PROGNOSTIC INFLAMMATORY INDEX PREDICTS THE PROGNOSIS OF PATIENTS WITH COLORECTAL CANCER: BASED ON A LARGE CHINESE RETROSPECTIVE COHORT

Jinming Fu; Lei Zhang; Dapeng Li; Hao Huang; Tian Tian; Yupeng Liu; Ying Liu; Yuanyuan Zhang; Jing Xu; Shuhan Meng; Chenyang Jia; Simin Sun; Xue Li; Liyuan Zhao; Ding Zhang; Lixin Kang; Lijing Gao; Ting Zheng; Yanlong Liu; Yashuang Zhao

PROGNOSTIC INFLAMMATORY INDEX PREDICTS THE PROGNOSIS OF PATIENTS WITH COLORECTAL CANCER: BASED ON A LARGE CHINESE RETROSPECTIVE COHORT

Authors: Fu J.¹, Zhang L.¹, Li D.¹, Huang H.¹, Tian T.¹, Liu Y.¹, Liu Y.¹, Zhang Y.¹, Xu J.¹, Meng S.¹, Jia C.¹, Sun S.¹, Li X.¹, Zhao L.¹, Zhang D.¹, Kang L.¹, Gao L.¹, Zheng T.¹, Liu Y.², Zhao Y.¹
Affiliations:
1. Department of Epidemiology, College of Public Health, Harbin Medical University
2. Department of Colorectal Surgery, Harbin Medical University Cancer Hospital, Harbin Medical University
Issue: Vol 10 (2021): Материалы XVII Международной Бурденковской научной конференции 22-24 апреля 2021 года
Pages: 562-564
Section: Medicine without frontiers
URL: https://new.vestnik-surgery.com/index.php/2415-7805/article/view/6703

Cite item

Full Text

Abstract
Full Text
About the authors
References
Statistics

Abstract

Relevance: Host inflammation is a critical component of tumor progression and its status can be indicated by peripheral blood cell counts.

The research objective: We aimed to construct a comprehensively prognostic inflammatory index (PII) based on preoperative peripheral blood cell counts and further evaluate its prognostic value for patients with colorectal cancer (CRC).

Methods: A total of 4154 patients with stage II and III CRC were included in this study. The PII was constructed by integrating all the peripheral blood cell counts associated with prognosis. Cox analyses were performed to evaluate the association between PII and overall survival (OS) and disease-free survival (DFS) of patients with CRC.

Results: In the cohort, multivariate Cox analyses indicated that high OS-PII (>4.27) was significantly associated with worse OS (HR: 1.330, 95% CI: 1.189-1.489, p<0.001); and high DFS-PII (>4.47) was also significantly associated with worse DFS (HR: 1.366, 95% CI: 1.206-1.548, p<0.001). The nomograms achieved good accuracy in predicting both OS and DFS, with a C-index of 0.718 and 0.700. Time-dependent ROC analyses showed that both OS-PII and DFS-PII have a stable prognostic performance at various follow-up times. The prognostic value of tumor-node-metastasis staging could be enhanced by combiningby combining it with either OS-PII or DFS-PII.

Conclusion: We demonstrated that PIIs are independent prognostic predictors for CRC patients, and the nomograms based on PIIs can be recommended for personalized survival prediction of patients with CRC.

Keywords

colorectal cancer, peripheral blood cell, prognostic inflammatory index, prognosis, nomograms

Full Text

Introduction

In the world, colorectal cancer (CRC) ranks second in terms of mortality, with an estimated 881,000 deaths in 2018 [1]. It is now recognized that inflammation is a hallmark of cancer and closely related to tumor progression [2-5]. Host inflammation response can suppress the antitumor function of adaptive immunity and break the balance between the immune system and malignant tumors, thereby causing the poor prognosis of patients [6]. Here, we separately assessed the relationship between all the six types of preoperative peripheral blood cell counts and the prognosis of CRC patients and constructed a prognostic inflammatory index (PII) by integrating the blood cell counts associated with prognosis.

Research objective

A total of 4392 primary stage II and III CRC patients confirmed by pathological diagnosis were enrolled in this study.

Methods

Study population

In this single-center, large sample retrospective cohort study, patients were underwent radical resection surgery and obtained from the Third Affiliated Hospital of Harbin Medical University between January 2007 and December 2013. 238 patients who met the following exclusion criteria were excluded: patients with age less than 18 years (n=1); missing data on preoperative peripheral blood cell counts (n=38); received neoadjuvant chemotherapy or other radiotherapy/chemotherapy before surgery (n=98); and patients lost to follow-up within 3 months (n=101).

Data collection

Data of patients’ demographic and clinicopathological characteristics were obtained from retrospective medical records. Patients were followed up regularly according to NCCN guidelines. The last time of follow-up for the cohorts was January 22, 2019.

Construction of the PII

Prognostic factors of interest for constructing the PII were platelet, neutrophil, lymphocyte, monocyte, eosinophil, and basophil counts. Significant prognostic factors (p<0.10) in the univariate analyses were then entered into multivariate Cox proportional hazards models. Finally, OS-PII and DFS-PII were constructed using blood cell counts associated with OS and DFS, with weights given by the corresponding coefficients from the multivariate Cox model.

Statistical analyses

The prognostic value of clinicopathological characteristics and PIIs were estimated using univariate and multivariate Cox analyses. X-tile 3.6.1 software [7] was used to determine the optimal cut-off values for the PIIs. The nomograms for possible prognostic factors associated with OS and DFS were developed to predict the probability of 1-, 3- and 5-year survival recurrence/metastasis in for CRC patients. The prediction accuracy of nomograms was evaluated by the C-index [8]. Time-dependent receiver operating characteristic (ROC) analyses were performed to compare the prognostic abilities of the PIIs, TNM staging, their combination, and biomarkers previously reported (NLR, PLR, LMR, and SII) [9]. All statistical analyses were performed with SPSS 24.0 (SPSS Inc., Chicago, IL, USA) and R 3.6.2 software (Institute for Statistics and Mathematics, Vienna, Austria). Two sided p<0.05 was considered statistically significant.

Results and discussion

The OS-PII was constructed using platelet, lymphocyte, neutrophil, monocyte and eosinophil counts with weights given by the corresponding coefficients from the multivariate Cox model: (1.878×Platelet) + (1.370×Lymphocyte) + (0.251×Neutrophil) + (4.570×Monocyte) + (2.094×Eosinophil). The DFS-PII was constructed using platelet, neutrophil, monocyte, and eosinophil counts with weights given by the corresponding coefficients from the multivariate Cox model: (2.370×Platelet) + (0.415×Neutrophil) + (2.600×Monocyte) + (2.437×Eosinophil). X-tile 3.6.1 software was used to determine the optimal cut-off values for OS-PII and DFS-PII, which were 4.27 and 4.47, respectively. Patients were separated into low PII groups (OS-PII ≤4.27; DFS-PII ≤4.47) and high PII groups (OS-PII >4.27; DFS-PII >4.47) for further study.

The OS-PII was statistically associated with the OS of CRC (HR: 1.330, 95% CI: 1.189-1.489, p<0.001), DFS-PII was also an independent prognostic predictor for patients with CRC (HR: 1.366, 95% CI: 1.206-1.548, p<0.001) by the multivariate Cox analysis. Next, we assessed the association of OS-PII and DFS-PII with prognosis according to different TNM staging. The stratification by a combination of PIIs and TNM staging divided patients into 4 risk groups (RG): RG1 (low PIIs and stage II), RG2 (high PIIs and stage II), RG3 (low PIIs and stage III), and RG4 (high PIIs and stage III). Multivariate Cox models demonstrated that, compared with patients in the RG1 group, the prognosis of patients in the RG2, RG3, and RG4 groups became worse and worse (Figure 1; p for trend <0.001). Both OS-PII and DFS-PII had a stable prognostic performance at various follow-up times, and their AUCs tended to be higher than the biomarkers previously reported (NLR, PLR, LMR, and SII) throughout the observation period (Figure 2).

Figure 1. Risk stratification combining PIIs and TNM staging in relation to overall survival and disease-free survival of CRC in the training cohort. Kaplan-Meier curves of four risk groups for overall survival (A) and disease-free survival (B). Multivariate Cox analyses of the four risk groups for overall survival (C) and disease-free survival (D) adjusting for the significant clinicopathological factors in relation to overall survival and disease-free survival.

Figure 2. The time-dependent AUCs of PIIs, TNM staging, a combination of PIIs and TNM staging, NLR, PLR, LMR, and SII in the training cohort. Time-dependent AUCs presented the sequential trends of PIIs, TNM staging, a model of PIIs and TNM staging, NLR, PLR, LMR, and SII for overall survival prediction (A) and disease-free survival prediction (B). The horizontal axis represents the months after radical resection, and the vertical axis represents the estimated area under the ROC curves for survival at the time of interest.

The nomograms based on OS-PII and DFS-PII were generated for personalized survival prediction of CRC patients. The concordance index (C-index) for OS and DFS prediction were 0.718 (95% CI: 0.704-0.731) and 0.700 (95% CI: 0.684-0.716), respectively and similar results were observed when we used bootstrapping for internal validation (0.714 and 0.694). Compared with AJCC (American Joint Committee on Cancer) system, nomograms had higher C-index.

The CRC 5-year relative survival ranges from greater than 90% in patients with stage I to slightly greater than 10% in patients with stage IV [10]. Although TNM staging provides valuable prognostic information, the outcome of individual patients is not predicted accurately. This is a drawback for patients with stage II and III CRC in particular but also reminds us of the importance of developing well-performed markers. These significant biomarkers can help identify populations at high risk for recurrence or death in stage II and III patients. We developed the PIIs in a large sample, including patients with stage II and III CRC with sufficient follow-up. After stratification of patients by a combination of PIIs and TNM staging, we found that both OS-PII and DFS-PII had the ability to independently identify high-risk populations in the same TNM staging. Our study also has limitations that a single-center retrospective cohort of Chinese, and patients with stage II and stage III CRC were included only.

Conclusion

We constructed a novel PII by integrating all the preoperative peripheral blood cell counts associated with prognosis and systematically analyzed the role of PIIs in the prognosis of CRC. Our study demonstrates that both OS-PII and DFS-PII are independent factors for predicting the prognosis of CRC and could enhance the prognostic ability of TNM staging by combination. The nomograms based on OS and DFS can be recommended for the personalized survival prediction of patients with CRC.

References

Bray, F., J. Ferlay, I. Soerjomataram, et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin, 2018. 68(6): p. 394-424.
Balkwill, F. and A. Mantovani, Inflammation and cancer: back to Virchow? Lancet, 2001. 357(9255): p. 539-45.
Hanahan, D. and R.A. Weinberg, Hallmarks of cancer: the next generation. Cell, 2011. 144(5): p. 646-74.
Greten, F.R. and S.I. Grivennikov, Inflammation and Cancer: Triggers, Mechanisms, and Consequences. Immunity, 2019. 51(1): p. 27-41.
Trinchieri, G., Cancer and inflammation: an old intuition with rapidly evolving new concepts. Annu Rev Immunol, 2012. 30: p. 677-706.
Mantovani, A., P. Allavena, A. Sica, et al., Cancer-related inflammation. Nature, 2008. 454(7203): p. 436-44.
Camp, R.L., M. Dolled-Filhart, and D.L. Rimm, X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res, 2004. 10(21): p. 7252-9.
Balachandran, V.P., M. Gonen, J.J. Smith, et al., Nomograms in oncology: more than meets the eye. Lancet Oncol, 2015. 16(4): p. e173-80.
Kamarudin, A.N., T. Cox, and R. Kolamunnage-Dona, Time-dependent ROC curve analysis in medical research: current methods and applications. BMC Med Res Methodol, 2017. 17(1): p. 53.
Brenner, H., M. Kloor, and C.P. Pox, Colorectal cancer. Lancet, 2014. 383(9927): p. 1490-1502.

Username
Password
Remember me

Forgot password?	Register

Username
Password
Remember me

Forgot password?	Register

PROGNOSTIC INFLAMMATORY INDEX PREDICTS THE PROGNOSIS OF PATIENTS WITH COLORECTAL CANCER: BASED ON A LARGE CHINESE RETROSPECTIVE COHORT

Full Text

Abstract

Keywords

Full Text

About the authors

References

This website uses cookies