Everywell will be the agent for seeling Clairity Breast mammography software results to consumers, according to a press release.
In the future, will patients be able to upload a pathology needle biopsy DICOM and get AI assistance? To be continued.
I'll clip the press release below.
https://radiologybusiness.com/topics/artificial-intelligence/health-and-wellness-company-begins-selling-mammography-ai-directly-consumers-249
- Marty Stempniak | October 09, 2026 | Radiology Business | Artificial Intelligence
- A health and wellness company has begun selling a key mammography artificial intelligence solution directly to consumers.
Austin, Texas-based Everlywell recently announced the nationwide rollout and commercial availability of Clairity Breast to its users. Founded in 2015, Everlywell sells lab testing kits and digital diagnostic solutions to patients, “transforming care” for some 60 million people.
The company said it is now offering Clairity Breast for $249 a pop, allowing women to upload their mammogram to the platform. Authorized by the Food and Drug Administration in 2025, the solution predicts patients’ five-year risk of developing breast cancer based on a routine screening.
"Women deserve clear answers about their health, and access to the most advanced tools in healthcare, not years from now, but today. That belief is why I built Everlywell, and it's why we're so proud to partner with Clairity Breast,” entrepreneur Julia Cheek, founder and CEO of Everlywell, said in a statement Oct. 1. “This partnership expands access to cutting-edge preventive care, and that means a woman can learn more about her future breast cancer risk from the mammogram she already gets and walk into her doctor's office with information she's never had before."
Paired with Everlywell’s nationwide reach, those involved believe the partnership will help give women more info about their health and risk of cancer. They note that about 85% of breast cancers occur in women with no family history of the disease.
First AI Tool with FDA De Novo for Breast Cancer Risk
Clairity was founded in 2020 by radiologist Constance “Connie” Lehman, MD, PhD, a Harvard professor of radiology and former chief of breast imaging at Massachusetts General. Last year its flagship product became the first AI-based tool to receive the FDA’s De Novo authorization for predicting breast cancer risk. See FDA home page and review here as DEN240047.
It has scored additional wins since then, including updated National Comprehensive Cancer Network guidelines introducing the product as an option for identifying women at increased risk beginning at age 35.
- [Note that this software predicts cancer risk based on a non-cancer mammogram; there are also software products to predict metastasis or relapse in women with a breast cancer H&E biopsy, e.g. "Oncotype via DICOM."]
- [I believe FDA does not allow sales of DTC RUO/LDT lab tests to consumers; a reason 23andMe got FDA review several years ago.]
The two companies note that Clairity Breast has had limited availability through health systems such as Beth Israel Deaconess Medical Center in Massachusetts and Invision Sally Jobe in Colorado. With the new distribution deal, patients across the U.S. with a qualifying recent mammogram from the past year can access the risk assessment tool via Everlywell’s platform. This includes review and ordering by a licensed provider in Everlywell’s affiliated clinician network and results reviewed and released with next steps.
“The information we need to understand a woman's breast cancer risk has been in her mammogram all along—we just haven't been able to interpret it until Clairity Breast. Now, it can also inform conversations with your healthcare provider about the best screening and risk reduction options for you,” Dr. Lehman said in the announcement.
The companies are encouraging individuals to share the results with their own providers to “inform her screening and risk reduction plan.” Clairity joins other cancer screening options in the Everlywell platform, including an at-home collection colon cancer screening kit and recently launched HPV self-exam option for cervical cancer screening.
STAT News also highlighted the new offering in a report published Thursday. It notes that this is part of a growing trend of preventive medicine products being consumerized online. However, physicians are expressing concern about the idea of putting the onus on individual patients to pay for a new test whose real-world impact has not yet been studied. Firewall at Stat Plus. Writing, "Clairity’s launch on Everlywell marks the first time an AI-based imaging algorithm has used telehealth to reach self-pay patients at a national scale — part of a growing trend of preventive medicine products being consumerized online." This STAT PLUS article is #5 in a series.
Bonus:
500 word AI summary of 22-page FDA Decision Summary for DEN240047.
FDA granted De Novo request DEN240047 for Clairity’s Allix5, an artificial intelligence software device that estimates five-year breast cancer risk from screening mammograms. The decision establishes a Class II device category, “radiological software device to predict future breast cancer risk,” under 21 CFR 892.8500, product code SEZ.
Intended use and workflow. Allix5 analyzes four standard, directly acquired 2D mammographic views from compatible Hologic Lorad Selenia and Selenia Dimensions systems. Synthetic 2D images are excluded. The software returns a numerical probability of breast cancer diagnosis or development within five years, without adjusting for competing mortality. Eligible patients have no known breast cancer at screening. Results are considered after the radiologist completes interpretation; the device is not intended to detect or diagnose cancer, recommend care, or serve as the sole basis for clinical decisions.
Clinical evidence. The pivotal retrospective study evaluated 77,511 examinations from 44,112 patients at five U.S. sites independent of algorithm development. Cancer outcomes were biopsy-confirmed, and statistical methods addressed incomplete follow-up and uncertainty about cancer onset. The study assessed both discrimination—distinguishing patients with different outcomes—and calibration—agreement between predicted and observed risk.
Overall five-year discrimination was AUC 0.70 (95% CI, 0.69–0.72), exceeding a prespecified performance goal of 0.64 informed by traditional risk models. After excluding cancers diagnosed within 90 days following an abnormal index screening assessment, AUC declined to 0.66 (95% CI, 0.66–0.68) but still met the goal. Both populations met the study’s aggregate calibration criterion. This was comparison against a performance benchmark, rather than a direct head-to-head demonstration of superiority over conventional models.
Important limitations. Acceptable aggregate calibration did not ensure accuracy across all risk levels or subgroups. Scores above 5% may substantially overestimate risk, particularly above 10%. Acceptable calibration was not demonstrated in Asian patients, patients younger than 50 or aged 65 and older, or patients with breast implants. Effectiveness in patients with implants was not established. FDA nevertheless found clinically acceptable performance in several other groups, including Black and White patients and patients with dense breasts.
Regulatory significance. FDA concluded that probable benefits outweighed risks with general and special controls. These require independent clinical validation, discrimination and calibration analyses, subgroup and reproducibility testing, software validation, appropriate labeling, and ongoing postmarket performance monitoring, including communication of current performance to users. The evidence supports risk prediction; the retrospective study did not establish that using Allix5 improves patient outcomes.
##
Pubs
AI found this 2022 publication but AI stressed it does not seem to be the same thing as the May 2026 FDA product ["Mirai" here as Yala et al, "Allix5" at FDA]:
Yala A, Mikhael PG, Strand F, et al. Multi-Institutional Validation of a Mammography-Based Breast Cancer Risk Model. Journal of Clinical Oncology. 2022;40(16):1732–1740. Published online November 12, 2021. doi:10.1200/JCO.21.01337. PMID: 34767469
AI also found this insurer quotation, which I insert:
Arkansas Blue Cross and Blue Shield. Mammography-Based AI (Image-only) Risk Prediction Models. Coverage Policy 2026013. Last reviewed September 2026; effective January 1, 2027.
Its Rationale section states that the literature search was updated through May 1, 2026, and reports:
“No published studies reported the validation of the mammography-based AI risk predication model, Clairity Allix5, only the pivotal study that led to the De Novo authorization.”
NCCN
AI found that NCCN seems to endorse the free population based risk calculator (with a cut off of 1.7% x 5 years) or AI to do a similar calculation. Quoting methods 4 and 5, "(4) those over 35 years of age with a 5 year risk of invasive breast cancer >1.7% per Gail model, (5) same "as calculated by an imaging based risk model."
https://www.nccn.org/professionals/physician_gls/pdf/breast-screening.pdf
