, Dong F, et al. A nomogram to predict postoperative overall and cancer specific survival in patients with primary parotid squamous cell carcinoma: a population based study. Sci Rep 2025;15:6774. [Crossref] [PubMed]
16. Chen Y, Duan C, Zhang X, et al. A nomogram for predicting cancer-specific survival in patients with rectal mucinous adenocarcinoma following surgery. Sci Rep 2025;15:6809. [Crossref] [PubMed]
17. Yu X, Deng Q, Gao X, et al. A prognostic nomogram for distant metastasis in thyroid cancer patients without lymph node metastasis. Front Endocrinol (Lausanne) 2025;16:1523785. [Crossref] [PubMed]
18. Xie L, Zhang Y, Niu X, et al. A nomogram for predicting cancer-specific survival in patients with locally advanced unresectable esophageal cancer: development and validation study. Front Immunol 2025;16:1524439. [Crossref] [PubMed]
19. Li Z, Lin Y, Cheng B, et al. Prognostic Model for Predicting Overall and Cancer-Specific Survival Among Patients With Cervical Squamous Cell Carcinoma: A SEER Based Study. Front Oncol 2021;11:651975. [Crossref] [PubMed]
20. Feng Y, Wang Y, Xie Y, et al. Nomograms predicting the overall survival and cancer-specific survival of patients with stage IIIC1 cervical cancer. BMC Cancer 2021;21:450. [Crossref] [PubMed]
21. Li E, Ni H. Prognostic nomogram for early-stage cervical cancer in the elderly: A SEER database analysis. Prev Med Rep 2024;41:102700. [Crossref] [PubMed]
22. Yu W, Huang L, Zhong Z, et al. A Nomogram-Based Risk Classification System Predicting the Overall Survival of Patients With Newly Diagnosed Stage IVB Cervix Uteri Carcinoma. Front Med (Lausanne) 2021;8:693567. [Crossref] [PubMed]
23. Liu Q, Li W, Xie M, et al. Development and validation of a SEER-based prognostic nomogram for cervical cancer patients below the age of 45 years. Bosn J Basic Med Sci 2021;21:620-31. [Crossref] [PubMed]
24. Chen X, Chen L, Zhu H, et al. Risk factors and prognostic predictors for Cervical Cancer patients with lung metastasis. J Cancer 2020;11:5880-9. [Crossref] [PubMed]
25. Pan X, Yang W, Wen Z, et al. Does adenocarcinoma have a worse prognosis than squamous cell carcinoma in patients with cervical cancer? A real-world study with a propensity score matching analysis. J Gynecol Oncol 2020;31:e80.
26. Akinyemi OA, Abodunrin FO, Andine TF, et al. Second Malignancies Following Primary Cervical Cancer Diagnosis: Analysis of the SEER Database. Cureus 2022;14:e26171. [Crossref] [PubMed]
27. Qian C, Liu H, Feng Y, et al. Clinical characteristics and risk of second primary lung cancer after cervical cancer: A population-based study. PLoS One 2020;15:e0231807. [Crossref] [PubMed]
28. Gardner AB, Charo LM, Mann AK, et al. Ovarian, uterine, and cervical cancer patients with distant metastases at diagnosis: most common locations and outcomes. Clin Exp Metastasis 2020;37:107-13. [Crossref] [PubMed]
29. Tewari KS, Monk BJ, Vergote I, et al. Survival with Cemiplimab in Recurrent Cervical Cancer. N Engl J Med 2022;386:544-55. [Crossref] [PubMed]
30. Colombo N, Dubot C, Lorusso D, et al. Pembrolizumab for Persistent, Recurrent, or Metastatic Cervical Cancer. N Engl J Med 2021;385:1856-67. [Crossref] [PubMed]
31. Abu-Rustum NR, Yashar CM, Arend R, et al. NCCN Guidelines® Insights: Cervical Cancer, Version 1.2024. J Natl Compr Canc Netw 2023;21:1224-33. [Crossref] [PubMed]