Koios DS is a software application designed to assist trained interpreting physicians in analyzing breast and thyroid ultrasound images. Koios DS software automatically classifies breast lesions suspicious for cancer based on image data into one of four ACR BI-RADS or European U1-U5 Classification System-aligned categories. Koios DS also categorizes thyroid nodules via the ACR TI-RADS or American Thyroid Association (ATA) risk stratification systems (RSSs) along with a cancer risk assessment presented as the Koios “AI Adapter.”
Product specifications Information source: Vendor
Last updated: Oct. 24, 2023
Product name Koios DS
Company Koios Medical, Inc.
Subspeciality Breast, Thyroid
Modality Ultrasound
Disease targeted Breast cancer, thyroid cancer
Key-features Lesion/nodule segmentation, lesion/nodule classification, auto-population of lesion/nodule descriptors, AI independent risk assessment, alignment to diagnostic classification systems
Suggested use During: perception aid (prompting all abnormalities/results/heatmaps), interactive decision support (shows abnormalities/results only on demand), report suggestion
After: diagnosis verification
Data characteristics
Population Adult (>= 22 years) female patients with soft tissue breast lesions and/or adult (>= 22 years) patients with thyroid nodules suspicious for cancer.
Input 2 orthogonal still ultrasound views of either breast lesions or thyroid nodules
Input format DICOM
Output BI-RADS/TI-RADS Categorization and Classifiers, Shape, Orientation, Confidence Level Indicator, Position, Size
Output format DICOM Secondary Capture and/or DICOM Structured Report, DICOM Client such as a scanner, selected PACS integrations
Integration Integration in standard reading environment (PACS), Integration CIS (Clinical Information System), Stand-alone third party application, Stand-alone webbased
Deployment Locally virtualized (virtual machine, docker)
Trigger for analysis Automatically, right after the image acquisition, On demand, triggered by a user through e.g. a button click, image upload, etc.
Processing time < 3 sec
Certified, Class IIb , MDR
510(k) cleared, Class II
Market presence
On market since 12-2021
Distribution channels GE Healthcare, RMS Medical Devices, Amplify SDK, Alma AI MARKETPLACE
Countries present (clinical, non-research use) 25
Paying clinical customers (institutes) >80
Research/test users (institutes) >50
Pricing model Subscription
Based on Number of users, Number of analyses
Peer reviewed papers on performance

  • Improving the Efficacy of ACR TI‑RADS Through Deep Learning‑Based Descriptor Augmentation (read)

  • A role for breast ultrasound Artificial Intelligence decision support in the evaluation of small invasive lobular carcinomas (read)

  • Toward AI-supported US Triage of Women with Palpable Breast Lumps in a Low-Resource Setting (read)

  • Can an Artificial Intelligence Decision Aid Decrease False-Positive Breast Biopsies? (read)

  • Impact of Original and Artificially Improved Artificial Intelligence-based Computer-Aided Diagnosis on Breast US Interpretation (read)

  • Impact of Data Presentation on Physician Performance Utilizing Artificial Intelligence-Based Computer-Aided Diagnosis and Decision Support Systems (read)

  • Should we ignore, follow, or biopsy? Impact of artificial intelligence decision support on breast ultrasound lesion assessment (read)

Non-peer reviewed papers on performance

  • SIIM2018 - Artificial Intelligence in Breast Ultrasound: Moving from Standalone Performance to the Physician-system Interface (read)

  • SPMB 2016 - Decision Quality Support in Diagnostic Breast Ultrasound through Artificial Intelligence (read)

  • Artificial intelligence in low- And middle-income countries: Innovating global health radiology (read)

  • RSNA 2020 - Mango - Decreasing Benign Breast Ultrasound Biopsies: Prospective Use of AI Decision Support (read)

  • SIIM 2021 - Barinov - Improving the Efficacy of TI-RADS Through Artificial Intelligence (read)

  • SBI ACR Breast Imaging Symposium 2020 - Cavallo - AI Analysis of Ultrasound Images Could Decrease Benign Breast Biopsies (read)

Other relevant papers