10.3 PET and PET/CT Systems and QC
Key Takeaways
PET detects near-simultaneous opposite annihilation photons (coincidence) to define lines of response (LORs); time-of-flight (TOF) uses arrival-time differences to localize events along the LOR
Daily blank scans and periodic normalization keep detector efficiencies matched; failed blank/normalization QC blocks accurate imaging
Well-counter (or equivalent dose calibrator/cross-calibration) pathways support quantitative SUV—activity concentration relative to injected dose and body size
CT provides attenuation correction and localization on PET/CT; metal, motion, and truncation distort AC maps
SUV pitfalls include wrong clock/dose times, residual syringe activity, hyperglycemia for FDG, and partial-volume underestimation of small lesions
10.3 PET and PET/CT Systems and QC
Quick Answer: PET = coincidence along LORs (± TOF). Daily blank scan, normalization, and solid SUV cross-calibration. CT AC fails with metal, motion, and truncation; SUV fails with timing, glucose, and partial volume.
Positron emission tomography (PET) images positron emitters (e.g., F-18, Ga-68, Rb-82). A positron annihilates with an electron, producing two 511 keV photons emitted roughly 180° apart. The scanner registers a true event when opposing detectors detect both photons within a short coincidence timing window.
Coincidence, LOR, and Time-of-Flight
| Concept | Meaning |
|---|---|
| Coincidence detection | Two 511 keV interactions accepted as a pair if within the timing window and energy windows |
| Line of response (LOR) | The line connecting the two detection points—annihilation is assumed somewhere along that line |
| True / scatter / random coincidences | Trues = one annihilation; scatter = deflected photon; randoms = unrelated photons within the window |
| Time-of-flight (TOF) | Measures small arrival-time difference to localize the event along the LOR, improving SNR especially in large patients |
PET does not use lead parallel-hole collimators for primary spatial encoding; electronic collimation via coincidence defines geometry. Physical septa appear in some historical 2-D systems (below).
2-D vs 3-D Acquisition Concepts
| Mode | Idea | Trade-off |
|---|---|---|
| 2-D | Interplane septa reduce out-of-plane scatter and randoms | Lower sensitivity |
| 3-D | Septa retracted/absent; accept more oblique LORs | Higher sensitivity; more scatter/randoms—needs good corrections |
Most modern whole-body PET/CT systems operate in 3-D with sophisticated scatter and randoms correction. Exam items may still contrast the concepts.
Core PET QC
Daily Blank Scan
A blank scan (or equivalent daily detector check using a transmission rod/source or manufacturer daily protocol) verifies that detectors respond uniformly enough for clinical use. It is analogous in spirit to a gamma-camera flood: catch dead blocks, unstable electronics, or temperature issues before patients.
| If blank fails | Action |
|---|---|
| Localized cold block | Service; do not image through a dead detector region |
| Global count loss | Check sources, HV, cooling, coincidence timing |
| Pass | Document; proceed with remaining daily checks |
Normalization
Normalization measures relative efficiency of each LOR or detector crystal pair so reconstruction can compensate for efficiency differences. It is performed on a schedule (and after major service) using manufacturer protocols with a known source geometry. Skipping normalization after crystal or module replacement yields patterned artifacts and bad quantitation.
Cross-Calibration / Well Counter for SUV
Standardized uptake value (SUV) compares lesion activity concentration to injected activity normalized by body weight (or lean body mass/BSA variants):
SUV = (activity concentration in ROI) / (injected activity / body mass)
(with consistent units and decay correction)
The PET scanner’s quantitative scale must match the dose calibrator used to assay the syringe. Departments perform cross-calibration (sometimes involving a well counter or phantom with known activity) so that 1 Bq/mL in the phantom reads correctly on PET. Without it, SUVs drift between cameras or after software updates.
CT Attenuation Correction on PET/CT
CT images are converted to a 511 keV attenuation map for PET reconstruction. Benefits: fast AC, anatomic fusion, and localization. Requirements: same patient position and good registration between CT and PET bed positions.
PET Artifacts and Pitfalls
| Problem | Effect on images / SUV |
|---|---|
| Metal implants | CT streak artifacts → incorrect μ → over/under-corrected emission |
| Patient motion | Misregistration; false defects or hot rims; wrong SUV |
| Truncation | Body outside CT FOV → incomplete μ-map |
| Residual activity / wrong injection time | Wrong decay correction → SUV error |
| Extravasation | Dose not fully in patient; SUV underestimated |
| High plasma glucose (FDG) | Competes with FDG; tumor SUV may fall |
| Partial-volume effect | Small lesions look colder than true concentration |
Timing discipline: record assay time, injection time, and scan start; assay residual syringe activity when protocol requires. Clock mismatch between hot lab and scanner is a classic quantitative error.
SUV Basics for the Exam
| Factor | Direction of effect (typical) |
|---|---|
| Longer uptake time (FDG) | Tumor SUV often rises over early intervals |
| Larger patient / wrong weight entry | SUV changes if weight wrong |
| Hotter reconstruction / sharper filter | Can raise max SUV |
| Larger ROI / mean vs max | Different numeric SUV |
Report which SUV metric (SUVmax, SUVmean) and reconstruction the site uses. Compare serial studies only when protocols match.
Randoms, Dead Time, and High-Activity Studies
At high count rates (e.g., Rb-82 MPI, high-activity dynamic studies), random coincidences and dead time rise. Systems estimate randoms (delayed window or singles-based methods) and apply dead-time correction, but extreme rates can still degrade quantitation. Follow manufacturer activity guidelines for injected dose and start times.
Daily / periodic PET QC summary (teaching):
| Test | Typical role |
|---|---|
| Blank / daily detector check | Catch failed blocks before patients |
| Normalization | Relative LOR/crystal efficiencies |
| Well-counter / dose-calibrator cross-cal | Absolute scale for SUV |
| Phantom (uniform / ACR-type) | Image quality, SUV recovery, artifact survey |
Worked SUV pitfall: Injected assay 10.0 mCi at 08:00; residual syringe 1.0 mCi at 08:05 not subtracted; scan at 09:00 with decay correction assuming 10.0 mCi in the patient. True net injected activity is lower → SUV systematically high. Always apply residual and correct clocks.
TOF note: TOF does not remove the need for AC or scatter correction; it improves effective sensitivity/SNR by constraining where along the LOR the event likely occurred.
Memory aid: PET QC Blank, Normalization, Cross-cal → “BNC” before trusting SUV.
In PET imaging, a line of response (LOR) is best described as:
The path of a single free positron before annihilation inside the crystal only
The lead collimator hole that projects 511 keV photons onto one PMT
The line joining two detectors that recorded a coincident annihilation photon pair
The CT x-ray path used solely for diagnostic Hounsfield unit measurement
A daily PET blank scan shows a persistent cold detector block. The most appropriate next step is:
Ignore the finding because reconstruction always interpolates missing data perfectly for SUV
Proceed with clinical FDG scans but avoid reporting SUVmax only
Widen coincidence timing windows until the blank appears uniform
Remove the system from clinical service until the detector problem is corrected and QC passes
Which factor most directly causes underestimation of FDG SUVmax in a small lung nodule even when calibration is correct?
Time-of-flight localization improving SNR in large patients
Partial-volume effect when the nodule is small relative to system resolution
Using CT-based attenuation correction instead of a blank scan for AC
Recording accurate injection time and residual syringe activity
Sections you finish are checked off in the contents.