Cracking the 'Undruggable': How Penn Scientists Aim to Stop Pancreatic Cancer Before It Starts

For decades, the standard approach to pancreatic cancer has been grimly reactive — diagnose late, treat aggressively, hope for the best. The results speak for themselves: a five-year survival rate stuck at 13%, the worst of any major cancer in America [1]. An estimated 52,740 Americans will die from the disease this year alone [2].

But a landmark study published March 12 in the journal Science offers a fundamentally different proposition: what if, instead of fighting pancreatic cancer after it takes hold, doctors could wipe out the microscopic seeds of the disease years before a tumor ever forms? [3]

Physician-scientists at the University of Pennsylvania's Perelman School of Medicine and Abramson Cancer Center have demonstrated, for the first time, that experimental drugs targeting the notorious KRAS oncogene can eliminate precancerous lesions in the pancreas of mice — nearly tripling survival compared to untreated animals and nearly doubling it compared to treatment initiated after cancer had already developed [4].

"I'm convinced that cancer interception will become the next frontier of cancer therapy," said Robert Vonderheide, MD, DPhil, director of the Abramson Cancer Center and co-corresponding author of the study [4].

The Deadliest Major Cancer

Pancreatic ductal adenocarcinoma (PDAC) — the most common form of pancreatic cancer — holds a uniquely terrible position in oncology. It is the third-leading cause of cancer death in the United States, and projections indicate it will overtake colorectal cancer to become the second-leading cause before 2030 [2]. While the combined five-year survival rate for all cancers has reached a milestone 70%, pancreatic cancer's rate has barely budged, stalling at 13% for three consecutive years [1].

The cruelty of the disease lies in its stealth. The pancreas sits deep in the abdomen, behind the stomach. By the time symptoms appear — jaundice, weight loss, abdominal pain — 80% of patients have tumors that have already spread beyond the organ. For the roughly 15% of patients diagnosed with localized disease, the five-year survival rate is a more encouraging 44%. For the majority diagnosed with metastatic disease, it is just 3% [5].

Source: SEER / Pancreatic Cancer Action Network

Data as of Mar 14, 2026CSV

This lethality has driven an urgent search for ways to catch the disease earlier. Recent advances include an AI-powered blood test (PanMETAI) that can detect pancreatic cancer at early stages with up to 94% accuracy, and a four-biomarker blood panel that distinguishes cancer cases from non-cancer cases 91.9% of the time [6][7]. But the Penn study goes further than detection — it proposes elimination of the precancerous process itself.

Targeting the Seeds of Cancer

Nearly all pancreatic cancers arise from microscopic precursor lesions called pancreatic intraepithelial neoplasias, or PanINs. These tiny clusters of abnormal cells harbor mutations in the KRAS gene — the single most common cancer-driving mutation across all tumor types, present in more than 90% of pancreatic cancers [3][4].

For 40 years, KRAS was considered "undruggable." The protein's smooth, featureless surface offered no obvious binding pocket for therapeutic molecules, and generations of researchers failed to crack it [8]. That changed in 2021 with the FDA approval of sotorasib, a drug targeting one specific KRAS variant called G12C. But G12C accounts for only 1–2% of pancreatic cancers, leaving the vast majority — driven by the G12D variant — untouched [8].

The Penn team used two experimental compounds developed by Revolution Medicines that represent a new class of KRAS inhibitors: RMC-9945, which selectively targets KRAS G12D (the most common variant in pancreatic cancer), and RMC-7977, which targets multiple RAS variants. Both drugs inhibit the active, GTP-bound form of the RAS protein, shutting down the signaling cascade that drives precancerous cells toward malignancy [3][4].

The Experiment: Treating Before Cancer Exists

The researchers employed a sophisticated, immunocompetent mouse model engineered at Penn that faithfully recapitulates the human progression from PanIN to invasive cancer. Critically, these mice retained fully functional immune systems, allowing the team to study how the drugs interact with the body's natural defenses — not just the cancer cells themselves [4].

The results were striking. Short-term treatment with either compound for 10 to 28 days produced dramatic reductions in PanIN burden. Long-term administration of RMC-7977 to mice bearing PanINs — but not yet tumors — tripled median overall survival compared to untreated controls [3][4].

The most compelling finding concerned timing. Mice that received the drug before tumors developed survived nearly twice as long as mice that began the same treatment after cancer had already formed. In absolute terms, the interception group achieved a median overall survival exceeding one year, compared to less than five months in untreated controls [3][4].

"This study shows us the power of being proactive, rather than reactive," said Minh Than, MD, PhD, a clinical and research fellow in hematology-oncology at Penn and lead author of the study [4].

From 'Undruggable' to Clinical Reality

The Penn study arrives at a moment of rapid progress in KRAS-targeted therapy. Revolution Medicines, which developed the experimental compounds used in the study, now has multiple drugs advancing through clinical trials. Zoldonrasib (RMC-9805), a KRAS G12D-selective inhibitor related to the preclinical tool compound RMC-9945, has demonstrated a 30% objective response rate and 80% disease control rate in patients with KRAS G12D-mutant pancreatic cancer in early clinical data [9]. The FDA granted it breakthrough therapy designation in January 2026 [10].

Source: GDELT Project

Data as of Mar 14, 2026CSV

Daraxonrasib (RMC-6236), the multi-selective RAS inhibitor related to RMC-7977, has also received FDA breakthrough therapy designation for previously treated metastatic pancreatic cancer and orphan drug designation. Its Phase 3 registrational trial (RASolute 302) for second-line metastatic PDAC is expected to report data later this year, while a separate trial (RASolute 304) is enrolling patients with resectable PDAC following surgery [11][12].

Meanwhile, Mirati Therapeutics (now part of Bristol Myers Squibb) has MRTX1133, another KRAS G12D inhibitor, in Phase 1/2 trials. Unlike earlier approaches, MRTX1133 is a non-covalent inhibitor that binds to both the active and inactive forms of KRAS G12D — the first targeted therapy of any kind to show promising tumor regression in mouse models of pancreatic cancer [13].

The Promise and Challenge of Cancer Interception

The concept underpinning the Penn study — cancer interception — represents a philosophical shift in oncology. Unlike prevention strategies such as vaccines (HPV) or lifestyle changes (smoking cessation), which aim to stop cancer from ever initiating, interception targets the earliest phases of a cell's journey toward malignancy — after abnormal changes have begun but before they become invasive cancer [14].

The analogy is colonoscopy: gastroenterologists routinely remove precancerous polyps from the colon, preventing colorectal cancer. No equivalent procedure exists for the pancreas. PanINs are too small to detect on imaging, and the pancreas cannot be easily accessed for biopsy or polypectomy [3].

"PanINs are now potential targets for cancer interception," said Ben Stanger, MD, PhD, the Hanna Wise Professor in Cancer Research and director of the Penn Pancreatic Cancer Research Center [4].

But translating the approach to humans presents formidable challenges. PanINs are invisible on current scans, so identifying which patients to treat requires relying on risk factors rather than imaging. The researchers envision starting with the highest-risk populations: individuals carrying BRCA1, BRCA2, or PALB2 gene mutations; those with hereditary pancreatitis; and patients with precancerous pancreatic cysts [3][4]. Eventually, the strategy could expand to people at intermediate risk.

There are also ethical complexities. Treating asymptomatic individuals with drugs that may have side effects requires a careful risk-benefit calculus. KRAS inhibitors, while increasingly well-tolerated, are not without adverse effects — and any interception strategy would require long-term, possibly indefinite administration [3].

A Broader Wave of Innovation

The Penn study is part of a broader surge in pancreatic cancer research that has produced several notable advances in early 2026. A comprehensive review published in CA: A Cancer Journal for Clinicians catalogued the expanding toolkit, from liquid biopsies and AI-powered diagnostics to novel immunotherapy combinations [15]. The Pancreatic Cancer Action Network's 2026 research outlook highlighted KRAS-targeted therapies as the most promising class of new drugs in a generation [16].

The convergence of these advances — better drugs, better diagnostics, and now a proof-of-concept for interception — offers a rare note of optimism in a field long defined by frustration. But researchers caution against premature celebration. The study was conducted in mice, not humans. Clinical trials testing the interception approach in high-risk patients have not yet begun.

And the disease remains brutally efficient. An estimated 67,530 Americans will be diagnosed with pancreatic cancer in 2026, and the vast majority will face the same grim odds that have defined the disease for decades [2]. The gap between a preclinical proof of concept in mice and a treatment that changes patient outcomes is one of the widest in medicine.

Still, for a cancer that has resisted nearly every therapeutic advance of the past half-century, the idea that its earliest seeds might be destroyed before they take root represents something genuinely new.

"The challenge now," Vonderheide said, "is careful translation" [4].