Biomarkers for oesophageal squamous cell carcinoma and the role of HPV: Multi‑omics approaches and current evidence (Review)

Introduction

According to the World Health Organization, oesophageal cancer (EC) is the sixth most common cancer worldwide, with a poor prognosis and difficult early diagnosis. EC is primarily classified into two histological types: Squamous cell carcinoma and adenocarcinoma, with squamous cell carcinoma being the predominant form (1). The overall 5-year survival rate for ESCC is <20% (2). The development of oesophageal squamous cell carcinoma (ESCC) is driven by a complex interaction of genetic and environmental factors, including diet, smoking, alcohol consumption, and potentially viral infections (3,4). Recent advances in multi-omics technologies have opened new opportunities for identifying prognostic biomarkers for ESCC. Concurrently, the possible role of human papillomavirus (HPV) infection as a contributing factor in ESCC development has garnered increasing attention. Despite extensive research on the link between HPV and ESCC, the underlying mechanisms remain incompletely understood. Consequently, current research is focused on a comprehensive exploration of the role of HPV in ESCC and the development of reliable detection methods. However, findings from studies investigating the relationship between HPV and ESCC show considerable variability.

HPV type 16 (HPV16) E6 has been shown to play a significant role in ESCC development (5), with HPV16 infection transforming basal esophageal cells and promoting cellular degeneration (6). Some studies, however, failed to detect high-risk HPV in ESCC samples (7), and in certain cases, HPV positivity was not significantly associated with clinicopathological features (8). These discrepancies may result from variations in study design, sample sources, testing methods, and geographical factors. Therefore, standardizing HPV testing protocols, particularly for ESCC, is essential. The aim of the present review is to outline multi-omics strategies for identifying prognostic markers of ESCC, explore the epidemiological and molecular mechanisms linking HPV infection to ESCC development, and evaluate the clinical applicability of current HPV detection methods. Ultimately, synthesizing these findings may provide novel insights into the diagnosis and treatment of ESCC, although some approaches discussed remain at an investigational stage.

Multiple factors promote ESCC formation

The onset of ESCC is closely linked to various factors, including lifestyle, environmental exposures, genetic susceptibility, dietary habits, and infections (Fig. 1).

Figure 1 Factors associated with ESCC pathogenesis and screening strategy. ESCC formation is caused by multiple factors including environment and lifestyle, genetics and microenvironment. HPV infection is a pivotal environmental factor that induces ESCC tumorigenesis through genome integration, oncogenic protein expression and immune escape. ESCC, oesophageal squamous cell carcinoma; HPV, human papillomavirus; HLA, human leukocyte antigen; TME, tumour microenvironment.

Lifestyle and environmental exposure are associated with ESCC

Environmental factors and lifestyle significantly contribute to the development of ESCC. Research indicates that smoking and alcohol consumption are major risk factors (9), particularly in regions where these behaviours are prevalent. Smoking directly damages esophageal tissue and increases cancer risk by impairing immune function and promoting chronic inflammation. Alcohol, especially spirits, is strongly linked to ESCC development, and the combined effects of alcohol and tobacco use further increase the risk of EC (8). Drinking hot beverages at temperatures above 65°C notably increases the risk of EC (10). Moreover, deficiencies in vitamins C and E, as well as folic acid, have been associated with ESCC onset, as these deficiencies may impair antioxidant defences, thus increasing the risk of cancer (11). Regarding environmental exposures, individuals chronically exposed to certain chemicals, such as asbestos and industrial toxins, particularly in environments with high levels of smoke and toxic gases, are at significantly greater risk of developing ESCC (12).

Genetic susceptibility is associated with ESCC

Genetic susceptibility is also a key factor in ESCC development. Variants in the human leukocyte antigen gene region, for example, are recognized as risk factors for ESCC, potentially influencing the immune response of an individual to tumour cells (13). Tumour microenvironment alterations are closely linked to the progression of the disease. Cytokines secreted by tumour-infiltrating cells form a complex network that promotes tumourigenesis through mechanisms such as inflammation, immune editing, and immune escape (14). Chronic inflammation is particularly critical, as it can induce genetic mutations and foster tumourigenesis by altering the metabolic state and microenvironment of cells (15). Future research should focus on understanding how genetic variants regulate the tumour microenvironment in ESCC and how these mechanisms can be harnessed to develop novel therapeutic strategies (16).

HPV infection is associated with ESCC

The association between HPV infection and ESCC has garnered considerable attention, although findings vary substantially across different studies (7,17). A comprehensive analysis of data from Asian countries revealed an overall HPV prevalence of 18.2% in ESCC, with a significant increase in cancer risk linked to HPV infection [odds ration (OR)=3.81; 95% confidence interval (CI), 2.84-5.11; P<0.001] (18).

Geographical and methodological heterogeneity is evident in studying the association between HPV infection and ESCC formation. Several studies have highlighted substantial variation in HPV positivity rates among patients with ESCC In some high-prevalence regions, the incidence of HPV infection is positively correlated with the incidence of ESCC, suggesting that HPV may be a significant contributing factor to the disease (19,20). For example, a meta-analysis of 26 studies involving 3,429 cases determined that the pooled prevalence of HPV16 infection was 38.1% (95% CI, 28.3-47.9%). The infection rate varied depending on the assay method (such as PCR primer selection) and the sample type (21,22). Conversely, negative results were consistently reported in studies from low-prevalence regions. A cohort study from a low-prevalence area found no cases of oesophageal squamous papilloma associated with squamous cell carcinoma (23), nor did it yield positive HPV RNA in situ hybridisation (ISH) results (24). Similarly, a Japanese study found no HPV DNA in any of the 31 EC samples (25), and a Korean study reported that all 129 oesophageal squamous cancer samples were HPV-negative (26). However, the HPV positivity rate was 9.6% (5/52 cases) in a Turkish study, with HPV type 39 being predominant (27). This heterogeneity underscores the need for standardized detection protocols and cautious interpretation of epidemiological data.

Different HPV types may play distinct roles in the pathogenesis of ESCC. HPV16 and HPV18, recognized as high-risk types, are linked to a variety of cancers, including ESCC. Higher frequencies of HPV16 infection have been observed in patients with ESCC and in those with Chagasic megaesophagus (CM)/ESCC (44.5 and 25.0%, respectively), and other high-risk types such as HPV31, HPV45, HPV51, HPV53, HPV56, HPV66, and HPV73 have also been detected (16,28). Additionally, co-infection with multiple HPV types has been documented, which may further increase the risk of cancer (29). Understanding the distribution of various HPV types and their association with cancer risk is essential for developing effective screening and prevention strategies.

Oncogenic mechanisms caused by HPV infection

HPV interacts with host cells in various ways, promoting its oncogenic effects. The primary oncogenic proteins, E6 and E7, disrupt normal cellular functions, particularly by interfering with cell cycle regulation. The E7 protein inhibits the function of retinoblastoma protein by binding to it, thereby inactivating the G1/S phase checkpoint and facilitating progression into the S phase. This process promotes viral replication and proliferation (30). Additionally, E6 contributes to immune evasion by promoting the degradation of p53 protein, inhibiting apoptosis, and enhancing cell survival, thereby creating a favorable environment for persistent viral infection (5,31). Therefore, it is important to distinguish between TP53 gene mutations and p53 protein degradation in ESCC clinical samples, namely, mutations alter the gene sequence to express mutated p53 protein, while HPV E6-mediated degradation affects the level of p53 protein post-translationally. This distinction has implications for both biomarker interpretation and therapeutic targeting. These disruptions in cell cycle regulation foster HPV replication and establish conditions conducive to tumour formation.

Genomic integration of HPV plays a pivotal role in its oncogenic mechanism. This typically occurs in high-risk HPV-infected cells, where the viral genome integrates into the host genome. This integration leads to continuous expression of HPV E6 and E7 proteins, potentially inducing instability and accumulating mutations in the host genome, further promoting carcinogenesis (32). Genes located near the HPV integration site are significantly upregulated, contributing to oncogenic phenotypes. Specific genomic regions associated with HPV integration are linked to key biological processes such as cell-cycle regulation, signal transduction, and DNA repair (33). This discovery underscores the critical role of HPV integration in oesophageal carcinogenesis.

The immune escape mechanism is another key aspect of the oncogenic process of HPV. HPV alters signalling pathways in host cells to promote immune evasion. A key strategy involves downregulating the expression of major histocompatibility complex class I on the cell surface, enabling infected cells to evade immune detection by CD8+ T cells (34). Furthermore, during cancer progression, the HPV oncoprotein E7 significantly reduces the expression of C-X-C motif chemokine ligand 14, a chemokine that recruits natural killer (NK) cells, thereby diminishing NK-cell and T-cell infiltration and promoting tumour progression (35). HPV may also influence the tumour immune microenvironment by altering macrophage polarization to a tumour-promoting phenotype (36). Moreover, HPV E7 can inhibit T-cell activity by downregulating Jumonji C histone demethylase 1B, leading to increased expression of the co-inhibitory molecule cytotoxic T lymphocyte-associated antigen 4, which further aids immune evasion and enhances tumourigenesis (37). This complex immune escape mechanism not only contributes to HPV-induced oncogenesis but also presents novel targets for cancer immunotherapy. Understanding the carcinogenic mechanisms is crucial for developing effective detection strategies, as different stages of viral integration and expression may require distinct diagnostic approaches.

Prognostic marker mining with multi-omics approaches

Mass screening and early detection are essential strategies for reducing the morbidity and mortality associated with ESCC. Given that ESCC is often asymptomatic in its early stages, with patients typically diagnosed at later stages, early detection is critical (7). Current screening strategies for EC include endoscopic screening (38), cytological testing, and biomarker testing. With advancements in molecular biology and genomics, researchers are increasingly exploring multi-omics approaches and artificial intelligence (AI)-assisted screening, which can offer more accurate risk assessments and facilitate earlier diagnosis (39).

Genomics

Whole-genome sequencing has revealed driver gene mutation profiles in ESCC, with TP53 mutations being particularly prevalent. These mutations disrupt cell cycle regulation and apoptotic pathways, thereby promoting tumourigenesis and progression. HPV18 E6 inhibits α-ketoglutarate-induced apoptosis in ESCC cells by promoting p53 protein degradation, which in turn downregulates malate dehydrogenase 1 (MDH1). MDH1 downregulates L-2-hydroxyglutaric acid (L-2HG) expression, thus preventing the increase of reactive oxygen species (ROS), as L-2HG is responsible for ROS accumulation (40). A meta-analysis demonstrated that patients with ESCC exhibiting high p53 protein expression had reduced overall survival (OS), regardless of tumour stage (41). Furthermore, the combined effects of HPV16 infection and TP53 mutations can influence ESCC prognosis. Among TP53 mutation-positive patients, those infected with HPV16 showed significantly improved OS compared with uninfected patients (median survival, 57 vs. 27 months), suggesting that HPV16 may serve as a prognostic marker for improved outcomes in this subset (42). Amplification of phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α, cyclin D1, and SH3 and multiple ankyrin repeat domains protein 2 has been linked to tumour aggressiveness (43), while downregulation of chromosome 20 open reading frame 54 was revealed to contribute to the development, metastasis, and poor prognosis of ESCC (44).

Transcriptomics

Analysis of mRNA expression profiles has revealed distinct gene expression patterns associated with patient prognosis. For instance, E6/E7 mRNA induces alterations in cell cycle, proliferation, invasion, metastasis, and other key cancer cell behaviours by modulating downstream tumour-related signalling pathways (45). In HPV-infected patients with ESCC, mRNA expression profiles showed significant upregulation of genes involved in the cell cycle and DNA replication, including cyclin A2, DSN1 component of MIS12 kinetochore complex, and minichromosome maintenance 10 replication initiation factor (46). Additionally, microRNA (miR)-25 was shown to accelerate ESCC progression by directly inhibiting B-cell translocation gene 2 expression, with high miR-25 expression significantly associated with lymph node metastasis (47). Doublecortin-like kinase 1, short isoform promoted ESCC progression by activating the MAPK/ERK/MMP2 signalling axis and inducing epithelial-mesenchymal transition, presenting a potential prognostic biomarker and therapeutic target, although functional validation is required (48). Expression levels of C-terminal binding protein 2 (CtBP2) and cyclin H/cyclin-dependent kinase 7 were higher in ESCC tissues with lymph node metastasis compared with tissues without, with CtBP2 also shown to promote ESCC proliferation via negative transcriptional regulation of p16 (INK4A). Therefore, targeting the CtBP2 axis may reduce ESCC cell migration and represent a potential novel therapeutic strategy for ESCC (49).

Proteomics

Advancements in proteomics have enabled the identification of differentially expressed proteins in clinical samples, which may play a pivotal role in tumourigenesis and progression. For example, integrin-related kinase (ILK) has been found to be significantly overexpressed in ESCC compared with normal tissues, with its expression levels strongly associated with patient prognosis. This suggests that ILK could serve as an important prognostic marker for ESCC (50), while its therapeutic potential remains to be explored. Additionally, high expression levels of heat shock protein 27 and pyruvate kinase M2 in ESCC are associated with tumour aggressiveness and poor prognosis (51). A multi-omics approach integrating public databases, proteomics, and immunohistochemistry (IHC) has revealed that interferon γ-inducible protein 30 (IFI30) regulates the JNK and P21/P16 pathways, thereby promoting tumourigenesis in ESCC. Therefore, IFI30 may represent a potential novel therapeutic target for ESCC (52), although functional validation in preclinical models and assessment of druggability are required before clinical translation. The programmed death-ligand 1 (PD-L1) antibody (rabbit anti-human PD-L1 monoclonal, 1:25, clone SP142) has demonstrated efficacy when used alongside immune checkpoint inhibitors (53,54). Furthermore, significant progress has been made in constructing protein-protein interaction networks. Through bioinformatics tools, it was found that proteosome 26S subunit, non-ATPase 2 promotes tumour cell proliferation by inhibiting autophagy in ESCC cells, with its expression levels being closely linked to patient prognosis (55).

Epigenomics

DNA methylation alterations affect genes involved in cell cycle regulation, DNA damage repair, and cancer-related signalling pathways in ESCC (56). A study identified 120 genes whose DNA methylation levels were inversely correlated with their mRNA expression, including 16 key genes such as sine oculis homeobox homolog 4, cellular retinoic acid-binding protein 2, and EH domain-containing protein 3, which were validated in ESCC samples. The expression of these genes was significantly associated with OS and disease-free survival (DFS), suggesting their potential as prognostic markers in patients with ESCC (57). Additionally, the methylation status of the paired box 1 (PAX1) gene has been linked to ESCC development, and PAX1 methylation detection has shown high sensitivity and specificity for early ESCC screening (58), making it a potential screening marker. The integration of HPV genomic integration sites with host gene methylation (epigenetic changes) may help predict the risk of invasive cancer. Circulating tumour DNA (ctDNA) mutations in TP53 and methylation of septin 9 and short stature homeobox 2 have been associated with reduced OS, DFS, and progression-free survival in individuals with HPV-negative cancer (59). Genome-wide methylation analyses have identified specific DNA methylation biomarkers that can distinguish tumour tissues from normal tissues in ESCC and its associated precancerous lesions (60). In patients with ESCC, hypermethylation is often associated with poorer survival prognosis, providing a novel basis for tumour risk assessment (61,62).

Metabolomics

The identification of characteristic metabolites in ESCC is a key component of metabolomics. Studies have shown that certain amino acid metabolites, such as phenylalanine and tyrosine, are significantly increased, while others, such as L-tryptophan and 5-hydroxyindoleacetic acid, are decreased in the serum of patients with EC, suggesting a close relationship between these metabolites and tumourigenesis and progression (63-65). A cohort study indicated that nicotinamide and phospholipid metabolism were dysregulated in patients with HPV-positive tumours, while HPV-negative tumours exhibited increased purine and pyrimidine metabolism. This metabolism profile conferred an NAD+ dependent survival advantage for the tumour cells, and targeting nicotinamide metabolism could specifically induce apoptosis in HPV+ cancer cells (66). Additionally, specific metabolites, including creatine and adenosine, play a significant role in the immunosuppressive microenvironment of EC, providing a theoretical basis for developing novel immunotherapeutic strategies (67). Alterations in various metabolites in body fluids or tissues are associated with cancer development and progression, and metabolomic analysis may uncover molecular biomarkers closely linked to ESCC development, capable of predicting cancer occurrence, progression, and prognosis.

Immunomics

A stronger immune response in the tumour microenvironment of HPV-positive patients, particularly the infiltration of CD8+ T cells, is associated with a favourable prognosis (68). Moreover, changes in immune-related protein concentrations are strongly associated with ESCC clinical outcomes. The p16 protein, commonly used as a surrogate marker for HPV infection, is closely linked to patient prognosis in EC. IHC expression of p16 shows a high positivity rate in patients with ESCC, especially in high-risk groups (19). HPV infection may upregulate host inflammatory factors, such as IL-6 and TNF-α, and detecting these markers through mass spectrometry or ELISA, in combination with HPV-positive results, may enhance the predictive accuracy of prognosis (69,70). Single-cell RNA sequencing has also revealed that specific immune cell subpopulations are significantly associated with tumour prognosis, providing a crucial foundation for the development of personalized immunotherapy regimens (71).

AI plus multi-omics

Gastrointestinal endoscopy exhibits limited sensitivity in detecting HPV-associated oropharyngeal cancer, but it can aid in distinguishing morphological features in ESCC. For instance, HPV-negative ESCC typically presents as flat erythematous lesions, while HPV-positive tumours, such as those with TP53 wild-type, may exhibit an exophytic growth pattern. Combining this with AI analysis may enhance early diagnostic accuracy (72). Radiomics effectively captures radio-typic features of HPV-positive tumours, such as intratumoural heterogeneity and infiltrative border patterns, demonstrating superior predictive performance for HPV-driven subtypes. This positions radiomics as a promising non-invasive biomarker for future HPV-stratified treatments (73). The integration of imaging feature identification, prognostic model development, and multi-omics data advances AI-driven early diagnosis and personalized therapy for ESCC. Currently, no authoritative guidelines recommend incorporating HPV testing into routine EC screening, and the multi-omics integration strategy remains in the preclinical research phase, lacking standardized screening protocols. Investigating the correlation between high-risk HPV subtypes and the geographical distribution of ESCC, developing integration models combining multi-omics data with imaging histology, and validating ctDNA methylation markers are promising research directions.

Methodological advances in HPV detection

In recent years, significant advancements in HPV detection methodologies have greatly enhanced the sensitivity and specificity of tests, providing more effective tools for clinical practice. The primary techniques used for HPV detection include polymerase chain reaction (PCR), IHC, genome sequencing, and liquid biopsy. The key developments in HPV testing methods over time are presented in Fig. 2.

Figure 2 Timeline of the core development of HPV testing methods. The detection methods for HPV testing have gone through four important stages, including early morphological and immunological stage, DNA testing stage, precision typing and RNA testing stage, and rapid portability and multi-omics methods stage. HPV, human papillomavirus; IHC, immunohistochemistry; HC2, Hybrid Capture 2; PCR, polymerase chain reaction; FQ-PCR, fluorescence quantitative PCR; NGS, next-generation sequencing; CRISPR, clustered regularly interspaced short palindromic repeats; AI, artificial intelligence.

Comparative analysis of traditional methods for HPV detection

Histocytological testing is commonly employed for diagnosis, involving the sectioning and staining of biopsy samples, followed by the examination of cellular morphology and tissue structure. This method enables the visualization of pathological features and is widely used in clinical settings. However, histocytological testing is limited in its ability to detect molecular features, particularly in early-stage lesions or poorly differentiated tumours, which can lead to misdiagnosis or missed diagnoses (74).

IHC plays a pivotal role in determining the degree of ESCC differentiation and its association with HPV infection. By using specific antibodies, IHC can detect the expression of key proteins in tumour cells, offering valuable insights into tumour biology and prognosis. For example, the overexpression of the p16 protein is frequently used as a marker for HPV infection, and its presence in ESCC is closely linked to HPV infection (75). Positive p16 expression is associated with a more favourable tumour prognosis, suggesting its potential value as a prognostic marker in clinical practice (76). However, IHC is highly dependent on the specificity and sensitivity of the antibodies used, which can result in false-positive or false-negative results. Additionally, the handling and fixation of the sample can influence the accuracy of IHC results. Therefore, it is important to interpret IHC findings in conjunction with clinicopathological features and other diagnostic tests to ensure greater diagnostic accuracy and reliability.

Molecular biology testing offers precise molecular insights, detecting gene mutations and viral DNA within cells and uncovering the molecular mechanisms behind tumours. ISH enables the direct observation of viral distribution and expression patterns in tissue sections, and the comparisons among three methods for ISH are depicted in Table I. Some studies have shown that HPV DNA-ISH is strongly associated with p16 IHC (77), providing significant pathological evidence for understanding the role of HPV in the development of ESCC (78). PCR and next-generation sequencing (NGS) are currently the two principal molecular biology techniques used to detect HPV and its association with ESCC. PCR, using primers that target specific HPV genes, efficiently amplifies viral DNA for subsequent analysis. For example, PCR has been utilized to detect high-risk HPV types, such as HPV16 and HPV18, in EC samples, with certain studies reporting a high positivity rate (17/101, 16.8%) (79). By contrast, NGS offers a broader genomic analysis, detecting mutations and gene expression profiles simultaneously. A North American study identified rare TP53 gene mutations associated with ESCC through NGS, opening novel avenues for targeted therapies (7).

Table I Multi-dimensional comparisons of three methods for ISH.

Table I

Multi-dimensional comparisons of three methods for ISH.

Method	Signal type	Equipment requirement	Genotyping capability	Tissue localization	(Refs.)
CISH	Permanent color development signal (brown/red)	Optical microscope	Limited (single/double type)	Precise (localized to nucleus)	(105,106)
FISH	Transient fluorescent signal (needs to be stored away from light)	Fluorescence microscopes and image analysis systems	Multi-type (multiprobe required)	Precise (fluorescent signal may bleach)	(107,108)
Broad-spectrum probe ISH	Color developed or fluorescent signal, depending on probe labelling method	Selection of microscope according to probe type (normal or fluorescence)	Multi-type (broad spectrum coverage)	Precise (probe coverage affects localization specificity)	(109,110)

[i] ISH, in situ hydridisation; CISH, chromogenic in situ hydridisation; FISH, fluorescence in situ hydridisation.

The advantages and limitations of IHC, PCR, and ISH in detecting HPV-related cancers are summarized in Table II. Combining IHC with HPV RNA ISH offers advantages over using IHC or PCR alone, particularly in guiding therapeutic decisions and making prognostic assessments (80). This combination enhances diagnostic sensitivity and specificity, especially in cases of IHC-positive and PCR-negative HPV. Subsequent ISH testing can confirm or rule out active HPV infection, thus reducing the risk of misdiagnosis (81).

Table II Multi-dimensional comparison of main traditional detection methods.

Table II

Multi-dimensional comparison of main traditional detection methods.

Methods	Sensitivity	Specificity	Genotyping capability	Tissue localization	(Refs.)
IHC	Moderate	Variable	None	Precisely locates the infected area	(111-113)
ISH	High (type-specific)	Moderate (dependent on probe design)	Limited (only partial subtypes)	Localizable viral DNA	(113-115)
PCR	Variable	High (type-specific)	Clear typing (multiple primers required)	Unorganized location information	(26,54,116)

[i] IHC, immunohistochemistry; ISH, in situ hydridisation; PCR, polymerase chain reaction.

Progress in novel technologies for HPV detection

Advances in high-throughput screening technologies have significantly facilitated the early detection of ESCC and the discovery of biomarkers. Recently, NGS technology has been widely adopted in cancer research, enabling rapid screening and characterization of HPV subtypes and their correlation with EC across large sample sizes and various specimen types (82,83). This technology can capture α-, β-, and γ-genus HPVs simultaneously, revealing multiple types and rare integrative events, providing a foundation for stratifying high-risk populations (84).

Droplet digital PCR can detect HPV DNA at concentrations as low as 0.1 copies/μl in plasma, making it ideal for early screening and postoperative surveillance, particularly in aggressive subsets of oesophageal adenocarcinomas, such as those with submucosal breakthroughs, where it exhibits a significantly higher detection rate (85). HPV E6/E7 mRNA assays, such as reverse transcription-quantitative PCR, reflect viral transcriptional activity and are more effective at distinguishing between latent and causative infections than DNA-based methods. These assays excel in differentiating latent infection from oncogenic status, with current research suggesting that methods capable of detecting viral transcription should be utilized to enhance the accuracy and sensitivity of testing (86,87). Digital PCR demonstrates 80% sensitivity (28/35) and 97% specificity (29/30) for HPV16 detection (88). Methylation-specific PCR (MSP) can measure the methylation levels of five tumour suppressor genes (RASSF1α, p16 (INK4a), TIMP3, and PCQAP/MED15) in salivary DNA (89). Single-cell sequencing has revealed that HPV genes of different types exhibit distinct integration preferences across various samples and disease stages (90).

The HPV K-mer Index Tversky Estimator is a novel detection algorithm that analyses K-mer data and employs Tversky indexes for DNA and RNA sequences. It offers a fast, sensitive alternative for detecting HPV in macrogenomic and transcriptomic datasets. Recognized as one of the quickest and most accurate methods for identifying HPV genotypes from virtually any NGS data, it also stands out for its simplicity. Furthermore, this method is highly scalable and can be adapted for detecting other microorganisms beyond HPV (91).

Positron emission computed tomography (PET/CT) has been utilized to assess HPV infection status in patients with EC. The sensitivity and specificity of PET/CT for detecting second primary malignant tumours (SPM) were 0.73 (95% CI, 0.49-0.88) and 0.99 (95% CI, 0.98-1.00) (92), respectively. Further subsite analysis indicated that the sensitivity and specificity for oesophageal SPM detection were 0.47 (0.30-0.64) and 0.99 (0.98-1.00), respectively. By combining imaging techniques with molecular biology approaches, researchers have been able to more accurately assess tumour biology and its response to treatment.

The development of novel nanosensors and biosensors is also paving the way for more advanced tools in the early screening of EC (93). Portable HPV detection devices based on ELISA and PCR technologies, enhanced by emerging nanotechnology and biosensor innovations, have significantly improved sensitivity and accuracy (94). Clustered regularly interspaced short palindromic repeats (CRISPR)-based biosensors, widely adopted in both basic and applied research, represent a promising novel approach for nucleic acid detection. Technologies such as CRISPR-associated protein 9 (CRISPR-Cas9), CRISPR-Cas12, and CRISPR-Cas13 show considerable potential for HPV detection (95). A biosensing platform that combines loop-mediated isothermal amplification with electrochemical analysis can differentiate between free and integrated HPV16 types by targeting critical components, including E7 mRNA and E2 viral gene transcripts, which are absent after integration (96). Additionally, electrochemical sensing technology has introduced novel methods for detecting circulating tumour cells, a technique valued for its high sensitivity and rapid detection capabilities (97).

AI and big data hold significant promise in the analysis of HPV test results. AI technology can process and analyze the large volumes of data generated by traditional testing methods, assisting in the automatic identification of potential lesion areas and enabling faster clinical decision-making (98). Big data analytics can also help researchers aggregate and analyze HPV infection data across multiple regions and populations, identifying trends in the prevalence of high-risk HPV types within specific groups (95). Furthermore, AI can be leveraged to develop novel HPV detection platforms that combine biosensors with CRISPR technology. These systems enable the rapid detection of HPV and provide real-time results for field testing, greatly enhancing the convenience and accessibility of testing (99). As bioinformatics and AI technologies continue to evolve, the integration of these tools with big data analysis has the potential to enable personalized medicine approaches in HPV detection and management.

Comparison of methodological advances in HPV testing

A systematic comparison of traditional and novel methods for HPV detection is presented, examining their principles, advantages, disadvantages, potential for multi-omics integration, and clinical applicability (Table III). Traditional techniques, such as PCR, offer high specificity but struggle to differentiate between free and integrated HPV DNA, while p16 IHC is cost-effective yet controversial in terms of specificity (100). Emerging technologies such as NGS can integrate host variant data across multiple genomes, albeit at a high cost, while CRISPR-Cas and microfluidics enable rapid detection. Liquid biopsy provides a non-invasive monitoring method. However, traditional methods require careful handling to avoid issues such as aerosol contamination in PCR, cross-contamination in Hybrid Capture 2, and false negatives due to low viral loads. NGS necessitates specialized bioinformatics analysis and stringent quality control, while the CRISPR-Cas system requires optimization to enhance typing resolution, and liquid biopsies exhibit limited sensitivity in early-stage lesions.

Table III Comparison of HPV detection methods: Traditional vs. novel.

Table III

Comparison of HPV detection methods: Traditional vs. novel.

A, Traditional methods
Method	Principle/Technology	Advantages	Disadvantages	Multi-omics integration potential	(Refs.)
PCR (single gene)	Amplification of HPV DNA	High specificity; detects high-risk HPV types	Cannot distinguish episomal vs. integrated HPV; false negatives at low viral load	Moderate (genomic data; requires validation with other omics)	(117-119)
Hybrid Capture 2	RNA probes hybridize to HPV DNA for chemiluminescent detection	High-throughput; detects 13 high-risk HPV types	No genotyping; risk of cross-contamination	Low (single-omics data)	(102,117)
Immunocytochemistry	Detection of p16 (INK4a) overexpression via antibodies	Associated with HPV oncogenic activity; low cost	Limited specificity (p16 can be upregulated in non-HPV lesions)	Moderate (protein-level data; combinable with genomic assays)	(115,120)
B, Novel methods
Next-generation sequencing	Whole-genome or targeted sequencing of HPV DNA/RNA	Detects HPV integration sites, mutations, and co-infections	High cost; complex bioinformatics analysis	High (integrates genomic, epigenomic, and host variation data)	(83,121)
Digital PCR	Absolute quantification of HPV DNA/RNA at single-molecule level	High sensitivity; precise viral load measurement	Expensive equipment; limited multiplexing capability	Moderate (quantitative data supports multi-omics validation)	(122-124)
CRISPR-Cas detection	CRISPR-Cas12a/9/3 enzymes target HPV DNA/RNA for signal amplification	Rapid (<1 h); portable; high specificity	Limited genotyping resolution; requires optimization	Moderate (requires protein-level validation)	(99,125,126)
Methylation-specific PCR	Detection of methylated HPV DNA (L1/L2) or host genes (FAM19A4/miR124-2)	Predicts progression to cancer; high specificity	Requires bisulfite conversion; variable thresholds across populations	High (epigenomic data enhances genomic/transcriptomic models)	(89,127,128)
Loop-mediated isothermal amplification	Isothermal amplification of HPV DNA with visual readout	Rapid; no specialized equipment needed	Limited genotyping; risk of primer dimerisation	Low (primarily genomic data)	(96,129)
Microfluidic biosensors	HPV DNA/RNA detection via electrochemical or optical signals on miniaturised chips	Portable; real-time results; low sample volume	Limited clinical validation; low multiplexing	Moderate (compatible with proteomic/metabolomic integration)	(130-132)
Liquid biopsy (ctDNA)	Detection of HPV DNA in ctDNA	Non-invasive; monitors treatment response	Low sensitivity in early-stage lesions	High (integrates genomic and transcriptomic profiling)	(59,133)
Single-cell sequencing	HPV integration and host transcriptome profiling at single-cell resolution	Reveals tumour heterogeneity and clonal evolution	Technically challenging; high cost	High (multi-omics data at cellular resolution)	(90,134)
AI-assisted cytology	Automated analysis of Pap smear images using deep learning	Reduces human error; improves screening speed	Requires large training datasets; high initial setup cost	Moderate (correlates with genomic/proteomic data)	(135-137)

[i] HPV, human papillomavirus; PCR, polymerase chain reaction; CRISPR-Cas, clustered regularly interspaced short palindromic repeats-CRISPR-associated proteins; ctDNA, circulating tumour DNA; AI, artificial intelligence.

AI-enhanced cytology improves screening efficiency but depends heavily on big data training. Emerging technologies significantly surpass traditional methods in terms of sensitivity and the integration of multi-omics data, particularly genomic, epigenomic, and single-cell data. However, clinical validation and cost constraints remain major challenges. To address these, a combined multi-omics approach, such as integrating DNA methylation testing with traditional HPV DNA testing and cytology, could comprehensively guide clinical decision-making, optimizing treatment plans and enhancing clinical efficacy by mitigating the limitations of individual technologies. Although not extensively discussed in the literature, HPV infection may induce aberrant methylation of host genes (such as p16 and adenomatous polyposis coli) via E6/E7 proteins. Combining this with MSP or pyrophosphate sequencing could lead to the development of an 'HPV infection + methylation profile' screening model, enhancing specificity and potentially reducing false positives in non-cancerous HPV carriers (22). To facilitate early diagnosis, enhancing the specificity of cancer screening through the combination of HPV DNA typing (genomic) with p16/Ki67 double staining (proteomic) is essential (101-103). Moreover, co-infection of HPV with Merkel cell polyomavirus significantly increases the risk of ESCC (OR=2.45), suggesting that multiplex PCR or macrogenomic sequencing could be used to assess infection-related cancer risk more comprehensively (104). A summary of the principles behind various HPV detection methods is provided in Fig. 3.

Figure 3 Schematic illustration of HPV detection principle. With the development of modern technology, there are multiple methods for HPV detection, mainly targeting the components or expression components of HPV. These methods include genomics, transcriptomics, epigenomics, proteomics, multi-omics, and cross-omics. HPV, human papillomavirus; DAB, 3,3'-diaminobenzidine; dPCR, digital polymerase chain reaction; qPCR, quantitative polymerase chain reaction; LAMP, loop-mediated isothermal amplification; CRISPR, clustered regularly interspaced short palindromic repeats; AI, artificial intelligence.

Conclusions

Based on their translational relevance, the biomarkers involved in this review can be categorized into three distinct groups including diagnostic biomarkers, prognostic biomarkers, and therapeutic biomarkers according to their intended clinical application. The first group comprises diagnostic biomarkers, which aid in the detection or confirmation of ESCC, such as PAX1 methylation for early screening. The second group contains prognostic biomarkers, which provide information on patient outcomes, including survival or recurrence risk, for example TP53 mutation status, p16 immunohistochemical expression, and DNA methylation signatures associated with overall survival. The third group consists of potential therapeutic targets, which are molecules or pathways that may be amenable to pharmacological intervention, such as ILK, IFI30, CtBP2, and components of the MAPK/ERK pathway. It is important to emphasize that diagnostic and prognostic biomarkers have demonstrated clinical utility, while the majority of therapeutic targets remain at a preclinical stage and require further functional validation and drug development efforts before clinical translation can be considered. The integrated application of multi-omics technologies provides a novel perspective for biomarker discovery in ESCC, enhancing our understanding of its complex biological features, particularly in studies related to HPV infection. By combining multi-level data from genomics, transcriptomics, proteomics, and other domains, this approach offers deeper insights into the mechanisms underlying ESCC. However, it is important to acknowledge that the majority of findings discussed in this review derive from retrospective or exploratory studies, these results require prospective validation in well-designed clinical cohorts. The heterogeneity in HPV detection rates across different studies underscores the need for standardized testing protocols before clinical implementation. Future research should focus on the following aspects: Firstly, developing standardized HPV testing methods with clear analytical and clinical validity; secondly, establishing the clinical utility of multi-omics biomarkers using prospective studies; thirdly, exploring the function of potential therapeutic targets in appropriate model systems; finally, investigating the epidemiological characteristics of HPV across different populations and geographical regions, and sources of heterogeneity. Although challenges remain, continued research and technological advances may eventually yield more precise and effective strategies for ESCC management. Currently, the role of HPV in ESCC and the application of multi-omics technologies have uncovered complex mechanisms of tumourigenesis, and they also highlighted avenues for further investigation that may ultimately benefit patients with ESCC.

Availability of data and materials

Not applicable.

Authors' contributions

LZo designed the scope and structure of the review, performed the structured literature searches, critically synthesised and interpreted the findings, and wrote major sections of the manuscript. KL and XZ jointly contributed to the design of the review framework, participated in the critical synthesis and interpretation of findings, and revised key sections of the manuscript. HZ, JZ, JX, XA, XQ, YY and LZh contributed to the structured literature searches, and helped draft and revise specific sections. All authors read and approved the final version of the manuscript. Data authentication is not applicable.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Abbreviations:

Acknowledgements

Not applicable.

Funding

The present review was supported by the National Natural Science Foundation of China (grant no. 82273044), the Science Fund for Distinguished Young Scholars of Fujian Province (grant no. 2021D034), the Science Fund from the Health Commission of Fujian (grant no. 2023GGB04), and the Nantong University Clinical Medicine Special Scientific Research Fund Project (grant no. 2024LZ014).

References