How PET/CT Combines Functional and Anatomical Imaging

How a PET/CT scanner combines radioactive-tracer distribution with CT anatomy, why attenuation correction matters, and how registration errors can make two healthy imaging systems produce a misleading fused image

PET/CT is really two imaging systems sharing one patient table.

Back to Biomed Basics

What This Page Explains

This page covers:

The Simple Version

PET/CT combines two different types of imaging. PET detects pairs of 511 keV gamma photons created when a positron emitted by a radiopharmaceutical annihilates with an electron. Those photon pairs are detected in coincidence by opposing PET detector elements, allowing the scanner to estimate where the annihilation occurred. Repeating that process over many events produces a three-dimensional map of radiotracer activity.

CT uses an external X-ray source and detector array to measure how strongly tissues attenuate X-rays. That produces a high-resolution anatomical image. The PET and CT datasets are then geometrically aligned so functional uptake can be displayed on top of anatomical structures. The CT data is also commonly converted into an attenuation map used during PET reconstruction so the system can compensate for PET photons that were absorbed or scattered before reaching the detectors.

A PET/CT study can therefore be wrong even when both scanners technically acquire images. Motion between CT and PET, table-position error, incorrect attenuation correction, CT artifacts, respiratory mismatch, or reconstruction problems can shift uptake or distort quantitative measurements. Preserve whether the abnormality appears on PET alone, CT alone, attenuation-corrected PET only, or the fused display, and whether a phantom reproduces the issue before assuming one detector system has failed.

Start With What PET Actually Measures

PET stands for:

Positron Emission Tomography.

The Patient Is the Radiation Source

A radiopharmaceutical is administered.

The radioactive isotope emits:

Positrons.

Positron

A positron is the antimatter counterpart of an electron.

It has:

The Positron Travels a Short Distance

After emission:

It loses energy in tissue.

Eventually It Encounters an Electron

The positron and electron annihilate.

Annihilation

Their mass is converted into energy.

Two Gamma Photons

The annihilation usually produces two photons of approximately:

511 keV each.

They Travel in Nearly Opposite Directions

This is the key PET event.

PET Does Not Need a Physical Collimator Like SPECT

Instead, it uses:

Coincidence detection.

Coincidence Detection

If two detectors on opposite sides of the scanner detect 511 keV photons nearly simultaneously:

The system assumes they came from the same annihilation event.

Line of Response

A line is drawn between those two detector positions.

The annihilation likely occurred somewhere along that line.

Thousands to Millions of Events

Produce enough information to reconstruct the radiotracer distribution.

Detector Ring

PET scanners contain detector elements arranged around the patient.

Modern PET Detectors

Commonly use scintillation crystals coupled to photodetectors.

Scintillator

Gamma photon interacts.

↓

Light is produced.

↓

Photodetector converts light into electrical signal.

Common Scintillator Materials

Depending on scanner generation, materials may include:

and other high-density scintillators.

Silicon Photomultipliers

Modern systems often use:

SiPMs, or silicon photomultipliers.

Older Systems

May use:

Photomultiplier tubes.

PET Detects Timing and Energy

The detector electronics determine:

Energy Window

Events near the expected 511 keV energy are accepted.

Scatter

A photon that scatters inside the patient may lose energy and change direction.

Energy Discrimination Helps Reject Some Scatter

But correction algorithms are still needed.

Coincidence Timing

The two photon detections must occur within a narrow timing window.

Random Coincidences

Two unrelated photons can sometimes happen to arrive within that window.

Those Must Be Estimated and Corrected

True Coincidence

Both photons came from the same annihilation.

Scatter Coincidence

One or both photons changed direction before detection.

Random Coincidence

Detections came from unrelated events.

PET Reconstruction Accounts for These Effects

Time-of-Flight PET

Modern PET can use:

Time of flight.

Basic Idea

The two photons should leave the annihilation at almost the same time.

If one detector sees its photon slightly earlier:

The annihilation was probably closer to that detector.

Tiny Timing Difference

Can provide additional localization information.

The Scanner Still Uses a Line of Response

But time-of-flight narrows where along that line the event probably occurred.

This Improves Signal Localization

Especially in larger patients.

PET Image

At the end of acquisition:

The scanner has a map of radiotracer concentration.

But PET Anatomy Is Relatively Blurry

Compared with CT.

This Is Why CT Is So Valuable

PET may show:

A bright area.

CT can show:

Which anatomical structure occupies that location.

CT Side of the System

The CT portion works like other CT scanners.

X-Ray Tube Rotates

Radiation passes through patient.

Detector array measures attenuation.

Reconstruction Produces

Anatomical slices.

CT Has Much Higher Spatial Resolution

than PET.

Fusion

The workstation overlays:

PET activity

onto:

CT anatomy.

Color Overlay

PET may be displayed using a color scale.

CT provides grayscale anatomy.

But Fusion Only Works If Coordinates Match

Shared Patient Table

One major advantage of PET/CT is that the patient remains on the same table.

CT Acquisition

Table moves through CT field.

PET Acquisition

Table then moves to PET bed positions.

Position Must Be Known Accurately

Table Encoder

The scanner tracks table location.

Registration Depends on Geometry

The system knows the relative physical position of:

Calibration Establishes This Relationship

PET/CT Alignment Test

Phantoms can verify whether an object appears in the same physical location in:

PET

and:

CT.

Misregistration

If PET uptake appears shifted relative to CT anatomy:

The images are:

Misregistered.

Possible Causes

Not Every Misregistration Is Hardware Failure

This is critical.

CT Is Fast

A CT scan may acquire a region in seconds.

PET Is Slow

PET acquisition may spend minutes at each bed position.

The Patient Breathes Throughout PET

Respiratory Mismatch

CT may capture the diaphragm at one respiratory phase.

PET averages activity over many phases.

Result

Structures near:

may appear shifted between PET and CT.

This Is Often Physiological Acquisition Mismatch

Not encoder failure.

Arms Can Move

If the patient changes arm position between acquisitions:

Fusion can be wrong.

Patient Can Slide

Even a small shift matters.

Table Sag

Patient weight can change table deflection.

System Calibration Accounts for Expected behavior

but excessive or abnormal movement can affect registration.

CT Is Also Used for Attenuation Correction

This is one of the most important PET/CT concepts.

PET Photons Must Escape the Patient

A 511 keV photon can be:

before reaching a detector.

Deep Events Are Harder to Detect

because photons pass through more tissue.

Without Correction

Deep tracer activity could appear artificially low.

Attenuation Map

The scanner needs to estimate how strongly each path attenuates 511 keV photons.

CT Already Measures X-Ray Attenuation

So the system uses CT data to create the PET attenuation map.

But CT Uses Much Lower Photon Energies

Diagnostic CT may use tube potentials such as:

120 kVp.

PET photons are:

511 keV.

The Scanner Must Convert

CT attenuation values are transformed mathematically into estimated attenuation coefficients appropriate for PET photon energy.

CT Numbers Become PET Correction Data

This is why a CT problem can change PET values.

Attenuation-Corrected PET

PET reconstruction applies the attenuation map to compensate for lost photons.

Non-Attenuation-Corrected PET

The system may also retain or generate:

NAC PET

data.

This Can Be Extremely Useful

If an abnormal hotspot appears only on:

Attenuation-corrected PET

but not on:

Non-corrected PET,

then attenuation correction deserves investigation.

Metal

Metal can produce extreme CT values and streak artifacts.

Those Values Feed the Attenuation Map

PET Artifact Can Result

Apparent increased or distorted uptake may appear near:

This Does Not Necessarily Mean Real Tracer Uptake

Contrast Media

CT contrast can also change attenuation values.

Modern Algorithms Account for Much of This

But contrast may influence quantitative PET depending on:

Again

PET and CT are not independent once attenuation correction is involved.

Truncation

If a patient's body extends beyond the CT field of view:

Part of the anatomy may be missing from the attenuation map.

PET Detector May See Activity There

But CT map does not represent it correctly.

Truncation Correction

Modern systems use algorithms to reduce this problem.

But Severe truncation can still matter.

PET Bed Positions

PET often acquires the patient in overlapping axial sections called:

Bed positions.

Table Stops

Data acquired.

Then table moves.

Adjacent Fields Overlap

This helps maintain continuous sensitivity.

Bed-Position Error

Could create:

Detector Sensitivity Is Not Identical Everywhere

Normalization corrects differences between detector elements.

PET Normalization

Similar concept to other detector systems.

Uniform source

or approved calibration procedure is used.

Detector Response Differences Are Characterized

Stable Differences Can Be Corrected

Unstable ones cannot be corrected reliably.

Detector Block Failure

Can produce:

depending on reconstruction and detector architecture.

PET Quality Control

May include:

Time-of-Flight Calibration

Timing differences between detector channels must be extremely well controlled.

Tiny Timing Error

Can shift time-of-flight localization.

This May Not Prevent Basic Imaging

but can degrade performance.

Energy Calibration

Each detector needs to identify the 511 keV photopeak correctly.

Gain Drift

Can move accepted events outside the energy window.

Counts Drop

Scanner May Still Acquire

But sensitivity changes.

PET Count Rate

The amount of detected activity depends on:

Low Counts

Increase image noise.

Not Automatically Hardware Failure

Clinical timing matters.

Radiopharmaceutical Decay

The isotope's half-life matters.

F-18

A commonly used PET isotope has a relatively short half-life.

Delays Matter

If acquisition starts much later:

Less activity remains.

PET Scanner Does Not Control the Biology

Tracer distribution depends on:

A Hot Area Is Not Automatically Equipment Artifact

Conversely

A strange PET pattern is not automatically disease.

Equipment and correction artifacts exist.

Standardized Uptake Value

PET often uses:

SUV, or Standardized Uptake Value.

SUV Attempts to Normalize Activity Concentration

Using factors such as:

depending on calculation method.

SUV Is Quantitative

That makes calibration important.

Dose Calibrator Connection

The administered activity may come from:

A dose calibrator measurement.

If That Value Is Wrong

PET SUV can be wrong.

Patient Weight

If entered incorrectly:

SUV can be wrong.

Injection Time

If incorrect:

Decay correction can be wrong.

Residual Activity

If significant activity remains in syringe:

Actual delivered dose differs from assumed dose.

PET Quantification Is a Chain

Not merely detector counts.

Cross-Calibration

PET scanner and dose calibrator may need appropriate cross-calibration.

If They Disagree

SUV values may shift systematically.

This Is Why Quantitative PET Requires More Than Pretty Images

CT-Based Localization

Suppose PET shows uptake.

Fusion places it on:

A lymph node.

But if registration is shifted:

It may appear on:

Adjacent muscle.

Clinical interpretation can change.

Registration Accuracy Matters

Hardware Registration

Scanner geometry establishes baseline alignment.

Patient Registration

Patient staying still maintains it.

Software Registration

Algorithms may make further adjustments.

All three matter.

PET/CT Phantom

A phantom containing known structures can test:

depending on design.

Known activity concentration

allows quantitative checks.

Phantom Removes Biology

Very important.

If Phantom Is Good

But one patient study is misregistered:

Patient motion/respiration becomes more likely.

If Every Phantom and Patient Is Shifted

System geometry becomes more suspicious.

PET/CT Calibration After Service

Replacing or adjusting:

may require registration calibration.

CT Tube Replacement

May require CT calibrations.

PET Detector Service

May require:

calibration.

Do Not Treat Hybrid System as One Calibration

Each subsystem has its own needs.

CT Failure

If CT is unavailable:

Some systems may not permit normal attenuation-corrected PET workflow.

PET Can Sometimes Acquire Without Diagnostic CT

depending on configuration and procedure.

But attenuation correction still needs an appropriate method.

PET Failure

CT may still operate independently.

Important Isolation Clue

If CT images normally but PET fails:

Shared patient/table systems still matter.

But CT detector and generator are less likely to be the PET detector problem.

CT Normal, Fusion Wrong

Registration becomes interesting.

PET Normal, CT Artifact

Attenuation-corrected PET may inherit a problem from CT.

Check NAC PET

Real-World Example: Uptake Shifted at Lung Base

PET uptake appears slightly displaced from CT lesion near diaphragm.

Phantom registration passes.

Other anatomy aligns normally.

Patient was breathing freely during PET while CT captured one respiratory phase.

This is likely respiratory misregistration, not table encoder failure.

Real-World Example: Hot Artifact Near Hip Prosthesis

Attenuation-corrected PET shows apparent increased uptake near metal.

Non-attenuation-corrected PET does not.

CT has severe metal streak.

The attenuation map is creating or amplifying the apparent abnormality.

Real-World Example: Entire PET Dataset Shifted

Every phantom shows PET data displaced several millimeters relative to CT.

Patient motion is removed from the equation.

Hybrid registration calibration is now highly suspicious.

Real-World Example: SUV Values Drift

Images look normal.

Routine quantitative phantom shows activity concentration reads consistently high.

PET detector QC passes.

Dose calibrator cross-calibration is also checked and found out of agreement.

The quantitative problem involves the measurement chain, not obvious image quality.

Real-World Example: One Detector Region Loses Sensitivity

PET normalization QC shows one detector region consistently low.

Clinical images develop a repeatable artifact.

CT remains normal.

The problem is localized to PET detector response.

Common Mistakes

Treating PET/CT as one detector system. It is two modalities sharing geometry, software, and workflow.

Blaming hardware for respiratory misregistration. PET and CT often represent different respiratory moments.

Ignoring non-attenuation-corrected PET. Comparing corrected and uncorrected PET can help identify attenuation-correction artifacts.

Assuming CT artifacts stay confined to CT. CT data commonly drives PET attenuation correction.

Treating a normal-looking PET image as proof of quantitative accuracy. SUV depends on scanner calibration, administered activity, patient data, and decay correction.

Ignoring table geometry. Accurate registration depends on patient and table coordinates.

A Useful PET/CT Framework

Think:

Radiopharmaceutical

↓

Positron Emission

↓

Annihilation

↓

511 keV Photon Pair

↓

PET Detector Coincidence

↓

PET Reconstruction

and separately:

X-Ray Tube

↓

Patient Attenuation

↓

CT Detector

↓

CT Reconstruction

Then the systems meet at:

Registration + CT-Based Attenuation Correction

↓

Fused PET/CT Image

Another Useful Troubleshooting Split

Ask:

Is the problem present on PET alone, CT alone, or only after fusion?

Then:

Is it present on attenuation-corrected PET and non-corrected PET?

Then:

Does a phantom reproduce the registration or quantitative error?

Those questions can separate:

very quickly.

What Did You Actually Prove?

If the fused PET/CT image appears aligned:

You proved:

The two datasets appear registered under that viewing condition.

You did not necessarily prove:

If CT is normal:

You proved:

The CT subsystem produced an acceptable CT dataset under that condition.

You did not prove:

The PET attenuation map generated from it is artifact-free.

If quantitative PET phantom testing, detector QC, CT QC, and PET/CT registration all pass:

You have much stronger evidence that the hybrid imaging chain is functioning properly under those controlled conditions.

Final Thoughts for Biomeds

PET/CT is not simply:

PET plus a CT picture.

The systems are connected.

PET tells you:

Where tracer activity is occurring.

CT tells you:

What anatomy is there.

Then CT also helps PET correct for photon attenuation.

That makes CT part of PET reconstruction.

Patient motion becomes a fusion problem.

Metal becomes an attenuation-correction problem.

Table position becomes a registration problem.

Dose-calibrator accuracy can become an SUV problem.

A detector that is technically producing images can still be quantitatively wrong.

The useful troubleshooting question is therefore not:

Is PET broken?

or:

Is CT broken?

It is:

At what point do the two imaging chains stop agreeing with each other?

Is PET data wrong?

Is CT data wrong?

Is the attenuation map wrong?

Did the patient move?

Did the datasets align incorrectly?

That is where hybrid imaging starts to make sense.

And, as always:

What did you actually prove?

— Jake

Important Note

PET detector materials, coincidence timing, energy windows, normalization, time-of-flight processing, CT protocols, attenuation-correction algorithms, PET/CT registration, SUV calculations, quantitative calibration, radiopharmaceutical handling, and QC procedures vary significantly by manufacturer, scanner model, software version, isotope, and clinical protocol. PET/CT involves ionizing radiation and radioactive materials. Follow current OEM documentation, nuclear-medicine and CT QC procedures, facility radiation-safety requirements, qualified medical-physics guidance, and authorized service scope when evaluating PET/CT performance.

Related Biomed Basics

Back to Biomed Basics