Surface what routine review can miss
Pancreatic lesions can be subtle on abdominal CT. PANCREASaver® provides an automated second read to bring suspicious findings forward for radiologist attention.
Why it matters
Support earlier detection without adding friction.
Built for the pressure points radiology teams face every day: subtle findings, crowded worklists and limited time.
Prioritize the cases that need a closer look
Automatically analyzed studies can be flagged for review, helping teams focus attention where it matters most.
Add confidence without adding clicks
PANCREASaver® fits the PACS workflow your team already knows, adding decision support without requiring a new viewing habit.
See what the AI sees
The same CT.
The same CT.
A different pair of eyes.
Pancreatic lesions hide in plain sight on routine abdominal CT. PANCREASaver® reviews the exact same images and marks what the human eye can overlook — automatically, in minutes, before the report is even started.
VS
Real clinical CT images with actual PANCREASaver® output overlay. Radiologists always review and confirm AI findings.
92.1%Sensitivity for tumors <2cmLancet Digital Health, 2020 ↗
0.95Nationwide multi-center AUCRadiology, 2023 · n=1,473 ↗
2,000+Cases interpreted (incl. clinical validation)Routine clinical use data
FDABreakthrough Device DesignationView regulatory status ↗
The product
An AI second reader that never blinks
PANCREASaver® is FDA Breakthrough-designated medical device software (SaMD). The moment an abdominal CT scan finishes, it analyzes the images, flags suspicious pancreatic findings, and sends the result back into the radiology workflow — supporting earlier review without changing the way your team works.
Fully automated, real-time reading
Analysis starts automatically when the scan completes — no extra steps for physicians. CNN deep learning and radiomics engines run in parallel.
Designed to surface small tumors
The cited study reported 92.1% sensitivity for lesions under 2cm. PANCREASaver® is intended to support radiologist review, not replace clinical judgment.
Clinically validated
5 international journal papers and a nationwide multicenter AUC of 0.95, led by the National Taiwan University Hospital team.
Reading workflow
From scan to alert, in minutes
1
CT scan
Patient undergoes a routine abdominal CT; images stream into the PACS as usual.
2
AI reads in real time
PANCREASaver® analyzes the full pancreas within minutes — no clicks required.
3
Auto-mark + alert
The lesion is locked and marked; a priority alert fires back into the PACS worklist.
4
Radiologist review
The radiologist confirms the marks and signs the final report — AI never replaces the read.
Seamless PACS integration
One integration. Every workflow.
PANCREASaver® was designed around the way radiology departments already work. It connects to your existing PACS — no new viewer, no extra clicks, no rip-and-replace.
Works with your existing PACS
DICOM-native integration connects to the PACS/RIS infrastructure you already run — the AI fits the system, not the other way around.
Vendor-neutral
Compatible with GE, Siemens, Philips, Canon and more — it runs on the scanners and viewers you already own.
Zero workflow change
Radiologists read exactly as they always have. AI works silently in the background and surfaces what matters.
Weeks, not months
Typically live in as little as 4 weeks with minimal IT lift — onboarding, training and go-live support included.
How the integration works
1
Connect
A DICOM link connects PANCREASaver® to your CT scanners and PACS.
2
Analyze
The AI triggers automatically the moment a scan completes — nothing to launch, nothing to click.
3
Alert
Suspicious studies are prioritized in the worklist with the lesion marked for the radiologist.
4
Report
Structured findings are delivered back into your reporting workflow, ready for review and sign-off.
Deployment options: On-premises or cloud · data stays in your environment · secure, auditable, HIPAA-ready architecture.
The real system
Not a mockup — a system running in hospitals today
From Stone Web Viewer integration to Bridge login, PANCREASaver® is deployed at medical centers, regional hospitals and screening centers — 2,000+ scans reviewed.
PANCREASaver®
Clinical evidence
Published. Peer-reviewed. Verified.
Five papers in leading journals — every figure on this page carries a source you can open.
The Lancet Digital Health 2020
A cross-racial study of the AI detection engine reported 92.1% sensitivity for tumors under 2cm and correctly recovered 11 of 12 cancers missed by radiologists. The study details and limitations are available in the original paper.
Radiology 2023
Nationwide multi-center performance with national-level AUC 0.95 — real clinical validation led by the NTUH team.
Real-world deployment
2,000+ cases interpreted in routine practice, including independent third-party validation — evidence from the field, not just the lab.
Learn more
Explore PANCREASaver® further
For Clinicians
The full evidence hall — AUC curves, regulatory approvals, and the integration workflow.
Go to Clinicians →Clinical Evidence
5 international journal papers with topics, abstracts and PubMed links.
Open the evidence →Regulatory Status
TFDA license, FDA Breakthrough, 510(k) status, and 10 US/TW patents — stated exactly as they are.
See regulatory status →
FAQ
Common questions, direct answers
Yes. PANCREASaver® is DICOM-native and integrates with the PACS/RIS infrastructure you already run — no new viewer and no replacement of existing systems.
Typically live in as little as 4 weeks, with minimal IT lift. On-premises and cloud deployment options are available, and our team handles onboarding, training and go-live support.
No. Radiologists read exactly as they always have — AI runs automatically in the background and surfaces prioritized findings in the worklist.
It holds FDA Breakthrough Device Designation; a 510(k) was submitted in July 2026 and is under review. It is TFDA-licensed in Taiwan (license no. 衛部醫器製字第007946號).
Contrast-enhanced abdominal CT, per the approved indications — the same scans your department already performs, with no extra acquisition protocol.
Voices from the field
What leading editors say about the science
Verbatim third-party statements on the peer-reviewed studies behind PANCREASaver®. Quotes are about the research — not the product — and each carries its original source.
“
…accurate detection of pancreatic cancer on CT scans, especially for tumors smaller than 2 cm. Early detection… greatly increases the chances of survival.
“
This deep learning–based tool had very good accuracy, and sensitivity approaching that of subspecialist radiologists… Tools such as this promise to benefit routine workflow and patient care.
“
These promising results show that CNN as a second reader can reduce misdiagnosis of pancreatic cancer and possibly lead to improved patient outcomes.
Want to see it in action?
Book a live demo and let the AI show you what it sees on a real CT.
Talk to us
Performance data sources: The Lancet Digital Health 2020 ｜ Radiology 2023 (PubMed) ｜ This product is a medical device — consult your physician before use.