Key point:
  1. 7 papers were included in this study, including 1711 patients;
  2. The results of the analysis were: the merge OS, HR=1.70, 95%CI 1.11-2.61; PFS, HR=1.34, 95%CI 0.47-3.81;
  3. Subgroup analysis included sample size, cut-off, treatment, variable and disease stage showed that a poor OS in patients with increased PLR;
  4. In patients with laryngeal cancer, an elevated PLR means a poorer prognosis;
  5. PLR was available preoperatively and could help physicians determine the prognosis of prognosis.
Introduction
Laryngeal carcinoma was a common tumor in otolaryngology department. In China, there were 26,300 new diagnosed patients and 14,500 new deaths caused by laryngeal cancer in 2015[1]。In United States, laryngeal squamous cell carcinoma (LSCC) accounted for 13,150 new patients and 3,710 new deaths in 2017 [2]。Literature had shown that for patients with local disease, the 5-year survival rate was about 75%, patients with regional invasion was 44%, patients with metastasis was about 35%[3]。Despite improvements in diagnostic techniques and surgical procedures, the American Cancer Society(ACS) noticed that for patients with laryngeal cancer the 5-year survival rate had been declining recently [4]。
Currently, it was considered important to accurately assess the prognosis of patients before surgery because it could help to select the appropriate treatment modality and thus promote the quality of life for patients. TNM staging was considered to be the finest predictor of long-term survival[5],however it could only be obtained after surgery, preoperatively available and easily detectable markers were needed for prognosis assessment, which was conducive to better personalized treatment.
It was now known that inflammation was a characteristic of tumors and exert a critical influence in the initiation and development of tumors [6, 7]. Inflammatory cells such as neutrophils could kill cancer cells directly through ADCC, T cells were the core of anti-tumor immunity, which could recognize and kill tumor cells.
In addition, Platelets were now known to release inflammatory mediators and mitotic materials into the TME (tumor microenvironment), and absorb more of them, leaded to platelet aggregation [9]. Through TGF-β/Smad pathway and NF-κB pathway, platelets could interact with tumor cells, and induce EMT of tumor cells, thereby promoted tumor metastasis[8]. It had been reported that beside peripheral blood NLR(Neutrophil-Lymphocyte Ration), PLR could also be used to predict the prognosis of a variety of solid tumors [10, 11],however, the prognostic role of PLR in laryngeal cancer remains controversial [12, 13].
Meta-analysis was now widely recognized as a powerful statistical tool for overcoming the limitations of individual studies on different sample sizes and for deriving best evaluations. Thus, the aim of this article was to quantitatively analyze the value of preoperative PLR in the prognosis of laryngeal cancer.
2. Material and methods
2.1 Retrieval methods
A comprehensive literature search was conducted through the following databases: PubMed, Web of Science, Ovid and The Cochrane Library. Search criteria updated to 1 May 2021. The searching words were: “PLR” (e.g., “platelet-to-lymphocyte ratio”, “platelet lymphocyte ration”, “PLR”) and “Laryngeal Neoplasm” (e.g., “Laryngeal Neoplasm”, “Larynx Neoplasms”, “Larynx Neoplasm”, “Cancer of Larynx”, “Larynx Cancers”, “Laryngeal Cancer”, “Laryngeal Cancers”, “Larynx Cancer” and “Cancer of the Larynx”). The references of the retrieved articles were also examined.
2.2 The criteria of inclusion and exclusion
The criteria for inclusion was below:(1)patients with laryngeal cancer were confirmed by pathology; (2) pathologically confirmed laryngeal cancer without evidence of metastasis or recurrence; (3) determination of PLR Value by Serum Test; (4) the relevance of PLR with OS and PFS was reported; (5) A multivariable / univariate proportional risk model adjusted for major clinical factors was used for statistical analysis; (6) NOS score (Newcastle Ottawa Scale) was more than six points are considered eligible;
The criteria for exclusion were below: (1) The types of articles were abstracts, reviews, correspondence, case reports and non-clinical studies; (2) non-English articles; (3) The necessary data to calculate HR and 95% CI were not provided; (4) Data was duplicated in these articles.
2.3 Extract data and assess quality
Data from the retrieved articles were extracted and evaluated for quality by the two separate authors. In the event of disagreement, the third author participated and a consensus was reached. Each article needed to record the following content: first author, country, published time, number of patients, follow-up time, thresholds, treatment options, tumor types and HRs with 95% CI. NOS score was composed of three parts: article selection, quality comparison, and result evaluation. A score of 6 or more was considered as high-quality research. Two authors separately evaluated the quality of included articles using NOS score.
2.4 Statistical method
In accordance with Parmer et al., HR and 95%CI or their approximations were obtained from the included articles. [12, 13]. P< 0.10 or I2 >50%, indicated notable heterogeneity. When significant heterogeneity was observed, the random-effect model was used. Or else, the fix-random effect model was adopted. The reasons of heterogeneity were discussed by subgroup analysis, sensitivity analysis and meta-regression.
Results
3.1 The retrieval results
A total of 88 articles were found in the preliminary search. After thoroughly investigation of these papers, our meta-analysis eventually included 7 studies involving 1711 patients published between 2016 and 2021. Figure1 summarized the process of research selection. All studies were from China. HRs and 95%CI were directly provided in these articles. HRs in 6 articles was calculated by the multivariable and univariable analysis and HRs in 1 article via univariable analysis. 4 articles enrolled > 200 cases, another 3 included < 200 cases. The cut-off values provided in these articles were inconsistent, between 106 to 189. The cut-off value in 4 articles ≤114, the value in another 3 articles > 114. 6 articles involved all disease stages, 1 article had advanced stage, all of them reported the interaction between PLR and OS.
Figure 1: Flow Chart for Selection of Qualified Documents.