Early detection of pancreatic cancer by evaluating combined serum biomarker: The key questions

Abstract

Rapid progression, late diagnosis, and poor prognosis are the factors that make pancreatic cancer highly lethal. Due to the lack of sensitive noninvasive screening tools, early detection is becoming challenging. This study demonstrates the diagnostic potential of the classical tumor marker carbohydrate antigen 19-9 with newer inflammatory markers (carcinoembryonic antigen, soluble intercellular adhesion molecule-1, chitinase-3-like protein-1) in combination in 51 pancreatic cancer patients and 51 matched healthy controls. It suggests potential utility in early detection by the multi-marker panel that stated higher sensitivity and specificity than single biomarkers. Limitations include small sample size, single-center design, absence of independent validation, and use of healthy rather than clinically relevant disease controls. There was no search of relationships between biomarkers and pre-analytical variability. These outcomes state that non-invasive early detection of pancreatic cancer may be improved by a combined biomarker approach. For confirming these results and guiding clinical application, larger, multi center, prospective studies are needed.

Key Words: Serum markers; Pancreatic cancer; Early diagnosis; Biomarkers; Carbohydrate antigen 19-9

Core Tip: For survival, early diagnosis of pancreatic cancer is crucial. However, it remains difficult due to its vague clinical presentation. In the early detection of pancreatic cancer, this study investigates the clinical utility of serum tumor markers, including carbohydrate antigen 19-9 and carcinoembryonic antigen. The diagnostic accuracy can be enhanced by combined analysis of these biomarkers which helps in timely intervention and improving patient outcomes.

INTRODUCTION

Pancreatic cancer being deadliest malignant disease marked by diagnostic delay, fast progression and poor prognosis. The five-year survival rate remains below 10% worldwide despite advancement in imaging and molecular diagnostics. It further reduces to 3% for patients with advanced metastatic disease. Serum tumor markers such as carbohydrate antigen 19-9 (CA19-9), carcinoembryonic antigen (CEA), chitinase-3-like protein-1 (CHI3 L1) and soluble intercellular adhesion molecule-1 (sICAM-1) are markedly elevated in malignant pancreatic carcinoma. It is due to lack of sensitive early screening tools and late-stage diagnosis[1]. The most common risk factor for pancreatic cancer is smoking as it is the cause of 20%-25% of all pancreatic tumors. Alcohol is a cause of chronic pancreatitis, which is a probable risk factor. Diabetes mellitus also increases the incidence of pancreatic cancer. Germline disorder is present in 5%-10% of patients with pancreatic cancer, while the others are thought to be caused by somatic mutations[2]. The four major mutated genes that are responsible for pancreatic cancer are KRAS, CDKN2A, TP53, and SMAD4. Primary events in development of pancreatic tumors include KRAS mutation and alterations in CDKN2A. High diagnostic ability for pancreatic cancer is provided by fine-needle aspiration guided-endoscopic ultrasonography (EUS). The only potentially curative treatment for pancreatic cancer is its surgical resection. After surgery, adjuvant chemotherapy with gemcitabine or S-1, an oral fluoropyrimidine derivative, is given[3,4]. Only localized malignant pancreatic cancer offers a chance to cure by surgical resection but 80%-85% of patients present with advanced unresectable disease which is even unresponsive to most chemotherapeutic agents[5,6].

WHAT ARE EARLY DETECTION METHODS FOR PANCREATIC CANCER

Many articles address an important clinical problem which is the need for precise, non-invasive methods for detection of pancreatic tumors. These are necessary for improving the survival of patients. There is marked diagnostic significance combining classical tumor markers (CA19-9) with newer inflammatory markers such as CEA, sICAM-1 and CHI3 L1 in screening of pancreatic tumors at an early stage. For pancreatic cancer, an overall mean sensitivity of CA19-9 is 81% and a mean specificity is of 90%. Additionally, baseline characteristics, for example age, gender, body mass index and lifestyle risk factors, used were stable across the groups which further support the internal accuracy for analysis[1,7,8]. In recommended diagnostic and monitoring therapy, CA19-9 kinetics are used in conjunction with radio-imaging computed tomography or EUS. Additionally, the invasive tests used for definitive diagnosis include endoscopic retrograde cholangiopancreatography, laparoscopy and EUS-guided fine needle aspiration[8]. The sensitivity of ¹⁸F-fluorodeoxyglucose positron emission tomography for detecting malignant pancreatic tumors is about 71%-100% and its specificity is almost 64%-90%[9].

WHAT ARE THE RECENT LIMITATIONS WHICH PREVENT THE WIDESPREAD CLINICAL IMPLICATIONS OF PANCREATIC TUMOR MARKERS FOR DIAGNOSIS

The study does not explore the associations among the selected biomarkers. A more comprehensive statistical or biological analysis would help show whether each marker contributes to a distinctive diagnostic value. Without considering their association, the effectiveness of the selected biomarker panel remains ambiguous[1].

The study included just 51 patients with pancreatic cancer and a group of 51 healthy people, which limits test sensitivity and generalizability. Lack of independent cohort validation further reduces the reliability in its reproducibility[1]. CA19-9 can give false negative results. Approximately 5%-10% of Caucasian population produce considerably low levels of CA19-9 because its expression depends on Lewis’s antigen. The false negative cases included patients with either of these three phenotypes: (1) Le^a-b-; (2) Le^a+b-; and (3) Le^a-b+. Thus, research that focuses solely on CA19-9 has the potential to underdiagnose a portion of the pancreatic cancer population[10,11].

CA19-9 can be increased in various types of adenocarcinomas such as advanced gastric carcinomas[12,13]. Total 67% patients with bile duct cancer, 49% patients with hepatocellular cancer, 34% patients with colorectal cancer, 41% patients with gastric cancer and 22% patients with esophageal cancer are also found to have elevated CA19-9 levels[14].

Several benign conditions, such as in patients with biliary tract infection, levels of CA19-9 are more than 1000 ku/L. Other conditions that can cause rise in serum CA19-9 levels include pancreatitis (chronic and acute), cholangitis, liver cirrhosis and obstructive jaundice. Treatment of cholangitis and appropriate decompression of the common bile duct return the elevated levels to normal. Therefore, the levels of CA19-9 should be measured after biliary decompression in patients presenting with obstructive jaundice[1,8,15].

Biomarker reference values can fluctuate across labs and study assays. Variation in storage, sample handling, technical setup, and conditions could interfere with the accuracy of results. However, these pre-analytical and analytical parameters are not discussed in this article[16]. The sensitivity/specificity of biomarkers might be different in people with different environmental exposure and genetic background. There is a requirement of measuring biomarker analysis in diverse populations with varying genetic profiles and environmental influences to confirm its generalizability and wider practical relevance[17].

DO COMBINED SERUM BIOMARKERS OFFER A SUPERIOR RESPONSE OVER SINGLE BIOMARKERS IN DIAGNOSIS OF PANCREATIC CANCER AT AN EARLY STAGE

Ney et al[18] modelling technique outclassed CA19-9, individual biomarkers and indices developed in separating the patients with non-specific symptoms from those with pancreatic ductal adenocarcinoma (PDAC) with prevailing algorithms. It has implications for improving its early detection in individuals at risk.

Blood-based multi-biomarker panels for the diagnosis of PDAC exhibit greater advantage in comparison with single biomarkers across all patient control cohorts. Kane et al[19] in his study demonstrated that the area under the curve (AUC) value for multi-biomarker panels (AUC = 0.898; 95%CI: 0.88-0.91) was greatly better than all single biomarkers (AUC = 0.803; 95%CI: 0.78-0.83; P < 0.0001). Thus, CA19-9 alone exhibits insignificant performance in diagnosis and is less efficient in mixed control cohorts, but after its combination with a panel of other multiple biomarkers, it yields a reliable diagnostic performance.

A panel of C-X-C motif chemokine ligand 10 when integrated with CA19-9 proposes improved diagnostic precision in distinguishing PDAC from benign or inflammatory pancreatic lesions. Additionally, inflammatory biomarkers including interleukin-8 and interleukin-15 exhibited marked diagnostic accuracy in distinguishing PDAC from benign diseases with biliary involvement such as cholangitis[20,21].

HOW CAN EMERGING FUTURE ADVANCES ENHANCE THE EARLY DETECTION AND DIAGNOSTIC APPROACH TO PANCREATIC CANCER

Large scale, prospective, and multi-center techniques are required for confirmation of diagnostic accuracy and prognosis of pancreatic cancer which help in understanding disease progression. Integration of biomarkers with artificial intelligence (AI)-driven system or liquid biopsy methods could increase diagnostic power.

Liquid biopsy can also be done for detection of PDAC. It includes circulating tumor cells, circulating tumor DNA (ctDNA), non-coding RNAs and extracellular vesicles or exosomes[22]. Circulating tumor cells are detectable even at the precancerous lesion stage of PDAC[22,23]. Circulating epithelial cells (CECs) have the ability to detect patients earlier in the disease process because hematogenous dissemination may occur before tumor formation[24-26]. CECs are present in many patients including benign, premalignant, or malignant pancreatic lesions. Thus, the manifestation of CECs in patients reporting with pancreatic lesions neither confirms the pancreatic malignancy nor can it be used as prognostic factor for patients with PDAC[27].

AI technology can promptly detect high-risk population via medical images, clinical examinations, biomarkers, and other attributes, then, can do early screening of pancreatic cancer lesions[14,28]. No universally accepted non-invasive diagnostic model exists for pancreatic cancer. Longitudinal studies examining dynamic changes in these biomarkers could further enhance clinical utility. It is helpful in evaluating disease progression or prognosis. Only healthy people and patients with pancreatic cancer were compared to the study. Diagnostic accuracy is inflated when tested against healthy volunteers rather than clinically relevant controls, like patients with obstructive jaundice or chronic pancreatitis. A spectrum bias is introduced by this study design. The differentiation between malignant pancreatic cancer and benign pancreatic or biliary disorders in clinical cases must be distinguished by biomarkers[29,30].

EVOLVING PROSPECTS IN PANCREATIC CANCER BIOMARKER RESEARCH

The incidence and prevalence of pancreatic cancer is continuously raising worldwide. With these increasing trends, PDAC can emerge as the second most lethal cancer in Western countries by the end of 2030[31,32].

Glypican-1 is a cell-surface proteoglycan which shed into circulating exosomes, specifically enriched on cancer-cell-derived exosomes. It is a highly sensitive and specific biomarker for detection of pancreatic cancer and can even detect cancer at the earliest stages of disease which represents a paradigm shift in exosome-based diagnostics[33].

Detection of resectable pancreatic cancers is also possible by multi-analyte blood tests. It is a combination of protein biomarkers with cell-free DNA mutation analysis. High specificity, which offers a practically feasible non-invasive screening platform to detect resectable pancreatic cancers, is provided by these protein biomarkers[34].

There are three complementary pillars: Serum proteomics, liquid biopsy technologies, and AI-enhanced diagnostic algorithms, in advancements of early detection of cancer and each of which offers incremental diagnostic value when combined[35].

Prospective feasibility studies are another method for detection of subclinical cancers, including pancreatic cancer. It combines cell-free DNA analysis with positron emission tomography-computed tomography imaging. It can detect even before conventional clinical presentation[36].

Serum thrombospondin-2 is a favorable innovative biomarker for the detection of pancreatic cancer. It can also differentiate between PDAC and chronic pancreatitis. This is a significant difference that CA19-9 alone often fails to attain[37].

A real-time molecular window into tumor heterogeneity and clonal dynamics is provided by ctDNA analysis, with potential applications in treatment response evaluation and minimal residual disease detection[38].

New-onset diabetes mellitus has also been prospectively accompanying with a noticeably elevated risk of succeeding in pancreatic cancer diagnosis. It suggests that, in high-risk population, the combination of metabolic biomarkers and serum tumor markers is helpful to enhance early detection[39].

Profiling of serum and pancreatic juice on base of proteomics has been recognized as a spectrum of secreted tumor-associated antigens and oncofetal proteins other than CA19-9[40].

MicroRNA signatures, which involve miR-196a, miR-217, and miR-196b, have explained a diagnostic utility as circulating biomarkers in pancreatic cancer. Its expression profiles discriminate between malignant and benign pancreatic pathology[41].

Stability in the circulation is better offered by exosome-encapsulated microRNAs rather than free miRNAs and for liquid biopsy-based cancer detection panels, it may serve as significantly reproducible analytes[42].

Cross-sectional imaging derived Radiomic features when combined with serum miRNA classifiers has exhibited promise for enhancing risk stratification in patients presenting with unspecified pancreatic lesions[43].

There is a potential of ctDNA and circulating epithelial cell analysis as biomarkers of occult disease in patterns of early hematogenous dissemination in pancreatic cancer, often preceding radiologically detectable tumor formation[44].

Multi-omic panels integrating metabolomic, proteomic, and genomic data are also under active investigation as comprehensive diagnostic frameworks because they can understand the molecular complexity and inter-patient heterogeneity of PDAC[35,45].

DISCUSSION

Dual challenge involved in the diagnostic model of pancreatic cancer: The biological complication of the microenvironment of tumor and the medical emergency imposed by the disease’s prognosis. Although, the CA19-9 is the most broadly used serum tumor biomarker for pancreatic cancer, still it has well-known restrictions in sensitivity and specificity, especially in the setting of benign pancreatic tumors and biliary conditions. The development and authorization of complementary diagnostic techniques of an urgent necessity for early diagnosis[46,47].

The addition of inflammatory markers such as CEA, sICAM-1, and YKL-40 in multi-marker panels shows a practical attitude to concentrating the essential deficiencies of single-marker approaches. Published research constantly explains that multi-marker panels are better than single-marker approach in distinguishing PDAC from benign disease across wide practical control populations[19,48].

A serious methodological observation emerging from a careful evaluation of the current literature is that most biomarker studies have been conducted in single centers, with retrospective designs and restricted sample sizes, consequently limiting the external validity of stated diagnostic performance[49].

The phenomenon of spectrum bias is when diagnostic accuracy is exaggerated by comparing cancer patients against healthy control groups instead of clinically related controls such as those with obstructive jaundice, chronic pancreatitis, or other benign biliary conditions. It is a universal and unappreciated confounder in the present biomarker literature[29,50].

Another serious and commonly ignored confounder in biomarker research is pre-analytical alterability. Substantial systematic errors, into measured biomarker levels, can be introduced by differences in sample collection protocols, freeze-thaw cycles, centrifugation parameters, and storage duration. It threatens reproducibility across associations[51,52].

The serum levels of CA19-9 are known to be greatly elevated even up to two years before a proper medical diagnosis of pancreatic cancer. It highlights the biological window that exists for early detection of PDAC if biomarker surveillance is applied prospectively in suitably recognized high-risk population[53].

Standardization of biobanking protocols and adoption of internationally recognized pre-analytical guidelines are necessary fundamentals for producing reproducible and clinically actionable biomarker data that are proficient in supporting approval and clinical guideline formation[54].

A transformative development in pancreatic cancer diagnostics can be achieved by combining AI and deep learning into biomarker data analysis[55,56].

In spite of their promise, the interpretability of many AI-based diagnostic algorithms remains significant concerns that must be addressed before clinical deployment[56].

From the perspective of molecular oncology, KRAS mutations are present in over 90% of pancreatic adenocarcinomas that can be detected as ctDNA in peripheral blood. These are present months to years before even clinical presentation. It represents an important transformative early detection target if sensitivity of assay can be sufficiently higher[57].

The gold standard test for tissue-based confirmation of pancreatic malignancy is endoscopic ultrasound-guided fine needle aspiration technology. In diagnostic practice, serum tumor markers and radio imaging results give us complementary, rather than competitive information. This information should be analyzed in an integrated algorithmic manner[58].

Recent clinical guidelines published by the American Gastroenterological Association and American College of Gastroenterology suggest systematic surveillance of high-risk population. It includes those individuals with germline mutations in BRCA2, CDKN2A, PALB2, and ATM, and the individuals with familial pancreatic cancer syndromes. It highlights the significance of risk-stratified biomarker screening systems[54,59].

Longitudinal biomarker research that trace active variations in serum tumor markers from baseline through disease progression are critically needed, as most existing evidence is derived from cross-sectional analyses that capture only a single temporal snapshot of tumor biology[46,49].

Eventually, the rational integration of multiple biomarker analytes such as serum proteins, circulating nucleic acids with imaging-derived features, and clinical risk factors are the main gateways for non-invasive pancreatic cancer early detection. The combined and prospectively validated diagnostic frameworks are valid across wide, and multi-racial patient populations[57,60].

CONCLUSION

In conclusion, this study provides us with vital support for the use of combined serum biomarkers (CA19-9, CEA, sICAM-1, CHI3 L1) to improve the diagnosis and statistical power of pancreatic cancer (Figure 1). In early stages of pancreatic cancer, the multi-marker approach is more valuable than single biomarkers because it provides higher sensitivity and specificity. However, the limited sample size of study, retrospective approach, and lack of external verification decrease practicability of findings. Further research should mean reproducing on large-scale and using prospective designs to validate these findings. While admitting its limitations, the study makes an important contribution to the non-invasive cancer diagnostics strategies. It also gives as a potential approach for early detection of pancreatic cancer.

Figure 1 Flow chart for early detection of pancreatic cancer using combined serum biomarker. CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen.