Sharper Images, Lower Dose
A software calibration method for coincidence discrimination in PET imaging
The dose–noise tradeoff in PET
PET imaging is governed by a hard tradeoff between image quality and the radiation dose given to the patient. A substantial fraction of the events a scanner records are not true signal: they are random and scattered coincidences that add noise to the image.
Removing them is one of the central problems in PET reconstruction. Current scatter and random correction methods are statistical — they estimate a background rather than identify each false event — and the residual error from that estimation sets a practical floor on how far dose can be reduced without degrading the image.
The Physivitis method
A software calibration that improves the accuracy with which a PET system distinguishes true coincidences from random and scattered ones. It sits in the signal-processing chain, takes the detector data the scanner already produces, and outputs a better-discriminated set of coincidence events for reconstruction.
In simulation, it reduced the error in identifying random coincidences from approximately 11% to approximately 1%.
Software-only
Requires no change to detector hardware, crystals or electronics. Delivered as software or firmware, operating on data the scanner already acquires.
Calibrated once per scanner
Set up using a single calibration against a standard reference source. Applies to all subsequent scans without per-scan re-tuning.
Drop-in to the existing pipeline
Augments the discrimination stage rather than replacing reconstruction — compatible with an OEM's established imaging chain.
What it enables
Because the method is software and retrofittable, the same benefit can in principle reach an installed base of scanners, not only new hardware — the difference between a feature and a fleet-wide upgrade for an OEM.
Inside the document
What you will read
- ✓Why PET imaging faces a hard tradeoff between image quality and patient radiation dose
- ✓How random and scattered coincidences limit achievable image quality at a given dose
- ✓Where current scatter and random correction methods reach their statistical floor
- ✓What the Physivitis A1-PET method does and where it sits in the signal-processing chain
- ✓Simulation results: random coincidence identification error reduced from ~11% to ~1%
- ✓Two interchangeable benefits: lower patient dose or sharper images / faster scans
- ✓Why software retrofittability means the same gain can reach an entire installed scanner base
- ✓Honest evidence status: simulation-validated, not yet tested on hardware or patients
Who it is for
PET scanner OEMs
A software-retrofittable improvement that can be offered as a fleet-wide upgrade, not just a new-hardware feature.
Clinical research groups
Early access to a method addressing the centre of PET image quality and patient dose, seeking validation on real scanner data.
Hospital imaging departments
Lower administered tracer dose for repeat-imaging protocols, or higher image quality within existing dose constraints.
Radiology technology partners
A patent-pending calibration framework applicable across detection technologies, with the PET application as the first vertical.
Honest framing
This is a simulation-validated method, patent pending, not yet tested on hardware or patients, and not a cleared medical device. The 11%-to-1% figure is a modelled result. Nothing in the document should be read as a clinical claim.
Get the technical overview
Enter your details below. The download link will be sent to your work email within a few minutes.
Interested in validating the method on real scanner data?
Physivitis is seeking imaging OEM and clinical research partners. Download the overview or get in touch.