Why a SPECT image looks noisy, and what to check first

By Vvon Engineering Team | Published 2026-06-24 | Updated 2026-09-13

An equipment and workflow review of low-count or noisy SPECT images, covering acquisition, collimation, calibration, reconstruction, motion and environment.

Nuclear medicine RFQ checklist (TXT)

SPECT noise and artefacts can reflect counting statistics, patient factors, acquisition, correction or reconstruction. Review the raw data, QC and protocol with the responsible clinical and physics team before changing settings. Similar appearances do not necessarily have the same cause.

Start with the counts

Was the prescribed activity actually administered? Check the dose calibrator record, and check the residual left in the syringe. On small doses the residual matters.
Extravasation. Look at the injection site. Activity sitting in the arm is not in the organ, and this is one possible contributor that the clinical team should assess.
Time per projection and number of projections. Compare against the written protocol, not against memory. Protocols get edited.
Matrix and zoom. A finer matrix spreads the same counts over more pixels. If someone increased the matrix to make the image look sharper, they made it noisier.
Orbit. For a parallel-hole collimator, increased source-to-collimator distance mainly degrades spatial resolution; sensitivity behaviour depends on geometry and cannot be described as a universal simultaneous count loss. Body contouring is there for a reason.
Patient size. A larger patient attenuates more. That is not a fault, but it changes what a reasonable acquisition looks like, and the protocol should say so.

Next, the collimator and the energy window

Confirm the fitted collimators and their compatibility with the radionuclide, protocol and calibration. A different sensitivity/resolution tradeoff may be appropriate, but a general-purpose collimator is not an automatic fix. Any protocol change needs physics and clinical validation, including its effect on quantification.

Now the energy window. Check that the peak has not drifted, that the window width is what the protocol specifies, and, most importantly, that the isotope selected in the protocol is the isotope in the patient. A window set for one photopeak and an acquisition of another is a spectacular way to lose counts, and it happens on the day a department runs an unfamiliar study.

Uniformity and geometry

Review flood results and trends, together with the validity of the correction data for the current detector configuration. Uniformity errors can create tomographic artefacts. Correction maps should follow manufacturer and QA requirements; an old map is not inherently worse than no correction.

Then centre of rotation and head alignment. A drifted centre of rotation blurs the reconstruction in a way that looks like poor statistics. On a dual head, misalignment between the two detectors produces the same appearance. Both are measured, not judged by eye, and both should have a record showing when they were last verified.

In a scintillation-camera detector, photomultiplier problems can affect flood images. CZT systems have different detector electronics and QC methods. Use the manufacturer-specific diagnostics rather than apply a photomultiplier explanation to every SPECT system.

Reconstruction settings

This is the commonest software cause and the easiest to test, because you can reprocess the same raw data repeatedly.

Iterations and subsets. Their product drives convergence, and past a point it amplifies noise rather than recovering detail. If the department raised the iteration count to sharpen images, this is very likely the answer.
Filter cut-off. A cut-off set too high on a low-count study passes through exactly the high-frequency variation you are complaining about. Filters have to suit the count level, not the anatomy alone.
Filtered back projection without adequate smoothing produces streaks that are often called noise. They are not noise, they are the algorithm.
Resolution recovery. Useful, and capable of producing edge overshoot and a mottled texture when applied at the wrong strength.
Attenuation correction misregistration on a SPECT/CT system. A shift between the CT and the emission data produces apparent defects and uneven noise. Check the fused images before blaming the detector.

Last, the patient and the room

Review appropriate projection cine, sinogram and reconstructed/fused images for motion and positioning artefacts. Motion is not visible only in raw data, and correction or repeat imaging must follow the responsible clinical decision. Do not repeat exposure solely to test an equipment hypothesis.

And the room itself, which matters more in this climate than the textbooks suggest. Detector electronics drift when they run warm, and an air conditioning unit that failed overnight will show up as a system that peaks differently in the morning. Humidity and unstable mains are the two environmental causes worth ruling in early, because they explain a pattern of problems rather than a single bad study.

Verify proposed technical corrections with suitable phantom data and the approved acceptance criteria. Record the configuration, baseline, results and authorisation before clinical use. A phantom check supports the investigation but cannot guarantee a fixed turnaround or resolve every patient-specific cause.

Technical references

Have a project to scope?

Share your application, equipment requirements and procurement timeline. Our engineers can help clarify the configuration and scope for your enquiry.

Discuss your requirements

Browse buying and selection guides | Back to all Insights