Research progress on early diagnostic markers for pancreatic cancer.

Introduction
Pancreatic cancer (PC) is a major digestive system malignancy, accounting for approximately 7% of all cancer-related deaths and representing one of the most aggressive gastrointestinal cancers. The overall 5-year survival rate for PC remains as low as 9%, and PC-related mortality is projected to double by 2060. When stratified by country, PC is the 3rd and 6th leading cause of cancer-related death in the United States and China, respectively, representing a major contributor to cancer mortality in both nations [1, 2]. PC can be roughly classified into pancreatic head cancer and pancreatic body-tail cancer based on its anatomical location. When categorized by histological type, over 90% of PC are pancreatic ductal adenocarcinoma (PDAC). Globally, the incidence and mortality rates of PDAC rank 12th and 6th respectively, accounting for 2.6% of all cancer cases and 4.8% of cancer-related deaths worldwide. More critically, PDAC has become the third leading cause of cancer-related death in the United States and is projected to rise to the second by 2030; in China, PC has attracted widespread attention from researchers and patients due to its extremely poor prognosis, with a 5-year survival rate of less than 10% [3–5]. Based on whether PC is caused by external environmental factors and is intervenable, its risk factors are classified into modifiable and non-modifiable categories. Modifiable factors mainly include smoking, alcohol consumption, obesity, and infections. Non-modifiable factors, on the other hand, include age, family history, genetic susceptibility, and diseases such as diabetes mellitus and chronic pancreatitis, the latter being a well-established major risk factor for PC [6].
However, traditional biomarkers exhibit significant limitations in sensitivity and specificity. While numerous reviews have addressed PC biomarkers, many focus on specific categories or provide broad overviews encompassing therapy and prognosis. This review specifically targets the critical challenge of early diagnosis. Our goal is to offer a comprehensive and up-to-date overview, ranging from established serum markers to the most advanced liquid biopsy technologies. More importantly, we critically evaluate the evolving paradigm of multi-parameter assays and discuss their practical application within integrated screening strategies for high-risk populations—a perspective essential for advancing the field. The overall strategy for early detection, integrating liquid biopsy and imaging, is illustrated in Fig. 1. This narrative review aims to synthesize and critically evaluate the current landscape of biomarkers for early PC detection, with a focus on their integration into multimodal screening strategies.

This integrated strategy begins with identifying high-risk populations. Initial screening is conducted through liquid biopsy, which combines both traditional and novel biomarkers. A positive result from the liquid biopsy prompts confirmatory imaging studies, such as endoscopic ultrasound (EUS). The ultimate goal of this sequential approach is to achieve early diagnosis and enable timely intervention, thereby improving patient prognosis.

Traditional tumor markers

CA 19 − 9 and DUPAN-2
Carbohydrate Antigen 19 − 9 (CA 19 − 9), or sialylated Lewis a antigen (sLea), ranks among the most widely utilized clinical diagnostic markers. It is secreted by exocrine epithelial cells and usually exists on the surface of red blood cells as an important component of glycoproteins and mucins. The U.S. Food and Drug Administration (FDA) has officially approved CA 19 − 9 as the first biomarker for the clinical diagnosis of PDAC [7]. In addition, changes in CA 19 − 9 levels can also be used to assess treatment efficacy and predict patient prognosis. According to a meta-analysis by Ye C, Sadula A, Ren S, et al., CA 19 − 9 is associated with preoperative treatment outcomes, and either normalization of CA 19 − 9 or a reduction of more than 50% from the baseline level is a positive prognostic factor for survival in patients with PC [8]. DUPAN-2 (sialylated Lewis c antigen, sLe-c) is expressed in epithelial cells of the digestive tract, pancreaticobiliary system, and respiratory tract. When detecting DUPAN-2 in serum, the positive rate of this marker is significantly higher in PC, biliary tract cancer, and liver cancer, while it is relatively lower in gastrointestinal tumors such as esophageal cancer, gastric cancer, and colorectal cancer. Of greater clinical significance, serum DUPAN-2 levels remain low in the majority of patients with chronic and acute pancreatitis. This characteristic renders DUPAN-2 a valuable biomarker for differentiating PC from benign pancreatic disorders [9]. From a biological perspective, DUPAN-2—a direct precursor of CA 19 − 9—is also elevated in PC patients, providing a theoretical basis for its use in assessing treatment response and prognosis [10]. However, it is important to note that the clinical application of DUPAN-2 is significantly limited by regional variability; it is primarily used in Japan and has not been adopted as standard practice in Western countries. In this context, the key clinical relevance of DUPAN-2 lies in its role as a specific and viable alternative for Lewis antigen-negative patients, who do not benefit from CA 19 − 9 testing [11]. Although CA 19 − 9 is not suitable for screening due to its low sensitivity and specificity (ranging from 41% to 86% and 30% to 100%, respectively), an elevation in its level indicates disease progression and increased risk of recurrence. Therefore, it remains the most well-established and widely used biomarker for monitoring the progression of PDAC and treatment response [5]. The application of CA 19 − 9 requires attention to the following two points: first, patients negative for Lewis a antigen may have false-negative CA 19 − 9 results; second, CA 19 − 9 can also be elevated in conditions such as biliary obstruction or pancreatitis, leading to false-positive results. Given these limitations, although this marker can be used to assess the resectability, diagnosis, and prognosis of PC, over-reliance on CA 19 − 9 alone should be avoided in clinical practice. Therefore, it is recommended to [perform CA 19 − 9 detection] after the resolution of inflammation or following biliary decompression surgery to avoid false-positive results. Cell Migration-Inducing Protein (CEMIP) is an important tumor-associated specific antigen, and it exhibits significant application value in various aspects of PC, including early diagnosis, determination of metastatic potential, prediction of invasive characteristics, and evaluation of clinical prognosis [12, 13]. Combining CEMIP and CA 19 − 9 for combined detection may enable the establishment of an efficient and accurate screening platform for PC, providing reliable scientific evidence for the early intervention of the disease. Therefore, despite its widespread use, CA 19 − 9 is not suitable for standalone early detection due to its suboptimal sensitivity and specificity, highlighting the necessity for combination strategies. While traditional biomarkers like CA 19 − 9 remain integral to clinical practice, their standalone utility for early detection is severely limited by issues of specificity and sensitivity. The future of PC diagnostics does not lie in discarding these markers, but in strategically integrating them with novel biomarkers within multi-parameter panels to enhance overall diagnostic accuracy and clinical utility.

CA 125 and CEA
Carbohydrate Antigen 125 (CA 125) is a high-molecular-weight mucin-like protein [14], Its antigen is located on MUC16, and elevated serum CA 125 is commonly used as a diagnostic biomarker in approximately 85% of ovarian cancer patients. However, increased levels of this marker are also observed in liver cirrhosis and uterine fibroids, and are associated with pelvic inflammatory disease and colorectal cancer [15]. Its levels are elevated in approximately half of PC patients, and the combined use of CA 125 with CA 19 − 9 increases sensitivity by 6% compared to the use of CA 19 − 9 alone. Therefore, it is not affected by jaundice. However, since CA 125 is produced by serosal epithelial cells, serum CA 125 levels will increase in the presence of serous fluid—a condition that is quite common in patients with PC [16]. Meanwhile, as a transmembrane glycoprotein, CA 125 exhibits high expression in gynecological malignancies such as cervical cancer and ovarian cancer on the one hand; on the other hand, in PC, its serum levels increase significantly with the progression of the TNM staging, and a higher staging is closely associated with poor prognosis in patients [17, 18].
Carcinoembryonic Antigen (CEA) is an acidic glycoprotein, and it is a typical embryonic glycoprotein molecule. The significance of CEA lies in its wide presence in a variety of tumors, such as ovarian cancer, cervical cancer, lung cancer, and breast cancer. Therefore, CEA holds clinical significance for diagnosis and monitoring in the context of various tumors [19, 20]. CEA, the second most common biomarker for diagnosing PDAC, when used in combination with other biomarkers (CA 19 − 9 and CA 125), will improve the diagnostic accuracy for PC patients. Its diagnostic sensitivity is only 44.2%, and the specificity is 84.8% [21]. It was the first tumor marker used for PC detection and plays an important role in the monitoring and prognosis of colorectal cancer. Therefore, due to its low sensitivity, CEA is not clinically suitable for being used alone in the diagnosis of PC [22]. Consequently, CEA serves primarily as an auxiliary marker in multi-parameter panels rather than a standalone diagnostic tool for PC.

CA50
Carbohydrate Antigen 50 (CA 50) is a ganglioside glycoprotein whose epitope is considered similar to that of CA 19 − 9. Although it is not widely recognized as a standalone global marker, CA 50 serves as a valuable biomarker for distinguishing PC from benign pancreatobiliary diseases. A distinctive feature of CA 50 is that, unlike CA 19 − 9, its high expression can also be observed in extraintestinal malignancies [23, 24]. Consequently, its primary potential lies in being part of a multi-marker panel to improve diagnostic accuracy. Overall, while biomarkers such as CA 50, CA 242, and DUPAN-2 provide valuable insights in specific diagnostic contexts, their limited standalone global prominence underscores the necessity of a combinatorial approach, with CA 19 − 9 remaining the central serological benchmark around which new diagnostic panels are developed.

CA242
Carbohydrate Antigen 242 (CA 242) is a sialylated mucin glycoprotein. Although its independent role is limited globally, it holds significant clinical value due to its high specificity. Compared to CA 19 − 9, CA 242 has a lower false-positive rate because its levels typically do not increase significantly in pancreatitis and benign biliary diseases [20, 25]. This characteristic makes it particularly useful for differential diagnosis. The primary utility of CA 242 is realized when it is combined with established markers such as CA 19 − 9, utilizing different monoclonal antibodies to develop a more robust combined detection model [7]. Thus, CA 242 exemplifies how a marker with superior specificity, even if lacking standalone sensitivity, can be strategically employed to enhance the overall reliability of a multi-parameter diagnostic assay.

Novel biomarkers
Liquid biopsy, a non-invasive technique for analyzing blood and other body fluids, has garnered significant attention in recent years due to its diagnostic potential [26]. In the latest research progress, new biomarkers that have been identified to show promise in the early diagnosis of PC include: Circulating Tumor Cells (CTCs), Circulating Tumor DNA (ctDNA), microRNAs (miRNAs), specific protein biomarkers, and metabolism-related biomarkers. All these important new biomarkers can be used for high-efficiency diagnosis of PC through liquid biopsy [27].

CTCs
CTCs are shed from primary or metastatic tumors into the bloodstream. Relevant data show that compared with healthy individuals, the number of CTC subtypes and total concentration in the peripheral blood of patients with PDAC are significantly increased. Detection of CTC surface or intracellular biomarkers—including epithelial markers, mRNA, and DNA mutations—enables real-time cancer monitoring and accurate assessment of tumor genomics [7].
In addition to the traditional strategy of capturing CTCs based on the expression of epithelial cell markers (such as epithelial cell adhesion molecule), emerging studies have expanded to other cell surface markers, including folate receptors and the mesenchymal cell marker vimentin. These novel biomarkers exhibit favorable diagnostic efficacy in distinguishing between PDAC, benign pancreatic diseases, and healthy individuals. Therefore, the combined detection of folate receptor-positive or vimentin-positive CTCs with CA 19 − 9 can further improve the diagnostic efficacy of [28].
Studies have shown that CTCs can be detected through a variety of methods and used as early diagnostic biomarkers, including immunofluorescence, chromosome in situ hybridization (FISH) staining system (NE-iFISH), and immunostaining fluorescence in situ hybridization (SE-iFISH). Among these methods, NE-iFISH exhibits a 90% CTC detection rate in PC patientsSE-iFISH, on the other hand, possesses a high sensitivity of 88% and a high specificity of 90%. Therefore, CTCs are present in the circulation of PC patients, providing useful information for diagnosis, staging, and prognosis, and can even serve as novel personalized therapeutic targets [29].
In addition, CTCs are primarily associated with the spread of tumors to distant sites. Before advanced metastasis occurs in PC, the concentration of CTCs in peripheral blood is usually low. Given that the liver is the most common early metastatic organ for this disease, Chapman and Waxman et al. [30]. Proposed the detection of CTCs in portal vein blood, as intrahepatic metastasis is associated with portal vein CTC count in patients with advanced PC. In patients with PC, since the liver is the first organ to which PC metastasizes, the detection rate of CTCs in the portal vein is significantly higher than that in peripheral blood. The detection of CTCs is technically challenging and not yet standardized for clinical use. It typically requires specialized blood collection tubes designed to preserve nucleic acids and cells (e.g., CellSave® tubes for the FDA-cleared CellSearch® system), followed by enrichment steps. These enrichment strategies can be based on physical properties (e.g., size, density via ISET or microfluidic devices) or biological properties (e.g., immunomagnetic capture using antibodies against epithelial cell adhesion molecule [EpCAM]) [31]. Subsequently, enriched CTCs are identified and characterized using immunocytochemical staining, fluorescence in situ hybridization (FISH), or molecular analysis. Additionally, the combined use of portal vein CTC count and serum CA 19 − 9 level can enhance the diagnostic efficacy of PC. Relevantly, the use of mesenchymal markers—alongside epithelial markers—to isolate CTCs can further improve diagnostic performance [30].
While CTCs hold significant diagnostic and prognostic potential, their low abundance in early-stage disease and the absence of standardized detection protocols continue to pose major challenges for clinical translation.

ctDNA
Circulating tumor DNA (ctDNA) comprises cell-free DNA fragments released by tumor cells, averaging 160 base pairs in length with a short half-life (15 min to 2.5 h). This property allows ctDNA to function as a real-time, dynamic biomarker that accurately reflects the tumor state at specific time points [32]. ctDNA can be released not only through the death and necrosis of tumor cells but also through the active transport of cell membranes [32, 33].
In oncology, ctDNA has two potential applications. First, it can assess the response to treatment regimens through dynamic changes in its levels. Second, it has high sensitivity for detecting residual lesions after treatment or those left behind, enabling early detection of recurrence. It has been confirmed in several cancers that the amount of ctDNA is associated with tumor staging; for example, ctDNA can be detected in 47% of patients with any type of early-stage cancer, whereas it can be detected in 82% of patients with advanced-stage cancer [34].
ctDNA can also identify KRAS and TP53 gene mutations, which are of great significance for evaluating treatment effectiveness, determining drug resistance mechanisms, and guiding treatment adjustments [35]. Beyond these genetic alterations, epigenetic modifications of ctDNA, particularly DNA methylation, have emerged as a highly promising class of biomarkers for early detection. KRAS mutations are present in approximately 90% of PDAC, but the sensitivity of ctDNA sequencing in PDAC needs to be improved to fully capture the expected number of KRAS mutations [36].
Therefore, ctDNA can be distinguished from normal circulating DNA by detecting specific oncogenic mutations of the primary tumor using highly sensitive and specific technologies, such as droplet digital polymerase chain reaction (ddPCR) [37]. Although the sensitivity of ctDNA for the early diagnosis of PC is usually lower than that of CA 19 − 9, its combination with CA 19 − 9 significantly increases the diagnostic sensitivity to 91% [7]. Second, when ctDNA is used in combination with protein biomarkers, it can significantly improve the accuracy of diagnosis, increasing the sensitivity to 64% and the specificity to as high as 99.5% [36].
Cell-free DNA (cfDNA), referred to as circulating cell-free DNA, consists of DNA fragments isolated from the extracellular space that are present in the blood or other body fluids. It is mainly derived from normal cell turnover, cell apoptosis, necrosis, and tumor cells. As a detection carrier, cfDNA can support the detection of various tumor-specific genetic alterations, such as mutations, microsatellite instability (MSI), loss of heterozygosity (LOH), and abnormal methylation [38].
Therefore, both CTCs and circulating tumor DNA (ctDNA) can serve as real-time biomarkers, and their combination with folate receptor-positive (FR+) CTCs, vimentin-positive (Vim+) CTCs, or CA 19 − 9 will improve diagnostic sensitivity. Beyond the genetic sequence, the epigenetic dimension of ctDNA—particularly DNA methylation—holds immense promise for early cancer detection and addresses the challenge of low mutation abundance. DNA methylation involves the covalent addition of a methyl group to cytosine residues in CpG dinucleotides, catalyzed by DNA methyltransferases (DNMTs) [39]. In cancer, this process becomes dysregulated, resulting in global hypomethylation and site-specific hypermethylation of tumor suppressor gene promoters [40]. These aberrant methylation patterns are stable, cancer-specific hallmarks that provide an abundant epigenetic signal and can be robustly detected in ctDNA, complementing mutation-based analyzes [41]. For example, hypermethylation of specific cfDNA promoters in plasma or serum has shown potential as a prognostic marker detectable across all stages of PDAC [42]. The integration of mutation and methylation profiling in ctDNA panels represents a powerful strategy to enhance sensitivity and specificity for early PDAC detection.
The primary challenge for ctDNA in early detection remains the biological scarcity of the analyte itself, as tumor DNA shedding is often minimal at stages when intervention is most effective [43]. This fundamental limitation necessitates ongoing innovation in ultra-sensitive detection technologies to capture these elusive genetic and epigenetic signals.

RNA biomarkers
Numerous emerging RNA biomarkers, such as messenger RNA (mRNA), long non-coding RNA (lncRNA), microRNA (miRNA), and circular RNA (circRNA), not only participate in the process of tumor development but also demonstrate the ability to serve as biomarkers for early cancer diagnosis. Relevant studies have indicated that members of the miR family, such as miR-21, miR-10b, miR-221, miR-210, miR-155, and miR-196a, serve as clear evidence—their expression levels are significantly upregulated in PC tissues and can be detected in early lesion areas. In contrast, miR-34a/b, let-7, miR-146a, and miR-126 exhibit downregulated characteristics, which suggests they may be involved in core oncogenic signaling pathways [44]. In the tumor microenvironment, non-coding RNAs can be distributed between cancer cells, stromal cells, and immune cells via small extracellular vesicles (sEVs), thereby promoting intercellular communication and influencing key cancer hallmarks, such as angiogenesis, immune system evasion, and metastatic dissemination. Studies have found that sEVs are rich in a variety of RNAs, including a large number of mRNAs and miRNAs, as well as lipid components of the phospholipid bilayer, such as cholesterol, phosphatidylcholine, and diglyceride. Among them, sEV-derived miRNA-21 has long been one of the most commonly used biomarkers for diagnosing PDAC, with a diagnostic accuracy of approximately 82.6% or 83% when used as a single serum sEV biomarker. When miRNA-21 is used in combination with miRNA-210 or CA 19 − 9, the diagnostic accuracy can be increased to 90% [45, 46]. Among them, Li et al. identified a novel piRNA, namely piR-162,725, which can significantly improve the detection rate of PDAC patients when used in combination with the serum marker CA 19 − 9. Therefore, in addition to serving as diagnostic biomarkers, sEV-ncRNAs are also key mediators of cancer progression [45, 46]. PiRNAs belong to small non-coding RNAs (sncRNAs), and sncRNAs exhibit complex roles in the pathogenesis of PDAC, including regulating key cellular processes such as proliferation, apoptosis, and metastasis. Additionally, a potential mechanism of sncRNAs is alternative splicing, which is regulated by tissue type. Splicing is a rapid response process; in cases of chemoresistance, it is a major factor in drug response. Therefore, sncRNAs hold potential as both promising diagnostic biomarkers and targets for targeted therapy [47]. In addition, miRNAs are not only closely associated with tumorigenesis and progression but also linked to drug resistance, tumor metastasis, angiogenesis, cancer recurrence, and poor clinical outcomes [48]. miRNAs belong to non-coding RNAs and are mainly involved in the post-transcriptional regulation of gene expression. Due to their high stability and frequent enrichment in serum exosomes, this characteristic makes them of great significance for the early detection of certain malignant tumors. In fact, a study by Madhavan and his colleagues found that many miRNAs, such as miR-1246, miR-3976, miR-4306, and miR-4644, were significantly highly expressed in the serum exosomes of 83% of PC patients. When distinguishing between malignant pancreatic lesions, benign pancreatic cysts, and normal pancreatic tissue, the sensitivity and specificity reached as high as 81% and 94%, respectively. Additionally, compared with serum exosomes from healthy donors, over 95% of the serum exosomes in the study were derived from patients with PC [49].
Circular RNAs (circRNAs) are a class of closed-loop RNAs, where the connection between their 3’ ends and 5’ ends is generated by the splicing of their precursor mRNAs (pre-mRNAs). CircRNAs not only play a central role in the pathogenesis of PDAC but also act as important participants in other tumors, such as esophageal squamous cell carcinoma, hepatocellular carcinoma, and gastric cancer. When the combination of circRNAs is integrated with CA19-9 levels, the diagnostic performance is significantly enhanced, with an AUC value of 0.95, a specificity of 0.96, and a sensitivity of 0.82 [50]. Recent studies have shown that circRNAs associated with PDAC, such as circ_0030235, circPRKD3 (hsa_circ_0000992), circ_0007534, ciRS-7, circRTN4 (hsa_circ_0001006), and circ-IARS, are positively correlated with the metastasis, invasion, and proliferation of PC. These circRNAs may serve as diagnostic biomarkers for PC [51–53].
Dysregulation of long non-coding RNAs (lncRNAs) exhibits oncogenic or tumor-suppressive capabilities and may contribute to the precise diagnosis and personalized treatment of cancers, including PDAC [51]. The field of RNA biomarkers is rapidly expanding and holds tremendous potential. However, transitioning to clinical practice requires rigorous validation of specific RNA panels through large-scale prospective studies, as well as the development of robust, standardized detection assays. For example, HOTAIR, HOTTIP, MALAT-1, and PVT1 have been confirmed to be promising diagnostic and prognostic biomarkers, as well as novel therapeutic targets [51, 54].

Exosomes
Exosomes are double-membraned extracellular vesicles (40–100 nm in diameter) found in various body fluids including blood, saliva, urine, and cerebrospinal fluid [7]. They are small extracellular vesicles secreted by various types of cells and play an important role in intercellular information transmission [55]. Critically, exosomes serve as important carriers of a diverse array of biomarkers. In addition to prproteins, they are enriched with various RNA species, including miRNAs, circRNAs, and lncRNAs, as discussed in the previous section, thereby functioning as a molecular reservoir for liquid biopsy. In recent years, researchers have conducted more in-depth studies on the characteristics and applications of exosomes. For instance, using proteomic techniques, Melo et al. identified a unique surface marker specific to PC cells, namely glypican 1 (GPC1). GPC1 exhibits high expression and high specificity in tumor cells, and it has been confirmed to serve as an important indicator for the early diagnosis of PC. Additionally, its expression level is closely associated with disease progression [56]. In addition, the combination of GPC1 with exosomal GPC1, CD82, and serum CA 19 − 9 levels can significantly improve diagnostic efficacy. Meanwhile, the combination of exosomal markers (EGFR, EpCAM, WNT2, and GPC1) also exhibits excellent diagnostic efficacy [57]. Among these, the combination of tumor mutant genes—Kras, TP53, CDKN2A, and SMAD4—yields higher diagnostic accuracy for PDAC [58]. Castillo et al. identified six PDAC-specific exosomal surface proteins, including CLDN4, EPCAM, CD151, LGALS3BP, HIST2H2BE, and HIST2H2BF. These proteins collectively constitute the tumor-specific surface markers of PDAC exosomes and have been proposed as a potential biomarker panel for diagnosing PDAC [3]. Despite the promising potential of exosomal biomarkers, challenges related to the standardization of isolation methods, accurate quantification, and the precise definition of tumor-specific cargo must be resolved before they can be widely implemented in clinical settings. Novel liquid biopsy biomarkers, including CTCs, ctDNA, and exosomal cargo, represent a paradigm shift toward non-invasive, dynamic monitoring. However, their translation into routine clinical practice is hindered by technological heterogeneity, lack of standardization, and the biological challenge of low analyte abundance in early-stage disease. Their greatest potential likely lies not as standalone tests but as complementary components within integrated diagnostic algorithms.

Specific protein biomarkers
The mucin family comprises high-molecular-weight glycoproteins categorized into transmembrane (e.g., MUC1, MUC4, MUC16) and secretory (e.g., MUC7, MUC8, MUC9, MUC20) subtypes [59]. Overexpression and abnormal glycosylation of MUC1 occur in many cancers, including PDAC. However, in cancer cells, abnormal glycosylation and increased MUC1 expression typically contribute to processes such as tumor invasion, systemic spread to body sites, angiogenesis, and cell death. MUC1 has important clinical significance and has become a key marker for the detection and prognosis of various cancers, especially epithelial adenocarcinomas found in organs such as the lung, liver, colon, breast, pancreas, and ovary [60]. Additionally, MUC1 promotes the expression of multidrug resistance (MDR) genes, including ABCC1 which encodes the MRP1 protein, thereby facilitating the efflux of chemotherapeutic drugs from cancer cells [61]. Therefore, MUC1 can not only serve as a cancer biomarker but also contribute to PDAC carcinogenesis; thus, it may represent an opportunity for developing therapeutic targets, including MUC1-based cancer vaccines for cancer treatment [62]. Although mucins like MUC1 and MUC4 are compelling biomarker candidates due to their overexpression and functional roles in PDAC, their complexity and heterogeneity in glycosylation present significant hurdles for developing reliable immunoassays.
MUC16 is a transmembrane mucin that bears the CA125 epitope on the surface of epithelial cells. It promotes cancer cell proliferation by inhibiting cell death, facilitates cancer cell adhesion and metastasis, and is associated with tumor immune escape. In PC, MUC16 enhances PDAC metastasis through activation by MMP-7, supports disease progression and aggressive subtypes by regulating oncogenic signaling, and is associated with tumorigenesis. Through phosphorylation, CA 125 is released from MUC16 and secreted into the serum [60]. Additionally, MUC16 is involved in PC stem cell enrichment and chemoresistance by regulating JAK2 and Lmo2 [63]. Jonckheere, N. et al. propose that MUC4 is a ligand for ErbB2 and a target of the TGF-β pathway, and human epidermal growth factor receptor 2 (HER2) may play an important role in MUC4-promoted PC progression. This is because HER2, which belongs to the ErbB family of receptor tyrosine kinases, colocalizes with MUC4 on the cell surface and in the cytoplasm [64]. The mechanism is as follows: interfering with RNA silencing of MUC4 through transient or stable expression of MUC4 targets leads to the downregulation of HER2, accompanied by a decrease in its phosphorylated form (pY1248-HER2). This reduces the ability of HER2 to regulate proliferation and metastasis by activating downstream mitogen-activated protein kinase (MAPK) and phosphoinositide-3-kinase/Akt (PI3K/Akt) pathways [64]. MUC4 or MUC5AC is not expressed in the normal pancreas, whereas MUC1, MUC3, and MUC6 are expressed from the normal pancreas to the cancerous state [65].
Therefore, MUC5AC plays two important roles in the diagnosis of PC. First, when MUC5AC is used in combination with CA 19 − 9, serum MUC5AC helps distinguish PDAC from other benign diseases, such as chronic pancreatitis. Second, when MUC5AC is added to CA 19 − 9, both specificity and sensitivity are significantly improved compared with previous results in differentiating cancer from chronic or normal pancreatitis (specificity increased from 43% to 83%, and sensitivity increased from 79% to 83%) [65].
Apolipoproteins (e.g., APOE, APOA2), insulin-like growth factor-binding proteins 2 and 3 (IGFBP-2, IGFBP-3), and osteopontin are of great significance in distinguishing PC from other diseases and serve as potential biomarkers for early-stage PC. Therefore, combining these biomarkers with CA 19 − 9 will improve the detection efficiency [16]. Macrophage inhibitory cytokine-1 (MIC-1) and alcohol dehydrogenase (ADH) have been used for the diagnosis of early-stage PDAC [66]. Studies by Jelski et al. have highlighted the potential of ADH isoenzymes as serological markers for PC. Among the various classes of ADH isoenzymes (I–IV), only class III ADH exhibited significantly higher activity in the serum of PC patients compared to healthy controls or individuals with benign pancreatic diseases [67–69]. This finding suggests that class III ADH may serve as a specific biomarker for distinguishing PC from other conditions. However, further validation in larger, prospective cohorts is necessary to assess its clinical utility in early detection panels. Protein biomarkers such as mucins (MUC1, MUC5AC), apolipoproteins, and enzymes like ADH provide mechanistic insights into tumor biology and show promise for improving diagnostic specificity, particularly in distinguishing PDAC from chronic pancreatitis. Nevertheless, their complexity and heterogeneity—especially in the glycosylation patterns of mucins—pose significant challenges for developing robust and reproducible immunoassays. Clinical implementation will depend on overcoming these analytical obstacles and rigorously validating their additive value within multi-marker panels.

High-risk populations and early screening strategies
Early PC diagnosis remains particularly challenging in the general population, making targeted screening of high-risk individuals a more cost-effective and pragmatic approach. Well-defined high-risk groups primarily include individuals with a family history of PC; carriers of known genetic susceptibility mutations (e.g., in BRCA1/2, PALB2, CDKN2A (p16), or Lynch syndrome-associated genes); those with new-onset diabetes; and patients with chronic pancreatitis [70].

Evolving screening strategies for high-risk populations
Screening strategies for these groups have advanced beyond reliance on single serum biomarkers to multimodal surveillance that integrates imaging and liquid biopsy.

Role of imaging modalities
Endoscopic ultrasound (EUS) is currently the most sensitive imaging technique for detecting small pancreatic lesions and serves as the cornerstone for screening high-risk individuals. It also enables fine-needle aspiration for histological confirmation [71]. In contrast, although PET-CT is highly sensitive for detecting metastases, its sensitivity for small, early-stage localized lesions is limited. Combined with its high cost, PET-CT is generally not recommended as a first-line screening tool.

Advantages and role of liquid biopsy
Frequent imaging is invasive and costly. Liquid biopsy technologies provide a significant advantage in this context by serving as potential “triggers” to initiate or intensify imaging investigations [72].

Applications of biomarkers in screening
Traditional biomarkers, such as CA19-9, exhibit limited sensitivity in this population; however, dynamic changes in their levels may still offer valuable supportive information. Novel liquid biopsy markers hold significant potential.

Circulating tumor DNA (ctDNA)
ctDNA analysis can detect somatic mutations associated with PC, such as those in the KRAS gene, potentially indicating malignancy before radiological detection of a visible lesion. Additionally, it provides valuable insights into the tumor’s molecular characteristics [73].

Exosomes and MiRNAs
These markers provide a non-invasive method for accessing tumor molecular information. Specific exosomal proteins (e.g., GPC1) and miRNA expression profiles (e.g., miR-21, miR-155) represent promising tools for risk stratification and for guiding further imaging studies [74]. Focusing screening efforts on well-defined high-risk populations is a pragmatic and cost-effective strategy. The emerging paradigm of using liquid biopsy as a minimally invasive “trigger” to triage individuals for confirmatory imaging (e.g., EUS) represents the most feasible approach for population-based early detection. However, the successful implementation of this strategy depends on the development of validated, cost-effective multi-analyte panels and clear clinical decision pathways, which have yet to be established.
The landscape of PC biomarkers is rapidly evolving from reliance on single, imperfect serological tests toward a multi-modal, multi-analyte integration paradigm. The critical appraisal presented herein underscores that no single biomarker is sufficient; the future of diagnostics lies in intelligently combined panels. The transition from promising research to clinical impact now faces two major challenges: rigorous prospective validation in relevant cohorts and the development of standardized, scalable detection platforms.

Prioritizing biomarkers for clinical translation
Based on current evidence, biomarkers can be tentatively categorized according to their readiness for clinical implementation: (1) Near-Clinical: CA19-9 (for monitoring), ctDNA for mutation tracking in high-risk groups, and miRNA panels (e.g., miR-21) undergoing ongoing validation [75, 76]; (2) Promising but Requiring Validation: exosomal markers (GPC1, protein panels), CTCs with mesenchymal markers, and circRNA signatures [77, 78]; (3) Exploratory: ADH isoenzymes, VNN1, S100A8, and other novel proteins that require further cohort studies [79]. A hierarchical, multi-parameter approach combining the best-validated markers from each category is therefore most likely to yield the earliest clinically viable screening panels.

Conclusion and future directions
A prospective screening paradigm involves initial, relatively low-frequency liquid biopsy testing (e.g., for ctDNA or exosomal markers), with positive results prompting more invasive but definitive EUS examination [80]. A key future research direction is the development of integrated risk prediction models that combine genetic background, liquid biopsy findings, and imaging data. Through such a dynamic, multimodal strategy, we can realistically aim to achieve true early diagnosis and timely intervention in high-risk individuals, ultimately improving patient outcomes. This integrated, risk-adapted approach represents the most promising frontier in early detection of PC. However, its successful implementation requires the creation of robust risk prediction models and validation of cost-effective screening algorithms through large, prospective clinical trials.

Other biomarkers
Molecules such as VNN1 and S100A8 have shown potential value in distinguishing PDAC from benign conditions or metabolic diseases such as type 2 diabetes [81]. For instance, VNN1 overexpression is associated with oxidative stress in PC cells and may help differentiate PDAC from diabetes [82]. S100A8 has been linked to cancer-associated cachexia [83]. However, these markers remain exploratory and require further validation in early-diagnosis contexts (Table 1).

Conclusions
In summary, the quest for the early detection of PC remains a formidable and largely unmet challenge in oncology. This review has critically synthesized the biomarker landscape, revealing a clear trajectory: the limitations of traditional, standalone biomarkers such as CA19-9 have necessitated a shift toward novel liquid biopsy markers and integrated diagnostic frameworks. While ctDNA, CTCs, exosomal cargo, and non-coding RNAs hold immense theoretical promise, their practical clinical translation is still in its infancy, hindered by issues of sensitivity in early stages, methodological variability, and a lack of standardized, validated panels. Current evidence unequivocally demonstrates that overcoming the complexity of early PDAC will not be achieved by a single “magic bullet” but through rational, evidence-based combinations.
Looking ahead, several key strategies will be critical to overcoming the stalemate in early diagnosis.

Towards multi-omics integrated models
The pursuit of a single “magic bullet” biomarker is likely to be futile. Future research should focus on developing multi-omics diagnostic models that computationally integrate diverse data streams. This approach includes combining genetic alterations (e.g., KRAS mutations in ctDNA), epigenetic markers (e.g., DNA methylation patterns), transcriptomic signatures (e.g., miRNA and circRNA profiles), proteomic data (e.g., exosomal proteins), and established serum markers (e.g., CA19-9). The application of artificial intelligence and machine learning to analyze these complex datasets is poised to unlock powerful, high-accuracy predictive tools capable of identifying early-stage PDAC with unprecedented reliability.

Prospective validation in high-risk cohorts
The promising sensitivity and specificity reported for many novel biomarkers are often based on case-control studies. There is an urgent need to validate these biomarkers in large-scale, prospective, longitudinal cohorts, particularly among asymptomatic high-risk individuals (e.g., those with genetic susceptibility or new-onset diabetes). Such studies are essential to confirm real-world screening performance, establish clinical decision thresholds, and ultimately demonstrate a reduction in PDAC-specific mortality—the true benchmark of success.

Technical standardization and automation
The heterogeneity in methodologies for isolating and analyzing liquid biopsy components (e.g., CTC enrichment platforms, exosome isolation kits, ctDNA assay protocols) remains a significant barrier to clinical adoption. A primary focus should be the development of standardized, automated, and cost-effective platforms to ensure reproducibility and scalability across various clinical laboratories.

Refining the multi-modal screening paradigm
The strategy of using liquid biopsy as a “trigger” for confirmatory imaging, as illustrated in this review, represents the most pragmatic path forward. Future research should focus on optimizing this pipeline by identifying the most effective and cost-efficient combination of liquid biopsy markers to indicate a positive screen, which would then prompt a definitive but more invasive examination with EUS. Establishing evidence-based guidelines for this risk-adapted, multi-modal approach is essential for its implementation in clinical practice.
In conclusion, while the challenge remains formidable, the path forward has become increasingly clear. The convergence of liquid biopsy, multi-omics integration, and AI-driven analytics is fundamentally reshaping the early detection landscape. By steadfastly pursuing these approaches through collaborative, interdisciplinary efforts, we can anticipate a future in which timely diagnosis of PC becomes the norm rather than the exception, ultimately improving the prognosis for this devastating disease. The concerted advancement along these avenues promises to convert the early detection of PC from an insurmountable challenge into a manageable clinical routine, thereby saving lives.