Literature search and study identification
The initial exploration of databases yielded a total of 53 entries, comprising 22 from PubMed, 1 from Cochrane, 30 from Embase and 1 from additional source. A visual representation of the selection progression is provided in Figure 1, while the search methodology and algorithm are detailed in Supplementary Table S1.
Upon the removal of duplicates, 2 studies were excluded based on title, and an additional 11 were excluded based on abstract assessments. Subsequently, 21 studies underwent a thorough review of their full texts, resulting in the exclusion of 9 study due to specific reasons as outlined in Supplementary Table S2. Ultimately, 13 studies were considered suitable for qualitative analysis and 12 studies met the criteria for quantitative analysis.(11-22, 25)
Characteristics of the included studies
Table 1 provides a succinct overview of the studies included in this meta-analysis (including one for qualitative analysis). It is noteworthy that all the studies incorporated in this analysis adopted a retrospective research design. Furthermore, all these studies were conducted in Asian countries, specifically in Japan, Taiwan, China, and Korea, with the exception of one study from Morocco. The publication dates of these studies span a decade, ranging from 2013 to 2024.
In total, the quantitative analysis involved 2037 patients recruited across studies within a recruitment timeline spanning from 2000 to 2020. The mean age of the patients varied from 43.2 to 66.5 years, with the male population constituting a proportion ranging from 0% to 97.3%. The duration of follow-up spanned from 10 to 62 months. The meta-analysis encompassed multiple cancer types, including 2 oral squamous cell carcinoma, 4 nasopharyngeal carcinoma, 1 esophageal cancer, 1 endometrioid endometrial carcinoma, 1 uterine cervix invasive squamous cell carcinoma, and 2 non-small cell lung carcinomas. A total of 8 studies reported cancer stage according to the AJCC TNM staging system, while 2 studies reported stage according to the International Federation of Gynecology and Obstetrics Staging System. All studies focused on the association between pretreatment NTR and survival outcomes. The reported aHR with 95% CIs for survival outcomes, including OS, DFS, and DMFS, were directly extracted from the included studies. (Table 2)
Quality Assessment
The present meta-analysis encompassed a comprehensive compilation of 13 studies, each undergoing meticulous scrutiny for methodological rigor utilizing the Newcastle-Ottawa scale. The results of this evaluation revealed that 10 out of the 13 eligible studies achieved scores equal to or exceeding 7 points, indicating fair quality across the majority of the studies. The remaining three studies received 5-6 points on the scale. Supplementary Figure S1 offers a more detailed visualization of the insights derived from the studies included in this analysis.
Meta-analysis results
Overall Survival
Seven studies provided aHRs relevant to OS.(13, 14, 16, 17, 19, 21, 26) The outcomes of the meta-analysis revealed a significant association, indicating that higher NTR was correlated with worse OS (aHR 2.21, 95% CI 1.63 to 2.99) in a low-heterogeneity context (I2 = 10.15%, p = 0.35; Figure 2). Although overall heterogeneity was low, subgroup analyses were still conducted to explore the impact of NTR in different tumor types. For nasopharyngeal carcinoma, NTR was significantly associated with worse OS (aHR 2.64, 95% CI 1.38 to 5.05; Figure 2) with low heterogeneity (I2 = 0%, p = 0.76; Figure 2).(13, 21) In studies focusing on patients with oral cavity SCC, NTR was significantly associated with worse OS (aHR 4.60, 95% CI 2.03 to 10.47; Figure 2) with minimal heterogeneity (I2 = 0%, p = 0.87; Figure 2).(14, 17) In each subgroup, heterogeneity was even lower, and the pooled hazard ratio remained consistent when classified by different cancer types. The results of various stratified analyses (including region, NTR cut-off values, median SUV-L, median SUV-T) for the impact of NTR on overall survival were shown in Table 3.
Disease-free survival
Incorporating data from 7 studies that offered aHRs pertinent to DFS, the outcomes of the meta-analysis unveiled a significant correlation.(11-14, 18, 19, 21) Specifically, the meta-analysis showcased that NTR was also notably linked to worse DFS (aHR 3.27, 95% CI 2.12 to 5.05; Figure 3) (I2 = 23.71%, p = 0.25; Figure 3).
Distant metastasis-free survival
Incorporating data from five studies that provided aHRs relevant to DMFS, the outcomes of the meta-analysis revealed a significant and affirmative correlation.(15, 16, 20, 21, 25) Specifically, the meta-analysis demonstrated that NTR was notably linked to worse DMFS (aHR 2.07, 95% CI 1.55 to 2.78), with minimal heterogeneity (I2 = 0%, p = 0.62; Figure 4). While the overall heterogeneity was low, subgroup analyses were conducted to explore the impact of NTR in different tumor types. In studies focusing on patients with nasopharyngeal carcinoma, NTR was significantly associated with worse DMFS (aHR 2.21, 95% CI 1.48 to 3.30; Figure 4) with minimal heterogeneity (I2 = 0%, p = 0.51; Figure 4).(15, 21, 25)
Publication bias and Sensitivity Analysis
Funnel plots and Egger’s test were conducted to examine publication bias (Figure S2-S4). Publication bias existed in the analysis of OS (p = 0.001) DFS (p= 0.01). By the Trim and Fill test, three potential missing studies were found in the OS analysis, and the recalculated pooled HR for OS was 1.95 (95% CI: 1.36–2.79); two potential missing studies were identified in the analysis of DFS and the recalculated pooled HR was 3.09 (95% CI: 2.00, 4.77) (Figure S2-S4). The results of these tests indicated that publication bias did not influence the final results of the meta-analysis. The results of the sensitivity analysis showed that removing individual studies did not affect the results of the meta-analysis during the analysis of the relationship between NTR and OS, DFS, DMFS (Figure 5), indicating the reliability of our results.