Letter to the Editor: Tumor markers in pancreatic juice: How and when to collect?

Abstract

Pancreatic cancer is regarded as one of the most aggressive tumors, with approximately 80% of affected individuals diagnosed at an advanced stage. Its delayed diagnosis results in a 5-year survival rate of only 11%. Notably, only 20% of patients are eligible for surgical resection at the time of diagnosis. Therefore, a pre-operative assessment is critical for defining prognosis, predicting expected survival and guiding the most appropriate treatment. In their prospective case-control study, Trogrlić et al demonstrated that levels of carbohydrate antigen 19-9 in pancreatic juice, collected during endoscopic retrograde cholangio-pancreatography (ERCP) or during the Whipple procedure, were significantly higher in patients who did not survive beyond 5 years from surgery, than in long term survivors. However, collecting pancreatic juice by ERCP or during surgical procedures may not be ideal due to potential procedure-related adverse events. Endoscopic ultrasound constitutes a feasible alternative. Accurate timing of pancreatic juice collection is also crucial, as early prognostic evaluation should ideally represent the gold standard. Fecal samples may constitute a promising source for tumor marker identification. However, certain issues, such as protein stability in stools, remain unresolved. Liquid biopsies could serve as a promising diagnostic tool.

Key Words: Pancreatic cancer; Pancreas; Pancreatic juice; Endoscopic retrograde cholangio-pancreatography; Tumor markers

Core Tip: Analysis of pancreatic juice collection is an interesting way to predict prognosis and natural history of pancreatic cancer. In this article, we discuss some key problems, such as the best method for juice sampling and the best timing. Future perspectives have been discussed.

TO THE EDITOR

Pancreatic cancer represents only the 3% of all malignant cancers, but it is considered as one of the most aggressive tumors, since about the 80% of affected individuals are diagnosed at an advanced stage, leading to a 5-year survival rate of 11%. Indeed, only 20% of patients are deemed fit for surgical radical resection at the time of diagnosis[1-3].

This is the reason why a pre-operatory assessment is fundamental to define prognosis, predict expected survival and guide the most suitable treatment. Computed tomography (CT) is the first and most commonly used imaging technique, with a positive predictive value of 45%-79% for non resectable tumors. Also, magnetic resonance imaging had comparable accuracy for such parameters[4-6].

On the other hand, endoscopic ultrasound (EUS) imaging should not only be taken into account for diagnostic, but also for therapeutic purposes. Yang et al[7] analyzed data from 1554 patients across 30 studies, among which only 9 directly compared EUS with CT. The pooled analysis showed that EUS achieved a sensitivity of 72% (95%CI: 67%-77%), while CT demonstrated a 63% (95%CI: 58%-67%) sensitivity. The specificity was 89% for EUS (95%CI: 86%-92%), and 92% for CT (95%CI: 90%-94%). In the subgroup of studies with direct comparison between EUS and CT, the difference became more pronounced: CT showed a lower sensitivity (48%; 95%CI: 40%-56%) when compared with EUS (69%; 95%CI: 61%-77%). Moreover, EUS demonstrated superior visualization of vascular invasion. Li et al[8] conducted a meta-analysis including 726 patients overall evaluating the diagnostic accuracy of EUS for lymph node and vascular staging in patients with in-situ pancreatic cancer. Preoperative EUS results were compared with intra-operative findings and with final histopathological results on surgical specimen. This study reported that EUS sensitivity and specificity of EUS of 72% and 90%, respectively, for defining T1-T2 stage, and 90% and 72% for defining T3-T4 stages.

Pancreatic juice is the most direct biological fluid for the evaluation of pancreatic diseases. It can be analyzed for cytological evaluation and biomarkers assessment. For instance, cytology examination of undetermined pancreatic duct stenosis has shown a sensitivity of 45.0%, a specificity of 84.6%, and an overall accuracy of 67.3% in differentiating benign stenosis from malignant stenosis[9]. Undoubtedly, these results are deceiving and insufficient for routine clinical use, thus encouraging the development of new methods, such as proteomics and transcriptomics[10].

Carbohydrate antigen 19-9 (CA19-9) is considered the most useful prognostic marker for pancreatic cancer in clinical practice. However, serum CA19-9 levels exhibit low sensitivity and specificity, as elevated values can also be detected in benign conditions such as diabetes mellitus, benign common bile dilation, biliary obstruction, acute and chronic pancreatitis[1]. At the same time, elevated CA19-9 levels are related to aggressive forms of pancreatic cancer. De Rosa et al[11] demonstrated that preoperative high levels of CA19-9 and larger tumor diameter were significantly associated with unresectable pancreatic cancer at laparotomy. Furthermore, preoperative CA19-9 levels < 100 U/mL were associated with a longer adjusted median survival and improved 5-year survival rates[12-14]. In their prospective case-control study published in World Journal of Gastrointestinal Oncology, Trogrlić et al[15] found that CA19-9 levels in pancreatic juice, collected either during endoscopic retrograde cholangio-pancreatography (ERCP) or during Whipple procedure, were higher in patients who died within 5 years from surgery, when compared to long term survivors. These findings strengthen CA19-9’s role as a valuable prognostic marker, while also raising questions about the clinical application of pancreatic juice analysis. There is no doubt that pancreatic juice represents the primary source of tumor markers secreted by pancreatic cancers, making it potentially the most sensitive biological fluid for their detection. However, careful consideration must be given to both the method and the timing of juice collection.

In this study, a subset of patients underwent pancreatic juice collection during ERCP procedure. This approach represents the most physiological method to provide pancreatic juice specimens; however particular caution is required due to the risk of procedure-related complications. Indeed, ERCP is an endoscopic technique that allows cannulation of the pancreatic duct, but it is currently reserved almost exclusively for therapeutic purposes. Diagnostic ERCP is no longer recommended due to the significant risk of complications which may be severe, notably acute pancreatitis[16]. As already mentioned, ERCP has an almost exclusive therapeutic role. In the context of pancreatic adenocarcinoma, the most common indication is bile drainage in cases of malignant bile duct stenosis caused by the tumor itself. As a consequence, most of the patients affected by pancreatic cancers undergoing ERCP are often in a locally advanced stage of disease, meaning that ERCP is performed with palliative intent. This limits the utility of pancreatic juice analysis obtained via ERCP for prognostic assessment. EUS with fine needle aspiration (FNA) or fine needle biopsy may be a valuable alternative, since these techniques are usually performed at earlier stages of pancreatic cancer evaluation, both for diagnosis and staging, and have shown promising results. For instance, analysis of TP53 and SMAD4 mutations in FNA-derived pancreatic juice samples revealed a sensitivity of 61.1% and a specificity of 95.7%[17,18]. In another study, the assessment of CA19-9 levels in pancreatic juice obtained via FNA showed an area under the curve of 0.86 with a sensitivity of 69.8% and specificity of 85.4%[19]. On the other hand, some technical issues should be acknowledged when using EUS. While aspiration of fluid from pancreatic cysts is relatively easy to perform and is already standard practice in the evaluation of cystic neoplasms[20], puncture and aspiration from the main pancreatic duct, especially when not dilated, is more cumbersome and is not devoid of possible complications including pancreatitis, bleeding and hemosuccus pancreaticus[21]. Therefore, EUS-guided pancreatic juice collection appears to be a promising approach in selected patients, but its feasibility and safety still require careful evaluation.

Regarding timing, most patients underwent pancreatic juice collection at the time of surgical resection, namely when the tumor has already been deemed suitable for radical surgery. This may represent a potential bias in prognostic analysis. The performance of CA19-9 on surgical samples has been already elucidated. For instance, CA19-9 has been proved to be effective for predicting post-surgical disease-free survival and in monitoring disease recurrence in low-risk patients[22]. The main limitation, however, remains that this approach requires the availability of a surgical specimen. Likewise, pancreatic juice collection during ERCP may indicate that the disease has already reached an advanced stage. A comparison of the suggested methods for pancreatic juice samples is reported in Table 1.

Table 1 Main characteristics, advantages and disadvantages of methods to sample pancreatic juice.

	ERCP	EUS	Surgery	Stools	Liquid biopsy
Invasiveness and adverse events	+++	++	++++	-	-
Current applicability in clinical practice	+	+	+	Only experimental	Only experimental
Timing	Only for therapeutic purposes	At diagnosis	Only for therapeutic purposes	At any time	At any time

ERCP: Endoscopic retrograde cholangio-pancreatography; EUS: Endoscopic ultrasound.

Ideally, an optimal prognostic marker should be able to predict prognosis at an early stage of disease, rather than only when a therapeutic intervention is performed, even though currently such an approach as gold standard is not widely applicable and may represent a future direction. From this perspective, both ERCP and surgery may appear to be suboptimal for pancreatic juice collection. Given that pancreatic secretions are released into the intestinal lumen, it is theoretically plausible that pancreatic proteins and antigens could be detected in stool samples. To the best of our knowledge, no studies have specifically evaluated fecal CA19-9 in pancreatic adenocarcinoma.

Additional evidence comes from genomic and proteomic studies investigating mutations in DNA sequence involved in pancreatic carcinogenesis. Kisiel et al[23] found areas under the receiver operating characteristic curves of 0.73 for methylated BMP3, 0.75 for mutant KRAS, and 0.85 for their combination in stool samples, with methylated BMP3 alone showing a sensitivity of 90%. In another study, KRAS mutations detected in feces demonstrated a sensitivity of 81.8% and a specificity of 81.5% for pancreatic cancer detection; KRAS mutations were detected in the 81.8% of patients[24].

Screening also represents a critical issue, particularly in high-risk populations such as patients with chronic pancreatitis. In these individuals, invasive procedures may be excessively burdensome, making non-invasive approaches highly desirable. In this context, liquid biopsy emerges as a promising alternative[25].

In conclusion, fecal samples may represent an intriguing source for tumor marker identification, although important issues remain, such as the stability of proteins in stools. Pancreatic secretions are rich in cancer-specific proteins, making them a promising tool for identifying novel prognostic biomarkers in pancreatic - including not only CA19-9 but also carcinoembryonic antigen-lactate dehydrogenase, mucins, oncogenes such as KRAS and TP53, and microRNAs[26]. Nevertheless, despite encouraging preliminary data, the field still requires further development.

In this scenario, liquid biopsy may also represent a promising diagnostic tool[27,28]. This approach is based on the detection of circulating tumor cells, circulating tumor DNA (ctDNA), noncoding RNAs, and extracellular vesicles, or exosomes[29]. From a prognostic point of view, the presence of detectable ctDNA in patients with metastatic disease has been associated with significantly worse progression-free and overall survival compared with patients without detectable ctDNA[30]. However, despite some encouraging findings, the path is yet to be paved.

In conclusion, current methods for pancreatic juice collection, namely ERCP and EUS-FNA, are too invasive and are associated with a significant risk of complication, limiting their use in routine clinical practice. Therefore, the development of novel molecular-based techniques using sample collected with non-invasive methods is essential to provide meaningful clinical benefit. Future research should focus on identifying simpler, earlier, safer and more reproducible sampling strategies for prognostic stratification in pancreatic cancer.

Unfortunately, we believe that, at present, routinary analysis of pancreatic juice is not feasible due to procedure-related risks. Only future developments (such as stool proteomics or liquid biopsy) are likely to represent a true breakthrough in prognostic assessment.