Radiology

AI Cancer Detection Software for Radiology

Marks places on a scan that may be cancer, and in some cases gives a score for how suspicious they look.

This is a bigger claim than triage, and the bar is correspondingly higher.

These products point at a specific place on the image and say look here. Some go further and attach a score for how suspicious the spot appears. Mammography is the best known use, but the same device type covers lung and other cancers.

The honest framing is a second pair of eyes that never gets tired and never gets bored, attached to a colleague who never gets to make the final call. The radiologist signs the report.

Two failure shapes matter. A mark on something harmless costs a patient a callback, extra imaging, and a bad fortnight. A missed mark costs far more. Any vendor worth talking to will show you both numbers without being pushed.

Clearances per Year

What the FDA Actually Signed Off On

FDA product code
QDQ
FDA name for it
Radiological Computer Assisted Detection/Diagnosis Software For Lesions Suspicious For Cancer
Rule number
21 CFR 892.2090
Risk class
Class 2
Medical field
Radiology
The FDA's own wording: A radiological computer assisted detection and diagnostic software for suspected lesions is an image processing device intended to aid in the detection, localization, and characterization of lesions suspicious for cancer on acquired medical images (e.g., mammography , MR, CT, ultrasound, radiography). The device detects, identifies and characterizes lesions suspicious for cancer based on features or information extracted from the images, and may provide information about the presence, location, and characteristics of the lesion to the user. Primary diagnostic and patient management decisions are made by the clinical user.

Before You Sign Anything

  • Ask for the false positive rate per study, not per lesion. Per study is the number your patients experience.
  • Ask whether the reading room saw the marks before or after forming an opinion. Concurrent and second-read setups produce genuinely different results.
  • Ask how it performs on dense breast tissue, or the equivalent hard case in your use. Average accuracy hides the cases you bought it for.
  • Ask what the retraining schedule is and how you will be told. A model that changes after purchase is a real event, not a footnote.

Who Holds the Most Clearances Here

Companies with the most clearances under AI Cancer Detection Software for Radiology
CompanyCleared products
Screenpoint Medical B.V.7
Therapixel5
DeepHealth, Inc.5
Lunit, Inc.5
Hologic, Inc.4
Icad, Inc.4
BeamWorks, Inc.1
See-Mode Technologies Pty, Ltd.1

Every Cleared Product, Newest First

Each row links to the FDA's own record. The number in the last column is the clearance number, which is the single best thing to ask a vendor for.

All 39 products cleared under AI Cancer Detection Software for Radiology
ProductCompanyClearedFDA record
Transpara (2.1.0) Screenpoint Medical B.V. August 14, 2026 K260898
CadAI-B Dx BeamWorks, Inc. August 12, 2026 K260086
See-Mode Augmented Reporting Tool, Breast (SMART-B) See-Mode Technologies Pty, Ltd. July 28, 2026 K260303
MammoScreen® (5) Therapixel June 26, 2026 K260714
Saige-Dx DeepHealth, Inc. June 15, 2026 K253825
DeepHealth ProstateAI Quantib B.V. April 29, 2026 K253682
Lunit INSIGHT MMG (v1.1.10) Lunit, Inc. April 23, 2026 K260320
Lunit INSIGHT DBT (V1.2) Lunit, Inc. March 26, 2026 K253796
Median LCS (internal name) / eyonis LCS (trade name) (1.0) Median Technologies February 6, 2026 K251474
Saige-Dx DeepHealth, Inc. August 11, 2025 K251873
Genius AI Detection 2.0 Hologic, Inc. July 31, 2025 K243341
MammoScreen® (4) Therapixel July 3, 2025 K243679
QP-Prostate® CAD Quibim S.L. March 18, 2025 K242683
Prostate MR AI (VA10A) Siemens Healthcare GmbH March 5, 2025 K241770
Saige-Dx (3.1.0) DeepHealth, Inc. December 19, 2024 K243688
Transpara (2.1.0) Screenpoint Medical B.V. November 25, 2024 K241831
Saige-Dx DeepHealth, Inc. November 18, 2024 K241747
ProFound Detection (V4.0) Icad, Inc. November 8, 2024 K240417
Lunit INSIGHT DBT v1.1 Lunit, Inc. October 4, 2024 K242652
See-Mode Augmented Reporting Tool, Thyroid (SMART-T) See-Mode Technologies Pte, Ltd. September 9, 2024 K240697
MammoScreen® (3) Therapixel August 1, 2024 K240301
Lunit INSIGHT DBT Lunit, Inc. November 6, 2023 K231470
Genius AI Detection 2.0 with CC-MLO Correlation Hologic, Inc. May 23, 2023 K230096
Genius AI Detection 2.0 Hologic, Inc. October 6, 2022 K221449
Transpara 1.7.2 Screenpoint Medical B.V. August 3, 2022 K221347
ProstatID Scanmed, LLC July 8, 2022 K212783
Saige-Dx DeepHealth, Inc. May 12, 2022 K220105
BU-CAD Taihao Medical, Inc. December 21, 2021 K210670
MammoScreen 2.0 Therapixel November 26, 2021 K211541
Lunit INSIGHT MMG Lunit, Inc. November 17, 2021 K211678
Transpara 1.7.0 Screenpoint Medical B.V. June 2, 2021 K210404
ProFound AI Software V3.0 Icad, Inc. March 12, 2021 K203822
Genius AI Detection Hologic, Inc. November 18, 2020 K201019
MammoScreen Therapixel March 25, 2020 K192854
Transpara Screenpoint Medical B.V. March 5, 2020 K193229
Transpara Screenpoint Medical B.V. December 10, 2019 K192287
ProFound AI Software V2.1 Icad, Inc. October 4, 2019 K191994
PowerLook Tomo Detection V2 Software Icad, Inc. December 6, 2018 K182373
Transpara Screenpoint Medical B.V. November 21, 2018 K181704

Clearance data comes from the FDA's public 510(k) and device classification records through openFDA. A clearance means the FDA agreed the product is close enough to something already on the market, not that the FDA tested it in your hospital. It says nothing about accuracy in your patient mix, your scanners, or your workflow. Confirm the current status and the exact cleared indication on the FDA record before you buy.