Exploring the Nuances of Postoperative Considerations in Pancreatic Cancer

Syed Nabeel Zafar, MD, MPH

After I elucidate the intricacies and potential risks of the Whipple procedure to my patients with pancreatic cancer, they invariably pose the crucial question, “What can I expect afterward?” Their query underscores the primary aim of this surgical intervention—to extend life. Regrettably, I often find myself without a definitive answer. Despite remarkable advancements in cancer care, we remain uncertain about the postoperative survival rates for patients with pancreatic cancer.

Patients undergoing resection for pancreatic ductal adenocarcinoma (PDAC), which constitutes approximately 90% of pancreatic cancers, exhibit varying survival outcomes.¹ Although average survival figures differ across study data, investigators of 1 study reported that 3-year survival rates ranged from 20% to 34% for patients with successfully resected tumors.2 Notably, the 5-year survival rate for Whipple surgery after pancreatic cancer is 20% to 25%.³ The challenge lies in predicting where an individual patient will fall within this spectrum. Will they experience an early recurrence, nullifying the surgical benefits, or will they become a long-term survivor? The absence of reliable prognostic data compels us to approach most nonmetastatic cases uniformly with a combination of chemotherapy and surgery.

The field of pancreatic cancer prognostication is severely constrained, which is particularly concerning given that pancreatic cancer is poised to become the second leading cause of cancer-related deaths in the United States by 2030.⁴ This research deficit manifests in our clinical approach to patients with pancreatic cancer, which lacks the complexity and individualization seen in breast cancer management, as breast cancer is a field that benefits from extensive research funding. Unlike breast cancer treatments—which are tailored based on various factors such as patient age, tumor characteristics, and molecular profile—pancreatic cancer treatments remain largely homogeneous across stages I through III, with minor variations. This oversimplified approach stems from our inability to effectively stratify patients based on long-term prognosis.

In a recent study, we evaluated the predictive accuracy of existing models that incorporate clinicopathologic factors and biomarkers, such as CA 19-9 levels and tumor stage. Findings from our analysis, conducted across 6 high-volume academic centers in the United States, revealed that these models possess approximately a 60% accuracy rate in predicting survival following pancreatectomy.⁵

Such limited predictive capability carries serious implications. In a retrospective study at our institution examining 66,326 patients undergoing the Whipple procedure across 1500 US cancer centers, data showed that approximately 25% of patients died within 1 year post diagnosis, despite under- going surgery. Given that patients with metastatic disease typically survive for up to 1 year with chemotherapy alone, surgery conferred no survival advantage for a significant subset of patients. Clearly, there is an urgent need for improvement.

Two major obstacles hinder the development of more accurate predictive models: data availability and analytical methodologies employed in previous studies. In our ongoing research, funded by the National Cancer Institute, we are leveraging machine learning techniques to analyze electronic health record (EHR) data from over 80 centers in the United States. Although EHRs offer extensive longitudinal patient information ideal for studying cancer outcomes, they also have limitations. Our current efforts involve incorporating additional biological and socioeconomic variables known to influence outcomes.

Over the past 2 years, in collaboration with University of Wisconsin informatics teams and funded by the Pancreas Cancer Task Force, we have established a unique data registry for patients with PDAC. This repository integrates EHR data from University of Wisconsin Health, socioeconomic data from the American Community Survey, surgical outcomes from the National Surgical Quality Improvement Program, genomic data, and cancer-specific information from the North American Association of Central Cancer Registries. Hosted on a cloud-based platform, the registry facilitates interdisciplinary pancreatic cancer research beyond prognostic modeling.

We are expanding this registry to encompass pathology, radiology, and other pertinent markers, and invite collaborators to join us in address-
ing these research questions. Furthermore, we plan to extend this pioneering registry to multiple academic sites, aiming to create the most comprehensive digital database specific to PDAC.

Our overarching goal is to devise a risk-stratification method capable of identifying distinct disease trajectories in patients with PDAC. This stratification will facilitate more informed treatment decisions, enhance clinician-patient communication, and pave the way for personalized treatment strategies tailored to individual patient needs.

Editor’s Note: This piece features references to unpublished data.

References

1. Stoffel EM, Brand RE, Goggins M. Pancreatic cancer: changing epidemiology and new approaches to risk assessment, early detection, and prevention. Gastroenterology. 2023;164(5):752- 765. doi:10.1053/j.gastro.2023.02.012
2. Spore LM, Dencker EE, Kvanner EA, et al. Perioperative factors associated with survival following surgery for pancreatic cancer – a nationwide analysis of 473 cases from Denmark. BMC Surg. 2024;24(1):76. doi:10.1186/s12893-024-02369-4
3. Whipple surgery for pancreatic cancer. UPMC Hillman Cancer Center. Accessed May 6, 2024. bit.ly/44zuxKT
4. Rosenzweig A. New paper highlights urgent need for pancreatic cancer progress. Pancreatic Cancer Action Network. April 7, 2021. Accessed May 6, 2024. bit.ly/3JQ4tSh
5. Marcinak CT, Parker WF, Parikh AA, et al. Accuracy of models to prognosticate survival after surgery for pancreatic cancer in the era of neoadjuvant therapy. J Surg Oncol. 2023;128(2):280-288. doi:10.1002/jso.27287