Predicting fibroblast activation protein overexpression in the overall survival rate of cancer patients: a systematic review and meta-analysis

Abstract

Introduction

Fibroblast activation protein-alpha (FAP-α) is a vital surface marker of cancer-associated fibroblasts, and its high expression is associated with distant metastasis. This systematic review and meta-analysis were performed to predict the role of FAP-α overexpression in the overall survival rate of cancer patients.

Methods

This systematic review and meta-analysis were first performed on studies published in online databases, including PubMed, Scopus, and Web of Sciences, based on the PRISMA framework. We focused on all cancer patients with reported FAP-α expression levels and survival rate outcomes. Two reviewers independently conducted study selection and data extraction, and discrepancies were resolved through group discussion. Pooled hazard ratios (HRs) were then calculated by the fixed-effect and random-effects models to determine the association between FAP-α with crude HR (univariable), adjusted HR (multivariable), progression-free survival, and disease-free survival.

Results

Forty-one studies were included in the systematic review, and 25 were included in the meta-analysis. Meta-analysis showed that high FAP-α expression was associated with the pooled HR for overall survival (HR = 1.49, 95% CI: 1.19–1.85, P < 0.001) and HR of 1.53 in studies that reported the adjusted effect of high FAP-α expression on overall survival (HR = 1.53, 95% CI: 1.16–2.03, P = 0.003), and HR of 1.36 for disease-free survival (HR = 1.36, 95% CI: 0.750–2.469, P = 0.31). In addition, lymph node metastasis and distant metastasis were significant factors and had pooled HRs of 2.053 (95% CI: 1.603–2.630, P < 0.001) and 2.630 (95% CI: 1.902–3.637, P < 0.001), respectively.

Conclusions

Our results showed that cancer patients with FAP-α overexpression had a significant association with poor overall survival. Incorporating FAP-α testing into cancer diagnostic protocols can help identify high-risk patients requiring more critical treatment interventions.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13643-025-02929-6.

Introduction

Recent advances have shown that the tumor microenvironment (TME) consists of various cell types, including stromal cells, immune cells, and blood vessels, which play critical roles in cancer cell growth, invasion, and metastasis [1, 2]. In this case, Fibroblast activation protein-α (FAP-α) is a transmembrane serine protease, a cell-surface protein expressed primarily in fetal mesenchymal tissues, stromal fibroblasts, and wounded tissues, which can help preserve tissue homeostasis in skeletal muscle [3]. Moreover, FAP-α is critically important in tumor progression and cancer metastasis, and its expression has been detected in various cancer cells [4–6]. Our previous systematic review and meta-analysis reported a significant association between FAP-α overexpression and metastasis [7]. FAP-α overexpression increased vascular, lymph node, lymphatic vessel, and distant metastasis in various cancers [7].

Previous studies have also reported conflicting findings regarding the association between FAP-α overexpression and survival in cancer patients. Unlike some studies, most of the literature showed FAP-α expression has an association with shorter overall survival (OS), disease-free survival (DFS), progression-free survival (PFS), and cancer-specific survival rates [8, 9]. Furthermore, studies have shown that the prognostic relevance of FAP-α can act as an independent predictor in multivariate analyses [10]. Thus, the lack of consistency in previous studies highlights the need for a systematic review and meta-analysis of the impact of FAP-α overexpression on patient survival rates.

Methods

Protocol and registration

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist guides this systematic review and meta-analysis [11]. The protocol was assessed and registered with the Tehran University of Medical Sciences Research Committee (Register number: 1403–1-417–72,108).

Eligibility criteria

We selected studies using the Population, Exposure, Comparison, Outcome, and Study Design (PECOS) framework based on predefined eligibility criteria. The interested population included cancer patients, regardless of cancer type, age, or sex, with FAP-α expression. Studies that evaluated FAP-α overexpression in cancer tissue met the exposure criteria. The comparison group consisted of cancer patients with low or negative FAP-α expression. Furthermore, the primary outcome of this meta-analysis was the overall survival, hazard ratio (HR), or survival rate of the patients, regardless of the type of statistical analysis used.

Only studies that were published in English and peer-reviewed journals were included. The study design criteria allowed all designs except case reports, case series, and letters to the editor. However, gray literature, such as conference abstracts, theses, and unpublished data, was excluded.

Information sources

The sources of the included articles were multiple electronic databases, such as PubMed, Scopus, and Web of Sciences. We gathered the studies using relevant search strategies for each database. The search was conducted in two steps. In the first step, block 1 included related keywords with FAP-α, and the terms were linked with an “OR” Boolean. Likewise, in the second block, words associated with survival outcomes were combined with"OR."These two search blocks were combined with AND in the next step, creating a final search strategy (Supplementary 1). In addition, no publication date and language restrictions were applied during the search. The relevant articles'reference lists were manually screened for additional studies.

Study selection

The study selection process involved multiple stages, including duplicate elimination, screening, eligibility assessment at full-text assessment, and selection for review. Database search results were imported into EndNote version × 9 at initial screening. Duplicates were identified and removed. Two independent reviewers (Z. A, and K. P) screened the remaining records based on titles and abstracts. Studies that did not meet the inclusion criteria were excluded at this stage. Two independent reviewers evaluated and reviewed full-text articles of potentially relevant studies (Z. A, and K. P). Any disagreements regarding study eligibility were resolved through discussion with the third reviewer to make the final decision.

Data extraction

Studies that met the eligibility criteria were included in the systematic review. The data extraction was conducted from all studies for systematic review, and studies providing interesting quantitative data were selected for meta-analysis. Data extraction was undertaken initially through a pilot phase involving ten studies. Two independent authors (M.J. and Z.A.) extracted the data during this phase. The research team administrators (V.K. and A.M.A.) assessed the checklist's interobserver reliability, accuracy, and comprehensiveness. The checklist was refined based on the administrators'feedback and group discussions. The revised form was then used to extract data for the pilot-assessed study and the remaining studies. Following this, data were independently collected by two reviewers (Z.A, and M.J). Variables included study details, participant data, cancer type, sample size, FAP-α detection methods, and survival outcomes. In addition, extraction was performed in duplicate, and any discrepancies were resolved through discussion with the third reviewer consulted (V.K). In the studies where data were incomplete, we emailed the corresponding author and/or the article's first author to request the missing information.

Risk of bias assessment

The risk of bias was assessed for studies included in the meta-analysis. The Joanna Briggs Institute (JBI) Critical Appraisal Checklist was used for risk of bias assessment and was designed explicitly for cohort studies. This tool evaluates key methodological aspects, including study design, exposure and outcome measurement validity, confounding control, and completeness of follow-up. Based on the checklist guideline, each parameter was classified as Yes (low risk), No (high risk), Unclear (insufficient data), or Not Applicable [12].

Statistical analysis

After data extraction, data were synthesized for meta-analysis. In the first step, we standardized variable units across studies. Besides, all reported HR were adjusted for the hazards of the high FAP-α group compared to the low FAP-α group based on study reports. Then, we took the logarithm of the HRs along with their 95% confidence intervals (CIs) due to the asymmetry in the HR range [13]. The primary method involved pooling HRs for overall survival. The effect size was calculated by fixed-effect and random-effects models based on the between-study variability [14].

The homogeneity and heterogeneity among studies were assessed by Cochran's Q test and I² statistics that quantify the percentage of total variation due to heterogeneity rather than chance. For the categorization of I², an I² value of 25%, 50%, and 75% was defined to indicate low, moderate, and high levels of heterogeneity, respectively [15]. In addition, Begg's [16] and Egger's [17] tests were used to detect publication bias, and the funnel plot was reported as a visual representation of potential asymmetry [18]. Moreover, the Trim and Fill method was used to estimate and adjust for missing studies. In cases of heterogeneity (I² ≥ 50% and significant Q test P-value) in studies, sensitivity analysis was conducted to assess outcome variations based on potential heterogeneity factors. A P-value less than 0.05 was considered statistically significant, and STATA version 17.0 was used for all analyses with 2-tailed p-values.

Results

General data

Figure 1 shows that 5,757 articles were included in the first search for this systematic review. In the next step, 513 studies were identified as duplicates. Following the evaluation of titles, abstracts, and keywords, 4,371 articles were excluded due to irrelevant patient populations (n = 2,896), exposures (n = 864), outcomes (n = 589), or study design (n = 22). Furthermore, 873 articles that initially met the inclusion criteria were re-evaluated with full-text assessment, 41 articles were included in this systematic review, and 25 studies were considered in this meta-analysis. Table 1 presents the characteristics of the included articles.

Fig. 1.

PRISMA flowchart illustrating the selection process of studies, from initial identification (5,757 articles) to final inclusion in the systematic review (41 articles) and meta-analysis (25 studies)

Table 1.

Characteristics of the included studies in the systematic review

First author name (year)	Country	Follow up duration	Sample size	Mean age	Gender	Cancer type	Tumor stage	Patients under treatment	FAP detection method, Cut off of FAP	Number of patients with high FAP level	Main results
Ariga, N. (2001) [41]	Japan	-	112	-	-	Breast cancer	-	-	IHC is focally positive (10%), positive (10% to 50%), or strongly positive (50%)	61	Higher expression of FAP/seprase in invasive ductal carcinoma cases is associated with longer OS and DFS Multivariate analysis has shown FAP/seprase expression as an independent prognostic factor for survival
Borchert, S. (2023) [46]	Germany		77	64.6 (Mean), 65.2 (Median), Range (37.6–82.9)	64 (male), 13 (female)	Lung cancer	-		IHC, negative (score 0) = 0%; low (score 1) = 1–10%; moderate (score 2) = 11–50%; high (score 3) = > 50%	46	Two-year OS was 51.7% in the high FAP expression compared to 29.6% in the low expression (cutoff: 95 counts; p = 0.012; HR: 1.93; 95% CI: 1.15–3.24). Median OS in high FAP expression was 18.7 months (n = 46; 95% CI: 14.4–30.9) vs. 16.0 months (n = 27; 95% CI: 15.2–22.4) in the low FAP
Calvete, J. (2019) [52]	Spain	mean: 51.4 ± 48.8, Range: 1–192 m	121	68.1 ± 9.25 (SD), range (44–89)	118 male and 3 female	Bladder cancer		Radical cystectomy with lymph node dissection/chemotherapy	IHC, Cutoff points, or an automated scoring system were not used. When at least one core was positive, the results of the two cores were combined as positive	76	Cytoplasmic immunostaining of FAP in CAFs is associated with worse DFS HR = 1.68; P =.048
Chen, L. (2018) [19]	China	In 85 cases, 3 Y	122	≥ 55 Y (75) & < 55 Y [47]	68 male and 17 female	Lung cancer	I-III	Surgical resection	IHC, grade 0, absent or < 1% staining stroma; grade 1, 1–10% positive staining; grade 2, 11–50% positivity; grade 3, > 50% positive staining. High expression was a grade > 2 (FAP-α positivity > 50%)	38	Cox proportional hazards model showed higher FAP density had a negative correlation with survival FAP density can serve as an independent prognostic factor for predicting 3-year OS and DFS
Chen. L. (2017) [20]	China	-	60	-	-	Colorectal cancer	-	-	IHC, low FAP expression to 1%−50% positive staining, high FAP expression to more than 50% positive staining	-	Results with Kaplan Meier showed lower survival in participants with high FAP
Coto-Llerena, M. (2020) [51]	Switzerland	-	92 (19 normal/92 tumor)	< 59 = 11, 60–69 = 22, 70–79 = 28, > 80 = 31	54 male and 38 Female	Colorectal cancer	I-IV	-	IHC samples containing < 10% of positive cells as low FAP and at least more than 10% of positive cells as high FAP. Intensities 0 or 1 were considered low, whereas samples with intensities 2 or 3 were considered high	72	FAP was up-regulated in colorectal cancer tumors at both mRNA and protein levels, and its expression was associated with advanced stages and poor survival High stromal FAP levels were associated with aggressive disease progression and worse survival
Da Silva, A.C. (2021) [55]	Brazil	48 M	Borderline ovarian tumors (n = 17) and malignant ovarian tumors (n = 28)	21–71 Y, Median (48)	All female	Ovarian cancer	I-IV	-	IHC, 0 (no staining), 1 (weak staining), 2 (moderate staining), and 3 (strong staining)	23	Stronger stromal FAP immunostaining (grades 2 and 3) was more frequently observed in epithelial ovarian cancer compared to borderline ovarian tumors (p = 0.0331). There was no significant association between stromal FAP immunostaining and OS or DFS in patients with epithelial ovarian cancer
Hemida, A.S. (2022) [54]	Egypt	-	70	62.857 (mean) ± 7.057 (SD)/26 < 60, 44 > 60	65 male and 5 female	Bladder carcinoma	Early/advanced	surgery	IHC, immunostaining intensity as (0, no; 1, weak; 2, moderate; 3, strong staining). Proportion of immunostaining (0, ≤ 25%; 1, > 26% & ≤ 50%; 2, (> 51% & ≤ 75%; 3, > 75%). The final FAP expression level was estimated (staining intensity score) x (proportion of immunostaining score). Mean value (1–3) was reported as low expression and (4–9) as high expression	27	FAP expression was found in 67.1% of urothelial carcinoma specimens and was significantly associated with advanced tumor stage, muscle invasion, and mitoses. However, there was no significant correlation between FAP expression and OS in the studied cohort
Henry, L.R. (2007) [49]	USA	35 M (1050 days)	138	28–90 Y, Median (68.5)	67 male and 71 female	Colorectal cancer	I-IV	chemotherapy	IHC, Grade 0: complete absence or weak FAP immunostaining in < 1% of tumor stroma; grade 1 + was focal positivity in 1% to 10% of stromal cells; grade 2 + was positive FAP in 11% to 50% of stromal cells; and grade 3 + : FAP in > 50% of stromal cells. Groups scored with 0 or 1 staining were compared with those with more outstanding (2 or 3)	101	In patients with metastatic disease, those with low FAP intensity had significantly longer OS compared to those with high FAP intensity (671 days vs. 428 days; P = 0.042). Patients with low FAP expression had better DFS, although results were not statistically significant (P = 0.23)
Im, S.B. (2022) [34]	South Korea	Median: 122 M (0–148.3)	135	39 ≥ 70 Y	Male: 80 and Female: 35	Colorectal cancer	I-III	-	IHC, grade 1, weak staining in < 50% or moderate in < 20% of stromal cells; grade 2, weak staining in ≥ 50%, moderate in 20%–49%, or strong staining in < 20%; and grade 3, moderate staining in ≥ 50% or strong staining in ≥ 20% IHC grades 1–2 were negative, and grade 3 was positive	32	Elevated FAP expression in CAFs has been associated with worse outcomes in terms of OS and disease recurrence. Specifically, high FAP expression correlated with poor 10-year OS
Jia, J. (2014) [21]	China	-	120	-	Female	Breast cancer	TNM1-4	-	IHC, Western blotting, RT-qPCR	-	Increased FAP-α expression is significantly associated with poor outcomes and reduced patient survival
Jiang, K. (2023) [22]	China	-	35	-	-	Renal cancer	-	-	IHC, Western blotting, RT-qPCR	-	Higher FAP expression correlated with poorer prognoses, as demonstrated by Kaplan–Meier survival analysis
Jung, Y.Y. (2015) [35]	South Korea	73.1 ± 28.9	939	566 (≤ 50 Y) & 373 (> 50 Y)	Female	Breast cancer	T1-T2/T3	-	IHC, 0, negative or weak immunostaining in < 1% of the tumor/stroma; 1, focal expression in 1–10% of the tumor/stroma; 2, positive staining in 11–50% of the tumor/stroma; and 3, positive staining in 51–100% of the tumor/stroma. ≥ 2 = high FAP	345	In breast cancer with fibrous stroma, FAP positivity was associated with significantly shorter DFS (63 m vs. 128 m) and OS (65 m vs. 130 m), though the differences were not statistically significant (DFS: P = 0.525; OS: P = 0.483). In adipose stroma type, FAP expression showed no significant impact on survival outcomes (DFS: P = 0.750; OS: p = 0.991)
Kawase, T. (2015) [42]	Japan		48	71.5 ± 1.3 (SD)	28 (male), 20 (female)	Pancreatic cancer	TNM1-4	post operation chemotherapy (n = 37)	Western blotting, IHC, Stromal, fibroblast FAP expression was graded by the number of positive cells per 1000 stromal fibroblasts for three randomly selected views (weak < 350, 350 ≤ moderate < 650, and 650 ≤ strong)	29	Stromal expression of FAP was detected in 98% of specimens, with a significant association between moderate to strong FAP expression and lower cumulative survival rates (352 days vs. 497 days, p = 0.006). Multivariate analysis confirmed moderate to strong FAP expression as a prognostic factor
Kim, H. M. (2015) [36]	South Korea	-	132	68 (≤ 50 Y) and 64 (> 50 Y)	female	Breast cancer	-	-	IHC, as 0 = negative or weak immunostaining in < 1% of the tumor/stroma, 1 = focal expression in 1–10% of the tumor/stroma, 2 = positive in 11–50% of the tumor/stroma, or 3 = positive in 51–100% of the tumor/stroma; scores of 2–3 were defined as positive	11	In the univariate analysis of the impact of CAF-related protein expression on patient prognosis, no CAF-related protein was significantly associated with shorter OS
Kim, H.M. (2016) [37]	South Korea	-	194	40.1 ± 12.3 (SD)	Female	Breast cancer	-	-	IHC, 0, negative; 1, less than 30% positive; 2, more than 30% positive. The Immunostaining intensity is defined as: 0, negative; 1, weak; 2, moderate; 3, strong. The scores for the proportion of stained cells and immunostaining intensity were multiplied, and staining was defined as positive when the final score was > 1	21	No significant correlation was found between FAP expression and OS in the univariate analysis
Knipper, K. (2023) [47]	Germany	med: 18 M (range 3–98)	321	104 (< 65 Y), 217 (≥ 65 Y)	157 (male) and 164 (female)	Pancreatic cancer	pT1-pT4	Surgery/radiochemotherapy/chemotherapy	IHC, the cutoff was defined as the median for FAP, and the 45th percentile defined as the cutoff for SMA. Values lower or equal to the cutoff were defined as low	160	High FAP expression did not significantly impact OS in pancreatic ductal adenocarcinoma patients (P = 0.208)
Lee, P. J. (2022) [56]	Taiwan	Min: 28 M	249	med: 48.96 (21.48–81.35)/mean: 49.34 ± 10.75	Male (103) 72% and female (40) 28%	Nasopharyngeal carcinoma	T1/2-T3/4	-	IHC, staining intensity ("−,""+,""+ +"or"+ + +"), and extent (positive percentage relative to total area investigated). Higher FAP expression (FAP score ≥ 80)	53	In nasopharyngeal carcinoma, higher levels of FAP in stromal fibroblasts were associated with poorer metastasis-free survival. Patients with elevated FAP expression had a significantly reduced metastasis-free survival compared to participants with lower FAP levels
Li, F. (2020) [23]	China	median: 49 Y (2–21)	121	64 ± 7.9 (SD)	95 (male) and 26 (female)	Esophageal cancer	pTN0-3 M0	post op chemo/radiotherapy	IHC, percentage scoring of Immunoreactive tumor cells was 0 (0–5%), 1 (6–25%), 2 (26–50%), 3 (51–75%), and 4 (> 75%). Staining intensity was 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). A final Immunoreactivity score was obtained for each case by multiplying the percentage and intensity score. 0 was negative (-),1–4 was weakly positive (+), 5–8 was moderately positive (+ +), and 9–12 was strongly positive (+ + +)	45	In esophageal squamous cell carcinoma, stromal FAP expression is significantly associated with poor survival outcomes. Higher stromal FAP levels were Linked with poorer survival. In univariable analysis, HR was 2.009 (95% CI 1.259–3.205; P = 0.003), and multivariable analysis reported HR of 1.833 (95% CI 1.144–2.937; P = 0.012)
Li, M. (2020) [23]	China		151	-	Female	Ovarian Cancer			IHC/gene expression analysis, percentage of positive cells, and intensity of staining of FAP antibody were first calculated and then divided into these three major categories: ≤ 3, negative or weak; > 3 and ≤ 6, Moderate; > 6, strong	118	In high-grade serous ovarian carcinoma, low FAP expression is associated with better prognosis. Cox regression analysis revealed that patients with low FAP expression had a notable survival advantage in OS (P = 0.005) and progression-free survival (P = 0.008). At 12 months, the survival rate was 91.1% for the low FAP group compared to 84.4% for the high FAP group. By 50 months, survival rates dropped to 31.9% for the low FAP group and 21.4% for the high FAP group
Liao, Y. (2013) [24]	China	Median: 30 M (8–40)	59	63.5 (Median)	Male (47) and Female (12)	Lung cancer	I-III	Surgery	IHC, grade 0, absent or < 1%; grade 1, 1–10%; grade 2, 11–50%; grade 3, (50%) and intensity (0, none; + 1, light; + 2, moderate; + 3, intense)	45	The study detected FAP in 76% of specimens, and its high expression was linked to poorer tumor differentiation (P = 0.06). Increased FAP staining percentage and intensity were associated with worse OS (percentage, P = 0.0087; intensity, P = 0.05)
Lyu, S.I. (2024) [48]	Germany	Mean: 66 M (max: 10 Y)	216	119 (55%) < 60 Y, 97 (45%) > 60 Y	Male (171) and Female (45)	Oropharyngeal squamous carcinoma	I-IV	Surgery (n = 132;61%)/radiochemotherapy (n = 84;39%)	IHC, the cutoff for high expression was defined as the median of the patient population	105	Periostin and α-SMA: High expression levels were linked to worse OS in univariate analyses (Periostin, P = 0.009; α-SMA, P = 0.014). FAP and PDGFRβ: No significant association with patient survival was observed (FAP, P = 0.557; PDGFRβ, P = 0.114)
Mhawech-Fauceglia, P. (2014) [50]	USA	-	338	61 (med): 24–89	All female	Ovarian Cancer	1 to 4	Chemotherapy	Tumor microarray construction and IHC, The percentage was assessed as follows: 0%, < 10%, 11–50%, 51–100%, and the intensity as 0, weak (1 +), moderate (2 +), and strong (3 +)	277	FAP + stroma was associated with a nearly significant reduction in OS (P = 0.0685). FAP + stroma was significantly associated with a higher rate of recurrence (P = 0.0247)
Moreno-Ruiz, P. (2021) [53]	Sweden		351	67 (mean): 42–84	Male (174, 133 FAP+) and Female (177, 124 FAP+)	Lung cancer	I-IV	Surgery	IHC, Intensity of FAP staining, was scored in each core on a 4-grade scale. For dichotomization of FAP, patients-based scores < 2 were considered as"Low"(25% of the cohort), and scores ≥ 2 were noted as"High"(75% of the total cohort)	257	High stromal FAP expression was identified as an independent poor prognostic marker in the overall study population (HR 1.481; 95% CI 1.012–2.167, P = 0.023). High stromal FAP expression was an independent marker of poor prognosis, specifically in adenocarcinoma (HR 1.720; 95% CI 1.126–2.627, P = 0.012)
Nam, Y. (2022) [38]	South Korea	med; 87 m	453	63 (med)	Male (309) and Female (144)	Lung cancer	I-IV	Surgery	Immunoreactivity was assessed using the semi-quantitative H-score (range 0–300), which was derived through the summation of each staining intensity (0–3) multiplied by the percentage (0–100). When the average value of the quantified result of 2 scores was above 0.5 of staining intensity, the result was defined as positive immunoreactivity	312	No significant association was reported to OS or DFS
Park, C. K. (2016) [39]	South Korea		628 (total)	58 (< 50), 46 (≥ 65 Y)	Female	Breast cancer	I/II-III		0, negative or weak immunostaining in < 1%; 1, focal expression in 1–10%; 2, positive in 11–50%; 3, positive in 51–100% of tumor/stroma. The stained slides were evaluated on the entire tumor area, and scores of 2 or higher were regarded as positive	134	FAP expression was not significantly associated with DFS or OS in invasive breast cancer cases, with p-values of 0.206 and 0.391, respectively
Rong, X. X. (2022) [25]	China		31	-		Gastric cancer		Immune checkpoint therapy	IHC, PET/CT, median value of FAP expression was defined as the cutoff value	-	High FAP expression was associated with poor prognosis in gastric cancer High FAP expression was associated with decreased OS (HR, 2.06; 95% CI, 1.43–2.97; p < 0.001)
Saigusa, S. (2011) [43]	Japan	43 M (14–105)	52	62 (Mean), 64 (Median): range (37–78 years)	Male (42) and Female (10)	Rectal cancer	I/II-III	CRT/surgery	ELISA, qRT-PCR, IHC	7	Kaplan Meier reported OS and recurrence-free survival in participants with high FAP levels were significantly lower than participants with lower ones (P = 0.031, P = 0.004, respectively). However, OS and recurrence-free survival was longer in participants with a high level of serum FAP level compared with participants with a lower level of serum FAP level
Shi, J. (2020) [26]	China	5–10 Y	92	61 (≤ 65 years) & 33 (> 65 years)	Male (51) and Female (43)	Lung cancer	I-IV	-	IHC/western/qRT-PCR, grade 0 (absent or < 1%), grade 1 (1–25%), grade 2 (26–50%), grade 3 (51–75%), and grade 4 (76–100%). The staining intensity was graded (0, 1 +, 2 +, and 3 +). Immunoreactivity score was determined by multiplying positive cells'intensity score and percentages (ranging from 0 to 12). The cutoff threshold was at 3, and a score of > 3 was considered a high FAP expression, and ≤ 3 was a low FAP expression	78	Kaplan Meier showed that patients with higher FAP expression in tumor stroma had a significantly worse prognosis (P = 0.019). In contrast, the prognosis associated with FAP expression in tumor cytoplasm was not statistically significant (P = 0.4)
Son, G.M. (2019) [40]	South Korea	> 5 Y	147	42 ≥ 70 and 105 < 70	Male (88) and Female (59)	Colorectal cancer	I-IV	Surgery	IHC, grade 1, weak staining in < 50% or moderate staining in < 20% of stromal cells; grade 2, weak staining in ≥ 50%, moderate staining in 20% to 49%, or strong staining in < 20%; and grade 3, moderate staining in ≥ 50% or strong staining in ≥ 20%. IHC grades 1 to 2 were considered negative, and grade 3 was counted as positive	27	In the intratumoral stroma, higher levels of immature CAFs and FSP-1 were significantly associated with poorer prognosis (P = 0.02 and P = 0.03, respectively), while α-SMA and FAP showed no significant impact. At the invasive front, immature CAFs were linked to worse outcomes and increased risk (P = 0.03 and P = 0.04), but α-SMA, FAP, and FSP-1 did not show significant associations
Song, Z. (2016) [27]	China	-	102	68 < 60, 34 ≥ 60	Female	Ovarian cancer	-	Surgery	IHC was treated as a low expression group with a positive rate ≤ 95%. Otherwise, it was included in the high-expression group,	61	Kaplan–Meier analysis revealed that patients with higher FAP expression had markedly shorter OS compared to those with lower FAP expression,
Takagi, K. (2023) [44]	Japan	-	67	70.3 ± 9.2 Y (SD)	Male (40: 60%) and Female (27: 40%)	Ampullary carcinoma	-	-	IHC, median values (1.2) for FAP were used as cutoff values to define the low and high groups	33	Univariate analysis identified high expression of a-SMA (HR: 9.48, P = 0.005) and FAP (HR: 4.50, P = 0.03) as significant predictors of disease-specific survival. Multivariable analysis adjusted for potential confounders found that high a-SMA remained a significant predictor (HR: 8.54, P = 0.03), while the significance of high FAP changed to insignificance (HR: 2.92, P = 0.21)
Tong, Y. (2022) [28]	China	> 5 Y	171	135 (< 65 Y), 36 (≥ 65 Y)	Male (128) and Female (43)	Gastric cancer	II-III	NCT/Surgery	IHC, the score was 0 for < 1% positive area; score 1:1%−25% positive area; score 2: 74% positive area; score 4: 75%−100% positive area. The IHC results were negative (total scores 0 and 1 +) and positive(total scores 2 + and 3 +)	68	Post-treatment expression of FAP was significantly associated with OS (P = 0.011). Specifically, high post-treatment FAP (HR: 1.843, P = 0.013) levels were linked to poorer OS. In contrast, pre-treatment biomarkers did not show significant predictive value for OS. Multivariable analysis indicated that post-treatment FAP (HR: 0.755, P = 0.439) did not retain statistical significance as a predictor of OS
Waki, Y. (2023) [45]	Japan	22.1 M (2.3–147.7)	37	FAP+: 71 (Median), Range (63.5 −76)/FAP-: 70 (Median), range (64.3–75.5)	Male (24) and Female (13)	Intrahepatic cholangiocarcinoma	I-IV	Surgery/Chemotherapy	IHC, the area percentages of FAP were calculated as the ratio of the positive-stained areas relative to the total areas. The 25th percentile values of FAP were selected as the optimal cutoff values	27	The median percentages of FAP expression were 15.5% in the peripheral region and 17.8% in the intratumoral region. High FAP expression in the intratumoral region was significantly linked to worse OS and DFS. Multivariate analysis revealed that high intratumoral FAP expression was a significant risk factor for poorer OS (HR: 2.450, P = 0.049) and DFS (HR: 2.743, P = 0.034)
Wang, H. (2014) [29]	China	up to 5 Y	84	54.13 (mean: 31–83) (49 < 54 Y: 17 FAP +/35 ≥ 54 Y: 15 FAP +)	Male (54) and Female (30)	Oral squamous cell carcinomas	I-IV	-	In IHC, staining scores of 0–4 and 5–6 were considered low and high expression, respectively	53	Kaplan–Meier survival analysis demonstrated a significant correlation between FAP expression levels and overall survival times (P = 0.005)
Wen, X. (2017) [30]	China	-	105	64 (mean)	-	Gastric cancer	I-IV	Surgery	IHC, 1 for < 30% positive, 2 for 30%–60% positive, and 3 for > 60% positive. A score of 2 and 3 was assigned to high FAP expression, and a score of 1: low FAP expression	44	High FAP expression was observed in 55.2% of gastric cancer patients who died, compared to 25.5% of survivors. Kaplan–Meier analysis showed significantly shorter OS confirmed high FAP expression as an independent risk factor for poor survival (HR: 1.943, 95% CI: 1.083–3.484)
Wen, Z. (2019) [57]	China	-	56	37 < 65 and 19 > 65	24 male and 32 female	Pancreatic cancer	I/II -III/IV	Surgery	IHC, flow cytometry, western blot; dyeing area ≤ 10% defined as 0 points; 11% ≤ 25% as 1 point; > 26% ≤ 50% as 2 points; > 51% as 3 points. A negative staining intensity was scored as 0 points, weak staining as 1 point, intermediate staining as 2 points, and strong staining as 3 points. The classification of slice staining was divided according to the sum of the stained area and staining intensity score: ≤ 3 indicated FAP −; > 3 indicated FAP +	34	Patients with high FAP expression had a significantly shorter mean survival time of 13 m (95% CI: 9.31–18.69) compared to those with low FAP expression, who had a median survival time of 34 m (95% CI: 8.84–59.16). The log-rank test indicated a statistically significant difference in survival (P = 0.044). In the univariate analysis, high FAP expression was associated with a significantly increased risk of poor survival (HR: 3.629, 95% CI: 1.479–8.904, P = 0.005). Multivariate analysis, high FAP expression remained a significant prognostic factor for worse survival (HR: 3.013, 95% CI: 1.240–7.319, P = 0.015)
Zhang, M. (2015) [31]	China	-	199	53 < 60,75 > 60	F	Ovarian cancer	I—III	Surgery	IHC, In situ mRNA hybridization, 0, no staining; 1, staining of < 10% of cancer cells; 2, staining of 11–50% of cancer cells; and 3, staining of > 50% of cancer cells. Staining intensity is represented with the following scale: 0, negative; 1, weak; 2, moderate; and 3, strong. Based on intensity and extent of staining, IHC results were scored between 0 and 3: 0, negative; 1, weak; 2, moderate; and 3, strong	110	Univariate analysis of 128 patients, increased levels of seprase were significantly associated with a decreased probability of DFS (P = 0.03)
Zhao, Y. (2024) [32]	China	-	84	32 < 60, 52 > 60	66 male and 18 female	Lung	-	-	IHC/Western blot, 0, no staining; 1, weak intensity; 2, moderate intensity; 3, high intensity. Scores of 0 and 1: were low expression, and scores of 2 and 3 were high expression	52	Overall survival of participants with high FAP was significantly lower than patients with low FAP (P = 0.023)
Zhao, Z. (2023) [33]	China	> 6Y	171	135 (78.9%) < 65,36 ≥ 65	128 male and 43 female	Gastric		NCT/surgery	IHC (0–3 points for negative staining, yellowish, light brown, and dark brown, respectively), and the range of positivity (1–4 points for 0–25%, 26–50%, 51–75%, and 76–100%), and final scores calculated by multiplying together to a range of 0–12	68	In univariable analysis for OS, FAP expression was significantly associated with poor prognosis (OR = 1.843, P = 0.013). Mean survival of the participants with high FAP was considerably lower than patients with low FAP (P = 0.011)
Zou, B. (2018) [9]	China	33.2 M (Median): 1.1–80.2	138	51 < 60, 13 > 60	116 male and 22 female	Hepatocellular Carcinoma	I, II-IV		IHC/Western blot/RT-PCR, X-tile analysis. Optimal cutoff points of the relative IODs were based on patients'outcomes	64	Mean OS was significantly shorter in those with high FAP expression (34.8 m) compared to patients with low FAP expression (71.1 months, p < 0.0001, log-rank test)

•IHC Immunohistochemistry, •OS Overall survival, •DFS Disease-free survival, •HR Hazard ratio, •CAFs cancer-associated fibroblasts, •FAP Fibroblast activation protein, •SD Standard deviation, •NCT Neoadjuvant Chemotherapy, •CI confidence Interval, •M Months, •Y Years, •IHC immunohistochemistry

Systematic review

Table 1 shows the study characteristics. The studies were published between 2001 and 2024. The sample size was 908 (min = 31, max = 939). Among the reviewed articles, all of them were conducted in cohort design, and 16 studies had sample sizes of less than 100 participants. There were 17 studies between 101 and 200 participants, 2 studies between 201 and 300 participants, and 6 studies with sample sizes of more than 300 participants. The included studies were published in China (18 studies) [9, 19–33], South Korea (7 studies) [34–40], Japan (5 studies) [41–45], Germany (3 studies) [46–48], the USA (2 studies) [49, 50]. Other countries, Switzerland [51], Spain [52], Sweden [53], Egypt [54], Brazil [55], and Taiwan [56], were each study 1.

The mean/median age of participants was distributed between 51 and 70 years old, and cancer types of participants included lung cancer (7 studies), breast cancer (6 studies), colorectal cancer (5 studies), ovarian cancer (5 studies), gastric cancer (4 studies), pancreatic cancer (3 studies), oral squamous cell carcinomas (2 studies), rectal cancer (2 studies), bladder cancer (2 studies), hepatocarcinoma (2 studies), ampullary carcinoma (1 study), esophageal cancer (1 study), and nasopharyngeal carcinoma (1 study) (Table 1).

Surgical interventions were the most common treatment, among 23 studies that described treatment; 10 studies just investigated surgery as treatment [19, 24, 27, 30, 31, 38, 40, 53, 54, 57], 9 studies combined surgery and chemotherapy [23, 28, 33, 42, 43, 45, 47, 48, 52] and chemotherapy-only treatments were observed in 4 studies [25, 49, 50, 56]. In addition, 41 studies reported the FAP-α detection method. All of the included studies employed immunohistochemistry as an assessment method of FAP-α expression level, and 12 studies used a combination of methods, including immunohistochemistry with other techniques such as western blotting, RT-qPCR, ELISA, PET/CT, and flow cytometry (Table 1).

Risk of bias assessment

The risk of bias assessment showed that most studies had a low risk of bias in group comparability, exposure measurement, and outcome reliability. However, a high risk of bias was observed in confounder control and follow-up completeness, with several studies lacking proper adjustments or having incomplete follow-up data and unclear risk in handling missing data and follow-up sufficiency (Table 2).

Table 2.

Risk of bias assessment using the joanna briggs institute (JBI) critical appraisal checklist for cohort studies [9, 22–57]

Meta-analysis results

The pooled HR of the included studies is reported in Table 3 and Fig. 2. The overall crude HR (univariable analysis) of high FAP-α in poor survival of participants was 1.49 (95% CI: 1.19 − 1.85); in other words, the hazard of participants with high FAP-α was 1.49 times or 49% higher than participants with low FAP-α (P < 0.001). Also, the pooled adjusted HR was 1.53 (95% CI: 1.16 − 2.03, P = 0.003), the pooled HR of disease-free survival of participants was 1.36 (95% CI: 0.750 − 2.469, P = 0.311), and the pooled HR of progression-free survival of participants was 6.62 (95% CI: 5.84 − 7.50, P < 0.001). In addition, the pooled HR of poor survival in participants with lymph node metastasis (LNM) was 2.053 (95% CI: 1.603 − 2.630, P < 0.001) compared with participants without LNM, and the HR for participants with distant metastasis (DM) was 2.63 time (95% CI: 1.902 − 3.637, P < 0.001) compare with participants without DM (Table 3, Fig. 2) (Supplementary 2).

Table 3.

Pooled hazard ratios (HRs) for overall survival, disease-free survival (DFS), lymph node metastasis (LNM), and distant metastasis (DM) based on high FAP expression compared to low FAP expression

Associated factor	Number of studies	Sample size	HR	95% Confidence Interval	P-Value	I2 (%)
Crude FAP (univariable)	18	2452	1.486	(1.194—1.847)	< 0.001	74.31
Adjusted FAP (Multivariable)	16	2550	1.534	(1.158—2.034)	0.003	70.74
DFS	6	1018	1.361	(0.750—2.469)	0.311	84.30
PFS	2	112	6.619	(5.844, 7.496)	< 0.001	98.93
LNM	12	2341	2.053	(1.603—2.630)	< 0.001	29.26
DM	7	1168	2.630	(1.902—3.637)	< 0.001	20.59

Fig. 2.

Forest plot illustrating the pooled HRs for A crude overall survival, B adjusted overall survival, C disease-free survival, D in patients with lymph node metastasis, and E distant metastasis, F progression-free survival among participants with high vs. low FAP expression HR

Heterogeneity assessment

The heterogeneity across the studies that examined crude HR with FAP-α was moderate to high, with I² = 74.31% and a significant Q-test for heterogeneity (Q = 64.68, P = 0.00). In addition, studies that reported adjusted analyses (I² = 70.74%, Q = 45.25, P = 0.00), PFS (I² = 98.93%, Q = 93.41, P < 0.001), and DFS (I² = 84.30%, Q = 23.44, P = 0.001) were statistically significant. However, studies reporting an association between LNM with HR (I² = 29.26%, Q = 16.89, P = 0.11) and DM with HR (I² = 20.59%, Q = 6.48, P = 0.37) showed non-significant heterogeneity (Table 3).

Publication bias

The publication bias assessment for studies that reported crude HR suggests significant small-study effects with the Egger test (beta1 = 2.09, P = 0.009), Begg's test (Kendall's score = 55.00, P = 0.041), and for studies that reported PFS (Egger test beta1 = −18.33, P < 0.001, Begg's test Kendall's score = not assessed due to insufficient observation); however, the trim-and-fill analysis reports that the effect size remains consistent. Also, publication bias assessment showed insignificant publication bias for studies that assessed adjusted HR (Egger test beta1 = −0.18, P = 0.866, Begg's test Kendall's score = 12.00, P = 0.620), for studies that reported DFS (Egger test beta1 = −1.17, P = 0.662, Begg's test Kendall's score = −1.00, P = 1.000), for studies that assessed the association of LNM with HR (Egger test beta1 = −1.35, P = 0.185, Begg's test Kendall's score = −14.00, P = 0.373), and DM with HR (Egger test beta1 = 1.15, P = 0.439, Begg's test Kendall's score = 3.00, P = 0.7639) (Fig. 3).

Fig. 3.

Funnel plot illustrating the publication bias for A crude overall survival, B adjusted overall survival, C disease-free survival, in patients with lymph node metastasis (D), and E distant metastasis, F progression-free survival among participants with high versus low FAP expression HR

Subgroup analysis for heterogeneity

A subgroup analysis assessed potential heterogeneity in survival outcomes between the included studies. The subgroup analysis showed older participants (over 65) had higher HR (HR of age > 65 years = 1.73 vs. HR of age < 65 years = 1.45). Also, studies with 60 months or less follow-up periods reported a higher HR (HR of follow-up ≤ 60 months = 1.79 vs. HR of follow-up > 60 months = 1.01). Studies conducted on untreated participants showed a lower HR (HR for untreated participants = 1.43 vs. HR for treated participants = 1.52). Moreover, studies with more than 50% high FAP-α expression in participants reported higher HR (HR of ≤ 50% = 1.40, HR of > 50% = 1.67). The HR was significantly increased in studies in which more than half of the participants had nodal metastases compared to the studies in which less than half had nodal metastases (HR of ≤ 50% = 1.78, HR of > 50% = 1.30). In addition, cancer type reported significant heterogeneity and various HR were reported between cancer type subgroups (Table 4).

Table 4.

Subgroup analysis of hazard ratios (HRs) based on potential sources of heterogeneity for crude HR, adjusted HR, and disease-free survival

Subgroups	No. of studies	HR (95% of CI)	Heterogeneity I2 (%)	P-value heterogeneity	P-value between subgroups
Survival (crude, based on univariable analysis)
Age					< 0.001
65 or under	4	1.45 (1.05–1.99)	51.65	0.102
Over 65	6	1.73 (1.05–2.86)	81.91	< 0.001
Mean follow-up duration					0.013
60 months or lower	4	1.79 (1.28–2.50)	34.92	0.174
Over 60 month	2	1.01 (1.03–2.00)	0.00	0.420
Treatment in Participants					< 0.001
Not reported	4	1.43 (0.79–2.59)	87.16	< 0.001
Participants with treatment	13	1.52 (1.20–1.92)	67.08	0.001
High FAP %					0.003
50% or under	8	1.40 (0.96–2.05)	68.09)	0.003
Over 50%	8	1.67 (1.14–2.44)	82.04	< 0.001
Nodal metastasis in patients					< 0.001
50% or under	7	1.78 (1.29–2.47)	60.91	0.014
Over 50%	5	1.30 (0.67–2.50)	77.90	0.003
Cancer type					< 0.001
Gastrointestinal and head & neck cancers*	12	1.53 (1.16–2.03)	69.10	0.002
Ovarian cancer	1	2.10 (1.23–3.57)	-	-
Lung cancer	4	1.28 (0.78–2.09)	88.19	< 0.001
bladder cancer	1	1.68 (0.99–2.85)	-	-
Survival (adjusted, based on multivariable analysis)
Age					0.189
65 or under	4	1.47 (0.998–2.16)	50.88	0.103
Over 65	4	2.00 (1.44–2.79)	0.00	0.730
Mean follow-up duration					0.103
60 months or lower	5	1.57 (1.15–2.15)	39.58	0.151
Over 60 months	2	1.03 (0.77–1.39)	0.00	0.652
Treatment in Participants					< 0.001
Not reported	2	0.69 (0.05–9.70)	82.47	0.017
Participants with treatment	14	1.61 (1.25–2.09)	65.83	0.001
High FAP %					< 0.001
50% or under	8	1.25 (0.97–1.63)	37.46	0.128
Over 50%	8	1.84 (1.12–3.01)	78.41	< 0.001
Nodal metastasis in patients					< 0.001
50% or under	9	1.64 (1.03–2.59)	74.83	0.001
Over 50%	6	1.35 (0.88–2.07)	69.21	0.001
Cancer type					< 0.001
Gastrointestinal and head & neck cancers*	10	1.44 (1.06–1.95)	55.89	0.029
Ovarian cancer	1	3.77 (2.14–6.63)	-	-
Breast cancer	1	0.19 (0.05–0.70)	-	-
Lung cancer	4	1.65 (1.16–2.34)	44.69	0.151
DFS
Age					0.137
65 or under	3	1.44 (0.94–2.21)	5.38	0.126
Over 65	1	2.31 (0.97–5.48)	-	-
Mean follow-up duration					0.126
60 months or lower	2	1.88 (1.24–2.85)	0.00	0.542
Over 60 months	1	1.13 (0.83–1.53)	-	-
Treatment in Participants					< 0.001
Not reported	2	0.83 (0.11–0.13)	91.06	0.001
Participants with treatment	4	1.66 (1.01–2.50)	61.96	0.038
High FAP %					< 0.001
50% or under	4	2.12 (1.56, 2.88)	0.00	0.775
Over 50%	2	0.616 (0.17–2.24)	88.43	0.003
Cancer type					< 0.001
Gastrointestinal and head & neck cancers*	3	1.95 (1.34–2.84)	0.00	0.760
Breast cancer	1	0.30 (0.13–0.69)	-	-
Lung cancer	2	1.63 (0.747–3.55)	84.63	0.011

^* Pancreas, mouth, duodenum, liver, stomach, colon, esophagus, nasopharynx, oropharynx, cancer

Acronyms: HR Hazard Ratio, CI Confidence Interval, I² I-squared statistic (a measure of heterogeneity), FAP Fibroblast Activation Protein

The adjusted HR was greater in studies that were conducted with older participants (HR of age > 65 years = 2.00 vs. HR of age ≤ 65 years = 1.47), in studies that followed participants less than 60 months (HR of follow-up ≤ 60 months = 1.57 vs. HR of follow-up > 60 months = 1.03), and in studies that were conducted with under-treated participants (HR untreated participants = 0.69 vs. HR treated participants = 1.61). Also, it was higher in studies conducted with a high proportion of participants with high FAP-α levels reported higher HR (HR ≤ 50% = 1.25 vs. HR > 50% = 1.84). In addition, HR was significantly greater in studies in which more than half of the participants had nodal metastases than in studies in which less than half had nodal metastases (HR of ≤ 50% = 1.64, HR of > 50% = 1.35) (Table 4).

Moreover, subgroup analysis for DFS reported higher in studies conducted with older participants (HR age > 65 years = 2.31 vs. HR age ≤ 65 years = 1.44), in studies following participants less than 60 months (HR of follow-up ≤ 60 months = 1.88 vs. HR of follow-up > 60 months = 1.13), in studies that conducted with under-treatment participants (HR untreated participants = 0.83 vs. HR treated participants = 1.66), and in studies conducted with a high proportion of participants with high FAP-α level (HR of ≤ 50% = 2.12 vs. HR of > 50% = 0.62) reported a poor DFS and higher HR (Table 4). Moreover, while significant heterogeneity was reported in studies that determined the association between FAP and PFS, subgroup analysis could not be conducted due to the limited number of included studies (n = 2).

Discussion

This study evaluated the association between FAP-α overexpression and survival rate in various cancer types. The results showed that the HR of patients with FAP-α overexpression was 1.49 (univariable) and 1.53 (multivariable) higher than the ones with low FAP-α expression. As a result, the patients with FAP-α overexpression in various cancer types showed a significant association with poor overall survival rate. Moreover, LNM (2.05) and DM (2.63) increased the hazard ratios for the patients by more than twofold.

Similar to our results, Liu et al. [8] included a meta-analysis of 15 studies in which FAP-α overexpression in tumor tissues was significantly associated with poor overall survival and tumor progression, and the hazard ratio for overall survival was 2.18 (P = 0.004) [8]. They have also shown a direct correlation between FAP-α overexpression, increased tumor grade, and poor patient survival [8]. Graizel et al. [57] also conducted a meta-analysis including 11 studies and 1040 patients. Univariate Cox regression analysis revealed that the high CAF density was an adverse prognostic factor in 5-year survival with an odds ratio of 5.33 (P < 0.001) for women and 2.21 (p < 0.001) for men [58]. Furthermore, multivariate Cox regression analysis indicated that the patients with high CAF density compared to ones with low CAF density had an odds ratio of 2.74 (95% CI 2.22–3.38, P < 0.001) with a significantly higher death risk [58].

Previous studies showed that FAP-α overexpression was also observed in malignant cells and cancer stromal fibroblasts. Our last meta-analysis showed a significant association between FAP-α overexpression and cell metastasis, increasing metastasis to lymph nodes and distant tissues in various cancer types [7]. A strong association between FAP-α overexpression and clinical features was also observed in patients with different cancers [59]. A significant correlation was also seen between FAP-α overexpression in primary tumors and distant metastases [60]. Besides, a positive correlation of FAP-α overexpression with lymphatic vessel density in lung squamous cell carcinoma was observed in another study [19]. In this case, lacking FAP-α expression is essential for tumor inhibition, reducing tumor angiogenesis and changes in extracellular matrix regeneration [61, 62].

Our analyses have also shown low heterogeneity in LNM (29.26) and DM (20.59), which may be due to the consistency of the included research findings. The absence of publication bias and small study effects, comparable sample sizes, and the number of studies support the hypothesis that low heterogeneity can be due to the consistency of included studies'results rather than methodological differences. However, our analyses showed significant heterogeneity in univariable, multivariable, and DFS results, which can be based on several key variables. Older participants (> 65 years) had lower survival than younger participants (p < 0.001) in our subgroup analysis. This association can suggest that FAP-α overexpression may negatively affect the survival rate in older patients, possibly due to age-related patient vulnerabilities such as weaker immune responses or more underlying diseases. Furthermore, the follow-up period was identified as one of the significant variables in the results'heterogeneity. The studies with shorter follow-up periods (≤ 60 months) showed higher hazard ratios than the ones with more extended follow-up periods. It could be related to the survival of participants and better health conditions, including survivor bias. In other words, the patients with more prolonged follow-up had lower FAP-α expression.

Our subgroup analysis also revealed that untreated participants had a lower hazard ratio than treated participants, which might highlight the advanced stage of the disease in patients with FAP-α overexpression. Lymph node metastasis was also a significant factor, as studies with more than 50% of participants having nodal metastasis showed a lower survival rate. This finding may emphasize the importance of considering the patients'and tumor characteristics when evaluating FAP-α impact on disease prognosis. It has also been shown that the association between high FAP-α expression and clinical outcomes can be multifaceted in various tumors. For instance, patients with gastrointestinal and head and neck cancers experienced poorer survival than other types of cancer, such as ovarian and lung cancer (Table 4). This issue can be seen in various cancer types that had distinct prognostic effects reflecting tumor biology and the microenvironment, suggesting FAP-α overexpression in the tumor can be a potential key marker [24, 63, 64].

In summary, FAP-α can be a suitable marker for developing novel and effective treatments. Targeted therapy with targeted therapeutic approaches, such as FAP inhibitors, may reduce tumor progression and metastasis and improve clinical outcomes. Regular monitoring of FAP-α levels can also be a valuable tool to assess disease progression, treatment effectiveness, and patient monitoring.

Limitations

We observed moderate to high heterogeneity in analyses of univariable and multivariable hazard ratios and disease-free survival rates, which may affect the reliability of the pooled estimates. A few studies in some subgroups may also contribute to the observed low heterogeneity, which could be misleading. The studies varied in sample size, and most studies reporting small sample sizes may introduce bias and errors in the study's power. Moreover, most studies relied on single-point FAP measurements, limiting the assessment of the temporal effects of FAP changes on survival rates.

Conclusion

This meta-analysis shows that cancer cells with FAP-α overexpression, along with lymph node and distant metastasis, can significantly increase the risk of cancer patients'poor survival rate. These findings support the potential importance of FAP-α as a predictor of patient prognosis and survival, a promising therapeutic approach to developing strategies.

Authors’ contributions

Conceptualization: Amirhoushang Poorkhani, Majid Janani, Vahid Khori, Ali Mohammad Alizadeh. Data curation: Majid Janani, Zahra Akbari, Khalil Pourkhalili, Saeed Golfiroozi, Taghi Amiriani, Arash Tahmasebifar, Farahnazsadat Ahmadi, Yalda Jorjanisorkhankalateh. Formal analysis: Majid Janani, Zahra Akbari, Khalil Pourkhalili, Ali Mohammad Alizadeh. Investigation: Amirhoushang Poorkhani, Mirmohammadhosseinaali Sharifiandavaei, Majid Janani, Zahra Akbari, Khalil Pourkhalili, Saeed Golfiroozi, Taghi Amiriani, Arash Tahmasebifar, Farahnazsadat Ahmadi, Yalda Jorjanisorkhankalateh. Methodology: Amirhoushang Poorkhani, Mirmohammadhosseinaali Sharifiandavaei, Majid Janani, Zahra Akbari, Khalil Pourkhalili, Saeed Golfiroozi, Taghi Amiriani, Arash Tahmasebifar, Farahnazsadat Ahmadi, Yalda Jorjanisorkhankalateh. Project administration: Vahid Khori, Ali Mohammad Alizadeh. Resources: Vahid Khori, Ali Mohammad Alizadeh. Software: Majid Janani, Zahra Akbari, Khalil Pourkhalili. Supervision: Vahid Khori, Ali Mohammad Alizadeh. Validation: All authors validated and approved the final version.

Funding

Tehran University of Medical Sciences and Health Services, 72108, Ali Mohammad Alizadeh.

Data availability

The data supporting this study's findings are available on request from the corresponding author.

Declarations

Ethics approval and consent to participate

The study was conducted under relevant national and international guidelines and approved by the Tehran University of Medical Sciences Institutional Animal Care and Use Committee (NO: IR.TUMS.IKHC.REC.1403.112).

Competing interests

The authors have no conflicts of interest to declare and are responsible for the paper's content.