Metabolic Patterns in Pancreatic Cancer Could Lead to Earlier Diagnosis

Microscopic image of pancreatic cancer cells - Generated with Google Gemini AI

Pancreatic cancer, specifically pancreatic ductal adenocarcinoma (PDAC), remains a challenge to treat due to its complex biology and resistance mechanisms; drug delivery is also a challenge because of the dense tumor environment.¹ Despite this, breakthroughs in quantifying metabolic signatures may provide a glimmer of hope. Exploratory analysis utilizing quantitative mass spectrometry has uncovered significant metabolic signatures associated with pancreatic cancer.^2,3

This article focuses on metabolomics in pancreatic cancers and is the third and final installment of this series on how metabolomics may potentially transform cancer treatment. The previous articles in this series focused on metabolomics in gynecological and breast cancers.

“We crafted a small pilot study, taking 10 known pancreatic cancer plasma samples and comparing them with our entire database of almost 800 others,” Robert A. Nagourney, MD, the study’s lead author, explained. The samples included both cancerous and noncancerous specimens. “We processed these metabolic signatures using machine learning techniques—‘What is a variation on recursive partitioning and amalgamation’—to segregate groups,” he said. Nagourney teaches pharmacology
at the University of California at Irvine, is volunteer associate professor of gynecologic oncology, Department of Obstetrics & Gynecology, UCI School of Medicine, and he is medical and laboratory director at Rational Therapeutics, Inc, in Long Beach, California. He also founded Nagourney Cancer Institute.

Metabolomic Study on PDAC

In the study, a cohort control of healthy patients was used as a benchmark to compare patients with metabolic profiles linked to PDAC through blood tests.² The results from Partial Least-Squares Discriminant Analysis showed distinction between the metabolic profiles of patients with PDAC and those of the control patients. This analysis also examined the relationship between specific metabolite ratios and patient survival.

Further investigation into overall survival rates demonstrated correlations between certain metabolic ratios and longevity after diagnosis. For instance, the ratios of the amino acid glycine divided by the phosphatidylcholine (PC) ae 38:2 and the biogenic amine putrescine divided by the PC ae 32:0 emerged as significant predictors of survival outcomes. Analysis revealed a median survival of 15 months, with a mean of 25.8 months in the PDAC cohort. The Kaplan-Meier plot indicated that patients with a C4/C4:1 ratio above 6.87 had an HR of 0.34.

A focused analysis on patients with PDAC who survived longer than 36 months uncovered distinct prognostic metabolic signatures: the ratios Gly/PC ae C38:2 and Putrescine/PC aa C32:0, both of which exhibited perfect accuracy in predicting survival.

A subsequent confirmatory analysis involving 30 patients with PDAC—median age 65.5 years, evenly split between sexes— compared their metabolic data with age- and sex-matched control patients’ data.

“We’re not only able to measure the presence or absence of glutamine or tryptophan, but we can also quantify them using deuterated internal standards, and for the first time we can say, ‘Yes, these [amino acids] are there, and this is how much is there,’ giving us the luxury to develop algorithms,” Nagourney stated.

Clinician’s Perspective

There are huge differences among stages [in pancreatic cancers], but overall, the 5-year survival rate is approximately 12% for all pancreatic cancers, explained Janie Yue Zhang, MD, an assistant professor of medicine at University of Pittsburgh Medical Center in Pennsylvania, in an interview with Targeted Therapies in Oncology. “That 12% refers to long-term survivors, and we do not really know what predicts that someone will become a long-term survivor. Thus, this is an interesting question in medical oncology,” Zhang said.

Zhang added, “The issue with the concept of having a predictive blood test
is that every time you have a test like that, there is a false-positive rate and a false-negative rate.” Zhang highlighted results of a recent study published in the New England Journal of Medicine that assessed a blood test to screen for colorectal cancer.⁴ In a population undergoing average-risk screening, this blood test for cell-free DNA showed 83% sensitivity for detecting colorectal cancer, 90% specificity for identifying advanced neoplasia, and 13% sensitivity for advanced precancerous lesions. Zhang’s point was that these types of blood tests are unlikely to be 100% accurate. “False negatives are particularly dangerous because you may have just told somebody who has cancer that they do not have cancer, and false positives will lead people to receive unnecessary testing,” she explained. Zhang believed this would be a hard sell for use in a community setting due to the potential harm caused to patients by false negatives.

Zhang also addressed the small size of the cohort in the study, noting that validation is needed with a separate cohort completely independent from the first to prove that this biomarker is useful for all pancreatic cancers, not just the 30 people involved in this study. Thus, it is too early to say whether these metabolic signatures are an accurate predictor for PDAC, she said.

A Study Predicting 5-Year Probability

Investigators in another recent study looked at the link between microbial-related metabolites in the bloodstream and the risk of developing pancreatic cancer.³

By leveraging data from participants in a prostate, lung, colorectal, and ovarian (PLCO) cancer cohort, investigators sought to create a metabolomic model that could potentially enhance early risk assessment for pancreatic cancer.

In this study, investigators analyzed serum samples from 172 individuals diagnosed with pancreatic cancer within the past 5 years and 863 matched control participants. Using advanced metabolomics profiling techniques, they identified a panel of 14 microbial-related metabolites.

The model was developed using data from 5 PLCO cancer centers and subsequently validated with samples from 2 additional centers, followed by testing in 3 independent cohorts.

A 3-marker microbial-related metabolite panel demonstrated an area under the curve (AUC) of 0.64 for predicting a 5-year probability of pancreatic cancer. When combined with 5 additional nonmicrobial metabolites, the predictive power improved significantly, achieving an AUC of 0.79. Further enhancing the model, the integration of this metabolite panel with the established tumor marker CA19-9 yielded an impressive AUC of 0.86 for predicting a 2-year risk of pancreatic cancer, surpassing the predictive capability of CA19-9 alone.

Research on metabolomics for detecting and prognosticating pancreatic cancer is very much in its early stages. However, investigators are looking to gather more samples to advance their studies and validate findings with the goal of establishing a reliable biomarker signature that can be clinically actionable.

“Pancreatic cancer is virtually untreatable today, and we must move pancreatic cancer into the realm of treatable cancers,” Nagourney said. “Negotiations are in process with a major university center
to acquire samples, as we will need more samples to validate this for pancreatic cancer,” he concluded.

REFERENCES

1. Fang YT, Yang WW, Niu YR, Sun YK. Recent advances in targeted therapy for pancreatic adenocarcinoma. World J Gastrointest Oncol. 2023;15(4):571-595. doi:10.4251/wjgo. v15.i4.571

2. D’Amora P, Silva IDCG, Evans SS, et al. Diagnostic
and prognostic performance of metabolic signatures in pancreatic ductal adenocarcinoma: the clinical application of quantitative nextgen mass spectrometry. Metabolites. 2024;14(3):148. doi:10.3390/metabo14030148
3. Fahrmann JF, Irajizad E, Kenney A, et al. Contribution of the microbiome to a metabolomic signature predictive of risk for pancreatic cancer. Cancer Prev Res. 2023;16(suppl 1):P076. doi:10.1158/1940-6215.PrecPrev22-P076
4. Chung DC, Gray DM 2nd, Singh H, et al. A cell-free DNA blood-based test for colorectal cancer screening. N Engl J Med. 2024;390(11):973-983. doi:10.1056/nejmoa2304714