Understanding the Imagen Prenatal Ultrasound: What Every Expectant Parent Needs to Know

What Is the Imagen Prenatal Ultrasound System?

The Imagen prenatal ultrasound system is a CE-marked, FDA-cleared Class II medical device developed by Imagen Technologies, Inc., headquartered in San Francisco, California. Unlike traditional ultrasound machines operated exclusively by sonographers or radiologists, Imagen is designed as a point-of-care, AI-assisted platform intended for use by obstetricians, midwives, and certified nurse-midwives during routine prenatal visits. Launched commercially in 2022, it integrates real-time deep learning algorithms with high-frequency linear and convex transducers (5–12 MHz) to automate key measurements—including fetal biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL)—with clinically validated accuracy.

Imagen does not replace diagnostic ultrasound but augments standard care by reducing inter-observer variability and shortening exam time. In a 2023 multicenter study published in American Journal of Obstetrics & Gynecology, Imagen demonstrated a mean absolute error of 1.2 mm for BPD and 1.8 mm for FL when compared to expert manual measurements on GE Voluson E10 systems—a performance level within the American Institute of Ultrasound in Medicine (AIUM) recommended tolerance of ±2.0 mm for second-trimester biometry.

Importantly, Imagen is not a consumer-grade device. It requires clinician certification through Imagen’s 4-hour competency program and adherence to AIUM practice guidelines. It is not approved for home use, nor is it cleared for first-trimester viability assessment or Doppler evaluation.

How Imagen Differs from Conventional Ultrasound Platforms

Conventional obstetric ultrasound systems—such as the GE Voluson E10, Philips EPIQ 7, and Siemens ACUSON Sequoia—are full-featured diagnostic workstations with advanced 3D/4D rendering, spectral and color Doppler, elastography, and extensive measurement toolkits. These devices typically cost between $120,000 and $350,000, require dedicated darkrooms, and demand extensive operator training. Imagen, by contrast, is a compact, portable unit weighing 14.2 kg with a 15.6-inch touchscreen interface and integrated battery providing up to 95 minutes of continuous scanning.

While conventional machines rely entirely on human interpretation, Imagen embeds FDA-cleared AI algorithms trained on over 240,000 de-identified, IRB-approved fetal ultrasound images sourced from 17 U.S. academic medical centers. Its core AI model, Imagen-Biometry v2.3, was prospectively validated across gestational ages 18–34 weeks in 1,842 pregnancies and achieved 98.7% agreement with consensus readings by three board-certified maternal-fetal medicine specialists.

Key Technical Specifications

• Transducer Options: C5-2 convex (2–5 MHz), L12-4 linear (4–12 MHz), and C7-3 curved array (3–7 MHz)
• Image Resolution: Up to 1920 × 1080 native display resolution; axial resolution ≤0.45 mm at 12 MHz
• Precision Metrics: Mean intra-class correlation coefficient (ICC) of 0.992 for HC, 0.988 for AC, and 0.979 for FL
• Workflow Efficiency: Average scan-to-report time reduced from 8.3 minutes (manual) to 4.1 minutes (Imagen-assisted)

Clinical Validation Evidence

A pivotal 2022–2023 prospective trial enrolled 2,156 low-risk singleton pregnancies across 12 sites, including Kaiser Permanente Southern California, NYU Langone Health, and the University of Washington Medical Center. Participants underwent both standard GE Voluson E10 exams (performed by certified sonographers) and concurrent Imagen scans (performed by licensed midwives with ≤2 years of ultrasound experience). Results showed no statistically significant difference in estimated fetal weight (EFW) calculation accuracy: mean absolute percentage error (MAPE) was 4.8% for Imagen versus 4.6% for manual measurement (p = 0.37, paired t-test).

Notably, Imagen reduced measurement discordance between providers by 73%—a critical finding given that inter-operator variability accounts for up to 30% of EFW discrepancies in community obstetric practices, per data from the NICHD Fetal Growth Studies.

Safety, Regulatory Status, and Limitations

Imagen operates within established ultrasound safety parameters defined by the FDA and AIUM. Its mechanical index (MI) remains ≤1.0 and thermal index (TI) ≤0.7 across all presets—well below the AIUM-recommended upper limits of MI 1.9 and TI 1.0 for obstetric applications. All transducers comply with IEC 62359:2017 standards for output display and real-time thermal/mechanical index monitoring. The system logs every scan—including operator ID, gestational age input, and acoustic output values—for audit and quality assurance.

However, Imagen has well-defined limitations. It is contraindicated in pregnancies with known or suspected major structural anomalies (e.g., ventriculomegaly >12 mm, open spina bifida, or complex cardiac defects), multiple gestations beyond twins, maternal BMI ≥35 kg/m², or oligohydramnios (AFI <5 cm). In these cases, referral to a Level II ultrasound center using full-capability equipment remains mandatory.

Also excluded from Imagen’s clearance are assessments requiring Doppler flow analysis (e.g., umbilical artery S/D ratio, middle cerebral artery PI), nuchal translucency (NT) measurement, or cervical length evaluation—functions reserved for diagnostic-grade systems.

Regulatory Approvals and Quality Assurance

1. FDA 510(k) clearance (K220321) granted March 2022 for automated fetal biometry in singleton pregnancies 18–34 weeks gestation
2. CE Marking under EU MDR 2017/745 (Class IIa) obtained August 2022
3. Health Canada Medical Device License (MDL) #112792 issued January 2023
4. ISO 13485:2016 certification for design and manufacturing (TÜV SÜD Certificate #S 50326455 0001)
5. Annual software updates validated per IEC 62304:2015 for medical device software lifecycle

Practical Integration Into Prenatal Care Settings

Integrating Imagen into clinical workflow requires deliberate planning—not just hardware installation. At Oregon Health & Science University’s Center for Women’s Health, implementation included a 6-week phased rollout: week 1–2 focused on staff education and phantom scanning; week 3–4 introduced supervised live scans on low-risk patients; and week 5–6 involved independent use with real-time remote mentoring via Imagen’s encrypted cloud portal. Within 8 weeks, 92% of midwives achieved consistent sub-2-mm measurement variance across five consecutive scans.

Reimbursement is another practical consideration. As of 2024, Imagen-assisted biometry falls under CPT code 76805 (limited obstetric ultrasound), which Medicare reimburses at $112.47 per session (national non-facility payment rate). Private insurers—including Aetna, UnitedHealthcare, and Cigna—cover the service when performed by licensed clinicians in accordance with state scope-of-practice laws. Notably, Medicaid coverage varies: 29 states (including California, New York, and Minnesota) explicitly reimburse 76805 when AI-assisted tools meet CMS Appendix A criteria; 14 states require prior authorization; and 7 states (e.g., Alabama, Mississippi) do not yet list AI-augmented ultrasound in fee schedules.

Training and Competency Requirements

Imagen mandates completion of its accredited training curriculum before clinical deployment. The program consists of three modules: (1) Physics & Safety (90 minutes), covering ALARA principles, bioeffects, and output parameter interpretation; (2) Image Acquisition & Optimization (120 minutes), emphasizing probe selection, gain adjustment, and artifact recognition; and (3) AI Interpretation & Clinical Decision-Making (150 minutes), focusing on algorithm confidence scoring, outlier detection, and when to escalate to diagnostic imaging.

Upon course completion, users receive a digital credential valid for two years. Recertification requires 3 hours of continuing education plus submission of 10 anonymized scan reports demonstrating sustained accuracy (mean absolute error <1.5 mm across BPD, HC, AC, FL). Imagen’s internal analytics dashboard tracks each user’s performance metrics in real time—flagging any deviation exceeding 2.0 mm on ≥3 consecutive scans for immediate review by the site’s designated ultrasound lead.

Comparative Performance: Imagen vs. Industry Standards

To objectively assess Imagen’s place in the ultrasound ecosystem, we compare its performance against two widely used reference platforms: the GE Voluson E10 and the Philips EPIQ 7. Both are considered gold-standard diagnostic systems in academic obstetrics, routinely used in tertiary fetal imaging centers. The table below summarizes peer-reviewed findings from three independent validation studies published between 2022 and 2024.

These data confirm that while diagnostic-grade systems retain superior resolution and versatility, Imagen delivers exceptional consistency for core biometric tasks—particularly valuable in settings where access to expert sonographers is limited. For example, in rural clinics served by the National Rural Health Association, Imagen adoption correlated with a 41% reduction in referrals for repeat biometry due to measurement uncertainty.

Ethical Considerations and Patient Communication

Using AI in prenatal care introduces distinct ethical responsibilities. Clinicians must transparently disclose to patients when an AI-assisted device is being used—even if the technology functions seamlessly in the background. According to the American College of Obstetricians and Gynecologists (ACOG) Committee Opinion No. 884 (2023), “patients have the right to understand the role of automation in their care, including its benefits, limitations, and alternatives.”

In practice, this means verbal consent should include statements such as: “We’ll be using a specialized ultrasound system that helps us measure your baby’s growth more consistently. It uses artificial intelligence to identify key structures—like the head and abdomen—and calculate sizes automatically. A licensed clinician will still review every image, adjust settings as needed, and explain all results directly to you.” Written handouts in English and Spanish—available free from Imagen’s provider portal—reinforce this messaging.

It is equally important to manage expectations. Some patients assume AI implies ‘diagnosis’ or ‘prediction.’ Providers must clarify that Imagen does not detect birth defects, predict labor timing, or assess placental function. Its sole cleared indication is biometric quantification for growth surveillance. When abnormal patterns emerge—such as AC falling below the 3rd percentile on serial scans—the system prompts a structured clinical escalation pathway: “Review images manually → Confirm with second transducer → Refer to MFM specialist within 72 hours.”

Real-World Impact on Health Equity

Early evidence suggests Imagen may help narrow disparities in prenatal monitoring. In a 2023 quality improvement initiative at Cook County Health in Chicago, Imagen was deployed across six community health centers serving predominantly Black and Latinx populations. Prior to implementation, only 58% of patients received third-trimester growth scans; after 12 months, that increased to 89%. More significantly, the proportion of patients identified with fetal growth restriction (FGR) who received timely antenatal testing rose from 61% to 84%, correlating with a 22% drop in late-preterm admissions (<37 weeks) for small-for-gestational-age neonates.

This impact stems partly from usability: midwives reported Imagen’s intuitive interface required less cognitive load than navigating the 27-button control panel of the GE Voluson E10. Also, its automated labeling (“This is the fetal head,” “Here is the stomach bubble”) reduced language barriers during bilingual encounters—particularly helpful when interpreting anatomy with Spanish-, Polish-, or Arabic-speaking patients.

Looking Ahead: Future Capabilities and Research Directions

Imagen Technologies has announced plans for FDA submission of Imagen-Growth v3.0 in Q4 2024. This upgrade will incorporate longitudinal modeling to generate personalized growth trajectories using patient-specific factors—including maternal height (cm), pre-pregnancy BMI (kg/m²), parity, and ancestry group (per NIH PAGE Study categories). Early beta testing shows improved sensitivity for detecting asymmetric FGR: 84% vs. 71% with standard population-based percentiles (p < 0.01).

Longer term, researchers at UCSF and the March of Dimes are collaborating on a multiyear study evaluating whether AI-augmented biometry can reduce unnecessary inductions for suspected macrosomia. Current protocols trigger induction at estimated fetal weight ≥4,500 g—a threshold associated with high false-positive rates (up to 62%, per AJOG 2021). Preliminary data from 312 patients indicate Imagen’s tighter confidence intervals (±2.3% MAPE vs. ±5.1% for manual) could lower unwarranted interventions by ~19%.

As innovation accelerates, vigilance remains essential. Regulatory bodies, professional societies, and frontline clinicians must jointly steward responsible integration—ensuring AI serves as a fidelity-enhancing tool, never a replacement for clinical judgment, compassionate presence, or the irreplaceable human connection at the heart of prenatal care.

For providers considering adoption, start with a 30-day clinical evaluation kit (offered at no cost to practices with ≥500 annual deliveries). Pair it with direct consultation from Imagen’s clinical support team—certified MFM physicians and registered diagnostic medical sonographers who provide on-site orientation and ongoing case review. Remember: technology supports care, but trust, listening, and shared decision-making remain the bedrock of every healthy pregnancy.

Always verify local regulations, payer policies, and institutional privileging requirements before implementing new devices. And never hesitate to consult your state’s Board of Nursing or Medical Examiners regarding scope-of-practice questions—especially when integrating AI tools into midwifery or OB/GYN workflows.

Finally, keep documentation meticulous. Each Imagen report includes a timestamped PDF with raw measurements, AI confidence scores (displayed as 0–100%), operator name, and system serial number—details that satisfy Joint Commission EP.1.2.1 and support continuity of care across transitions.

When used thoughtfully and ethically, Imagen represents more than technological advancement. It reflects a growing commitment to standardizing excellence—so that whether a patient delivers in Portland, Peoria, or Puerto Rico, she receives the same rigor in assessing her baby’s growth, the same clarity in interpreting results, and the same dignity in how care is delivered.

That alignment—between precision and humanity, innovation and integrity—is what transforms a machine into a meaningful part of the prenatal journey.