Introduction

Gastrointestinal cancer is common cancer that includes colorectal cancer (CRC) and gastric cancer (GC), esophageal cancer (EC), liver cancer, cholangiocarcinoma, and pancreatic cancer. According to the Global Cancer Epidemiological Statistics published in 2018, CRC was the fourth most common cancer and the second leading cause of cancer-related death, and GC was the third leading cause of cancer-related death [1]. Despite the continuous improvement of the economy and techniques, radical resection remains the main strategy for the treatment of gastrointestinal cancer [2,3,4,5]. However, the mortality of patients after surgery is still high, with a 5-year survival rate of nearly 50% [6,7,8]. Some studies have demonstrated that patients with low body mass index (BMI), low albumin (Alb) levels, and high inflammatory conditions have a higher risk of postoperative complications and poor survival [9,10,11,12]. More sensitive prognostic indicators are needed to instruct doctors to take measures in advance and to improve the survival and quality of life of gastrointestinal cancer patients.

The advanced lung cancer inflammation index (ALI) is a new marker. It is calculated by the patient's BMI, serum Alb level, and neutrophil-to-lymphocyte ratio (NLR) (ALI = BMI × ALB/NLR) [13]. Some studies have revealed that BMI has a prognostic value for malignant diseases [14, 15]. Lee J et al. performed a 16-study meta-analysis and revealed that overweight CRC patients had worse survival [15]. Alb is synthesized in the liver and is the main component of total serum protein in the body, which mainly reflects nutrition status [16]. Gonzalez-Trejo S reported that serum albumin was a prognostic factor for CRC [17]. NLR is the ratio of neutrophils to lymphocytes. When the body is in an inflammatory state, neutrophils are elevated, and lymphocytes are decreased [18, 19]. Moreover, high levels of neutrophils can promote tumor progression and inhibit the antitumor effects of lymphocytes [20]. Thus, the NLR was established as an inflammation indicator and could be considered a balance between pro-tumor status and antitumor status. Therefore, ALI could reflect the inflammation and nutrition state. Currently, some studies have reported that ALI can predict the survival of many cancers, including CRC [21,22,23,24,25,26,27], GC [28, 29], EC [30, 31], liver cancer [32], cholangiocarcinoma [33], pancreatic cancer [34, 35], and B-cell lymphoma [36].

Many studies have discussed the relationship between ALI and gastrointestinal cancer. Pian G, Yin CZ, and Barth DA et al. demonstrated that there was no significant difference in ALI between patients receiving CRC surgery or GC surgery [22, 28, 34]. Others thought that a low ALI level would lead to a poor prognosis for gastrointestinal cancer patients [21, 23,24,25,26,27, 29,30,31,32,33]. Therefore, we aimed to explore the exact prognostic ability of ALI for patients with gastrointestinal cancer undergoing surgery.

Methods

We conducted this current meta-analysis in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [37]. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist was submitted as a supplementary document. The registration ID of this meta-analysis on PROSPERO is CRD42022362548, and the link is https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022362548.

Literature search

We searched studies in four databases including PubMed, Embase, the Cochrane Library, and CNKI on Dec 10, 2022. ALI was searched as “Advanced lung cancer inflammation index” OR “ALI” OR “BMI x ALB / NLR” OR “BMI x serum albumin / NLR” OR “body mass index x serum albumin / neutrophil-to lymphocyte”. As for gastrointestinal cancer, the search strategy was “gastrointestinal cancer” OR “gastrointestinal neoplasms” OR “colon cancer” OR “rectal cancer” OR “colorectal cancer” OR “rectum cancer” OR “colorectal neoplasm” OR “colon neoplasm” OR “rectal neoplasm” OR “rectum neoplasm” OR “colorectal carcinoma” OR “colon carcinoma” OR “rectum carcinoma” OR “rectal carcinoma” OR “CRC” OR “gastric cancer” OR “gastric carcinoma” OR “gastric neoplasms” OR “stomach cancer” OR “stomach carcinoma” OR “stomach neoplasms” OR “liver cancer” OR “hepatocellular carcinoma cancer” OR “esophageal cancer” OR “esophageal neoplasm” OR “esophagus cancer” OR “esophagus neoplasm” OR “esophageal squamous cell carcinoma” OR “cholangiocarcinoma” OR “extrahepatic cholangiocarcinoma” OR “gallbladder cancer” OR “gallbladder neoplasms” OR “bile duct cancer” OR “bile duct neoplasms” OR “pancreatic cancer” OR “pancreatic carcinoma”. The search scope was limited to titles and abstracts. Language and study design had no limitations.

Inclusion and exclusion criteria

The inclusion criteria for our meta-analysis included: 1, Patients with gastrointestinal cancer (CRC, GC, EC, liver cancer, cholangiocarcinoma, or pancreatic cancer) who received radical or palliative intent surgery; 2, Patients were divided into the high ALI group and the low ALI group; and 3, Prognosis including OS, DFS, or CSS was reported (both effect value and survival curves were allowed). The exclusion criteria included: 1, Studies’ types were reviews, case reports, letters, conferences, comments, or preprint articles; and 2, Data was repeated or overlapped. When two studies had overlapped data, the study with a larger sample size would be included. The PICO framework was more intuitive and was shown in a supplementary document.

Study selection

According to the search strategy, two authors searched studies based on the search strategy in four databases independently. Titles and abstracts would be scanned first. Then, the full text was screened based on the inclusion and exclusion criteria. If there was a disagreement on study inclusion, a group discussion would be held to solve the resolution.

Data collection

Two authors collected studies’ characteristics and patients’ information independently, then the data were checked to reach a consistent. Characteristics of the studies included the first author, year, country, study period, sample size, the cut-off value of ALI, prognosis, included patients, follow-up, and Newcastle–Ottawa Scale (NOS) score. The prognosis included OS, DFS, and CSS. The baseline information included age, sex, carcinoembryonic antigen (CEA), carbohydrate antigen 19–9 (CA 19–9), preoperative anemia, chemotherapy, lymph node metastasis, vessel invasion, neural invasion, distant metastasis, histology, and postoperative complication.

Quality assessment

Our meta-analysis assessed the study quality in accordance with the Newcastle–Ottawa Scale (NOS) according to comparison selection, comparability between groups, and the determination of results [38]. High-quality studies have scores higher than 8 points, and medium-quality studies have scores of 7 or 8 points.

Statistical analysis

In this meta-analysis, we focused on the prognosis of gastrointestinal cancer patients. The hazard ratios (HRs) and 95% confidence intervals (CIs) of OS, DFS, or CSS were pooled for analysis. HRs from the multivariate analysis were preferred, without which univariate analysis would be replaced. I2 value and chi-squared test were used to evaluate statistical heterogeneity. According to the Cochrane handbook, the I2 < 30% was considered low heterogeneity, the I2 range from 30 to 60% was considered moderate heterogeneity, and the I2 > 60% was considered high heterogeneity. The fixed effects model was used when the I2 value ≤ 50%, and p < 0.05 was thought of as statistically significant. The random effects model was used when the I2 value > 50%, and p < 0.1 was thought of as statistically significant. Subgroup analysis was used to assess the risk of heterogeneity. As for sensitive analysis, each study was excluded at a time, and repeat meta-analyses were conducted. The funnel plot was used for assessing the publication bias. RevMan 5.3 (The Cochrane Collaboration, London, United Kingdom) was used for all data analysis.

Results

Study selection

We searched 229 studies in four databases according to the search strategy (68 studies in PubMed, 135 studies in Embase, 24 studies in the Cochrane Library, and 2 studies in CNKI). After duplicates removing, the titles and abstracts were scanned for initial selection. Then, 22 studies were left for full-text assessment. Finally, 14 studies after qualitative synthesis with sufficient data were included (Fig. 1).

Fig. 1
figure 1

Flowchart of study selection

Full size image

Baseline characteristics of studies

Fourteen studies involving 5091 patients were included in the current meta-analysis. Studies were conducted in China, Japan, Korea, and Austria ranged from 2014 to 2022. The cut-off values of ALI ranged from 18.0 to 43.5 in eleven studies, and the value in another study was 70.4. Twelve studies reported OS, eight studies reported DFS/ progression-free survival (PFS)/ relapse-free survival (RFS), and two studies reported CSS. More information, including study period, sample size, cancer type of included patients, follow-up, and NOS scores are also shown in Table 1.

Table 1 Characteristics of the studies included in the meta-analysis
Full size table

Baseline characteristics of patients

Compared to the low ALI group, the high ALI group had a lower proportion of older patients (OR = 0.74, I2 = 0%, 95% CI = 0.64 to 0.85, P < 0.0001) and males (OR = 0.49, I2 = 92%, 95% CI = 0.25 to 0.98, P = 0.04), less preoperative anemia (OR = 0.53, I2 = 0%, 95% CI = 0.36 to 0.78, P = 0.001), chemotherapy (OR = 0.75, I2 = 0%, 95% CI = 0.61 to 0.92, P = 0.006), less distant metastasis (OR = 0.42, I2 = 52%, 95% CI = 0.26 to 0.66, P = 0.0002). No significant differences were found in CEA levels, CA 19–9, lymph node metastasis, vessel invasion, neural invasion, histology, and postoperative complication (P ≥ 0.05 in the fixed effects model or P ≥ 0.1 in the random effects model) (Table 2).

Table 2 Summary of characteristics between High ALI group and Low ALI group
Full size table

Clinical impact of the preoperative ALI on the survival outcome

After pooling the HRs and 95% CIs of OS from fourteen studies, ALI was a prognostic predictor for OS (HR = 2.09, I2 = 92%, 95% CI = 1.01 to 1.76, P < 0.01), DFS (HR = 1.48, I2 = 83%, 95% CI = 1.18 to 1.87, P < 0.01) and CSS (HR = 1.28, I2 = 1%, 95% CI = 1.02 to 1.60, P = 0.03) in gastrointestinal cancer patients after surgery (Fig. 2). Subgroup analysis was conducted for CRC patients and GC patients independently. As for OS, we found that there was still a close association between ALI and CRC (HR = 2.26, I2 = 93%, 95% CI = 1.53 to 3.32, P < 0.01) and GC (HR = 1.51, I2 = 40%, 95% CI = 1.13 to 2.04, P = 0.006) (Fig. 3). As for DFS, the prognostic value of ALI was also shown in CRC (HR = 1.54, I2 = 85%, 95% CI = 1.14 to 2.07, P = 0.005) and GC (HR = 1.37, I2 = 0%, 95% CI = 1.09 to 1.73, P = 0.007) (Fig. 4).

Fig. 2
figure 2

a OS, b DFS, and c CSS of the low ALI group and the high ALI group. Abbreviations: OS, overall survival; DFS, disease-free survival; CSS, cancer-specific survival; ALI, advanced lung cancer inflammation index

Full size image
Fig. 3
figure 3

Subgroup analysis for OS. Abbreviations: OS, overall survival; ALI, advanced lung cancer inflammation index

Full size image
Fig. 4
figure 4

Subgroup analysis for DFS. Abbreviations: DFS, disease-free survival; ALI, advanced lung cancer inflammation index

Full size image

Sensitivity analysis

We excluded each study at a time for repeated analysis, and the exclusion of any one study did not significantly alter the results. The funnel plot was performed for assessing the publication bias of OS, DFS, and CSS (Fig. 5). Unfortunately, only a small publication bias was found in the CSS due to the symmetry of its funnel plot, which meant that the result of the CSS was reliable. The source of publication bias in OS and DFS might be that no published studies with a negative outcome or no correlation were found currently.

Fig. 5
figure 5

Funnel plots for a OS, b DFS, and c CSS

Full size image

Discussion

Our meta-analysis enrolled 5091 patients from fourteen studies. We found that a low ALI was an independent prognostic factor for OS, DFS, and CSS in gastrointestinal cancer patients. Further subgroup analysis reported that ALI had prognostic effects for both CRC and GC patients.

Gastrointestinal cancer is a malignant wasting disease that can be accompanied by obstruction or bleeding, which causes malnutrition [39]. Malnutrition was related to high incidences of poor survival and postoperative complications [40]. Therefore, some nutritional parameters including BMI [14, 15] and Alb levels [16, 17], have prognostic value for gastrointestinal cancer patients. ALI is a combined score that is calculated by BMI, Alb levels, and NLR. The prognostic value of ALI was first reported in non-small cell lung cancer [13]. It can reflect the nutrition status of the host [13].

Some studies have researched the relationship between ALI and gastrointestinal cancer. Most studies demonstrated that ALI could predict prognosis including OS and DFS for colorectal cancer patients [21, 23,24,25, 27]. Zhang X showed that preoperative ALI was an independent prognostic factor for GC patients undergoing curative gastrectomy [29]. ALI was also revealed to be a prognostic predictor for EC, liver cancer, and cholangiocarcinoma [30,31,32,33]. However, Pian G, Yin CZ, and Barth DA reported that there was no significant difference between ALI and DFS after gastrointestinal surgery [22, 28, 34]. Therefore, the extra relationship and potential mechanism need to be analyzed.

ALI not only predicts the prognosis of gastrointestinal cancer patients by reflecting the body's nutritional status but also the response to inflammatory conditions. It was reported that inflammation was an important process for the occurrence and development of gastrointestinal cancer [41]. When inflammatory factors and inflammatory cells were activated, new lymphatic vessels and blood vessels formed, which created a microenvironment conducive to the growth and differentiation of tumor cells. Tumors can also disrupt immune cell function, and then, tumor cells can more easily invade and metastasize. Many inflammatory indicators, including the albumin-to-globulin ratio (AGR), NLR, and systemic immune-inflammation index, are independent prognostic factors for gastrointestinal cancer [42,43,44,45,46,47]. ALI, an inflammatory indicator, might also have use for gastrointestinal cancer patients.

Current studies are controversial regarding the prognostic predictive value of ALI in patients with gastrointestinal cancers. Our meta-analysis was the first study to summarize the outcomes and mainly solved the inconsistency of ALI in DFS. Moreover, the included studies were relatively new, with the earliest one published in 2019. Then, we included all gastrointestinal cancer studies for analysis. In addition, almost all the included studies were published in Asia, which means that the results of our study are reliable for Asian individuals. However, the results could not be extrapolated worldwide because of the differences in metabolic profiles between different nations.

However, there were some other limitations in this meta-analysis. First, we only included fourteen retrospective studies without any randomized controlled trials or cohort studies. Second, the cutoff value of studies was not consistent, which might lead to error. Third, the potential association between ALI and curative or palliative surgery needs to be explored further.

In conclusion, gastrointestinal cancer patients with a low ALI had a higher risk of poor prognosis after surgery. ALI was an independent prognostic factor for both OS and DFS. Doctors need to pay more attention to patients with low ALI to improve their prognosis.