Questions &
Answers

Insilico PET® is an AI software platform that processes standard CT scan DICOM data and generates a synthetic functional imaging volume that is quantitatively and qualitatively equivalent to what a PET scanner would produce. A conventional PET scan requires injecting a radioactive tracer into the patient, waiting for it to distribute, and then scanning with expensive, specialised PET hardware. Insilico PET® eliminates all of those requirements — using deep learning to extract functional metabolic information directly from the structural CT data already being captured in routine clinical care.

The output is delivered as a standard DICOM series in standardised uptake value (SUV) units — the same metric used in conventional PET — so it integrates immediately into existing radiology reading workflows.

Across our validation studies, Insilico PET® demonstrates a 94% diagnostic concordance with ground-truth PET imaging when assessed by consultant radiologists. This has been validated across three clinical areas — oncology, neurology, and cardiology — and multiple patient cohorts spanning different scanner manufacturers, protocols, and demographics.

In addition to binary concordance, we measure SUV correlation coefficients, lesion detection sensitivity and specificity, and segmentation overlap (Dice coefficient) — all of which fall within clinically acceptable thresholds. Full validation data is available to qualified clinical and research partners upon request.

Insilico PET® is designed to be scanner-agnostic. It has been validated across CT systems from all major manufacturers (Siemens, GE, Philips, Canon) and supports a broad range of reconstruction protocols. The model includes a preprocessing normalisation layer that adjusts for acquisition variability, maintaining output consistency regardless of scanner type.

Specific CT protocol recommendations vary by indication (e.g. contrast vs. non-contrast for different applications), and our clinical team works with each partner institution to optimise protocol selection during the integration process.

Insilico PET® integrates via a DICOM-native, HL7-compatible REST API. Your PACS automatically routes eligible CT series to our cloud inference endpoint (or an on-premises deployment, depending on your data governance requirements). The synthetic PET output is returned as a DICOM series and stored directly back into your PACS, appearing alongside the original CT in the radiologist's reading worklist.

Integration typically takes 2–4 weeks, handled by our technical team with minimal input from your IT department. We have pre-built connectors for major PACS platforms including Sectra, Fujifilm Synapse, Philips IntelliSpace, and others. Full integration documentation is available on request.

RADiCAIT is currently in active pursuit of CE marking under the EU AI Act and EU MDR (Medical Device Regulation) frameworks for the EU/UK market, and has an active Pre-Submission engagement with the FDA for 510(k) clearance for the US market. Our regulatory strategy has been developed with specialist MedTech regulatory counsel.

In the UK, Insilico PET® is currently available for use within MHRA-supervised clinical research programmes and NHS innovation pathway trials. Deployment for routine clinical use outside of research protocols is subject to completion of the regulatory pathway. We actively update clinical partners on regulatory milestones. Please contact us for the latest regulatory status.

Patient data security is a foundational design principle of the Insilico PET® platform. All data is processed in compliance with UK GDPR, the Data Protection Act 2018, and, where applicable, HIPAA (US). CT images are de-identified (DICOM tag stripping) prior to transmission. Data is encrypted in transit (TLS 1.3) and at rest (AES-256).

We offer both cloud-hosted (UK data residency, NHS-grade cloud infrastructure) and on-premises deployment options for organisations with strict data governance requirements. We hold ISO 27001 certification and undergo annual third-party penetration testing. A full Data Processing Agreement (DPA) is provided as part of all partnership contracts.

End-to-end processing time from CT scan upload to synthetic PET output delivery is typically under 10 minutes. This includes PACS routing, DICOM de-identification, model inference, and return delivery to your PACS. Processing time is largely independent of scan length or slice count, as our inference engine is optimised for GPU-accelerated batch processing.

By comparison, scheduling and completing a conventional PET scan often takes anywhere from 5 days to several weeks, depending on availability. The practical difference in turnaround time has significant implications for time-critical diagnoses and treatment planning.

Yes. This is one of the most impactful use cases for Insilico PET®. Pharmaceutical companies can incorporate synthetic PET-derived metabolic response measurements as biomarker endpoints in Phase II and III clinical trials without the logistical constraints of conventional PET (radiotracer supply, restricted site lists, regulatory variance across geographies).

We work with pharmaceutical partners and clinical research organisations (CROs) to establish the appropriate analytical and regulatory framework for endpoint qualification. RADiCAIT provides centralised image analysis services, standardised read protocols, and statistical support as part of our pharma research offering. Please contact us to discuss your specific trial requirements.

Insilico PET® is offered on a usage-based subscription model, priced per scan processed. Pricing is tiered by volume, with preferential rates for NHS Trusts, academic medical centres, and research partnerships. Enterprise and site-licence arrangements are available for large health systems.

For pharmaceutical and CRO partners, project-based pricing is available, structured around the scope and duration of the clinical research programme. Pricing discussions are conducted directly with each partner to ensure the model reflects the specific deployment context and volume. Contact us for a tailored pricing proposal.

We actively welcome academic and NHS research collaborations. RADiCAIT provides research partners with access to our models, APIs, and technical team under formal research agreements. Collaboration formats include joint grant applications (UKRI, NIHR, Horizon Europe), co-supervised PhD studentships, visiting researcher programmes, and data-sharing agreements for model validation studies.

We are particularly interested in partnerships that help extend Insilico PET® to new indications, expand our validation datasets across under-represented populations, and develop explainability frameworks. To discuss a potential collaboration, please get in touch with a brief description of your research interest.