AI-powered handheld imaging that delivers submillimeter 3D scans, operated by trained healthcare assistants under registered sonographer supervision.
Patients wait weeks for MRI and CT appointments. Costs are prohibitive. Access requires specialist infrastructure that most of the world simply doesn't have.
Patients wait weeks or months for MRI and CT scans. Anxiety builds, disease progresses, and outcomes worsen while queues grow.
MRI costs £500–750 per scan. Machines cost £3M. Healthcare systems can't scale access fast enough to meet diagnostic demand.
Specialist radiographers, shielded rooms, fixed infrastructure. Not available in GP clinics, rural areas, or emergency zones.
A complete rethink of how diagnostic imaging works, from the hardware to the workflow to the output format.
Pairs with a standard smartphone. No trolleys, no fixed rooms. Pocket-sized and deployable in seconds anywhere a clinician goes.
Proprietary deep-learning algorithms reconstruct raw signals into high-fidelity 3D NIfTI volumes with submillimeter resolution in real time.
Guided acquisition workflows enable trained HCAs to acquire images under registered sonographer supervision. All clinical interpretation and reporting is performed by qualified professionals.
Same-day imaging pathways, at the point of care.
No mains power, no shielded room, no capital infrastructure. GP clinics, rural posts, fertility centres, field hospitals.
Following formal MHRA review (ref. CEC 179614), CUS is positioned as a non-medical device image acquisition workflow tool under UK MDR 2002.
CUS reconstructs full volumetric data in real time. These are direct outputs from the system: a before/after musculoskeletal pair showing structural change following physiotherapy.
Pre-physio
Baseline CUS scan showing soft-tissue and structural state of the knee joint prior to physiotherapy intervention showing posterior medial condyle effusion and lateral to the PCL effusion.
MSK module
Post-physio
Follow-up scan of the same patient after treatment, demonstrating resolution of both effusions as assessed by the reporting sonographer.
MSK moduleCUS uses a modular architecture. Each clinical application is a discrete AI model trained on validated datasets. Three modules are commercially live today, with 20+ in active development.
● Live ○ In development
A four-step workflow designed for trained healthcare assistants operating under registered sonographer supervision.
Open the CUS app and select the relevant anatomical module. No configuration required. The AI handles the rest.
On-screen prompts guide probe placement. The AI tracks position in real time and confirms adequate anatomical coverage before proceeding.
Raw signals are processed through the CUS deep-learning engine, producing a full 3D NIfTI volume in under three minutes.
Images are transmitted to PACS. A registered sonographer reviews image quality, performs all clinical measurements, and produces the clinical report.
CUS reconstructs at submillimeter (under 1mm) voxel resolution. The AI reconstruction engine uses proprietary deep-learning to produce 3D volumetric data from standard ultrasound signals. In two independent clinical workflows, measurements taken by registered sonographers from CUS-acquired images showed strong agreement with measurements from conventionally acquired images.
CUS image acquisition is designed to be performed by a trained Healthcare Assistant (HCA) working under registered sonographer supervision. All HCAs complete a formal training programme including classroom instruction and supervised scans before operating independently. The guided acquisition workflow walks the operator through probe placement step by step. All clinical measurements and reporting are performed by registered professionals.
CUS outputs standard NIfTI (.nii.gz) format, the same format used by MRI and CT systems. NIfTI files open natively in 3D Slicer, ITK-SNAP, FSLeyes, and most PACS systems. There is no proprietary lock-in and no new software required for radiologist review.
Following formal MHRA review (ref. CEC 179614), CUS is positioned as a non-medical device image acquisition workflow tool under UK MDR 2002. CUS acquires and formats image data; all clinical measurements, interpretation, and reporting are performed by registered sonographers and radiologists. Three modules (KUB, Prostate, and TA Pelvis) are available for deployment.
The core AI reconstruction algorithms are protected as trade secrets. The competitive advantage lives in the model weights and training methodology, not the hardware, making it extremely difficult to reverse engineer even if the probe hardware were replicated.
We're actively deploying across healthcare sites in the UK. Whether you want a demo, a partnership conversation, or have technical questions, we'd love to hear from you.
jason@carriertech.ukTell us about your setting and we'll arrange a live demonstration of the CUS system.