A beautiful day to be out
It was a quiet morning by the river. He felt well and went for a routine checkup.
The product
PANCREASaver® — AI-assisted pancreatic CT review
An AI-assisted medical device that analyzes contrast-enhanced abdominal CT and flags suspicious pancreatic findings for radiologist review. FDA Breakthrough Device Designation; Taiwan TFDA medical device license.
Real CT · PANCREASaver® AI overlay — sub-2 cm lesion auto-marked
The same scan, without AI
⚡ Automated analysis
Analysis starts after the scan completes and returns findings for radiologist review within the existing workflow.
🎯 Published small-lesion results
The cited study reported 92.1% sensitivity for tumors under 2 cm; a nationwide, multicenter study reported an AUC of 0.95.
🔗 PACS integration
DICOM-native and vendor-neutral; implementation is designed around your existing PACS/RIS environment.
92.1%
Sensitivity for tumors under 2 cm
0.95
Multicenter AUC
5500+
Cumulative patient cases (research & commercial use)
10
Patents (Taiwan and U.S.)
Seamless PACS integration
Built for the way you already work
Vendor-neutral and DICOM-native; PANCREASaver® is designed to fit your existing PACS/RIS workflow, with implementation timing set with each institution.
🔗 Works with your existing PACS
🏥 Vendor-neutral
⚡ Designed to minimize workflow changes
🗓️ Example implementation timeline: about 4 weeks
All-in-one evidence
From product to certification, all in one place.
Six themes covering the full picture of PANCREASaver® — swipe to explore.
Product
AI-assisted pancreatic cancer CT review
An automated end-to-end system that analyzes CT scans, flags suspicious findings, and integrates with hospital PACS/RIS.
Learn more
Clinical Data
Nationwide real-world validation of 1,473 cases, AUC 0.95
89.7% sensitivity, 92.8% specificity; 92.1% sensitivity for tumors under 2 cm — published in Radiology and The Lancet Digital Health.
View clinical data
Publications
Six international papers, with source links.
Real papers in The Lancet Digital Health, Radiology, BMC Cancer, and other journals — with summaries and PubMed links to the original articles.
Visit the publication hall
IP Portfolio
10 patents in Taiwan and the U.S., covering system and method claims.
System-level and method-level claims cover medical imaging analysis and machine-learning-based pancreatic cancer detection; see the official patent records for scope.
View IP portfolio
Regulatory
TFDA medical device license, FDA Breakthrough Device Designation
Taiwan TFDA medical device license No. 007946; FDA Breakthrough Device Designation; FDA Pending; not yet cleared for U.S. marketing.
See certification journey
Deployment
From contract to go-live, implementation in about 4 weeks.
Medical centers, regional hospitals, and health centers — cumulative use in more than 5,500 patient cases across research and commercial applications. Deployment and update options depend on the institution.
View deployment plans
Why early detection matters
Pancreatic cancer can be difficult to detect early.
Tumor stage and other clinical factors help guide treatment planning. PANCREASaver® is designed to support review of suspicious findings on abdominal CT.
13%
Overall five-year relative survival is about 13% overall and about 44% for localized disease (U.S. SEER data). Earlier detection can expand treatment options; individual outcomes vary.
<2cm
The cited study reported 92.1% sensitivity for tumors under 2 cm in its defined study population.
Found late = few options
By the time symptoms appear, the tumor has often spread. AI-assisted review of routine CT studies is designed to find the lesion before it becomes a late-stage diagnosis.
Why screening matters ↗
Figures as published in peer-reviewed literature and PanCAD.ai’s clinical validation. See the evidence center for sources.
Regulatory & trust
Licensed in Taiwan. FDA Breakthrough. Evidence published.
Every claim on this website is backed by an issuing authority, a publication, or a verifiable record.
🛂
FDA PendingNot yet cleared for U.S. marketing🧩
10 Patents4 U.S. · 6 Taiwan📄
6 PublicationsLancet Digital Health · Radiology🏆
9 AwardsRSNA Margulis Award 2023 recognition
Designed for You
Patients, clinicians, or health systems
Three audiences, three journeys. PANCREASaver® provides information and workflow support from screening through follow-up.
Patients & Families
Learn how pancreatic CT review works, what an AI flag means, and what questions to discuss with your clinician.
Enter Patient JourneyFor Clinicians
The interactive evidence hall: AUC 0.95, 92.1% sensitivity, documented regulatory status, and the integration workflow.
See EvidenceHealth Centers
Add an AI-assisted pancreatic CT review option to eligible programs, with workflow integration and radiology collaboration.
Explore Screening
Patient First
An ordinary day, with a finding that prompted a closer look
One morning, one checkup, one flagged finding — and a family making informed next steps.
A flagged finding for review
A routine abdominal CT produced a flagged finding for radiologist review.
The whole family at the table
The finding prompted timely follow-up. The family sat together, planning the next conversation with the care team.
The Care Journey
From screening to survivorship, every step supported
Scroll through five stations that explain how patients and physicians may use information along the care journey.
01
Screening: why check regularly
Early pancreatic cancer may cause few or nonspecific symptoms. Screening and imaging decisions should follow clinician guidance and applicable protocols.
02
Detection: AI flags a finding for review
After an abdominal CT, AI flags a finding for radiologist review. Published study results include 92.1% sensitivity for tumors under 2 cm; outcomes vary by patient and disease stage.
03
Diagnosis: the physician's double check
AI findings go to radiologists for review, and specialists confirm the diagnosis. AI does not replace physicians; it adds information for clinical review.
04
Treatment: options depend on the full picture
A 1 cm tumor and metastatic disease can lead to different treatment paths. Options depend on disease stage and individual factors.
05
Survivorship: live well, with family
After treatment comes regular follow-up. Patients and families can work with their care team on appropriate monitoring and support.
Milestones
From research bench to bedside
2016Research begins at National Taiwan University
Prof. Wang’s team launches early pancreatic cancer imaging AI research with Prof. Liao’s clinical team at NTU Hospital.
2019Multicenter clinical validation
Nationwide multicenter validation begins, building the dataset behind the AUC 0.95 national performance.
2023RSNA Margulis Award · FDA Breakthrough
PANCREASaver® wins the RSNA Margulis Award — the first for Taiwan — and receives U.S. FDA Breakthrough Device Designation.
2026TFDA license · FDA Pending
Licensed by Taiwan’s TFDA (衛部醫器製字第007946號); PANCREASaver® is not yet cleared for U.S. marketing.
News
The latest from PanCAD.ai
PANCREASaver® surpasses 5,500 patient cases across research and commercial applications
As of October 13, 2026, PanCAD.ai reports that PANCREASaver® has been used in more than 5,500 patient cases across research and commercial applications.
2026 Storm Media AI Healthcare Innovation Awards (AIHIA) — PANCREASaver® recognized ↗
Storm Media’s AI Healthcare Innovation Awards recognized PANCREASaver® among applications focused on healthcare innovation and earlier pancreatic cancer review.
Storm Media interview: PANCREASaver® and the path from research to clinical practice ↗
Director Kai-Chun Yang and Dr. Wei-Chih Liao discuss how PANCREASaver® enters clinical practice while fitting existing physician workflows.
PANCREASaver® presents latest real-world data at APDW
At Asian Pacific Digestive Week, the team presented real-world data from more than 2,000 CT studies analyzed in clinical practice.
FAQ
Questions about AI-assisted pancreatic CT review
These answers summarize published evidence and current product scope. They are for education; discuss personal decisions with a clinician.
Early pancreatic cancer may cause few or nonspecific symptoms, and small lesions can be difficult to detect on abdominal CT. Screening and imaging decisions should follow clinician guidance.
A CNN + radiomics dual-engine analyzes contrast-enhanced abdominal CT images and presents suspicious findings for radiologist review; PACS/RIS alerts depend on the institution’s integration setup.
No. The system analyzes existing abdominal CT images. Whether a scan is appropriate should be discussed with your physician.
No. An AI flag identifies a finding for closer review; the physician makes the diagnosis using the complete imaging and clinical information.
Five-year relative survival is about 13% overall and about 44% for localized disease (American Cancer Society and NCI SEER data). Earlier detection can expand treatment options; individual outcomes vary.
People with a family history or a known hereditary syndrome should discuss risk assessment with a specialist. Age, smoking, or diabetes alone does not mean routine pancreatic cancer screening is appropriate; asymptomatic average-risk adults should not undergo routine screening.
PANCREASaver is used in medical institutions and health checkup centers in Taiwan; see the For Health Systems page for current public information.
Source: USPSTF pancreatic cancer screening recommendation. Product information and study summaries are available on the Clinical Evidence page.
"
The day the report came, the doctor said: “Let’s review what we found and discuss the next steps.”
— Wife of a 45-year-old patient
Support earlier review. Stay informed.
For yourself, your family, or your patients — start with an informed conversation about abdominal CT.
Contact us