Computed Tomography Instrumentation – Lecture 3 Comprehensive Notes
Lecture Goals and Learning Outcomes
Ability to:- Describe specific requirements and operating parameters for CT x-ray sources.
Explain detector requirements and compare scintillation vs gas-filled detectors.
Outline construction/operation of scintillation and gas-filled detectors; list their pros/cons.
Diagrammatically represent 1st → 5th scanner generations and critique their advantages/disadvantages.
Compare 3rd- vs 4th-generation scanners in detail.
Discuss fan-beam acquisition impact on reconstruction algorithms.
Historical Context & Mechanical Demands
1st-generation scanners intuitively illustrate CT principle but had severe limitations (slow, low power, single detector).
Modern scanners differ radically:- Rotary gantry carries <= 1000 kg; rotates several rps => accelerations > 10 g.
Millimetre-scale spatial accuracy required despite high G-forces.
Radiation-related hardware focus: x-ray tube & detector.
CT X-Ray Sources
General Characteristics
Custom, high-performance tubes (e.g. Toshiba MegaCool, GE Performix Pro).
Power rating: 5–7 MJ heat storage (vs 0.3–0.5 MJ diagnostic) => physically larger & costlier; replaced ~ every 9–12 months.
Wide collimation (e.g. 10 mm → 40 mm beam) quadruples usable output.
Mounted with anode–cathode axis parallel to patient long axis; beam emitted perpendicular to patient axis.
Supplied by high-voltage, high-frequency generator mounted on gantry.
Operating Modes & Exposure Factors
Emission: mostly continuous; some systems still pulse 2–3 ms.
kVp range: 80–140 kV (approx 120 kV typical) => effective photon energy after filtration approx 70–80 keV.
Tube current (continuous): 70–320 mA average; peaks up to 800 mA.
Power example:- 140 kV, 300 mA => P = V I = 140 kV x 300 mA = 42 kW mostly dissipated as heat.
Latest tubes: 20–60 kW average, peaks <= 100 kW.
Question answer: at 100 kW & 140 kV => I = P/V = 100,000 W / 140,000 V approx 0.714 A = 714 mA.
Heat Load & Management
Transition to helical CT shortens scan times => need higher beam intensity & smaller focal spots (0.6–1.6 mm) => greater heat flux.
Strategies:- Rotating anodes up to 10,000 rpm with large, thick disks.
Enhanced cooling pathways (graphite backings, liquid metal bearings, advanced oils).
Heat capacity approx 6 MHU vs 1 MHU diagnostic.
Heat-unit formula: HU = kVp x mA x t (time in seconds).
Technique compensation (reduce kVp/mA/time) lowers HU but increases image noise.
Exposure Parameter Trade-offs
mA (or time):- Linearly scales photon number; no energy change.
Double mA => double photons.
Adjusted to minimise motion blur (shorter time), paediatric dose tailoring, or noise control (upwards arrow mAs downwards arrow noise).
kVp:- 15 % upwards arrow kVp approx doubling photon flux (similar to 2x mAs rule).
Too low kVp => photon starvation, high noise.
Higher kVp reduces photoelectric absorption, patient dose, and tube loading; improves detector signal.
Dual focal spots selectable:- Small spot => high resolution, lower max power.
Large spot => higher power, lower resolution (favours large FOV/high attenuation regions like abdomen).
Heel Effect & Over-Beaming
Heel effect: intensity greater on cathode side due to self-absorption in anode; mitigated by tube orientation vs gantry.
MDCT needs loosely collimated broad beam so penumbra lies outside active detectors => increases patient dose (over-beaming).
Modern tubes can magnetically deflect electron beam for focal-spot shift => doubles angular sampling without doubling detectors.
Generator Technology
High-frequency, 3-phase square-wave inverters:- <1 % beam intensity ripple (variation would mimic attenuation & corrupt mu values).
Compact enough to mount on gantry.
Beam Conditioning: Collimation & Filtration
Collimation
Two sets:- Pre-patient collimator: defines slice thickness / beam width nT (n rows x detector pitch).
Post-patient collimator: reduces scatter before detectors.
Adjustable lead blades; further power lost in shaping.
Filtration
Removes low-energy photons (dose only) & equalises beam spectrum to minimise beam-hardening artefacts.
Standard layers: aluminium + copper (few 0.1 mm).- Dual material provides smoother hardening across spectrum; Cu (higher Z) preferentially removes mid-range energies after Al pre-filter.
Order matters: higher-Z (Cu) usually closer to tube to absorb higher-energy tail after Al eliminates low E, but practical designs vary with heat load & bow-tie needs.
Bow-tie filter:- Thicker at edges, thinner centrally => spatially equalises patient dose and detector signal.
Results: lower peripheral dose, more uniform detector exposure.
Graphs show normalised air-kerma (K_a) vs fan angle for small/large filters.
Detector Technologies
Desired Detector Traits
High quantum efficiency (upwards arrow SNR, downwards arrow dose).
Fast temporal response (enable rapid acquisitions without after-glow artefact).
Stability & calibration retention.
Spectral independence to x-ray energy.
Large dynamic range (20-bit digitisation typical).
High packing (geometric) efficiency & reasonable cost.
1. Scintillation Crystal + Photomultiplier Tube (Historic)
80–100 % intrinsic detection efficiency (high density crystals, e.g. NaI(Tl), BGO, CaF₂).
Bulkiness of PMTs => poor packing density & resolution.
Pronounced after-glow and mechanical fragility.
High cost, reliability issues => obsolete in modern CT.
2. Gas-Filled Xenon Ionisation Chambers
Long narrow chambers (approx 10 cm x few mm); electrodes form anti-scatter grid.
25 atm Xe (Z = 54) => approx 50 % intrinsic efficiency.
Advantages:- Cheap, good packing (little dead space), built-in scatter rejection.
Directional sensitivity (must face tube focal spot).
Downsides:- Slow response, lower quantum efficiency vs scintillators.
Hard to fabricate multislice arrays => rare now, incompatible with 4th-gen geometry (directionality conflict).
3. Scintillation Crystal + Photodiode (Current Standard)
Photodiodes are compact, solid-state light transducers.
Typical element: 1 mm x 15 mm (axial).- Axial length sets slice-thickness upper bound.
Reflective gaps reduce cross-talk.
Materials: CsI, CdWO₄, GOS, YSO, GSO ceramics.
Performance:- >98 % intrinsic efficiency.
Sub-µs decay → minimal after-glow.
Excellent stability; ring artefacts rarely from drift.
Modular; plug-in clusters simplify replacement.
Computer & Control Subsystem
Tasks:- Gantry & couch motor control.
HV generator regulation & exposure modulation.
High-speed A/D of detector signals (>= 20-bit depth).
Real-time reconstruction pipeline.
CT Scanner Generations
First Generation (1972-1974)
Pencil-beam, single NaI(Tl)+PMT detector.
Rotate-translate geometry: 160 lateral steps x 180 deg (1 deg) => 4–5 min per slice.
Image matrix 180x180.
Excellent scatter rejection; severe time/dose/after-glow limits; head only (static) imaging; water bags to compress dynamic range.
Second Generation (approx 1973)
Narrow fan-beam (approx 10 deg) + approx 30 detectors.
Still rotate-translate but fewer rotational positions (factor approx detector count).
Scan time downwards arrow to approx 18 s (single breath-hold feasible).
Adoption of Korenblyum fan-beam mathematics.
PMT scintillation detectors; higher scatter than 1st gen despite grids.
Third Generation (approx 1976) – Rotate-Rotate
Wide fan-beam (30–40 deg) covers entire patient; 512–900 detectors on rotating arc.
Pure rotation (no translation) => scan time approx 0.5 s => respiratory motion eliminated.
Mixed detector types; anti-scatter grid possible.
Edge detectors act as reference (I₀) => real-time tube output normalisation.
Challenges: expensive, complex rotating cabling/housing.
Forms basis of most modern helical & multislice scanners.
Fourth Generation (approx 1978) – Rotate-Stationary / Rotate-Nutate
600–4800 stationary detectors form full ring; only x-ray tube rotates.
Nutating version wobbles ring to clear tube path but capture beam.
Requires omnidirectional scintillation detectors (gas chambers not usable).
Pros: simplified cabling; each detector acquires its own I₀ (good for drift correction).
Cons:- Smaller gantry bore; geometric blur reduces spatial resolution.
No anti-scatter grid => more scatter noise.
Needs very stable tube output because reference readings time-shifted among detectors.
Reconstruction cannot start until full rotation completed (every detector supplies complete fan set).
Fifth Generation – Electron Beam CT (EBCT)
Stationary-stationary: no mechanical motion.
Tungsten anode ring beneath patient in vacuum; electron beam magnetically steered approx 360 deg.
Above patient: arc of CdWO₄ detectors.
Newer EBCT: 4 e-beam tracks + dual detector rings => 8 slices simultaneously.
Performance: 50 ms scan, 17 fps, axial resolution 1.5–10 mm, in-plane 0.25–0.5 mm.
Main niche: cardiac imaging (cine, ECG-gated).
Sixth & Seventh Generations (Brief)
Slip-ring continuous rotation (late 1980s) => helical/spiral CT (6th gen).- GANTRY rotates continuously; couch translates continuously → volumetric acquisition without stopping.
Multislice / MDCT (late 1990s onward): multiple parallel detector rows (7th gen) => faster z-coverage, thin collimation.
Fan-Beam Reconstruction Considerations
Traditional filtered back-projection assumes parallel beams.
For fan-beam (3rd + gens): options1. Post-acquisition re-binning: sort raysums into equivalent parallel sets (slows workflow).
Direct fan-beam filtered back-projection (Korenblyum-based) => enables reconstruction to begin after first projection acquired.
Fan-beam algorithms must account for varying ray path lengths & divergence; errors produce artefacts if re-binning or weighting mishandled.
Key Technical & Practical Implications
High x-ray intensity met via high mA, high kVp (>120 kV), larger focal spots.
Rapid heat dissipation via large, fast rotating anodes + high heat capacity.
Spatial resolution in CT derived from detector collimation & matrix/FOV, not geometric unsharpness as in radiography.
Generator ripple <1 % crucial to avoid false attenuation errors.
Over-beaming and over-ranging (penumbra allowing full dose outside active sensors) increase patient dose in MDCT — mitigated by dynamic collimators & focal-spot deflection.
Bow-tie & dual-material filtration critical for dose optimisation and algorithmic accuracy (beam hardening correction).
Detector drift manifests as ring artefacts; modern scintillation + photodiode detectors virtually eliminate this via stability and daily calibration.
Representative Questions / Calculations Embedded in Lecture
Why are powerful tubes essential? Because wider beams, faster rotation, thinner collimation require far greater instantaneous photon flux to maintain SNR while minimising scan time.
Differentiate kVp (peak accelerating potential) vs keV (individual photon energy; spectrum peaks approx 1/3–1/2 kVp after filtration).
Current drawn at 100 kW & 140 kV: 714 mA.
Why high-Z gas (Xe) in detectors? Increases photoelectric & Compton interactions per unit length => boosts quantum detection efficiency.
Why two filtration materials and order relevance? Layered Z values tailor spectral shaping; Cu filters higher-energy tail, Al removes low-energy photons; sequence affects heat loading and beam quality at patient.
Why generator output ripple limited to 1 %? RMS fluctuation appears indistinguishable from patient attenuation, corrupting mu calculations.
Why gas detectors incompatible with 4th-gen geometry? Directional electrodes act as converging grid; orientation changes relative to focal spot during rotation.
Ethical, Clinical & Practical Notes
Higher kVp selected not only for SNR but to reduce patient dose by lowering photoelectric absorption in bone.
Dose penalty from over-beaming in MDCT requires balancing image quality vs radiation safety.
Slip-ring continuous rotation enables faster studies, reducing need for breath-holding and potentially lowering motion-induced repeats.
EBCT offers unique capability for dynamic cardiac imaging, but high cost and limited anatomical coverage reduce widespread adoption compared with MDCT.