First AI Software for Integrated Detection and Diagnosis
Median Technologies has received FDA 510(k) clearance for eyonis LCS, the first AI software combining both detection (CADe) and diagnostic assistance (CADx) for lung cancer screening on low-dose CT scans. The clearance marks a significant advance in integrating artificial intelligence into population-level cancer screening.
What sets eyonis LCS apart is its integrated approach: while most AI software is limited to either detecting nodules or classifying them, this is the first to do both in a single product. The software analyzes low-dose CT to characterize pulmonary nodules and produce a malignancy risk score, assisting the radiologist in clinical decision-making.
Validated Clinical Performance
The clearance was based on the RELIVE pivotal clinical trial (NCT06751576), which demonstrated statistically significant improvement (p=0.027) in diagnostic performance when eyonis is used alongside a radiologist, compared to a radiologist alone.
The performance numbers are robust:
- Sensitivity: 93.3% — detects the vast majority of cancers
- Specificity: 92.4% — low false-positive rate
- Negative Predictive Value: 99.9% — when the result is negative, the confidence that there is truly no cancer is extremely high (only 1 false positive per 1,000 exams)
Reimbursement and Commercialization
An important practical aspect: eyonis LCS is already reimbursable under Medicare, at values between $601 and $700 per exam, through CPT category III codes 0721T and 0722T. The company plans U.S. distribution through a partnership with Tempus and expects to obtain European CE marking in the second quarter of 2026.
For professionals working with radiology coding and billing, the existence of a specific CPT code and defined reimbursement value significantly facilitates clinical adoption. Integration with existing DICOM systems enables workflow incorporation without major infrastructure changes.
Impact on Lung Screening
Lung cancer remains the deadliest cancer worldwide. Low-dose CT screening programs have demonstrated mortality reduction but face challenges with inter-reader variability and high exam volumes. AI tools like eyonis LCS can standardize reading quality and increase diagnostic confidence.
Source: AuntMinnie and Diagnostic Imaging
