Development and external validation of a nomogram for predicting the survival outcomes of patients with Ki-67 positive grade 4 diffuse gliomas

doi:10.21203/rs.3.rs-2595501/v1

Download PDF

Article

Development and external validation of a nomogram for predicting the survival outcomes of patients with Ki-67 positive grade 4 diffuse gliomas

https://doi.org/10.21203/rs.3.rs-2595501/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Grade 4 diffuse gliomas is a highly malignant tumor with considerable health implications. Studies have investigated the immunohistochemical molecules associated with glioblastoma development. However, grade 4 diffuse gliomas in Ki-67 positive patients have not been conclusively investigated.

Methods

We retrospectively extracted data for 146 patients with Ki-67 positive grade 4 diffuse gliomas at the affiliated hospital of Qingdao University between 2020 and 2021. The data were analyzed using the R software. Statistically significant indicators were identified by COX regression analysis and used to establish the Nomogram. The nomogram was corrected by C-index, area under the curve (AUC), calibration curve and decision curve analyses (DCA). Finally, the model was externally validated using the Chinese Glioma Genome Atlas (CGGA) database. The experiment was approved by the Medical Ethics Committee of the Affiliated Hospital of Qingdao University. All methods were performed in accordance with relevant guidelines and regulations. The authors of this experiment promise that informed consent of all subjects and/or their legal guardians has been obtained for this research.

Results

Age, treatment and IDH were found to be significant. The models’ C-index was 0.743 while the area under the curve (AUC) value of the time-dependent ROC curve at 3- and 6-months were 0.832 and 0.829, respectively. These findings imply a good discriminatory ability. Finally, a nomogram was constructed and validated using validation and DCA curves.

Conclusion

Three risk factors (age, treatment and IDH) were identified to be independent prognostic factors in Ki-67 positive grade 4 diffuse gliomas patients. The model can be used to accurately assess the disease-specific survival rates of these patients and inform on treatment options.

Biological sciences/Cancer

Biological sciences/Neuroscience

grade 4 diffuse gliomas

Ki-67 positive

nomogram

survival

the time-dependent receiver operating characteristic (ROC)

Among all primary brain and central nervous system tumors, glioblastoma(GBM)is the most common primary malignant brain tumor, accounting for about 16%¹. The average age-adjusted incidence of GBM was 3.2 per 100,000^2,3. Although GBMs are almost exclusively found in the brain, they can also be found in the brainstem, cerebellum, and spinal cord. About 61% of primary gliomas occur in the four lobes of the brain: frontal (25%), temporal (20%), parietal (13%), and occipital (3%)⁴. Originally, GBM was thought to only originate from glial cells. However, there is evidence that it can develop from various cell types with neural stem cell-like properties. These cells are at multiple stages of differentiation, from stem cells to neurons to glia, and their phenotypic variations are largely determined by molecular alterations in signaling pathways rather than differences in cell type and origin⁵. Until 2021, WHO pointed out that grade 4 diffuse gliomas included the former IDH wild-type glioblastoma and grade 4 IDH mutant astrocytoma⁶.

Patients often present with symptoms of increased intracranial pressure, including headache and focal or progressive neurological deficits. Seizures are present in up to 25% of patients and may appear later in the disease in up to 50% of patients⁷.

Treatment of newly diagnosed grade 4 diffuse gliomas requires a multidisciplinary approach. The current standard of care consists of maximal safe surgical resection, followed by concurrent radiation therapy with temozolomide (TMZ) (Temodar®), an oral alkylating chemotherapeutic agent, followed by adjuvant chemotherapy with TMZ⁸. Since GBMs are often aggressive and are located in functional areas of the brain, including areas that control language, motor function, and sensation, extensive and complete surgical removal is difficult. Due to their highly invasive nature, radical resection of the primary tumor mass is not efficient because infiltrating tumor cells always remain in the surrounding brain, leading to later disease progression or recurrence⁹.

Ki-67 is a non-histone nuclear protein that is associated with ribosomal RNA (rRNA) transcription and is expressed when cells enter the mitotic cycle. In glioma, Ki-67 expressions are positively correlated with histological grade. Correlations between Ki-67 and the poor prognosis of different types of malignancies have been reported¹⁰.

Studies have investigated on the outcomes of grade 4 diffuse gliomas with Ki-67 positive, however, we were the first to establish a model of grade 4 diffuse gliomas in Ki-67 positive patients.

In this study, data for 146 cases of grade 4 diffuse gliomas in the brain were retrospectively extracted from the affiliated hospital of Qingdao University between 2020 and 2021. Variables that were independent prognostic factors were established using R language. Prognostic outcomes of patients were predicted using the nomogram. Finally, the model was calibrated using the calibration and DCA curves. Validation of the model was done using the Chinese Glioma Genome Atlas (CGGA) database.

Data sources

Data for grade 4 diffuse gliomas patients between 2020 and 2021 were retrospectively obtained from the affiliated hospital of Qingdao University. Chinese glioma Genome Atlas ( http://www.cgga.org.cn/ ) were used for external validation of the model.

Inclusion And Exclusion Criteria

The inclusion criteria for this study were: i. Ki-67 positive grade 4 diffuse gliomas patients; ii. A positive histopathological diagnosis for grade 4 diffuse gliomas performed between 2020 and 2021 at the pathology department of the affiliated hospital of Qingdao University; iii. Cases with all treatment information; iv. Cases with follow-up dates; v. The behavior recode was malignant; vi. Patients who underwent gross total resection. The exclusion criteria were: i. Patients who were lost to follow-up and ii. Patients whose tumor sizes were unknown (Fig. 1).

Data Variables

The collected data included age, sex, tumor size, treatment (including chemotherapy after surgery; surgery; radiation after surgery; radiation and chemotherapy after surgery), location (frontal lobe and other location), S100, P53, ATRX, Syn and IDH. All cases had well defined return dates.

Statistical analysis

Univariate Cox regression analysis was performed to identify the possible prognostic risk factors at a threshold of p ≤ 0.05. A prediction model was created based on all risk factors with p ≤ 0.05. The Akaike information criterion was used to select the best prediction model¹¹, which was visualized on a nomogram. Concordance index (C-index) and time-dependent receiver operating characteristic (ROC) curves were used to assess precision and discrimination abilities of the model^12,13. Comparisons between the actual probability and predicted outcomes of the model were visualized using calibration curves¹⁴. The 100 re-sampling method was used to evaluate the calibration curves and discrimination. The X-tile software was used to establish the best cutoff value. Statistical analyses were performed using R version 4.2.1 (http://www.r-project.org/).

Clinic-pathological characteristics of patients

A total of 146 patients with Ki-67 positive grade 4 diffuse gliomas were enrolled in this study (Table 1). The female to male ratio was 1.39:1. The ratio of patients aged above and below 66 years old was 1:1.863. The ratio of tumors that were above and below 37.5 mm in size was 2.318:1. The ratio of grade 4 diffuse gliomas in the frontal lobe to other locations was 1:2.537. The ratio of IDH negative to positive grade 4 diffuse gliomas was 7.588:1. All patients were diagnosed by positive histology.

Table 1

Baseline clinicopathological features and treatments for all patients.
Variables	Number of patients
Age	146
< 66	95
>=66	51
Sex	146
Female	61
Male	85
Size	146
< 37.5	44
>=37.5	102
Location	146
Frontal lobe	41
Other	105
Treatment	146
Chemotherapy after surgery	4
Radiotherapy after surgery	8
Surgery	62
Radiotherapy and Chemotherapy after surgery	72
S100	146
(-)	11
(+)	135
P53	146
(-)	16
(+)	130
ATRX	146
(-)	25
(+)	121
Syn	146
(-)	26
(+)	120
IDH	146
(-)	129
(+)	17

Univariate Analysis

Cox univariate analysis revealed that age, treatment and IDH were significant (Fig. 2). Other factors (p < 0.2) were included in the multivariate analysis.

Independent Prognostic Factors

Findings from the multivariate analysis were shown in Fig. 3. Age, treatment and IDH were found to be significant. Therefore, multivariate analysis included treatment [chemotherapy after surgery: hazard ratio (HR, 95% CI) = 0.2341 (0.05517–0.9935), p = 0.049; radiotherapy after surgery: hazard ratio (HR, 95% CI) = 0.5447 (0.22320–1.3294), p = 0.182; radiation and chemotherapy after surgery; hazard ratio (HR, 95% CI) = 0.2862 (0.17888–0.4579), p < 0.01], Age [ > = 66: hazard ratio (HR, 95% CI) = 2.4115 (1.55357–3.7433), p < 0.01] and IDH [positive: hazard ratio (HR, 95% CI) = 0.3892 (0.16185–0.9361), p = 0.0351]. The effects of all dependent prognostic factors on patient survival were presented by the Kaplan-Meier curve (Fig. 4).

Prognostic Nomogram For Overall Survival

The independent prognostic factors were shown in the nomogram (Fig. 5). In the nomogram, each variable was matched to a score on the scale, and the scores of individual variables added to obtain the total score. Finally, the survival rate was determined by corresponding the total score to the score scale. The largest and smallest proportions were treatment and age, respectively.

Internal validation of the nomogram

The C index for the cohort was 0.743. The time-dependent ROC curve showed that the area under the curve (AUC) values at 3- and 6-months were 0.832 and 0.829, respectively, implying good discriminatory abilities (Fig. 6). Calibration curves were used to indicate the comparisons between actual probability and predicted outcomes (Fig. 7). Figure 8 showed the decision curve of the cohort to predict the survival probability at 3 and 6 months. The blue line indicates the net benefit when all patients die, while the green line indicates the net benefit when none of the patients dies. When the curve of the model is between the blue and green lines, it indicates that the model is of great clinical value. We established that the prediction model has great net benefits for almost all threshold probabilities at 3-month (Fig. 8A) and 6-month (Fig. 8B).

External validation of the nomogram

In the CGGA database, the time-dependent ROC curve (Fig. 9) showed that the area under the curve (AUC) of 3 months and 6 months were 0.739 and 0.74, respectively. The effects of all dependent prognostic factors on patient survival were presented by the Kaplan-Meier curve (Fig. 10).

Grade 4 diffuse gliomas is a highly malignant tumor. Studies have shown that Ki-67 is a key molecule affecting the prognosis of grade 4 diffuse gliomas, so it is important to identify the risk factors that affect the disease prognosis in grade 4 diffuse gliomas patients with Ki-67 positive. For this study, high-quality clinical data is necessary. However, it is difficult to study grade 4 diffuse gliomas prospectively because the disease is initially difficult to diagnose. Therefore, we retrospectively extracted data on 146 cases of grade 4 diffuse gliomas with Ki-67 positive from the affiliated hospital of Qingdao university. The results of the multivariate analysis revealed that IDH, treatment and age were statistical significance factors in this study.

The clinical features and prognostic models for grade 4 diffuse gliomas with Ki-67 positive have not been conclusively established. The nomogram has been reported to have better predictive values than the traditional model^15,16. Therefore, we constructed a nomogram using data for 146 grade 4 diffuse gliomas patients with Ki-67 positive. The AUC, C-index, calibration curves and DCA curves were used to assess the discriminatory ability and accuracy of the model. Finally, the nomogram was externally validated using grade 4 diffuse gliomas data^17–19 from the CGGA database.

Age, an important prognostic factor in cancer²⁰, is also considered to be an independent predictor of glioblastoma prognosis²¹. In this study, multivariate Cox analysis showed that age is a significant prognostic factor. The HR for age ≥ 66 was 2.4115 compared to < 66 (p < 0.01). This is consistent with previous findings.

Tumor size was not a significant prognostic factor in multivariate Cox analysis. The HR for tumor size ≥ 37.5 mm was 1.1347 compared to < 37.5 mm (p = 0.5979). Some studies reported that tumor size was an independent predictor of prognosis in glioblastoma patients²². These differences in outcomes may be attributed to the small sample size.

In this study, multivariate Cox analysis showed that IDH was a significant prognostic factor. The HR for IDH (+) was 0.3892, compared to IDH (-) (p = 0.0351), consistent with previous findings. It has been reported that patients with IDH mutations have better prognostic outcomes²³. Specific point mutations in genes encoding isocitrate dehydrogenase (IDH) 1 or 2 are associated with favorable outcomes²⁴.

Results of multivariate analysis showed treatment to be statistically significant [Chemotherapy after surgery: hazard ratio (HR, 95% CI) = 0.2341 (0.05517–0.9935), p = 0.049; Radiotherapy after surgery: hazard ratio (HR, 95% CI) = 0.5447 (0.22320–1.3294), p = 0.182; Radiation and Chemotherapy after surgery; hazard ratio (HR, 95% CI) = 0.2862 (0.17888–0.4579), p < 0.01]. The treatment of newly diagnosed GBM requires a multidisciplinary approach. The current standard of care consists of maximal safe surgical resection followed by concurrent radiation therapy with temozolomide (TMZ) (Temodar®), an oral alkylating chemotherapy agent, followed by adjuvant chemotherapy with TMZ⁸. Extensive and complete surgical removal of grade 4 diffuse gliomas with Ki-67 positive is difficult because these tumors are often aggressive and are often located in functional areas of the brain, including areas that control language, motor function, and sensation.

This study is the first study to construct a nomogram in Ki-67 positive patients with grade 4 diffuse gliomas. Most importantly, nomogram helped clinicians predict the prognosis of patients and provide treatment plans. Therefore, this research had certain clinical application value.

Certainly, this study had certain limitations. Firstly, the time span of the included population was short. Finally, selective bias was inevitable for retrospective study. But this research also had many benefits.

We established that IDH, treatment and age are independent prognostic factors in Ki-67 positive grade 4 diffuse gliomas patients. The developed nomogram can predict patient's prognostic outcomes and inform on appropriate treatment measures.

GBM: Glioblastoma

C-index: concordance index

ROC: the time-dependent receiver operating characteristic

AUC: area under the curve

DCA: decision curves analysis

HR: hazard ratio

CGGA: Chinese Glioma Genome Atlas

RCas: Radiation and Chemotherapy after surgery

Ras: Radiation after surgery

Cas: Chemotherapy after surgery

WHO: World Health Organization

Authors' contributions

Liu H: the conception and design of the study.

Lun P: acquisition of data.

Zhao J: acquisition of data.

Wang F: acquisition of data.

Cai X: analysis and interpretation of data.

Sun P: drafting the article or revising it critically for important intellectual content and final approval of the version to be submitted.

Declaration of competing interest

All authors have no financial and personal relationships with other people or organizations that could inappropriately influence (bias) their work. Declarations of interest: none.

Data availability

Data will be made available on request. If you would like data from this study, please contact Hui Liu [email protected].

Ethics approval and consent to participate

The project was ethically approved by Medical Ethics Committee of Affiliated Hospital of Qingdao University.All participants were informed.

Consent for publication

This article agrees to be published.

Funding

This paper is supported by the project of Qingdao Science and Technology Demonstration and Guidance Project, the number of which is 22-3-7-smjk-2-nsh.

Acknowledgements

Thanks to everyone who contributed to this article.

Thakkar, J. P. et al. Epidemiologic and molecular prognostic review of glioblastoma. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 23, 1985-1996 (2014).
Ostrom, Q. T. et al. CBTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in the United States in 2008-2012. Neuro-oncology 17 Suppl 4, iv1-iv62 (2015).
Ostrom, Q. T. et al. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2007-2011. Neuro-oncology 16 Suppl 4, iv1-63 (2014).
Blissitt, P. A. Clinical practice guideline series update: care of the adult patient with a brain tumor. The Journal of neuroscience nursing : journal of the American Association of Neuroscience Nurses 46, 367-368 (2014).
Phillips, H. S. et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer cell 9, 157-173 (2006).
Louis, D. N. et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro-oncology 23, 1231-1251 (2021).
Perry, J. et al. The use of prophylactic anticonvulsants in patients with brain tumours-a systematic review. Current oncology (Toronto, Ont.) 13, 222-229 (2006).
Nabors, L. B. et al. Central Nervous System Cancers, Version 1.2015. Journal of the National Comprehensive Cancer Network : JNCCN 13, 1191-1202 (2015).
Wilson, T. A., Karajannis, M. A. & Harter, D. H. Glioblastoma multiforme: State of the art and future therapeutics. Surgical neurology international 5, 64 (2014).
Alkhaibary, A., Alassiri, A. H., AlSufiani, F. & Alharbi, M. A. Ki-67 labeling index in glioblastoma; does it really matter? Hematology/oncology and stem cell therapy 12, 82-88 (2019).
Harrell, F. E., Jr., Lee, K. L. & Mark, D. B. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in medicine 15, 361-387 (1996).
Hanley, J. A. & McNeil, B. J. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143, 29-36 (1982).
Wolbers, M., Koller, M. T., Witteman, J. C. & Steyerberg, E. W. Prognostic models with competing risks: methods and application to coronary risk prediction. Epidemiology (Cambridge, Mass.) 20, 555-561 (2009).
Coutant, C. et al. Comparison of models to predict nonsentinel lymph node status in breast cancer patients with metastatic sentinel lymph nodes: a prospective multicenter study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 27, 2800-2808 (2009).
Chen, S. et al. Development and Validation of a Nomogram for Predicting Survival in Male Patients With Breast Cancer. Frontiers in oncology 9, 361 (2019).
Martini, A. et al. A Nomogram to Predict Significant Estimated Glomerular Filtration Rate Reduction After Robotic Partial Nephrectomy. European urology 74, 833-839 (2018).
Zhao, Z. et al. Chinese Glioma Genome Atlas (CGGA): A Comprehensive Resource with Functional Genomic Data from Chinese Glioma Patients. Genomics, proteomics & bioinformatics 19, 1-12 (2021).
Wang, Y. et al. Localizing seizure-susceptible brain regions associated with low-grade gliomas using voxel-based lesion-symptom mapping. Neuro-oncology 17, 282-288 (2015).
Liu, X. et al. A radiomic signature as a non-invasive predictor of progression-free survival in patients with lower-grade gliomas. NeuroImage. Clinical 20, 1070-1077 (2018).
Schiffer, D. et al. Prognostic factors in oligodendroglioma. The Canadian journal of neurological sciences. Le journal canadien des sciences neurologiques 24, 313-319 (1997).
Wang, Z. et al. Development and validation of a nomogram with an autophagy-related gene signature for predicting survival in patients with glioblastoma. Aging 11, 12246-12269 (2019).
Liu, Z. Y. et al. Competing risk model to determine the prognostic factors and treatment strategies for elderly patients with glioblastoma. Scientific reports 11, 9321 (2021).
Olar, A. et al. IDH mutation status and role of WHO grade and mitotic index in overall survival in grade II-III diffuse gliomas. Acta neuropathologica 129, 585-596 (2015).
Wirsching, H. G., Galanis, E. & Weller, M. Glioblastoma. Handbook of clinical neurology 134, 381-397 (2016).

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Development and external validation of a nomogram for predicting the survival outcomes of patients with Ki-67 positive grade 4 diffuse gliomas

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Introduction

Materials And Methods

Data sources

Inclusion And Exclusion Criteria

Data Variables

Statistical analysis

Results

Clinic-pathological characteristics of patients

Univariate Analysis

Independent Prognostic Factors

Prognostic Nomogram For Overall Survival

Internal validation of the nomogram

External validation of the nomogram

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1