Pancreatic Cancer—Advances in the Last 50 Years

The last 5 decades have witnessed significant paradigm shifts in the management of pancreatic ductal adenocarcinoma (PDAC) borne by a concerted attempt at understanding its pathogenesis, accompanied by improvements in medical imaging, and a multidisciplinary approach comprising surgery, chemotherapy, and immunotherapy, to guide personalized decision-making. These advances have translated into modest, but real, improvements in survival and quality of life in recent decades 1. This special feature article will highlight some of the sentinel events across five domains, namely advances in our understanding of disease pathobiology, imaging, surgery, medical oncology, and pain management. There is a palpable sense of optimism that the knowledge acquired, thus far, will empower us to explore the optimal use of therapies to further improve patient outcomes in the years to come. Imaging plays a pivotal role in the early detection, staging, and management of PDAC. Over the past decades, innovations in imaging modalities and computational tools have transformed the diagnosis and management of PDAC. Multidetector CT (MDCT), introduced in the 1990s, remains the cornerstone for staging due to its high spatial resolution and reproducibility 2. However, modern dual-energy CT (DECT) is now being preferred for PDAC imaging due to its superior image contrast and lower artifacts, allowing visualization of small or isoattenuating tumors and superior assessment of vascular anatomy 3. Ghasempourabadi et al. reported DECT sensitivity of 96% and accuracy of 85% for tumor detection and 96% sensitivity with 95% accuracy for nodal staging 4. Recently introduced photon-counting detector CT (PCCT) further enhances spatial resolution and tumor conspicuity 5 Table 1. MRI, particularly with 3 T magnets and diffusion-weighted imaging (DWI), is valuable for screening high-risk individuals and detecting recurrence. MRI screening in the at-risk subjects demonstrated 85% sensitivity and 90% specificity 6. DWI improves early stage disease evaluation and distinguishes recurrence from post-surgical fibrosis, increasing sensitivity from 61.5% to 88.5% 7. The adoption of structured reporting templates using standardized lexicons, such as those from the Society of Abdominal Radiology, has had a major impact on improving result communication and staging clarity 8. Surgeons found surgical planning information significantly more accessible in 60%–98% of structured reports versus 32%–54% of nonstructured reports and sufficient information in 69%–98% versus 25%–43%, respectively 9. Hybrid imaging modalities like PET/CT and PET/MRI have enhanced functional imaging, aiding in the detection of occult tumors and treatment response assessment. PET/MRI demonstrated 100% sensitivity, 84.6% specificity, and 94.6% accuracy, outperforming PET/CT 10. Novel PET tracers targeting tumor-specific markers (e.g., 68Ga-FAPI) are further improving sensitivity for early disease detection. In the last decade, AI has emerged as a transformative tool in pancreatic imaging. Deep learning models can detect subtle imaging features predictive of malignancy, even when not visible on the scans. AI-based extraction of quantitative radiomics features from the pixel data has the potential to support personalized risk stratification and prognostication. Initial studies show AI achieving 90% sensitivity and 93%–96% specificity in detecting pancreatic cancer on CT 11-13. These imaging innovations contribute to earlier diagnosis, improved staging, and personalized management of PDAC, aligning with broader multidisciplinary advances in pancreatic cancer care. Pancreatectomy is now considered a prerequisite component of the attempt for the cure of pancreatic cancer. Even as recently as the 1970s, however, postoperative mortality rates exceeded 25%, leading prominent surgeons of the time like Crile and Thorbjarnarson to advocate for abandoning the Whipple entirely 14, 15. Shortly thereafter, John Cameron and other pioneers proved that it was possible to perform hundreds of pancreatectomies consecutively without any deaths 16-18. Today, postoperative mortality is less than 4% and approaches 1% at high-volume centers 19. Herein, we celebrate the advances in surgical technique and management that helped usher in this modern era of pancreatic surgery. This progress was initially made possible by the methodical study of the Halstedian technique through randomized trials (Table 2) 37. Many landmark trials have focused on mitigating pancreatic fistula, the Achilles heel of pancreatic surgery, and found that techniques such as ductal stenting, fibrin glue, and pancreaticogastrostomy do not improve postoperative pancreatic fistula rates 20, 24, 28. The prophylactic administration of somatostatin analogs has shown mixed results 31, 38. Other trials have also demonstrated that fistula rates between duct-to-mucosa and invagination pancreaticojejunostomy were equivalent aside from one notable trial which favored invagination 23, 26, 29, 32. A series of trials from Johns Hopkins have also defined the extent of lymphadenectomy after pancreatoduodenectomy 22, 39, 40. Several studies have also shown that classic and pylorus-preserving approaches were largely equivalent with regards to delayed gastric emptying 25, 27, 30. Other trials have also defined standardized pathways for postoperative care: the HYSLAR trial from Thomas Jefferson showed that hypertonic saline as perioperative maintenance fluid decreased complications 33. The WARP trial from the same institution also established an accelerated recovery pathway after the Whipple procedure, with a length-of-stay targeted at 5 days 34. Pancreatectomy is now safer than it has ever been and a short postoperative hospitalization is the norm rather than the exception (Figure 1). Timeline of major developments in pancreatic surgery. JHH, Johns Hopkins Hospital; MIS, minimally invasive surgery; PD, pancreatoduodenectomy. Another major area of improvement is in the progress made by our nonsurgical colleagues in avoiding “failure to rescue”. Post-pancreatectomy hemorrhage was once a rare but often fatal complication 41. Emergent exploration was traditionally the only option and carried a high mortality rate. Nowadays, however, endovascular embolization or stenting (often of the gastroduodenal artery) by interventional radiologists has supplanted laparotomy and substantially reduced mortality 42. Interventional radiologists may also help percutaneously drain abscesses and divert bile leaks. Gastroenterologists are similarly capable of endoscopically stenting biliary or pancreatic leaks 43. Many patients undergoing pancreatectomy are also cared for within specialized surgical intensive care units 44. These multidisciplinary advances have all contributed to the dramatic improvement in outcomes seen over the last 50 years. Lastly, pancreatic surgery also continues to evolve and adapt into minimally invasive techniques. Laparoscopic distal pancreatectomy was adopted first and has been associated with faster recovery with equivalent oncologic outcomes 35. Pancreatoduodenectomy has required the increased dexterity of the robotic platform before being translated into minimally invasive surgery, but robotic Whipple procedures can now be selectively performed by well-trained surgeons with similar morbidity as compared to the traditional open approach 36, 45-49. As the management of pancreatic cancer continues to become more nuanced with more diverse systemic treatments, pancreatic surgery will inevitably develop in tandem. Traditional paradigms for the treatment of localized pancreatic cancer include a surgery first approach for resectable disease. Over the last 20 years, multiple randomized trials have evaluated initially single agent gemcitabine chemotherapy compared to observation following successful surgical resection 50, 51, and more recently, combination chemotherapy regimens, for example, modified(m) FOLFIRINOX (5-fluorouracil, oxaliplatin, irinotecan, leucovorin), and gemcitabine and capecitabine, compared to gemcitabine have resulted in improvement in disease-free and overall survival 52-54. Although a distinct survival advantage for neoadjuvant systemic therapy followed by surgery compared to surgery first and adjuvant therapy has been demonstrated for borderline resectable PDAC 55, 56, the data for upfront resectable disease have been hard to discern 57-59. Multiple randomized trials are underway to address this question; it is very likely that the use of combination chemotherapy in earlier stage PDAC is favorably impacting outcomes by more patients with this disease receiving multimodality therapy, that is, surgery and combination chemotherapy, with the best order in which these therapeutic modalities are delivered being an important point of discussion. Embraced in these improvements in outcome is multidisciplinary discussion and management as part of tumor boards and the access to experienced expertise with this disease 60. Combination chemotherapy regimens with (m) FOLFIRINOX, gemcitabine, nab-paclitaxel, and NALIRIFOX (liposomal irinotecan, leucovorin, oxaliplatin, 5-fluorouracil) each have shown improved overall survival, improved progression free survival, and higher response rates compared to the current prior standard of gemcitabine for the first two regimens and to gemcitabine/nab-paclitaxel in the case of NALIRIFOX 61-63. Overall survival has incrementally increased in the last decade and will continue to have impact as these regimens are now routinely used in the setting of localized disease, both locally advanced and resectable pancreatic cancer. There have been major improvements in understanding the genomic, both somatic and germline, and immune landscape of pancreatic cancer over the last several decades 64. Very specifically, the identification of fundamental oncogenic drivers, including KRAS, TP53, Smad4, and Cdkn2a 65-67 along with other rarer genomic alterations, have provided major insights into the pathobiology and therapeutic opportunities for this disease. About 5%–10% of pancreatic cancers have a germline hereditary predisposition, including BRCA1/2, PALB2, ATM, and the various Lynch genes (MLH1, MSH2 MSH6, PMS2) that may have therapeutic actionability and can lead to the identification of a family at risk for other diseases and hence screening and early detection opportunities 68, 69, although the impact of screening specifically for pancreatic cancer even in genomically enriched subgroups can be challenging to discern and may result in detection at an earlier stage. For germline BRCA1/2, a maintenance poly-ADP-ribose polymerase inhibitor, Olaparib 70, demonstrated significant improvement in progression free survival following platinum-based therapy, compared to observation/placebo in metastatic PDAC. Further, the therapeutic potential of Kras mutations being near ubiquitous in pancreatic cancer and alterations in other parts of the mitogen activated kinase pathway (MAPK) in nearly 95% of individuals with this disease are about to be realized. KRAS inhibitors have entered the clinic, and early promising signals have been identified for the 1% of individuals with pancreatic cancer and Kras G12C mutations, and late-stage clinical trial data are anticipated in 2026, evaluating a panRAS inhibitor which has an activity in the more common KRAS alleles of Kras G12D, G12V, and G12R compared to standard combination chemotherapy in previously treated metastatic pancreatic cancer. These drugs, both panRAS and allele specific agents, are moving quickly in development to an earlier stage disease. For the small subset of individuals with no Kras mutation (KRAS wild-type), there are targeted therapeutics that impact various signaling pathways, including NRG-1 71, NTRK, BRAF, ROS, ALK, and others, and can have a meaningful impact on the disease. Specifically, the first bispecific in pancreatic cancer, zenocutuzumab 71, received a disease-specific regulatory approval in 2024. Similarly, for the small subset of individuals with mismatch repair deficiency/microsatellite instability (dMMR, MSI-H), immune checkpoint inhibitors have demonstrated a meaningful value in this disease. Collectively, the value and impact of precision-based medicine approaches have resulted because of routine germline testing and somatic next generating sequencing in this disease. Newer tools, including circulating tumor DNA (ctDNA) and RNA sequencing profiling, will further augment the value and identification of newer targets. Traditionally, pancreatic cancer has been generically considered as the prototype immunologically cold disease; however, the recognition that the treatment of this disease with neoantigen and oncogene targeted approaches can generate a potent immune response in subsets of individuals provides optimism for the coming decade 72, 73. In parallel with these above stated developments has been the moving away from a perspective of therapeutic nihilism and for more patients diagnosed with this disease having the option to consider treatment options as opposed to nonreferral. In the fields of gastroenterology and interventional radiology effective palliation of malignant biliary obstruction and duodenal/gastric outlet obstruction by the use of endoscopic 74 and percutaneous stenting can meaningfully impact symptomatology in this disease. In parallel, usage of pancreatic enzyme replacement therapy 75 to treat pancreatic exocrine insufficiency that accompanies PDAC 76, effective pain palliation 77 including the use of chemical splanchnicectomy 78/celiac plexus neurolysis within a multidisciplinary team approach can optimize symptom control and improve both quality and length of life. Although the pancreas was the first organ where transgenesis was successfully accomplished in 1984 79, the early GEM models of pancreatic cancer did not resemble the multistep progression of ductal adenocarcinoma, preceded by intraductal PanIN lesions, observed in humans 80, 81. In 2003, this barrier was successfully overcome through GEM models that expressed mutant Kras from the endogenous locus in the exocrine pancreas within the transcriptional domain of Pdx1 and later Ptf1/p48 82. These data confirmed the seminal role of mutant Ras in the initiation of exocrine pancreatic neoplasia, with cooperating driver alterations (Smad4, Cdkn2a/p16, p53) essentially altering the penetrance and natural history of this Ras-induced disease process 83-85. Subsequently, animal models where mutant Kras can be induced by doxycycline were developed 86, 87, which confirmed that the genetic extinction of Ras in established tumors had the potential to attenuate tumor growth. In many respects, these genetic studies provided the compelling rationale for pharmacological targeting of KRAS in PDAC, which has now become a reality in the past few years 88. More recently, dual recombinase models have also been developed which have allowed for asynchronous perturbation of genes within the epithelial and stromal compartments, respectively 89-91. In addition to models of PDAC that recapitulate the more common noncystic pathway to invasive cancer, mice that develop cystic lesions resembling intraductal papillary mucinous neoplasms have also been generated 92-94. Cumulatively, the development of these autochthonous models has led to significant developments in pancreatic cancer research, from mechanistic understanding of cancer cell intrinsic biology 95-99, the host (stromal and immune) response to multistep tumorigenesis 100-103, and the influence of environmental influences like diet and exercise 104-106. Further, these models are widely used in evaluating experimental therapies prior to human translation 107, 108. Although orthogonal preclinical models like murine and patient-derived organoids have been invaluable in accelerating translational research 109, 110, one can argue that GEM models represent one of the pivotal advances in this field with a lasting impact. Prior to the advent of dissociative single cell RNA sequencing (scRNA seq) and, more recently, spatial transcriptomics (ST), it was commonly believed that pancreatic cancers were relatively homogeneous, and that the host response was monolithic. This viewpoint is now obsolete, with scRNA-seq and ST approaches underscoring the profound tumor and host heterogeneity that characterizes this disease 111, 112. Although the full scope of discussion of PDAC heterogeneity is outside the scope of this commentary, a couple of vignettes will be highlighted. First, we have now recognized that the entity “cancer associated fibroblasts” or CAFs is an admixture of multiple CAF subtypes including myofibroblastic CAFs (myCAFs), inflammatory CAFs (iCAFs), and antigen presenting CAFs (apCAFs) 113-115. Each CAF subtype can be defined by distinct expression patterns on RNA and proteomics analyses 116, 117, as well as distinct cellular neighborhoods within the tumor microenvironment in which they can be found 118, 119. Further, we have also recognized the divergent roles of CAF subtypes in promoting or restraining tumor progression 120. Second, we have also recognized that beyond genetic heterogeneity, PDAC cells are characterized by transcriptomic heterogeneity with the recognition of at least two major cell states (classical and basal-like) 121, 122 and additional states defined based on further refinement of profiling data 123. Recent spatial data have also highlighted the profound intra-tumoral heterogeneity of epithelial lineage states in primary and metastatic tumors 124, as well as the susceptibility to distinct therapeutic regimens (such as chemotherapy vs. targeted therapies like KRAS inhibitors) 125, 126. The recognition of the cancer cell and stromal heterogeneity will continue to have profound effects on how more efficacious and selective therapeutic regimens are designed for this disease. In summary, the future for treatment of PDAC is brightly related to a renewed focus on this recalcitrant cancer, recognition of the importance of the value of multiagent cytotoxic chemotherapy for all disease stages, integration of targeted therapy options where feasible and indicated and recognition of the value and importance of multidisciplinary care. Leveraging RAS targeting and capitalizing on novel immune therapeutics are further poised to improve outcomes in PDAC. S. George Barreto: conceptualization, project administration, data curation, investigation, methodology, formal analysis, writing – original draft, writing – review and editing. Anirban Maitra: data curation, investigation, methodology, formal analysis, writing – original draft, writing – review and editing. Eileen M. O'Reilly: data curation, investigation, methodology, formal analysis, writing – original draft, writing – review and editing. Dushyant Sahani: data curation, investigation, methodology, visualization, formal analysis, writing – original draft, writing – review and editing. Richard Zheng: data curation, investigation, methodology, visualization, formal analysis, writing – original draft, writing – review and editing. Charles J. Yeo: data curation, investigation, methodology, visualization, formal analysis, writing – original draft, writing – review and editing. The authors acknowledge research support as follows: SGB support—Flinders Foundation Grant: 49358025. NHMRC Ideas by from the The and Anirban access by as part of the the of The authors have to Eileen M. received to of Society of The other authors no of The authors have to