Evolving Biomarkers in Pancreatic Cancer: Path to Prediction

Andrew S. Barbas, MD

A promising push for biomarkers in pancreatic cancer could generate new prognostic and predictive tools for clinicians. However, more research is needed to develop optimal methods for earlier identification of patients with one of the deadliest malignancies.

Pancreatic cancer (PC) is the most lethal tumor malignancy, the fourth leading cause of cancer- related mortality. It is expected to become the second-leading cause of cancer-related mortality in North America and Western Europe by 2030.¹ The average survival rate is 5 months after patients receive a diagnosis.¹ Of the PCs, pancreatic ductal adenocarcinoma (PDAC) is the most common, comprising 90% to 95% of PCs, and is an aggressive malignancy with a poor prognosis.^1,2 According to Andrew S. Barbas, MD, a kidney/pancreas transplant surgeon at Duke Health in Durham, NC, the scarcity of testing that can provide accurate results contributes to diagnoses in the advanced stages of the disease and consequently to poor prognoses.

“The early identification of high-risk patients and patients with PDAC is of utmost importance. In addition, the identification of molecules that are associated with tumor biology, aggressiveness, and metastatic potential is crucial to predict survival and to provide patients with personalized treatment regimens,” Barbas, who also is an assistant professor of surgery at the Duke University School of Medicine, told Targeted Therapies in Oncology™ (TTO).

Ideally, biomarkers should be obtained noninvasively and inexpensively and yield results with high sensitivity (particularly when used in early detection) and specificity.¹ Several biopsy and blood sample–derived biomarkers have been studied for their potential in PDAC diagnosis, prognosis, and prediction. However, many of these have not shown promising sensitivity or specificity,² such as carbohydrate antigen 19-9 (CA19-9).^1,3 Furthermore, in 5% to 10% of patients CA19-9 is undetectable because of a FUT3 mutation. Other carbohydrate antigens have been found to be inferior to CA19-9.¹

“More efficient noninvasive biomarkers are needed to facilitate early-stage diagnosis and interventions. Diagnostic panels that combine known biomarkers seem to be promising cost-effective and time-efficient alternatives to the discovery of newer biomarkers,” Barbas said.

Tanios Bekaii-Saab, MD, FACP, program leader of the Gastrointestinal Cancer Program at Mayo Clinic Cancer Center in Phoenix, Arizona, told TTO that “NRG1 fusions, KRAS G12C, NTRK fusions, MSI-H, and BRAF mutations, among others,” were the more promising of the biomarkers being studied. In terms of biomarkers, the complexity and genetic variation of PC likely will necessitate a panel of biomarkers rather than individual biomarkers.¹ For example, a study of 6-plex immunoassays found that the panels, including targets such as MIC-1, CEACAM-1, and REG-4, were shown to be complementary to CA19-9.³

Current Biomarker Research

At the 2021 American Society of Clinical Oncology (ASCO) Gastrointestinal Cancers Symposium, 2021 ASCO Quality Care Symposium, and ASCO 2021 Annual Meeting, several abstracts were presented on investigations of potential biomarkers in PC, many specifically in PDAC. Several of these abstracts focused on the prognostic capacity of biomarkers, and others focused on predictive capability or a combination of the 2 forms of prediction. Still other studies incorporated biomarkers into a clinical study design or investigated the uptake of biomarker use in clinical practice.

Predictive Biomarker Studies

In one such study, KRAS mutation status was explored via comprehensive genomic profiling (CGP) for patients with localized or locally advanced PC. The patients received neoadjuvant chemotherapy and chemoradiation before potential surgery or all intended therapy (AIT). CGP was performed on metastatic tumor tissue from patients with recurrent or metastatic disease following AIT, or on primary tumor tissue from patients who did not develop recurrent disease or did not have sufficient metastatic tissue. The completion of AIT did not appear to depend upon KRAS wild-type or mutation (P = 1.00), and OS was not associated with KRAS mutations. The patients with G12A, G12L, and G12R mutations all completed AIT.⁴

Hepatocyte growth factor (HGF) has been investigated for its role in the c-MET pathway involved in tumor cell invasion and survival. In this study (NCT01821729), the plasma levels of HGF were analyzed from patients with PDAC who were enrolled in a phase 2 trial of neoadjuvant FOLFIRINOX (folinic acid, fluorouracil, irinotecan, and oxaliplatin) and losartan prior to chemotherapy and then resection. Circulating HGF levels were sampled starting on the first day of treatment, with the median values used to define high and low levels. Patients with HGF levels in the lower half of the range had an overall survival (OS) rate of 38.4 months and in patients with higher HGF levels, OS was 19.3 months. On both uni- and multivariate analyses, the higher HGF levels were associated with a poorer OS (HR, 2.58; 95% CI, 1.16-5.70; P = .02). The study authors concluded that the circulating HGF is a biomarker for lower OS rates in patients with PDAC in the neoadjuvant setting.⁵

At the ASCO 2021 Annual Meeting, “data [from NCT029129496] were presented on the efficacy of a HER2/HER3 bispecific antibody, zenocutuzumab, in 10 patients with KRAS wild-type pancreatic cancer whose tumors harbored an NRG1 translocation,” James M. Cleary, MD, PhD, a coinvestigator on the trial, wrote in an email to TTO.

“Encouragingly, the objective response rate to zenocutuzumab in this population was 40%. This result reinforces the idea that KRAS wildtype pancreatic cancer is a molecular subpopulation that can sometimes benefit from targeted therapy,” noted Cleary, who is a senior physician at Dana-Farber Cancer Institute and an assistant professor of medicine at Harvard Medical School in Boston, Massachusetts.

Prognostic Biomarker Studies

In a retrospective analysis, blood samples from patients with gastrointestinal cancers were taken before, during, and after radiation therapy (RT) to analyze T-cell receptor (TCR) repertoire in relation to the occurrence of grade 4 lymphopenia during RT. The study found that a better OS occurred in patients with a higher TCR repertoire from the period starting at the absolute lymphocyte count nadir to 2 months after RT completion. Compared with patients who had a decreased TCR repertoire following RT, the patients with a higher TCR repertoire had an improved 3-year OS (22% vs 75%, respectively; P = .026). Hence, TCR repertoire after RT was found to be an independent prognostic factor for improved OS.⁷

In the retrospective COMPASS analysis (NCT02750657), OS and response to platinum-based therapy were studied in patients with familial pancreatic cancer (FPC), defined as family members with 2 or more first-degree relatives who had/have PDAC. The patients with FPC who received platinum-based therapy had a median OS of 16.9 months compared with a median OS of 9.6 months for the non-FPC patients (HR, 0.46; 95% CI, 0.37-0.58). In terms of treatment response, patients with FPC who received platinum-based therapy had a median OS of 19 months, whereas those who did not receive platinum therapy had a median OS of 15.5 months. Also, FPC and platinum-based therapy appeared to be prognostic on multivariable analysis. The investigators concluded that FPC is prognostic rather than predictive, with respect to platinum response.⁸

A study of patients (n = 50) with early-stage resectable PC investigated methylation markers as a means of potential disease prognosis. Prior to surgery, blood and tumor tissue samples were collected from the patients, and the respective methylation levels were profiled with a PC panel for 80,672 CpG sites. Additionally, methylation blocks were analyzed for a methylation profile. Seven of the blocks, differentially methylated (5 hyper- and 2 hypomethylated), were found to have a significant association with OS. From that information, a model of blocks was devised for OS prediction in patients with PC along with a risk score, using the median as a cutoff. Those patients determined to have a high risk score had a poorer OS compared with those in the lower risk score group (HR, 0.30; P = .0016). The risk score was also found to be associated with OS in the preoperative blood samples. Moreover, the risk score was demonstrated to be an independent prognostic factor after including clinical parameters and persisted across patients with differing CA-125 levels. The study authors concluded that their methylation block–devised risk score has the potential for use in PC prognosis.⁹

Prognostic and Predictive

The SLUG gene has been implicated in epithelial mesenchymal transition via metastasis promotion. In this study, whole transcriptome sequencing, next generation sequencing, and whole exome sequencing were used to analyze pancreatic tumors. Primary and metastatic tumors were grouped by level of SLUG expression. The highest quartile SLUG expression group more frequently had peritoneal -retroperitoneal-omentum metastasis when compared with the lowest quartile group. Liver and lung metastases occurred more frequently with the tumors in the lowest SLUG expression quartile. The study found differences in genetic mutations between the highest and lowest SLUG expression groups. In the lower SLUG expression group, Wnt was more frequently expressed. There were also differences between the SLUG groups in ATM and APC mutations. Additionally, the level of PD-L1 expression was higher in the highest SLUG quartile. Hence, the study showed that SLUG analysis could be used to characterize tumor progression, particularly metastasis, with potentially prognostic and predictive information.¹⁰

Other Investigations of Biomarkers and Biomarker Use

Biomarker testing in line with the National Comprehensive Cancer Network (NCCN) guidelines was evaluated across several cancers, including PC. The most common reason physicians gave for not pursuing bio-marker testing per NCCN guidelines was that testing was thought to offer no benefit or served no need. The study authors concluded that biomarker testing rates may improve with the approvals of targeted agents for patients with specific biomarker profiles.¹¹

“In addition to obtaining germline testing, I also order next generation sequencing (NGS) on the tumor (somatic DNA testing) on all of my patients with advanced pancreatic cancer. Although the vast majority of pancreatic cancer patients (approximately 95%) have a KRAS mutation, in the KRAS wild-type popu-lation we have observed targetable alterations such as ALK, ROS1, NRG1, and NTRK trans-locations. Because we’ve observed actionable alterations in KRAS wild-type pancreatic cancer patients, I try to molecularly characterize these patients as much as possible. If an NGS assay indicates a patient is KRAS wild-type, but does not show a clear driver mutation, I order an RNA-based sequencing test because these RNA-based assays can sometimes reveal translocation events that are missed by DNA-based NGS assays,” Cleary said.

“My hope is that the FDA approval of olaparib ushers in the beginning of the molecular treatment era of pancreatic adenocarcinoma,” Cleary said.¹² “Ongoing preclinical research is being conducted to identify molecularly defined populations of pancreatic cancer that are vulnerable to molecularly guided therapies. Importantly, funding organizations are partnering preclinical and clinical investigators so that these laboratory advances are quickly incorporated in clinical trials.”

Because of the urgency of identifying biomarkers in the prognosis and treatment of PDAC, biomarker research has been incorporated into the design of a phase 1 clinical trial of emerging therapeutic eryaspase (red blood cell-encapsulated asparaginase) in combination with modified FOLFIRINOX in patients with local or metastatic pancreatic adenocarcinoma (NCT04292743).¹³

Most of these studies concluded with the observation that further research is needed. “We are only at the top of the iceberg and more work is needed to identify additional targets,” Bekaii-Saab said.

However, “one of the biggest challenges in testing in pancreatic cancer is lack of sufficient tissue. Thankfully, liquid biopsy platforms have gained utility and can be used as an alternative in many instances,“ Bekaii-Saab said.

Similarly, Barbas observed that “multi-institutional collaborations that provide adequate sample sizes are essential in the evaluation of any novel biomarker or diagnostic panel.”

_REFERENCES:

_{1. Al-Shaheri FN, Alhamdani MSS, Bauer AS, et al. Blood biomark-ers for differential diagnosis and early detection of pancreatic cancer. Cancer Treat Rev. 2021;96:102193. doi:10.1016/j.ctrv.2021.102193}

_{2. Giannis D, Moris D, Barbas AS. Diagnostic, predictive and prognostic molecular biomarkers in pancreatic cancer: an overview for clinicians. Cancers (Basel). 2021;13(5):1071. doi:10.3390/cancers13051071}

_{3. Song J, Sokoll LJ, Pasay JJ, et al. Identification of serum biomarker pan-els for the early detection of pancreatic cancer. Cancer Epidemiol Biomarkers Prev. 2019;28(1):174-182. doi:10.1158/1055-9965.EPI-18-0483}

_{4. Shreenivas AV, Annunzio K, Kamgar M, et al. Impact of KRAS alter-ations in localized pancreatic cancer (PC). J Clin Oncol. 2021;39(suppl 3):431. doi:10.1200/JCO.2021.39.3_suppl.431}

_{5. Roberts HJ, Wo JYL, Yeap BY. The use of elevated circulating he-patocyte growth factor (HGF) level as a potential prognostic biomarker in locally advanced pancreatic cancer. J Clin Oncol. 2021;39(suppl 3):429. doi:10.1200/JCO.2021.39.3_suppl.429}

_{6. Schram AM, O’Reilly EM, O’Kane GM, et al. Efficacy and safety of zenocutuzumab in advanced pancreas cancer and other solid tu-mors harboring NRG1 fusions. J Clin Oncol. 2021; 39(suppl 15):3003. doi:10.1200/JCO.2021.39.15_suppl.3003}

_{7. Kim DW, Heather J, Lee G, et al. Increased T-cell receptor rep-ertoire diversity to predict better overall survival in gastrointestinal malignancies. J Clin Oncol. 2021;39(suppl 3):474. doi:10.1200/JCO.2021.39.3_suppl.474}

_{8. Kelly D, Borgida A, Perera S, et al. Familial pancreatic cancer (FPC) status as a clinical biomarker in patients receiving platinum-based systemic therapy: a case-control analysis. J Clin Oncol. 2021;39(suppl 3):430. doi:10.1200/JCO.2021.39.3_suppl.430}

_{9. Zhou J, Tang Y, Li Y, et al. Development of DNA methylation markers for predicting the prognosis of pancreatic cancer. J Clin Oncol. Published online May 28, 2021. doi:10.1200/JCO.2021.39.15_suppl.e16216}

_{10. Kawanishi N, Baca Y, Xiu J, et al. Molecular characterization of pancreatic cancers as seen in the SLUG gene revealing cancer progression. J Clin Oncol. 2021;39(suppl 3):433. doi:10.1200/JCO.2021.39.3_suppl.433}

_{11. Goldfarb SD, McLaurin KK, McAneny BL, et al. Biomarker testing patterns and identification of barriers to testing for homologous re-combination deficiencies across four advanced-stage solid tumors in a multicommunity practice setting. J Clin Oncol. 2021;39(suppl 28):16. doi:10.1200/JCO.2020.39.28_suppl.016}

_{12. FDA approves olaparib for gBRCAm metastatic pancreatic ade-nocarcinoma. FDA. News release. Updated December 30, 2019. Ac-cessed October 12, 2021. https://bit.ly/3FFjrqO13. Noel MS, Philip PA, Tejani MA, et al. A phase I dose escalation study of eryaspase in combination with modified FOLFIRINOX in locally advanced and metastatic pancreatic ductal adenocarcinoma. J Clin Oncol. Published online January 22, 2021. doi:10.1200/JCO.2021.39.3_suppl.TPS453}