Construction and validation a Nomogram for Predicting Cancer-Specific Survival for patients with Ependymoma: An Analysis of the Surveillance, Epidemiology, and End Results Database

doi:10.21203/rs.3.rs-4674733/v1

Download PDF

Article

Construction and validation a Nomogram for Predicting Cancer-Specific Survival for patients with Ependymoma: An Analysis of the Surveillance, Epidemiology, and End Results Database

https://doi.org/10.21203/rs.3.rs-4674733/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

The purpose of this study was to Construct and validate a novel nomogram to predict cancer-specific survival (CSS) in patients with ependymoma.

Methods

Clinical data from the Surveillance, Epidemiology, and End Results database spanning 2000 to 2018 were used for the analysis. Data were randomly categorized into development and validation groups in a 7:3 ratio. Univariate and multivariate Cox regression analyses and LASSO regression were conducted to identify independent risk factors for CSS. Predictive models were evaluated using calibration plots, concordance index (C-index), and area under the receiver operating characteristic curve (AUC). Additionally, decision curve analysis and clinical impact curves were conducted.

Results

The final sample comprised 2,340 patients. Multivariate analysis identified race, age, histological type, surgery, and tumor site as independent predictors of CSS. A nomogram incorporating these risk factors was developed to predict CSS in patients with ependymomas. The calibration plot shows strong agreement between predicted and actual values. The C-index for the training cohort was 0.746 (95% CI: 0.715–0.777), and for the validation cohort, it was 0.743 (95% CI: 0.698–0.788). Additionally, both AUC and decision curve analysis analyses indicated robust performance and clinical significance benefits. Kaplan–Meier curves further demonstrated the nomogram’s strong ability to predict patient outcomes.

Conclusions

Our nomogram may be of great value in predicting the prognosis of patients with ependymoma. This predictive model will assist doctors and patients in devising effective clinical strategies.

Biological sciences/Cancer/Cancer prevention

Biological sciences/Cancer/Cancer therapy

Ependymoma

Prognostic factors

prognostic model

SEER database

Ependymoma is a neuroepithelial tumor that typically grows along the entire neuraxis, often in association with the ventricles or central canal of the spinal cord (1). These tumors account for 1.6–1.8% of all primary central nervous system (CNS) tumors (2) and rank as the third most common CNS tumors in children, accounting for 6–12% of pediatric brain tumors compared with approximately 4% in adults (3). Men have an approximately 1.3 times higher likelihood of being affected than women. While approximately 90% of pediatric ependymomas occur intracranially, the spinal cord is the primary site for 66% of adult tumors (4).

In the case of ependymoma, the standard treatment is to remove as much of the tumor as safely possible. Due to its rarity, the literature on ependymomas is limited and typically involves a small patient cohort. The evidence regarding the adjuvant treatment approaches is currently low. Consequently, the prediction of ependymoma prognosis remains unclear. Accurately predicting patient survival can significantly improve health, overall quality of life, and lifespan. Consequently, creating a prognostic model to effectively estimate the survival of individuals with ependymomas is essential. A nomogram, which is a precise and user-friendly tool, can evaluate risk factors and predict tumor prognosis (5). The Surveillance,Epidemiology,and End Results (SEER) database in the United States collects comprehensive patient information, clinicopathological characteristics, and treatment data for about a third of the U.S. population, making it a valuable resource for studying cancer outcomes.

The research conducted by Hu et al. (6) in 2021 resulted in the creation of a nomogram designed to predict the overall survival (OS) outcomes for ependymoma patients. The predictive tool identified age, grade of the tumor, initial site of the tumor, surgery performed, tumor stage, and size of the tumor as distinct risk factors for ependymoma. However, their study did not explore the impact of chemotherapy or radiotherapy. The existing body of research presents conflicting findings regarding the effectiveness of chemotherapy and radiotherapy for treating ependymomas. One study (7) showed that performing multiple surgical procedures followed by a novel radiotherapeutic method is feasible, suggesting a potential improvement in outcomes for children.

On the contrary, a 2013 study (8) with 92 participants who had fully resected non-anaplastic ependymomas in the supratentorial area revealed that radiotherapy did not affect survival results. Similarly, in the Children’s Cancer Group 942 trial, Participants with ependymoma were randomly divided into two cohorts: one cohort received craniospinal irradiation (CSI) exclusively, while the other cohort was treated with CSI plus weekly vincristine and eight cycles of supplementary therapy including vincristine, prednisone, and lomustine. Upon reviewing the 10-year follow-up data, an overall survival (OS) rate of 39% was recorded for all patients, with no discernible differences noted between the two groups (9). However, in children without metastatic disease who did not receive upfront radiation therapy, a 5-year overall survival rate of 63.4% was observed, indicating a possible benefit of chemotherapy (10).

Our research involved integrating radiotherapy and chemotherapy data to screen for prognostic factors associated with ependymoma patients, with the SEER database serving as the basis for our statistical analysis.

2.1. Data Acquisition and Processing

Patient data were obtained from the National Cancer Institute SEER Project, encompassing all patients diagnosed with ependymomas between 2000 and 2021. The SEER database is an extensive cancer registry that includes about 28% of the U.S. population and 17 cancer registries (11). Demographic information, survival status data, and clinicopathological characteristics were obtained from the SEER database. Our study utilizes publicly accessible and anonymized clinicopathological data, thereby eliminating the necessity for ethical approval and consent from patients. Our approach conforms to the guidelines of the SEER database. We gathered demographic information (age, race, sex, and marital status), characteristics of the tumors (grade, size, type, and location), treatment details (surgery, radiotherapy, and chemotherapy), and follow-up details (survival time and event-free survival). Participants were included in the study based on a diagnosis of ependymoma. Exclusions were based on the following criteria: (1) surgical techniques not specified, (2) non-primary malignancies, (3) undefined tumor sizes, and (4) uncertain tumor locations.

2.2. Patient Characteristics and Eligibility Criteria

Patients were categorized by racial background into white, black, or Asian/Pacific Islander. The main tumor locations were categorized as the spine, ventricle, supratentorial, and subtentorial. Histological types were further categorized into three groups based on the SEER surgery codes 9391, 9392, and 9393, representing ependymomas, anaplastic ependymomas, and papillary ependymomas, respectively. We randomly distributed the patients into a training subset (70%) and a validation subset (30%) to achieve a balanced dataset.

2.3. Prognostic Factor Analysis and Model Development

Factors predictive of outcomes were ascertained through a univariate Cox regression analysis. Additionally, the LASSO method was applied for regression analysis to select variables linked to ependymoma prognosis. Subsequently, the training dataset was analyzed using a multivariate Cox regression to determine the independent risk factors. We created a predictive model in the form of a nomogram to project the 1-, 5-, and 8-year CSS rates in individuals with ependymomas, incorporating risk factors determined by univariate and multivariate Cox regression analyses. The nomogram assigned each variable to its corresponding line based on weight, which was then translated into points. The totol scores calculated from all variables in the nomogram, facilitating survival rate estimation at various time points. The C-index was employed to determine the model's ability to discriminate between different outcomes.

2.4. Model Evaluation and Validation

The model's validity was established by employing a receiver operating characteristic (ROC) curve. To ensure the nomogram's reliability, a calibration curve analysis was conducted using 1,000 bootstrap samples to compare observed outcomes with predicted values. Decision curve analysis (DCA) was conducted to measure the advantages in survival at 1, 5, and 8 years. Subsequently, the patients were stratified into two categories—low-risk and high-risk—based on their scores from the nomogram. We then used the Kaplan-Meier survival curves and the log-rank test to examine the variations in survival rates between these risk categories.

2.5. Statistical Analysis

Statistical analyses were conducted using the SPSS23.0 software (IBM, Chicago, IL, USA) and R software (version 3.4.1; http://www.Rproject.org). For the analysis, tumor size and age were categorized as continuous variables, with their non-normal distributions described using median and interquartile range measures. The preliminary evaluation of prognostic factors was performed through univariate Cox regression analysis. Variables that showed significance were advanced to a multivariate analysis, where a stepwise regression technique was employed to select the optimal model. A result was deemed statistically significant if the p-value was less than 0.05.

3.1. Baseline characteristics

Figure 1 presents the flowchart of the study. We identified 2,340 patients with ependymomas who met the screening criteria. Table 1 summarizes the clinical characteristics of the included patients. Men accounted for 50.8% (1189) and women accounted for 49.2% (1151) of all patients. The median age of the patients was 41 years (interquartile range, 22–56 years), including 1952 white (83.4%), 207 black (8.85%), and 181 Asians or Pacific Islanders (7.74%). Patients were randomly categorized into training (n = 1639) and validation (n = 701) sets, with relatively similar distributions between the training and validation sets regarding enrolled characteristics.

Table 1

Clinicopathological characteristics of patients with Ependymoma
Characteristics	ALL N = 2340	Testing n = 701	Training n = 1639	p
Race				0.999
White	1952 (83.4%)	585 (83.5%)	1367 (83.4%)
Black	207 (8.85%)	62 (8.84%)	145 (8.85%)
Asian or Pacific Islander	181 (7.74%)	54 (7.70%)	127 (7.75%)
Site				0.253
Spine	1210 (51.7%)	341 (48.6%)	869 (53.0%)
SUB	578 (24.7%)	180 (25.7%)	398 (24.3%)
SUP	288 (12.3%)	93 (13.3%)	195 (11.9%)
Ventricle(NOS)	264 (11.3%)	87 (12.4%)	177 (10.8%)
Surgery				0.614
No	199 (8.50%)	56 (7.99%)	143 (8.72%)
Yes	2141 (91.5%)	645 (92.0%)	1496 (91.3%)
Chemotherapy				0.755
No	2151 (91.9%)	642 (91.6%)	1509 (92.1%)
Yes	189 (8.08%)	59 (8.42%)	130 (7.93%)
Radiothearapy				0.065
0	1473 (62.9%)	421 (60.1%)	1052 (64.2%)
1	867 (37.1%)	280 (39.9%)	587 (35.8%)
Sex				0.978
Female	1151 (49.2%)	344 (49.1%)	807 (49.2%)
Male	1189 (50.8%)	357 (50.9%)	832 (50.8%)
Histologic				0.008
9391	1858 (79.4%)	530 (75.6%)	1328 (81.0%)
9392	444 (19.0%)	155 (22.1%)	289 (17.6%)
9393	38 (1.62%)	16 (2.28%)	22 (1.34%)
size	47.0 [25.0;999]	45.0 [25.0;999]	48.0 [25.0;999]	0.381
Age	41.0 [22.0;56.0]	41.0 [20.0;56.0]	41.0 [22.0;56.0]	0.457
Married				0.534
No	1067 (45.6%)	327 (46.6%)	740 (45.1%)
Yes	1273 (54.4%)	374 (53.4%)	899 (54.9%)
Alive				0.79
Yes	1937 (82.8%)	583 (83.2%)	1354 (82.6%)
No	403 (17.2%)	118 (16.8%)	285 (17.4%)

3.2. Establish the prognostic model

Through a univariate Cox regression analysis, we identified potential prognostic factors, finding that race, age, histological type, surgery, chemotherapy, and site (all p-values < 0.05) were significantly related to patient prognosis. The LASSO regression model was applied to choose the most relevant variables,with dashed lines marking the optimal value (using a minimum “min” value as the criterion, Fig. 2). After multivariate Cox regression analysis, we verified that race, age, histological type, surgery, marital status, and site (all p < 0.05) were verified as independent prognostic factors for CSS (Table 2). A nomogram was created for predicting CSS using the six independent prognostic factors (Fig. 3). By totaling the scores assigned to each variable, we can easily calculate the survival chances for a specific patient.

Table 2

Univariate and multivariate analyses on the risk factors of patients with ependymoma
Groups	Univariate analysis		Multivariable analysis
Groups	HR (95%CI)	P-value	HR (95%CI)	P-value
Race
Asian or Pacific Islander	1 [Reference]		1 [Reference]
Black	2.15 (1.196–3.865)	0.011	1.891 (1.047–3.418)	0.035
White	1.38 (0.831–2.29)	0.213	1.201 (0.719–2.007)	0.484
Age	1.016 (1.01–1.022)	< 0.0001	1.036 (1.029–1.044)	< 0.0001
Size	1 (1–1)	0.505
Histological type
Histologic9391	1 [Reference]		1 [Reference]
Histologic9392	2.484 (1.929-3.2)	< 0.0001	4.295 (3.088–5.973)	< 0.0001
Histologic9393	0.579 (0.144–2.332)	0.442	0.774 (0.191–3.136)	0.72
Sex
Female	1 [Reference]		1 [Reference]
Male	1.328 (1.05–1.678)	0.018	1.27 (1.001–1.611)	0.049
Married
No	1 [Reference]		1 [Reference]
Yes	1.089 (0.86–1.377)	0.479	0.654 (0.48–0.892)	0.007
Radiotherapy
No	1 [Reference]
Yes	1.167 (0.919–1.481)	0.206
Surgery
No	1 [Reference]
Yes	0.448 (0.323–0.622)	< 0.0001	0.501 (0.35–0.716)	< 0.0001
Chemotherapy
No	1 [Reference]
Yes	1.647 (1.156–2.346)	0.006
Site
Spinal	1 [Reference]		1 [Reference]
SUB	1.913 (1.445–2.533)	< 0.0001	2.104 (1.569–2.822)	< 0.0001
SUP	2.053 (1.457–2.892)	< 0.0001	1.653 (1.09–2.507)	0.018
Ventricle(NOS)	2.059 (1.437–2.95)	< 0.0001	1.934 (1.331–2.81)	0.001

3.3. Validation of the nomogram

The concordance index (C-index) of the nomogram prediction model in the development and validation samples was 0.746 (95% CI 0.715–0.777) and 0.743 (95% CI 0.698–0.788), respectively, demonstrating good discriminative power. The training set showed a 1-, 5-, and 8-year AUC for the nomograms of 0.803, 0.770, and 0.757, respectively. The validation set exhibited AUCs of 0.830, 0.742, and 0.715 for the 1-, 5-, and 8-year nomograms, respectively (Figs. 4A, B). Validation findings indicated that the nomogram had a minimum accuracy of 70%, especially when predicting early-term survival outcomes. In the calibration analysis, the training and validation sets displayed a close match between the projected and observed outcomes (Figs. 5A and 5B), validating the nomogram's reliability. After establishing the precision of the predictive model, we conducted decision curve analysis (DCA) for further scrutiny. This analysis indicated a broad threshold probability range in the histogram, highlighting the model's excellent clinical relevance for forecasting cause-specific survival (CSS) in ependymoma cases, with a substantial net benefit (Fig. 6A, B). Using nomogram scores, patients were assigned to either low-, moderate-, or high-risk groups. The Kaplan-Meier survival curves indicated that high-risk patients had decreased survival rates in contrast to the low-risk group (p < 0.0001) (Figs. 7A, B).

Ependymomas are CNS tumors with a relatively low morbidity rate, and comprise only 1.6–1.8% of all primary tumors (12). Owing to their low incidence, prognosis investigations for ependymomas remain limited, with most series involving a restricted number of patients and applied treatment, which remains controversial (13). Moreover, most ependymoma studies have primarily focused on children, with limited research on adult ependymomas. Therefore, analyzing factors affecting the survival time of patients with ependymomas across all age groups is important for guiding clinical work. A previous analysis has produced a nomogram for forecasting outcomes in patients with ependymoma; however, it presented a low C-index and excluded treatments such as radiotherapy and chemotherapy (6). Therefore, we constructed a clinical prognostic model using data from the SEER database to estimate survival rates.

This study’s clinical nomogram included five independent prognostic factors identified by means of univariate and multivariate Cox regression analyses: age, race, site, sex, surgery, histology, and marital status. Our findings indicated that age significantly influenced cause-specific survival (CSS) in ependymoma patients. Being a known risk factor for cancer, older adults are at an increased risk of complications following surgery, potentially impacting their overall prognosis (14). Furthermore, our research detected a substantial difference in cause-specific survival (CSS), with males exhibiting lower rates than females, corroborating past investigations (6). The etiology behind the higher incidence of ependymomas in men compared to women is still unclear.

Furthermore, race significantly affected patient survival, with individuals of Asian or Pacific descent having a longer CSS than whites, followed by blacks. These outcome disparities may stem from socioeconomic factors. A study examining various types of childhood cancer and racial disparities in black and white children highlighted the crucial role of socioeconomic status in increased cancer mortality among black children (15). In our investigation, we also considered the influence of marital status on cause-specific survival (CSS). Logistic regression analysis showed that marital status is an independent predictor of CSS. Several potential reasons could explain this finding, including increased support from family and spouses among married patients, contributing to better prognosis than single patients (16). As with previous studies, our research indicated that the histological classification is linked to the prognosis of individuals with ependymoma (17).

Additionally, tumors’ location in the spinal cord has been linked to better survival outcomes. Spinal ependymomas grow relatively slowly, while aggressive cases are rare (18). Nonetheless, in the case of brain tumors, the entanglement of cranial nerves and vascular structures might restrict the possibility of fully resecting the tumor. It is generally accepted that surgery is the preferred treatment method, with studies consistently showing that resection offers better outcomes than no resection (19–21).

It should be highlighted that our findings suggest that radiation therapy does not confer any survival advantages to ependymoma patients. Despite the common practice of incorporating post-operative radiotherapy into the standard treatment plan for adults with ependymoma following incomplete resection, its benefits have been minimal (8). Similarly, chemotherapy has not demonstrated efficacy in improving survival.

Our nomogram demonstrated greater effectiveness in discriminating and predicting survival than the previous staging approach. However, it is important to acknowledge the limitations of this study. First, there may be misclassification of information owing to variations in the expertise of researchers providing data. Second, while our study utilized data from the SEER database, certain variables, such as surgical margins, tumor recurrence, and genetic factors, were unavailable. Additionally, specifying the location for tumors identified as ventricular poses challenges as they cannot be assigned to supratentorial, infratentorial, or spinal categories for survival analysis.

Consequently, we categorized these cases separately as ‘not otherwise specified’ (NOS) ventricles. Finally, it is crucial to note that the constructed nomogram lacks external data for validation. Therefore, further external validation studies are required to assess the accuracy and reliability of the prediction models. Additionally, we plan to carry out future studies at our hospital to confirm the accuracy of this prediction model and promote its its use in clinical settings.

Despite these limitations, the nomogram exhibits good accuracy and reliability, bearing significant clinical significance. With its user-friendly design, the nomogram can be readily utilized by anyone after a brief learning period.

Our research team has formulated a new predictive tool in the form of a nomogram, aimed at estimating the overall survival (OS) of ependymoma patients. Additionally, we identified tumor type as a significant risk factor affecting these patients’ OS. Identifying risk factors and constructing nomograms can assist physicians in accurately assessing patients’ prognoses, addressing patient inquiries, and aiding patients in making informed clinical decisions.

SEER: Surveillance, Epidemiology, and End Results
CNS: Central Nervous System
CSS: Cancer-Specific Survival
AUC: Area Under the Curve
ROC: Receiver Operating Characteristic
DCA: Decision curve analysis

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and analyzed in the current study are available from the Seer database.

Competing interests

The authors declare that they have no competing interests.

Funding

Not applicable.

Acknowledgments

Not applicable.

Araki A, Chocholous M, Gojo J, Dorfer C, Czech T, Heinzl H, et al. Chromosome 1q gain and tenascin-C expression are candidate markers to define different risk groups in pediatric posterior fossa ependymoma. Acta Neuropathol Commun. 2016;4(1):88.
McGuire C, Sainani K, Fisher P. Incidence patterns for ependymoma: a surveillance, epidemiology, and end results study. Journal of neurosurgery. 2009;110(4):725–9.
Liu F, Dong H, Mei Z, Huang T. Investigation of miRNA and mRNA Co-expression Network in Ependymoma. Front Bioeng Biotechnol. 2020;8:177.
Zhang C, Ostrom Q, Semmes E, Ramaswamy V, Hansen H, Morimoto L, et al. Genetic predisposition to longer telomere length and risk of childhood, adolescent and adult-onset ependymoma. Acta neuropathologica communications. 2020;8(1):173.
He X, Lu M, Hu X, Li L, Zou C, Luo Y, et al. Osteosarcoma immune prognostic index can indicate the nature of indete rminate pulmonary nodules and predict the metachronous metastasis in o steosarcoma patients. Frontiers in oncology.12.
Jia Z, Yan Y, Wang J, Yang H, Zhan H, Chen Q, et al. Development and validation of prognostic nomogram in ependymoma: A retrospective analysis of the SEER database. Cancer Med. 2021;10(17):6140–8.
Swanson EL, Amdur RJ, Morris CG, Galloway TJ, Marcus RB, Pincus DW, et al. Intracranial ependymomas treated with radiotherapy: long-term results from a single institution. Journal of Neuro-Oncology. 2010;102(3):451–7.
Ghia AJ, Mahajan A, Allen PK, Armstrong TS, Lang FF, Jr., Gilbert MR, et al. Supratentorial gross-totally resected non-anaplastic ependymoma: population based patterns of care and outcomes analysis. J Neurooncol. 2013;115(3):513–20.
Evans A, Anderson J, Lefkowitz-Boudreaux I, Finlay J. Adjuvant chemotherapy of childhood posterior fossa ependymoma: cranio-spinal irradiation with or without adjuvant CCNU, vincristine, and prednisone: a Childrens Cancer Group study. Medical and pediatric oncology. 1996;27(1):8–14.
Grundy RG, Wilne SA, Weston CL, Robinson K, Lashford LS, Ironside J, et al. Primary postoperative chemotherapy without radiotherapy for intracranial ependymoma in children: the UKCCSG/SIOP prospective study. The Lancet Oncology. 2007;8(8):696–705.
Cronin KA, Ries LA, Edwards BK. The Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute. Cancer. 2014;120 Suppl 23:3755–7.
Liu F, Dong H, Mei Z, Huang T. Investigation of miRNA and mRNA Co-expression Network in Ependymoma. Front Bioeng Biotechnol.8:177.
Senetta R, Miracco C, Lanzafame S, Chiusa L, Caltabiano R, Galia A, et al. Epidermal growth factor receptor and caveolin-1 coexpression identifies adult supratentorial ependymomas with rapid unfavorable outcomes. Neuro Oncol. 2011;13(2):176–83.
Ruda R, Bruno F, Pellerino A, Soffietti R. Ependymoma: Evaluation and Management Updates. Curr Oncol Rep. 2022;24(8):985–93.
Kehm RD, Spector LG, Poynter JN, Vock DM, Altekruse SF, Osypuk TL. Does socioeconomic status account for racial and ethnic disparities in childhood cancer survival? Cancer. 2018;124(20):4090–7.
Goldzweig G, Andritsch E, Hubert A, Walach N, Perry S, Brenner B, et al. How relevant is marital status and gender variables in coping with colorectal cancer? A sample of middle-aged and older cancer survivors. Psycho-oncology. 2009;18(8):866–74.
Lopez-Rivera V, Dono A, Abdelkhaleq R, Sheth SA, Chen PR, Chandra A, et al. Treatment trends and overall survival in patients with grade II/III ependymoma: The role of tumor grade and location. Clinical neurology and neurosurgery. 2020;199:106282.
Swanson AA, Raghunathan A, Jenkins RB, Messing-Junger M, Pietsch T, Clarke MJ, et al. Spinal Cord Ependymomas With MYCN Amplification Show Aggressive Clinical Behavior. Journal of neuropathology and experimental neurology. 2019;78(9):791–7.
Rudà R, Gilbert M, Soffietti R. Ependymomas of the adult: molecular biology and treatment. Current opinion in neurology. 2008;21(6):754–61.
Reni M, Brandes AA, Vavassori V, Cavallo G, Casagrande F, Vastola F, et al. A multicenter study of the prognosis and treatment of adult brain ependymal tumors. Cancer. 2004;100(6):1221–9.
Rogers L, Pueschel J, Spetzler R, Shapiro W, Coons S, Thomas T, et al. Is gross-total resection sufficient treatment for posterior fossa ependymomas? Journal of neurosurgery. 2005;102(4):629–36.

No competing interests reported.

Download PDF

Reviewers invited by journal
10 Jul, 2024
Editor assigned by journal
10 Jul, 2024
Editor invited by journal
07 Jul, 2024
Submission checks completed at journal
05 Jul, 2024
First submitted to journal
02 Jul, 2024

You are reading this latest preprint version

Construction and validation a Nomogram for Predicting Cancer-Specific Survival for patients with Ependymoma: An Analysis of the Surveillance, Epidemiology, and End Results Database

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

1. Introduction

2. Materials and Methods

2.1. Data Acquisition and Processing

2.2. Patient Characteristics and Eligibility Criteria

2.3. Prognostic Factor Analysis and Model Development

2.4. Model Evaluation and Validation

2.5. Statistical Analysis

3. Results

3.1. Baseline characteristics

3.2. Establish the prognostic model

3.3. Validation of the nomogram

4. Discussion

5. Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1