VinDr-Mammo

by VinBigData Joint Stock Company  · Based in Vietnam → — AI-powered precision for mammography interpretation and breast cancer screening.
Radiology
Website
Claim this tool
Is this the maker? Claim the company →

Contact for pricing
Regulatory Status Disclosed

Overview

VinDr-Mammo is an advanced Artificial Intelligence (AI) powered Computer-Aided Diagnosis (CAD) tool developed by VinBigData Joint Stock Company. As a module within the comprehensive VinDr solution for medical image analysis, it is designed to assist radiologists in the rapid and precise interpretation of full-field digital mammograms. The tool leverages AI algorithms, including computer vision and deep learning, to classify mammography studies into three BI-RADS (Breast Imaging-Reporting and Data System) levels and four types of breast density. Furthermore, it can localize thirteen types of common abnormalities on mammography images. VinDr-Mammo integrates seamlessly into Picture Archiving and Communication Systems (PACS), providing exam-level output to flag suspicious cases and enable worklist prioritization for MQSA-qualified interpreting physicians.

Reviewed by Pouyan Golshani, MD — Interventional Radiologist

Key Features

  • Classifies mammography studies into 3 BI-RADS levels
  • Classifies 4 types of breast density
  • Localizes 13 types of common abnormalities on mammography
  • Integrates with Picture Archiving and Communication Systems (PACS)
  • Assists radiologists in fast and precise diagnoses
  • Utilizes advanced AI algorithms (computer vision, deep learning, image analysis)
  • Provides exam-level output for flagging suspicious cases
  • Enables radiology worklist prioritization

Use Cases

  • Breast cancer screening
  • Assisting radiologists in mammogram interpretation
  • Prioritizing suspicious cases in radiology workflows
  • Computer-aided detection (CADe) for breast abnormalities
  • Computer-aided diagnosis (CADx) for breast cancer

What Physicians Need to Know

Key Capabilities
Utilizes an AI algorithm to analyze 2D Full-Field Digital Mammography (FFDM) screening mammograms, flagging exams suggestive of suspicious findings at the exam-level. It classifies mammography studies into BI-RADS assessment categories (1-5) and breast density levels (A, B, C, or D), and localizes 13 types of common abnormalities including mass, calcification, asymmetries, and architectural distortion.
Integration Options
Designed to produce an exam-level output to a Picture Archiving and Communication System (PACS)/Workstation for flagging suspicious cases and enabling worklist prioritization. It integrates AI into PACS and supports two-way integration with medical digital applications like Hospital Information Systems (HIS) and PACS.
Compliance Status
FDA 510(k) cleared (K233108) as a Class II medical device, classified as Radiological computer aided triage and notification software. It is intended for Prescription Use and complies with relevant FDA regulations including registration, listing, labeling, and medical device reporting.
Pricing Model
Information regarding the pricing model (e.g., subscription, per-use) is not publicly available in the provided search results.
User Experience
Operates as a passive notification tool within a parallel workflow, serving as an assistant to radiologists by providing an objective opinion. It boasts a fast diagnosis time of less than 0.5 seconds per scan.
Implementation Complexity
Integrates into existing PACS/workstation environments and is intended for use with complete 2D FFDM mammography exams acquired using validated FFDM systems.
Evidence Base
Trained and validated on approximately 50,000 studies from Vietnamese hospitals, with bounding box annotation and disease labeling performed by experienced radiologists. It achieves over 80% accuracy for BI-RADS classification and over 85% average accuracy for parenchyma density classification and lesion localization. The VinDr-Mammo dataset, a large-scale public benchmark dataset of 5,000 four-view FFDM exams (20,000 images) with breast-level assessment and finding annotations, was double-read by experienced mammographers and de-identified to meet HIPAA and GDPR privacy regulations.
Physician Tip

VinDr-Mammo is a valuable AI tool for prioritizing mammography cases with suspicious findings, enhancing workflow efficiency. Physicians should remember that it is a triage and prioritization aid, not a diagnostic tool, and their comprehensive evaluation remains paramount. Always review each exam on an approved display according to the current standard of care, and do not rely on the AI to confirm diagnoses or replace full patient assessment. The tool's ability to classify BI-RADS and breast density, along with localizing abnormalities, can serve as a helpful 'second opinion' or initial filter.

VinDr-Mammo is designed for seamless integration with existing PACS and workstation environments. Its output is at the exam-level, facilitating worklist prioritization. The system also supports two-way integration with other medical digital applications like HIS, aiming for a comprehensive AI-integrated medical imaging solution.

Details

Category Oncology AI, Radiology & Imaging AI
Pricing Contact for pricing — N/A
DeploymentCloud; Integrates with on-premise PACS
Compliance
BAA AvailableUnknown AI-estimated
HIPAA CompliantUnknown AI-estimated
FDA Status 1 AI-estimated

VinDr-Mammo is a passive notification for prioritization-only, a parallel-workflow software tool used by MQSA qualified interpreting physicians to prioritize patients with suspicious findings in the medical care environment. It analyzes 2D FFDM screening mammograms and flags suspicious cases for worklist prioritization.
Integrations
EHR Not specified
Specialties Radiology

What the Web Says

VinDr-Mammo is an AI-powered solution for breast cancer detection in mammography, developed by VinBigdata. It aims to assist radiologists in identifying abnormalities, reducing false positives/negatives, and improving workflow efficiency. Reviews suggest it's a promising tool for early detection and workload management.

Overall: Positive

Strengths

  • Aids in early detection of breast cancer, potentially improving patient outcomes.
  • Reduces radiologist workload and improves reading efficiency.
  • Assists in identifying subtle abnormalities that might be missed by the human eye.
  • Can help standardize diagnostic quality across different radiologists.
  • Potentially reduces false positives and unnecessary biopsies.
  • Offers a second opinion, enhancing diagnostic confidence.

Limitations

  • Reliance on AI may lead to over-reliance and deskilling of radiologists.
  • Potential for algorithmic bias if not trained on diverse datasets.
  • Integration challenges with existing PACS and hospital IT systems.
  • Cost of implementation and ongoing maintenance could be a barrier for some institutions.
  • Requires validation in real-world clinical settings beyond initial studies.
  • Ethical considerations regarding AI in medical decision-making.

Based on reviews from: VinBigdata official website, Healthcare IT News articles, Radiology journals and research papers, Medical technology review sites, Physician forums and discussions

Last updated: 2026-07-20

Ratings & Reviews

No reviews yet. Be the first to review this tool!

Rate VinDr-Mammo

Clinical Value
Ease of Use
Integration
Support & Docs
Value for Money

Press & Coverage

FDA
510(k) Premarket Notification - VinDr-Mammo
The FDA has cleared VinDr-Mammo, a software tool that uses AI to analyze 2D FFDM screening mammograms and flag suspicious findings for prioritization by MQSA qualified interpreting physicians.
2024-05
Medimaging.net
AI-Based Algorithm Improves Accuracy of Breast Cancer Diagnoses
Researchers at the University of Eastern Finland developed MV-DEFEAT, a novel AI algorithm that significantly improves mammogram density assessment and breast cancer diagnosis accuracy, showing a 50.78% improvement in distinguishing between benign and malignant tumors on the VinDr-Mammo dataset.
2024-08
University of Eastern Finland
Novel AI algorithm assists in breast cancer screening
The University of Eastern Finland has announced the development of MV-DEFEAT, an AI-based algorithm designed to enhance mammogram density assessment and improve diagnostic precision in breast cancer screening.
2024-08
PubMed (Scientific Data)
VinDr-Mammo: A large-scale benchmark dataset for computer-aided diagnosis in full-field digital mammography
This article introduces VinDr-Mammo, a large-scale Vietnamese dataset of digital mammography with breast-level assessment and extensive lesion-level annotations, created to develop more robust and interpretable support systems in breast imaging.
2023-05
Research Communities (University of Edinburgh)
VinDr-Mammo: The largest public dataset of full-field digital mammography to evaluate and compare algorithmic support systems for breast cancer screening
VinDr-Mammo is presented as the largest public dataset of full-field digital mammography, offering 20,000 scans to evaluate and compare algorithmic support systems for breast cancer screening, with breast-level BI-RADS assessment and suspicious findings.
2023-03
PhysioNet
VinDr-Mammo: A large-scale benchmark dataset for computer-aided detection and diagnosis in full-field digital mammography
This publication announces the release of VinDr-Mammo, a large-scale benchmark dataset of 5,000 four-view full-field digital mammography exams with breast-level assessment and finding annotations, aimed at advancing computer-aided detection and diagnosis algorithms.
2022-03
iMerit
Top 10 Medical AI Training Datasets for 2026
VinDr-Mammo is highlighted as a valuable medical AI training dataset, released in 2022 by VinBigData and available via PhysioNet, containing 5,000 full-field digital mammography examinations annotated by experienced radiologists.
2026-01
Journal of Cultural Analysis and Social Change
Application of Artificial Intelligence in the Healthcare Sector and Legal Challenges: A Study in Vietnam
This study discusses the application of AI in healthcare in Vietnam, noting that the U.S. FDA has officially approved the VinDr application for market release in the United States, and VinDr can classify mammography images with high accuracy.
2025-11

Videos

Product demos, reviews, and walkthroughs for VinDr-Mammo.

Loading videos...

View all on YouTube

Frequently Asked Questions

VinDr-Mammo is an AI-powered Computer-Aided Diagnosis (CAD) tool designed to assist MQSA qualified interpreting physicians in mammography. It analyzes 2D FFDM screening mammograms to classify studies into BI-RADS levels and breast density types, and localizes common abnormalities, primarily for prioritizing suspicious cases in the worklist.
VinDr-Mammo's safety and performance have been evaluated and verified according to software specifications and applicable performance standards, including Federal Regulations and Guidance documents for medical device software. It is a passive notification, prioritization-only tool, indicating it has undergone a regulatory review process, likely FDA 510(k) clearance.
VinDr-Mammo is designed for seamless integration into Picture Archiving and Communication Systems (PACS) without disrupting standard clinical workflows. It produces exam-level outputs to the PACS/Workstation, flagging suspicious cases to enable worklist prioritization for interpreting physicians.
VinDr-Mammo functions as a prioritization-only tool and does not provide diagnostic care, meaning physicians remain solely responsible for final interpretations. Its underlying training dataset, while large, lacks pathology-confirmed ground truth and comprehensive clinical information, relying instead on radiologist consensus. Additionally, the dataset has limited samples for certain rare abnormalities, which could impact the tool's performance on those specific findings.
Yes, the field of AI in mammography is evolving, with various deep learning-based computer-assisted detection and diagnosis (CADe/x) tools available or under development. These alternatives aim to improve screening accuracy and reduce workload, often trained on diverse datasets.
Specific pricing details for the VinDr-Mammo software product, such as subscription fees or per-study costs, are not publicly available in the provided search results. The VinDr-Mammo dataset itself is publicly accessible for research and development purposes.
While specific product-level data handling protocols are not detailed, the underlying VinDr-Mammo dataset was created with patient privacy in mind, with all personally identifiable and protected health information removed. This process aligns with principles of regulations like HIPAA and GDPR, suggesting the product would also adhere to stringent data privacy and security standards.

Related Tools

Surgical Reality Viewer
Surgical Reality
Oncology AI
Surgical Reality offers CE- and FDA-certified AI-powered 3D medical imaging and surgical planning software that transforms CT scans into patient-specific 3D models for thoracic surgery, enhancing precision in complex lung procedures.
Cancer Copilot
Color Health
Oncology AI
Color Health's Cancer Copilot, built with OpenAI, helps oncologists and clinicians identify and order workup needs for recently diagnosed patients, aiming to speed up time to cancer treatment. This AI tool is designed to expand cancer expertise, speed decisions, and improve care, reducing a two-hour clinical task to 10 minutes with over 95% validated accuracy.
Cercare Medical Oncology Virtual Expert
Cercare Medical
Neurology AI
Cercare Medical's Oncology Virtual Expert is an AI-powered brain tumor segmentation module that provides fast, consistent, and semi-automated tumor analysis from standard MRI data, designed for on-premise deployment within existing clinical radiology workflows.
Synapse Lung Nodule AI
Fujifilm Corporation
Oncology AI
Fujifilm's Synapse Lung Nodule AI is an AI software algorithm for detecting pulmonary nodules on CT imaging, aiming to support early detection and treatment of lung cancer.
On request
RadOncAI
Informai (for RadOncAI product) / Global Network for AI in Radiation Oncology (for RadoncAI.net)
Oncology AI
RadOncAI is an AI solution by Informai designed to optimize and enhance radiation oncology workflows, improving efficiency and accuracy in treatment planning and delivery.
FAITH project (Fatigue Therapy – AI-supported Diagnosis and Therapy of Tumour-associated Fatigue Syndrome)
Fimo Health GmbH (Consortium Leader)
Mental & Behavioral Health AI
The FAITH project is developing an AI-based solution for the diagnosis and therapy of tumor-associated fatigue syndrome in cancer patients. It uses wearable sensors, a smartphone app, and artificial intelligence to provide individualized therapy recommendations and monitor mental health.

See all Oncology AI tools →

Suggest an Edit → | Last Verified: 2026-04-21 | First Added: 2026-04-21
AI Tool Finder
AI-powered search. Results may not be comprehensive.