Association between preoperative neutrophil-to-lymphocyte ratio and malignant intraductal papillary mucinous neoplasms: A systematic review and meta-analysis

Abstract

BACKGROUND

Intraductal papillary mucinous neoplasms (IPMN) are pancreatic cystic neoplasms with heterogeneous malignant potential. Accurate preoperative identification of malignant IPMN remains clinically challenging. The neutrophil-to-lymphocyte ratio (NLR) has been reported as a simple systemic inflammatory marker associated with tumor progression, but its value in malignant IPMN remains uncertain.

AIM

To determine the association between the preoperative NLR and malignant IPMN.

METHODS

A systematic literature search of Ovid MEDLINE, Web of Science, EMBASE, and the Cochrane Library was performed for studies published before August 30, 2024. Studies evaluating the association between the NLR and malignant IPMN were included. Two reviewers independently conducted study selection, data extraction, and quality assessment using the Newcastle-Ottawa Scale. A random-effects meta-analysis was performed. Subgroup and leave-one-out sensitivity analyses were also conducted.

RESULTS

Ten retrospective studies involving 1355 patients were included. An elevated preoperative NLR was significantly associated with malignant IPMN (odds ratio = 2.56, 95% confidence interval: 1.24-5.27; I² = 66%). Subgroup analyses revealed variation across study characteristics, and several subgroups had wide confidence intervals. Leave-one-out sensitivity analysis revealed that the pooled result remained statistically significant after the sequential exclusion of each study. Funnel plot inspection did not show obvious asymmetry, although the small number of studies limited the assessment of publication bias.

CONCLUSION

An elevated preoperative NLR is associated with malignant IPMN and may serve as an adjunctive marker.

Key Words: Neutrophil-to-lymphocyte ratio; Intraductal papillary mucinous neoplasms; Malignancy; Meta-analysis; Pancreatic cystic neoplasms

Core Tip: This meta-analysis suggests that an elevated preoperative neutrophil-to-lymphocyte ratio (NLR) is associated with malignant intraductal papillary mucinous neoplasm. As an inexpensive and readily available inflammatory marker, the NLR may assist in preoperative risk stratification. However, because of study heterogeneity and the nonspecific nature of the NLR, it should be considered an adjunctive marker rather than a standalone predictor.

INTRODUCTION

Pancreatic cystic neoplasms are fluid-filled lesions of the pancreas, and their reported incidence has increased substantially with the widespread use of cross-sectional imaging[1,2]. Intraductal papillary mucinous neoplasms (IPMN) are among the most common pancreatic cystic neoplasms with malignant potential[3,4]. Appropriate preoperative risk stratification is therefore clinically important, because management ranges from surveillance for low-risk lesions to surgical resection for lesions with a high likelihood of high-grade dysplasia or invasive carcinoma.

The neutrophil-to-lymphocyte ratio (NLR) is a readily available marker of systemic inflammation derived from routine blood testing. It has been investigated as a prognostic or predictive biomarker in a variety of malignancies, reflecting the broader link between inflammation and tumor progression[5]. In the context of IPMN, several studies have explored whether an elevated preoperative NLR is associated with malignant transformation, but the reported findings have been inconsistent. Some studies have suggested that the NLR may help identify malignant IPMN[6-8], whereas others have reported limited diagnostic value[9,10]. A systematic synthesis of the available evidence is therefore warranted. The aim of this study was to systematically evaluate the association between the preoperative NLR and malignant IPMN through a systematic review and meta-analysis.

MATERIALS AND METHODS

This systematic review and meta-analysis study was conducted in accordance with the PRISMA statement[11].

Search strategy

In the present study, a comprehensive literature search was performed across Ovid MEDLINE, Web of Science, EMBASE, and Cochrane Library databases for studies published up to August 30, 2024. The search terms included (“Neutrophil-to-Lymphocyte Ratio” OR “NLR”) AND (“pancreatic cyst” OR “pancreatic cystic neoplasm” OR “pancreatic cystic tumors” OR “intraductal papillary mucinous neoplasm” OR “IPMN”).

Eligibility criteria

Studies were included if they met all the following criteria: (1) Patients were diagnosed with IPMN; (2) Preoperative NLR values were reported; and (3) The study examined the relationship between the preoperative NLR and malignant IPMN. Exclusion criteria were as follows: (1) Editorials, letters, abstracts, case reports, and reviews; (2) Studies that did not provide separate data for IPMN; (3) Full-text articles not published in English; (4) Studies not reporting NLR cutoff values; and (5) Studies not reporting relevant outcomes.

Data extraction

The relevant data were extracted independently by two reviewers (Dong WW and Zhou LL) using a predefined standardized extraction form. The reviewers were not blinded to the study authors, institutions, or journal sources. To minimize subjective bias, predefined eligibility criteria were applied throughout the review process, and discrepancies were resolved through discussion and consensus, with consultation from an additional reviewer when necessary. Extracted data included the first author, publication year, country, study design, years of data collection, sample size, gender, age, benign group definition, benign group number, malignant group definition, malignant group number, NLR cutoff values, odds ratios (ORs) with 95% confidence intervals (CI), and type of analysis. Multivariate analyses were preferred when available. All ORs were standardized to reflect comparisons of high NLR vs low NLR.

Quality assessment

Methodological quality was independently assessed by two reviewers (Dong WW and Zhou LL) using the Newcastle-Ottawa Scale (NOS), which evaluates selection, comparability, and outcome domains. Studies with NOS scores of 6 or higher were considered high quality. The reviewers were not blinded to the study identifiers. Any disagreements in scoring were resolved by discussion and consensus.

Statistical analysis

Meta-analysis was performed using R version 4.3. Continuous and categorical variables were analyzed appropriately. Heterogeneity was assessed using the χ² test with a significance threshold of P < 0.1, and the I² statistic quantified heterogeneity (≥ 50% indicated significant heterogeneity)[12]. A fixed-effects model was used initially, but a random-effects model was applied in the presence of significant heterogeneity[13]. Descriptive methods were used when data were unsuitable for meta-analysis. Subgroup analyses were conducted according to country, NLR cut-off value, sample size, and definition of malignancy to explore potential sources of heterogeneity. Leave-one-out sensitivity analysis was performed by sequentially omitting each study to evaluate the robustness of the pooled estimate. Funnel plot inspection was used to assess potential publication bias, although this assessment was interpreted cautiously because of the limited number of included studies[14].

RESULTS

Results of the literature review

A total of 111 articles were retrieved, with 38 duplicates removed, leaving 73 articles. After screening the titles and abstracts, 42 articles were excluded. Full-text assessments were performed on the remaining 31 articles, excluding 7 studies that did not describe IPMN patients separately[15-21], 6 abstracts[22-27], 3 reviews[28-30], 2 non-English articles[31,32], 2 studies without NLR cut-off values[33,34], and 1 study that did not report the NLR-malignancy relationship[35]. Ultimately, 10 studies with 1355 patients were included in the meta-analysis[6-10,36-40]. The study selection process is shown in Figure 1.

Figure 1  PRISMA flow diagram of eligible studies selection.

Characteristic of the included studies

Table 1 shows the characteristics of included studies. Studies were published between 2015 and 2022, and all were retrospective in design. Seven studies were conducted in Japan[6,8-10,36-38], one in China[7], one in Italy[39], and one in the United States[40]. The study periods ranged from 1995 to 2020, and the sample sizes ranged from 50 to 333 patients. Overall, the pooled sample comprised 821 men (60.6%) and 534 women (39.4%), including 785 patients in the benign group and 570 in the malignant group. The reported NLR cut-off values ranged from 2.0 to 4.0. Seven studies reported multivariate analyses, and all the studies were of high methodological quality according to the NOS score.

Table 1 Characteristics of included studies.

Ref.	Country	Study type	Years of data collection	Sample size	Gener (M/F)	Age	Benign group	Number	Malignant group	Number	NLR cutoff value	OR (95%CI)	NOS
Arima et al[36], 2015	Japan	Retro	December 1995 to February 2015	80	51/25	67.9	IPMA	52	IPMC	28	2.074	3.439 (1.10-11.8)	6
Gemenetzis et al[40], 2017	United States	Retro	January 1997 to July 2015	272	138/134	68 (61-75)	Benign + HGD	201	IC	71	4	6.62 (2.23-20.9)	7
Hata et al[37], 2019	Japan	Retro	January 1996 to December 2017	205	146/80	68 (34-85)	LGD	69	HGD + IC	136	2.5	2.905 (1.266-6.667)	7
Ohno et al[38], 2019	Japan	Retro	May 1996 to December 2017	56	33/23	71 ± 10/74 ± 8	IPMA	21	IPMC	35	2.2	9.79 (2.06-45.6)	6
Yamashita et al[6], 2019	Japan	Retro	January 1996 to December 2016	79	55/24	67 ± 8/70 ± 9	IPMA	49	IPMC	30	2.63	14.3 (1.01-211.71)	7
Li et al[7], 2021	China	Retro	November 2010 to February 2018	333	210/123	63 (19-83)	LGD	198	HGD + IC	135	2	2.487 (1.310-4.722)	6
Serafini et al[39], 2021	Italy	Retro	January 2005 to December 2015	83	45/38	69 (62-76)	LGD + IGD	37	HGD + IC	46	2.38	0.779 (0.327-1.855)	8
Sugimachi et al[8], 2021	Japan	Retro	2005-2020	50	24/26	68.0 ± 1.6/73.9 ± 2.3/72.7 ± 2.1	LGD + IGD	23	HGD + IC	27	3.1	7.16 (1.23-41.6)	6
Suzuki et al[9], 2021	Japan	Retro	2007-2018	121	73/48	68.7 (44-86)	LGD + HGD	100	IC	21	2.42	1.444 (0.559-3.732)	7
Maruyama et al[10], 2022	Japan	Retro	January 2013 to October 2018	76	46/30	73 (69-76)	LGD	35	HGD + IC	41	3.27	0.55 (0.21-1.45)	7

Retro: Retrospective study; M: Male; F: Female; NLR: Neutrophil-to-lymphocyte ratio; OR: Odds ratios; CI: Confidence interval; NOS: Newcastle-Ottawa Quality Assessment Scale; IPMA: Intraductal papillary mucinous adenoma; HGD: High-grade dysplasia; LGD: Low-grade dysplasia; IGD: Intermediate-grade dysplasia; IPMC: Intraductal papillary mucinous carcinoma; IC: Invasive carcinoma.

Associations between the NLR and malignant IPMN

Pooled analysis revealed that an elevated preoperative NLR was significantly associated with malignant IPMN (OR = 2.56, 95%CI: 1.24-5.27; I² = 66%; P < 0.01) (Figure 2). Seven studies reported that the NLR was an independent risk factor associated with malignant IPMN.

Figure 2 Forest plot of the association between elevated neutrophil-to-lymphocyte ratio and malignant intraductal papillary mucinous neoplasms. OR: Odds ratio; CI: Confidence interval; SE: Standard error; HK: Hartung-Knapp.

Subgroup analysis

Because substantial heterogeneity was observed in the primary analysis, subgroup analyses were performed (Figure 3). In the Japan subgroup, the pooled OR was 2.82 (95%CI: 1.03-7.72; I² = 64%; P < 0.01), whereas in the non-Japan subgroup the pooled OR was 2.26 (95%CI: 0.17-30.25; I² = 79%; P < 0.01). When stratified by the NLR cut-off value, studies using a cut-off < 3 had an OR of 2.43 (95%CI: 1.12-5.28; I² = 53%; P < 0.01), whereas studies using a cut-off ≥ 3 had an OR of 2.74 (95%CI: 0.07-110.19; I² = 85%; P < 0.01). When malignancy was defined as invasive carcinoma alone, the pooled OR was 4.22 (95%CI: 1.42-12.55; I² = 45%; P < 0.12); when malignancy included high-grade dysplasia and invasive carcinoma, the pooled OR was 1.68 (95%CI: 0.51-5.52; I² = 71%; P < 0.01). In subgroup analyses by sample size, studies with fewer than 100 patients had an OR of 2.66 (95%CI: 0.65-10.97; I² = 74%; P < 0.01). However, studies with 100 or more patients had an OR of 2.67 (95%CI: 1.20-5.94; I² = 29%; P < 0.24). Overall, subgroup effect estimates varied across study characteristics, and several subgroups yielded wide confidence intervals, indicating limited precision. These subgroup findings should therefore be considered exploratory rather than definitive.

Figure 3 Subgroup analysis of the association between elevated neutrophil-to-lymphocyte ratio and malignant intraductal papillary mucinous neoplasms. A: Japan subgroup; B: Non-Japan subgroup; C: Neutrophil-to-lymphocyte ratio cut-off < 3 subgroup; D: Neutrophil-to-lymphocyte ratio cut-off ≥ 3 subgroup; E: Malignancy defined as invasive carcinoma alone; F: Malignancy defined as high-grade dysplasia and/or invasive carcinoma; G: Sample size < 100 subgroup; H: Sample size ≥ 100 subgroup. OR: Odds ratio; CI: Confidence interval; SE: Standard error; HK: Hartung-Knapp.

Sensitivity analysis

A leave-one-out sensitivity analysis was performed by sequentially omitting each included study and recalculating the pooled effect estimate (Figure 4). The association between an elevated preoperative NLR and malignant IPMN remained statistically significant across all iterations. Compared with the overall pooled OR of 2.56 (95%CI: 1.24-5.27), the pooled OR varied modestly from 2.25 (95%CI: 1.08-4.71) to 2.99 (95%CI: 1.54-5.80). These findings suggest that no individual study had a disproportionate influence on the overall results.

Figure 4 Leave-one-out sensitivity analysis of the association between elevated neutrophil-to-lymphocyte ratio and malignant intraductal papillary mucinous neoplasms. OR: Odds ratio; CI: Confidence interval; HK: Hartung-Knapp.

Publication bias

Funnel plot inspection did not show obvious asymmetry (Figure 5); however, because only ten studies were included, the reliability of publication bias assessment was limited.

Figure 5  Funnel plot for assessment of publication bias.

DISCUSSION

In this systematic review and meta-analysis, an elevated preoperative NLR was significantly associated with malignant IPMN. These findings suggest that systemic inflammatory status may be linked to the malignant potential of IPMN. However, the current evidence should be interpreted as demonstrating an association rather than definitive predictive performance, because all included studies were retrospective and substantial heterogeneity was present.

The observed association is biologically plausible. Systemic inflammation has long been linked to tumor initiation, progression, and invasion[41,42]. The NLR reflects the balance between neutrophil-mediated tumor-promoting inflammation and lymphocyte-mediated antitumor immune surveillance. In this context, an elevated NLR may capture a host inflammatory milieu that is more permissive for malignant transformation or more closely associated with invasive biological behavior.

Patients with IPMN often have no obvious symptoms, and many lesions are detected incidentally[43]. Surveillance is recommended for patients without high-risk features in several guidelines[40,44,45]. Imaging assessments and serum tumor markers such as carbohydrate antigen 19-9 are commonly used during follow-up[46]. Previous studies on IPMN have focused on imaging characteristics and cyst fluid analysis for predicting malignancy, but these methods have limitations, including variable diagnostic accuracy and invasiveness[47]. From a clinical perspective, the NLR is attractive because it is inexpensive, widely available, and routinely measured in preoperative blood tests. This practicality makes NLR a potentially useful adjunctive marker for preoperative risk stratification in patients with IPMN, particularly when considered alongside established clinical, radiological, and biochemical variables. Nevertheless, our findings do not support the use of the NLR as a standalone diagnostic or decision-making tool. The current evidence is insufficient to conclude that the NLR independently improves management beyond existing approaches such as imaging features, carbohydrate antigen 19-9, and cyst fluid or molecular analyses.

The moderate-to-high heterogeneity observed in this meta-analysis is likely multifactorial. First, the NLR cut-off values varied substantially across studies, ranging from 2.0 to 4.0, which may have affected the comparability and pooled effect estimates. Second, the definition of malignant IPMN was not uniform; some studies considered invasive carcinoma alone, whereas others combined high-grade dysplasia with invasive carcinoma. Third, most studies were conducted in Japan, which may limit the generalizability of the findings to other populations. Finally, all included studies used retrospective designs, which may have introduced selection bias and inconsistencies in case definitions or data collection. These factors likely contributed to the wide confidence intervals observed in several subgroup analyses.

Although several subgroup estimates were imprecise, the leave-one-out sensitivity analysis supported the robustness of the primary finding. The sequential exclusion of each study did not markedly alter the overall association, and statistical significance was maintained throughout every iteration. This suggests that the pooled result was not driven by any single study despite the observed heterogeneity.

Inflammation is linked to tumor development, and systemic inflammatory markers have been widely investigated in cancer progression[48]. Importantly, the NLR is a nonspecific marker of systemic inflammation and may be influenced by a range of malignant and nonmalignant conditions. An elevated NLR has been associated with a worse prognosis in cancers such as colorectal, lung, and gastric cancers[49-51]. Therefore, although our findings support an association between an elevated NLR and malignant IPMN, its specificity for this disease process is limited. Its potential clinical value likely lies in incorporation into multimodal predictive models rather than isolated interpretation.

Several limitations should be acknowledged. First, all included studies were retrospective, which limits causal inference and may introduce selection bias. Second, although ten studies were included, the total number of studies remained relatively small, particularly for subgroup analyses, resulting in limited precision in some comparisons. Third, substantial heterogeneity was present across studies, likely related to differences in NLR cut-off values, definitions of malignancy, and study populations. Fourth, the NLR is a nonspecific inflammatory marker and may be affected by a range of clinical conditions unrelated to IPMN malignancy. Finally, the reviewers were not blinded to study identifiers during study selection, data extraction, or quality assessment, which may have introduced subjective bias despite independent review and consensus procedures.

Future research should focus on prospective validation studies using standardized NLR thresholds and harmonized definitions of malignant IPMN. It will also be important to evaluate whether the NLR provides incremental predictive value when combined with established markers and imaging-based models in contemporary, multicenter cohorts.

CONCLUSION

An elevated preoperative NLR is associated with malignant IPMN. However, given the heterogeneity across studies, the retrospective nature of the available evidence, and the nonspecific nature of the NLR, it should be considered an adjunctive marker rather than a standalone predictor. Future prospective studies using standardized NLR thresholds and multimodal predictive models are needed to clarify its clinical utility.