The diagnostic utility of glycosaminoglycans (GAGs) in the early detection of cancer: a systematic review

Research question and aim of study

The research question was formulated based on the Population, Intervention, Comparison Outcome (PICO) framework: Is glycosaminoglycans (GAGs) diagnostic test accurate in the detection of early-stage cancer? This systematic review aims to assess the sensitivity, specificity, and overall diagnostic accuracy of using GAGs in detecting early-stage cancer. Moreover, to identify any potential bias or inconsistency in the use of GAGs as a diagnostic tool for cancer detection. The protocol was registered in OSF Registries with registration DOI: https://doi.org/10.17605/OSF.IO/VCN9W.

Search strategy

Four databases (EBSCO, PubMed, Scopus and Ovid) were used to identify publications that satisfy our eligibility criteria. Different combinations of keywords were entered in each database search engine. The search results were limited across all databases to the years 2008–2023. Additionally, we manually searched the databases for related articles that may have been missed in the electronic search. The Boolean operators utilize the following combination (“Diagnosis” AND (“Cancer” OR “Tumor”) AND (“GAG” OR “Glycosaminoglycans”) AND (“blood” OR “plasma” OR “serum” OR “urine” OR “saliva”)

Inclusion and exclusion criteria

The following criteria were employed to screen and assessed the search results:

Inclusion criteria

• Sample Size: Studies with a sample size of at least 50 participants were included.

• Language: Studies that were published in English.

• Outcome measures: Studies that reported sensitivity and specificity values.

• Study design: Randomized controlled trials (RCTs), case-control studies, cohort studies, and diagnostic studies.

Exclusion criteria

• Publication Date: Studies published before 2008.

• Studies that involved in vitro or animal models

• Congress and meetings publications

• Systematic reviews

• Non-English publications

• Non-GAG (proteins, receptors, genes) indicators.

Study selection

Study selection was conducted in two stages: identification and screening. The identification stage entailed entering the keyword combinations in the databases’ search engines with the publication date limited to 15 years back (2008–2023). The results attained from each database totaled 402 articles. Additionally, we added five articles that we retrieved through a manual search. The resulting 407 records were then imported to Endnote, a reference management tool, where 71 articles were removed for duplication. In the screening stage, titles and abstracts were independently reviewed by two authors based on the predefined eligibility criteria. Any discrepancies were resolved by discussion with a third reviewer. Only articles that satisfied the specified criteria were included.

Full texts of potentially eligible studies were retrieved and assessed for inclusion based on study design, population, intervention, outcome and language. A total of 11 studies met the inclusion criteria and were included in the final qualitative synthesis (see PRISMA Flowchart–Fig. 1).

We anticipate that the publications we have selected will align with phase 2 of the Early Detection Research Network’s (EDRN) five-stage biomarker development process, which includes clinical tests and validation (Robey, 2004). Phase 2 is concerned with the confirmation of GAG through smaller-scale clinical testing before transitioning to a larger-scale clinical study, which is delineated in phase 3. This analytical strategy ensures that the established markers, GAGs, are dependable, repeatable, and have potential in clinical trials.

Quality assessment and data extraction

The 11 articles were analyzed and appraised using Critical Appraisal Skills Programme (CASP) checklists appropriate for each study design: cohort and case-control. Each article was given a grade score representing the strength of the article based on the questions of the CASP checklist. The articles were graded: good evidence, average evidence, or poor evidence (Table S1–CASP Checklist).

The data from the articles was extracted in a tabular form. The following details were included: author’s names, year of publication, and overall strength of the article (Table 1–Evidence Table).

The methodological quality of the included studies was assessed using validated tools appropriate for each study design. The Newcastle-Ottawa Scale (NOS) was used to evaluate the quality of the 11 observational studies: 6 case-control studies (Gatto et al., 2016, 2018; Biskup et al., 2021; Silva et al., 2018; Aghcheli et al., 2012; Xing et al., 2008) and 5 cohort studies (El-Mezayen et al., 2013; Gatto et al., 2022; Zhang et al., 2019; Creaney et al., 2013; Rangel et al., 2015).

The quality ratings for the NOS assessment were categorized as good, fair, or poor according to the Agency for Healthcare Research and Quality (AHRQ) standards as follows:

• Good quality: three or four stars in the selection domain AND one or two stars in the comparability domain AND two or three stars in the outcome/exposure domain.

• Fair quality: two stars in the selection domain AND one or two stars in the comparability domain AND two or three stars in the outcome/exposure domain.

• Poor quality: 0 or one star in the selection domain OR 0 stars in the comparability domain OR 0 or one star in the outcome/exposure domain (Tables 2A and 2B).

Table 2A:

Quality assessment of case-control studies and cross-sectional using the Newcastle-Ottawa Scale (NOS).

Article	Selection				Comparability	Outcome			Quality score
	1	2	3	4	5	6	7	8
Gatto et al. (2016)									Good (6)
Gatto et al. (2018)									Good (7)
Biskup et al. (2021)									Good (6)
Silva et al. (2018)									Good (6)
Aghcheli et al. (2012)									Fair (6)
Xing et al. (2008)									Fair (6)

DOI: 10.7717/peerj.18486/table-2a

Note:

: ** ; * ; : 0.

**: 2 points; *: 1 point; X: 0 point

Table 2B:

Quality assessment of cohort studies using the Newcastle-Ottawa Scale (NOS).

Article	Selection				Comparability	Outcome			Quality score
	1	2	3	4	5	6	7	8
El-Mezayen et al. (2013)									Good (7)
Gatto et al. (2022)									Good (7)
Zhang et al. (2019)									Good (6)
Creaney et al. (2013)									Good (7)
Rangel et al. (2015)									Good (6)

DOI: 10.7717/peerj.18486/table-2b

Note:

: ** ; * ; : 0.

**: 2 points; *: 1 point; X: 0 point

Heparan sulfate

Chondroitin sulfate

Hyaluronic acid

Heparan sulfate

HS, a complex glycosaminoglycan, is essential for the progression of malignancy. Gatto et al. (2016) investigated the diagnostic potential of the overall glycosaminoglycan (CS/HS) ratio in metastatic clear renal cell carcinoma (RCC). The role of GAGs in the regulation of biosynthesis and the remodeling of the extracellular matrix, both of which contribute to tumorigenesis, was underscored by this approach, which was supported by a large sample size and independent validation. The study did, however, recognize its limitations, which include the potential for confounding factors like renal inflammation and the possibility that altered expression of individual GAGs could also be a characteristic of malignancy. The study’s findings regarding the diagnostic potential of GAGs were substantiated by analyses of clinical outcomes in validation cohorts, despite these limitations, indicating alignment with plasma marker and combined marker scores (Gatto et al., 2016). This implies that, although the aggregate GAG ratio may have diagnostic value, additional research is required to completely understand the contribution of individual GAG components, including HS and its modifications, to tumor progression and their potential as biomarkers for early detection and monitoring. Gatto et al. (2018) conducted a subsequent study that expanded upon the diagnostic potential of GAGs in renal cell carcinoma (RCC), in addition to the findings from Gatto et al. (2016). The GAG score derived from plasma, which has a high specificity (94.7%) and sensitivity (93.5%) in distinguishing RCC from healthy controls. The biomarker’s potential for RCC diagnosis is evident in the high AUC values (0.999 in the discovery cohort and 0.991 in the validation cohort). It is crucial to note that the GAG score was not influenced by the tumor burden, which implies that plasma GAG alterations may be influenced by broader biological mechanisms, such as immune or stromal responses. Furthermore, the inclusion of total chondroitin sulfate (CS tot) as an independent prognostic factor for both overall survival (OS) and recurrence-free survival (RSF) demonstrated. The C Stot-based risk stratification enabled more precise predictions, indicating that plasma GAG components such as CS tot could provide substantial clinical value in patient management. This minimally invasive biomarker improves risk assessment and prognosis.

Overall, these results suggest that the plasma GAG score, particularly when combined with CS tot, has substantial potential as a diagnostic and prognostic tool in RCC. Nevertheless, the retrospective design and the variability resulting from the use of different sampling methodologies are among the limitations, underscoring the necessity of standardized assays and comprehensive assessments of baseline plasma GAG levels in a variety of populations. Despite the fact that additional prospective studies are necessary to determine their therapeutic relevance, the findings underscore the significant potential of plasma GAGs as significant diagnostic and prognostic instruments in RCC (Gatto et al., 2018).

Chondroitin sulfate

Chondroitin sulfate (CS) disaccharide profiles are emerging as potential diagnostic and prognostic biomarkers for various cancers. In ovarian cancer, a novel serum marker called “CS-bio,” based on the relative abundance of specific CS disaccharides, shows promise for early-stage detection (Gatto et al., 2018). This finding is particularly significant due to the limitations of the established marker CA-125 in early-stage disease. The observed increase in CS-4S and decrease in CS-0S disaccharides in ovarian cancer patients, likely due to elevated CHST11 activity, corresponds with prior research demonstrating increased CS disulfation in ovarian cancer tissues (Oliveira-Ferrer et al., 2015; Vallen et al., 2012). The promising results of CS-bio in ovarian cancer highlight the potential of CS disaccharide analysis in cancer detection. Nevertheless, its applicability might be specific to certain cancer types. In prostate cancer, for instance, overall CS levels did not differ significantly between patients and controls (Silva et al., 2018), suggesting a more nuanced role for CS in different cancers. However, a prospective study on renal cell carcinoma (RCC) demonstrated changes in plasma and urine free GAGomes, including CS disaccharides, associated with tumor recurrence after surgery (Gatto et al., 2016). The high predictive value of this multi-analyte approach suggests that analyzing CS disaccharide profiles from various sources could be valuable for developing robust cancer biomarkers across different malignancies.

Hyaluronic acid

Hyaluronic acid (HA) is a glycosaminoglycan with a multifaceted role in cancer progression. A study on breast cancer explored a multi-biomarker composite score (MBCS) incorporating HA, CA 15.3, hyaluronidase, NAG, glucuronic acid, and glucosamine (El-Mezayen et al., 2013). This MBCS showed significantly improved accuracy for detecting metastatic breast cancer, with an AUC value of 0.901 compared to the already established individual marker CA 15.3 with an AUC of 0.694 (Cheung, Graves & Robertson, 2000) suggesting a potential for enhanced diagnostic power through a combined approach. Among these markers, HA appeared as a potential marker for diagnosing metastasis, with a high AUC value of 0.792. HA also plays a crucial role in breast cancer progression. Increased HA production by tumor cells and the stroma creates a supportive microenvironment, facilitating tumor cell migration and hindering immune cell infiltration through an HA pericellular coat (Necas et al., 2008; Yabushita et al., 2004). Degradation of this coat by hyaluronidase can enhance immune attack but also generate fragments that promote angiogenesis (Ravi et al., 2000). A low molecular weight form of hyaluronan (~6 kDa HA) was identified in colorectal cancer (CRC) tissues by Zhang et al. (2019). Concentrations of ~6 kDa HA were found to be significantly elevated in cancerous tissues when compared to adjacent normal mucosa, suggesting its possible involvement in development and progression of CRC. Additionally, serum levels of ~6 kDa HA were significantly higher in CRC patients than in healthy controls, suggesting its potential utility as a non-invasive biomarker for early detection and monitoring of the disease. The increased concentrations of ~6 kDa hyaluronan in cancer tissues and its correlation with tumor stage, grade, and metastasis highlight its connection with disease progression. Moreover, the identification of ~6 kDa hyaluronan in serum samples suggests its potential role as a non-invasive biomarker for early detection and monitoring of colorectal cancer. The diagnostic accuracy of ~6 kDa hyaluronan, indicated by the high AUC value of 0.931, exceeds that of conventional markers like CEA (0.707) and CA 19-9 (0.677), as confirmed in previous metanalyses (Liu et al., 2014; Fernandez-Fernandez, Tejero & Tieso, 1995). The sensitivity and specificity of ~6 kDa hyaluronan, recorded at 95.65% and 71.43%, respectively, highlight its potential as a significant diagnostic tool. Furthermore, the association of ~6 kDa hyaluronan with lymphatic metastasis underscores its promise in predicting disease progression and guiding treatment decisions. Aghcheli et al. (2012) found that serum hyaluronic acid (HA) and laminin have moderate diagnostic accuracy for upper gastrointestinal (UGI) cancers. HA had a sensitivity of 80% and an AUC of 0.708 for esophageal squamous cell carcinoma, while laminin showed an AUC of 0.828 and a sensitivity of 90% for gastric cardia cancer. These findings suggest that neither biomarker is considerably raised, although they could still aid in distinguishing UGI cancer cases from controls (Aghcheli et al., 2012). Hyaluronic acid displayed a fairly good diagnostic accuracy for esophageal squamous cell carcinoma and gastric noncardiac cancer, whereas laminin showed comparatively lower effectiveness. However, laminin showed stronger utility for gastric cardia cancer, suggesting it may be a subtype-specific biomarker. HA and laminin are not exclusive to UGI cancers and can be elevated in non-malignant conditions such as liver disease and inflammation, but their diagnostic performance suggests they may be used in detecting high-risk patients. The study conducted by Creaney et al. (2013) confirms the diagnostic significance of effusion hyaluronic acid (HA) and mesothelin for malignant mesothelioma (MM). Soluble mesothelin, a well-characterized biomarker for MM identified in 2003 (Robinson et al., 2003). However, in recent meta-analyses, mesothelin alone has too low for early diagnosis, suggesting the need for further biomarker research (Hollevoet et al., 2012). The findings indicated that effusion HA and mesothelin levels were considerably higher in MM patients, which is consistent with earlier findings (Roboz et al., 1985; Pettersson et al., 1988; Grigoriu et al., 2009). The study found that a combination of effusion HA and mesothelin, both in the effusion and serum, yielded superior diagnostic accuracy compared to mesothelin alone, achieving an AUC of 0.92. This indicates that the integration of multiple biomarkers can improve the detection of MM, thereby overcoming a significant limitation associated with the use of individual markers. Serum HA proved to be non-diagnostic; however, effusion HA displayed a biphasic distribution in MM patients, with levels surpassing 75 mg/L associated with improved survival outcomes. This contrasts with various cancers such as prostate, gastric and breast cancers, where increased HA typically signifies a poor prognosis (Ghatak et al., 2010; Jang et al., 2011; Montgomery et al., 2012). In MM, elevated effusion hyaluronic acid levels were associated with a median survival of 18 months, in contrast to 12.6 months for individuals with reduced hyaluronic acid levels (p = 0.004). The reasons for this phenomenon remain unclear; however, it may relate to the distinct mechanisms by which MM tumors synthesize HA in comparison to other cancers, indicating that HA could play a protective role by inhibiting tumor dissemination within the pleura. The results of this study have important clinical implications for the management of MM patients. The combined use of effusion HA and mesothelin could improve the accuracy of diagnosis, leading to earlier detection and more effective treatment. The study conducted by Xing et al.’s (2008) demonstrates that serum HA levels were significantly elevated in oral cancer patients compared to individuals with benign tumors and controls, suggesting its potential as a tumor biomarker. However, the absence of specificity, due to increased HA in non-malignant conditions such as liver disease and rheumatoid arthritis, limits its diagnostic utility (Engström-Laurent et al., 1985; Engström-Laurent & Hällgren, 1985). The study found a correlation between serum HA levels and disease stage, with higher HA concentrations in stages III and IV compared to stages I and II. This suggests that HA may reflect tumor progression and metastatic status, as seen in other malignancies (Toole, 2004). However, HA did not exhibit significant differences between earlier and advanced stages of oral cancer (stages I & II vs. III & IV), and there was considerable heterogeneity in HA values among patients. This indicates that although HA may increase with tumor burden, it is not a reliable marker for early diagnosis or as an accurate method for tumor staging. Changes in HA levels did not show a significant correlation with treatment response, suggesting limited prognostic value in assessing therapeutic results. While serum HA may serve as a supplementary diagnostic tool for oral cancer, its effectiveness in staging and monitoring treatment outcomes remains unclear and requires additional research. A study by Rangel et al. (2015) on lung cancer demonstrated that hyaluronan (HA) levels are significantly elevated in lung cancer tissues, particularly in squamous cell carcinoma (SCC), where HA expression correlates with advanced tumor stages and poor survival outcomes. Notably, HA concentration serves as a prognostic marker, with levels above 692.1 mg/mg associated with reduced median survival. In addition to tissue analysis, they found that HA levels in sputum from lung cancer patients were significantly higher than in cancer-free individuals and healthy controls. This study established HA’s diagnostic potential, as ROC analysis indicated an area under the curve (AUC) of 0.821, providing 66% sensitivity and 80% specificity for distinguishing lung cancer from healthy individuals. These findings suggest that sputum HA levels can effectively identify patients with lung cancer, highlighting its potential as a non-invasive biomarker (Rangel et al., 2015). However, challenges remain in translating these findings to clinical practice. The generalizability of some studies is limited due to small sample sizes or lack of information on control groups. Additionally, while the multi-biomarker approach shows promise, its clinical utility and cost-effectiveness need further validation. Larger-scale studies and rigorous validation are necessary to confirm the clinical utility of HA as a diagnostic and prognostic biomarker, paving the way for its potential as a therapeutic target in various cancers.

Limitation of this study

The heterogeneity of GAG structures and lack of standardized analytical techniques hinder the accurate identification and quantification of cancer-specific GAG profiles. Developing standardized methods for GAG extraction, purification, and analysis is essential for improving the reproducibility and reliability of biomarker studies. Additionally, the relatively small sample sizes and varying methodologies of the included studies highlight the need for larger-scale, prospective studies to validate the diagnostic and prognostic utility of GAGs in diverse patient populations. Moreover, our selected articles are primarily focused on certain areas of the world, therefore rendering the current results evaluated in this review not fully representative of the global population. This is a limitation that needs to be addressed, as biomarker performance may vary across different demographics due to genetic, environmental, and lifestyle factors. Future studies should aim to include larger, and more diverse populations to ensure the generalizability of the findings. Despite these challenges, the potential of GAGs as biomarkers for early cancer detection and disease monitoring remains significant. Further research should focus on developing advanced analytical techniques, potentially incorporating machine learning and artificial intelligence, to refine GAG profiling and improve diagnostic accuracy. A deeper understanding of the mechanisms underlying GAG dysregulation in cancer is crucial for identifying novel therapeutic targets and developing personalized treatment strategies.