Lunit
Revolutionizing cancer care with precision AI technology.. [Paid]
Last verified:
What is Lunit?
Lunit is a medical AI software company focused on conquering cancer through AI-powered solutions for early detection and precision oncology. Their cancer AI ecosystem transforms medical imaging and precision oncology into clinical intelligence, spanning screening, diagnosis, treatment decisions, and drug development.
The platform offers two main product lines: Lunit INSIGHT for radiology (including INSIGHT CXR for chest X-rays, INSIGHT MMG for 2D mammography, and INSIGHT DBT for 3D tomosynthesis) and Lunit SCOPE for precision oncology (including SCOPE IO for immuno-oncology biomarker analysis and SCOPE GP for genotype prediction). These tools help radiologists detect cancer earlier with 95-100% accuracy across 11 major thoracic abnormalities, improve breast cancer detection especially in dense breasts, and enable biopharma companies to identify tumor-agnostic immune biomarkers from H&E slides without additional stains.
Lunit is designed for radiologists, breast imaging specialists, pulmonologists, biopharma companies, clinical researchers, and healthcare institutions. Over 10,000 medical institutions in 65+ countries use Lunit AI, with partnerships including Microsoft, GE Healthcare, Philips, AstraZeneca, and Guardant Health. The FDA-cleared and CE-marked solutions integrate seamlessly into PACS and viewer workflows.
Lunit pricing
Pricing model: Freemium
The website does not publicly disclose specific pricing details, free tier information, or paid plan costs. Lunit INSIGHT and Lunit SCOPE products appear to be enterprise solutions sold through direct contact with the company. Pricing is likely customized based on institution size, deployment type (on-premise or cloud), and specific product modules. The company offers SecondRead™ AI as an optional out-of-pocket service for patients. For pricing information, healthcare institutions and biopharma companies must contact Lunit directly.
Lunit pros
- FDA-cleared and CE-marked under EU MDR for clinical use
- 95-100% AUC accuracy detecting 11 major thoracic abnormalities
- 18% higher accuracy in extremely dense breasts for mammography
- Reduces radiologist workload by 36.2% while maintaining 95% sensitivity
- Trained on diverse global datasets from US, UK, and Korea
- Over 10,000 medical institutions worldwide adopting Lunit AI
- Seamless PACS and viewer integration for workflow efficiency
- Highlights key areas of interest to reduce cognitive load
- Reduces false positives and unnecessary callbacks
- Supports single-reading and double-reading workflows
- Current-prior comparison with automated progression tracking
- 96-99% accuracy in chest abnormality and breast cancer detection
- Trained on 50,000+ whole-slide images with 10M+ cell annotations
- Identifies interval cancers that were previously overlooked
- Normal flagging feature streamlines high-volume chest X-ray reading
Lunit cons
- Not available for sale or distribution in all countries
- Lunit SCOPE IO is For Research Use Only, not for diagnostic procedures
- No publicly disclosed pricing information on website
- Does not learn live from client cases in real-time
- Score scale is not linear (90 is not three times more suspicious than 30)
- AI predicts malignancy likelihood, not severity or treatment outcome
- Requires integration with existing PACS infrastructure
- INSIGHT Manager is a non-medical device only
Frequently asked questions about Lunit
Does a higher AI score mean a worse prognosis?
No. The AI mammography software is trained to predict the likelihood that a lesion is cancerous, not its severity, clinical implications, or treatment outcome. While invasive cancers may receive higher scores than noninvasive ones, that is not always the case. If the AI is confident that a malignant lesion exists, it assigns the score accordingly, regardless of subtype. Note: The scale is not linear (e.g., a score of 90 is not three times more suspicious than a score of 30).
How is Lunit INSIGHT CXR deployed?
It can be deployed in two ways: On premise, where all data is analyzed on the client's infrastructure and servers, or on Cloud, with only a component of the software with the function to anonymize data and route to Lunit's AI Engine on Azure.
Does the AI learn live from my cases or update in real-time?
The AI does not learn live from client cases. Patient images are not used for real-time training updates. The models are trained on diverse global datasets and do not update based on individual institution data.
How does the AI handle benign lesions?
The AI is trained to predict the likelihood that a lesion is cancerous. It assigns malignancy likelihood scores to lesions, and benign lesions typically receive lower scores. The system helps radiologists distinguish between suspicious and non-suspicious findings to reduce false positives.
Is there a performance difference between different modality manufacturers?
Lunit INSIGHT DBT was trained on diverse datasets from the US, UK, and Korea and delivers consistent performance across racial and ethnic groups. The AI is designed to work with various modality manufacturers, though specific performance differences by manufacturer are not detailed on the website.
What abnormalities does Lunit INSIGHT CXR detect?
Lunit INSIGHT CXR detects 11 major thoracic abnormalities including acute bone fractures, nodules, pneumothorax, consolidation, and pleural effusion. It achieves 95-100% AUC accuracy across these abnormality types and supports tuberculosis screening.
What cancer types does Lunit SCOPE IO cover?
Lunit SCOPE IO covers 16 types of cancer including lung, head and neck, esophagus, cervix, stomach, colorectum, breast, pancreas, bladder, melanoma, kidney, liver, ovary, prostate, biliary tract, and endometrium. It performs quantitative immune phenotyping and classifies immune phenotypes as Inflamed, Excluded, or Desert.
How does SecondRead™ AI work?
SecondRead™ AI, powered by Lunit INSIGHT, gives patients more choice in breast screening by offering an optional out-of-pocket service. It provides enhanced AI analysis during annual screening, supporting radiologists by identifying areas that may benefit from closer review. All results are reviewed and communicated by the radiologist, and full program support including training and marketing materials is provided.
What is the accuracy of Lunit INSIGHT MMG in dense breasts?
Lunit INSIGHT MMG has 18% higher accuracy in extremely dense breasts and 8% higher accuracy in all breast density categories, enhancing detection across tissue types. This makes it particularly valuable for breast cancer screening in patients with dense breast tissue where cancer detection is typically more challenging.
How quickly does Lunit SCOPE IO generate analysis results?
It takes only three days to get AI analysis results from Lunit SCOPE IO. The platform analyzes whole slide images to detect cancer area, stroma, and tumor-infiltrating lymphocytes (TILs) from H&E slides, instantly classifying three immune phenotypes.