Introduction
CT: imaging technique using X-rays from multiple angles to create cross-sectional images. Invention: Godfrey Hounsfield and Allan Cormack (1971, Nobel Prize 1979). Advantage: fast (seconds), excellent bone and lung detail, widely available. Limitation: ionizing radiation, limited soft tissue contrast (vs. MRI). Volume: ~90 million CT scans annually in US alone. Evolution: single-slice (1970s) → multi-detector (64-320 slices) → photon-counting (2020s).
"CT transformed medicine by making the invisible visible. For the first time, physicians could see inside the living body in cross-section,no surgery required. It remains the workhorse of diagnostic imaging." -- Diagnostic radiologist
X-ray Physics
X-ray Production
X-ray tube: heated cathode emits electrons, accelerated toward anode. Bremsstrahlung: electrons decelerated by anode atoms produce X-rays (continuous spectrum). Characteristic radiation: electrons knock out inner-shell electrons (discrete energy peaks). Tube voltage (kVp): 80-140 kV (determines X-ray energy/penetration). Tube current (mA): controls X-ray quantity (beam intensity).
X-ray Interactions with Tissue
Photoelectric absorption: X-ray absorbed completely (dominant at low energies, high Z). Compton scattering: X-ray scattered with partial energy transfer (dominant at CT energies). Attenuation: exponential decrease through tissue (I = I0 × e^(-µx)). µ (linear attenuation coefficient): depends on tissue density, atomic number, X-ray energy.
Tissue Attenuation
| Material | HU Value | Appearance |
|---|---|---|
| Air | -1000 | Black |
| Fat | -50 to -100 | Dark gray |
| Water | 0 | Gray |
| Soft tissue | +20 to +80 | Light gray |
| Bone | +400 to +1000 | White |
| Metal | >+1000 | Bright white |
Scanner Design and Components
Gantry
X-ray tube: rotates around patient (0.25-0.5 seconds per rotation). Detector array: opposite tube, receives transmitted X-rays. Slip-ring technology: continuous rotation (no cables to wind/unwind). Bore: 70-80 cm diameter (patient lies inside). Weight: several tons (precision engineering).
Detector Technology
Solid-state: ceramic scintillator + photodiode. Rows: 16 to 320 detector rows (more = faster, wider coverage). Element size: 0.5-1.0 mm (determines resolution). Data rate: gigabytes per second (massive data throughput). Photon-counting: emerging technology (direct conversion, no scintillator).
Patient Table
Movement: precise translation through gantry (mm accuracy). Speed: coordinated with gantry rotation (helical pitch). Weight limit: 200-300 kg (bariatric patients). Carbon fiber: radiolucent, strong, lightweight. Positioning: automated for reproducible scans.
Computer System
Data acquisition: raw detector data collected. Reconstruction: filtered back-projection or iterative reconstruction. Display: DICOM format, multiplanar reformats. Post-processing: 3D rendering, virtual endoscopy, perfusion maps. Storage: PACS integration.
Image Reconstruction
Filtered Back-Projection (FBP)
Classic method: mathematically project X-ray data back to create image. Filter: high-pass filter removes blurring (ramp filter). Speed: fast computation. Quality: well-understood, reliable. Limitation: noisy at low doses.
Iterative Reconstruction (IR)
Approach: forward-project estimate, compare to measured data, refine. Types: statistical (model photon noise), model-based (physics modeling). Advantage: 30-60% dose reduction at equivalent image quality. Disadvantage: computationally intensive (seconds per image). Status: standard on modern scanners.
Deep Learning Reconstruction
AI-based: neural network trained on low-dose and high-dose image pairs. Speed: faster than iterative reconstruction. Quality: approaching or exceeding iterative methods. Advantage: significant dose reduction potential. Status: FDA-cleared products available (TrueFidelity, AiCE).
Matrix and Resolution
Matrix: typically 512 × 512 pixels per slice. Pixel size: FOV / matrix (e.g., 350 mm / 512 = 0.68 mm). Slice thickness: 0.5-5 mm (thinner = more detail, more noise). Spatial resolution: 0.3-0.5 mm (high-resolution mode). Contrast resolution: ability to distinguish similar tissues (depends on dose).
Hounsfield Units and Contrast
Hounsfield Scale
HU = 1000 × (µ_tissue - µ_water) / (µ_water - µ_air)Water = 0 HU (reference)Air = -1000 HUDense bone = +1000 HURange displayed: -1000 to +3000+ HUWindowing
Window width: range of HU displayed (narrow = high contrast, wide = low contrast). Window level: center of displayed HU range. Presets: brain (W:80, L:40), lung (W:1500, L:-600), bone (W:2000, L:400), abdomen (W:350, L:40). Purpose: optimize visualization for specific tissues.
Tissue Characterization
Fat: -50 to -100 HU (diagnostic for lipoma, fat-containing lesions). Fluid: 0-20 HU (simple cysts). Hemorrhage: 50-70 HU (acute blood higher than tissue). Calcium: >100 HU (vascular calcification, kidney stones). Clinical: HU measurement helps characterize lesions without biopsy.
Helical and Multi-Detector CT
Helical (Spiral) Scanning
Concept: continuous rotation + table movement (helical path through patient). Advantage: fast volumetric coverage (entire chest in 2-5 seconds). Interpolation: data between rotations calculated mathematically. Pitch: table distance per rotation / beam width (higher pitch = faster, lower quality).
Multi-Detector CT (MDCT)
Multiple rows: 16, 64, 128, 256, 320 detector rows. Coverage: wider z-axis coverage per rotation. Speed: 64-slice can cover entire body in ~10 seconds. Isotropic voxels: equal dimensions in all directions (enables high-quality reformats). Current standard: 64-128 detector rows.
Dual-Energy CT
Concept: acquire data at two different X-ray energies simultaneously. Methods: dual-source (two tubes), rapid kVp switching, dual-layer detector. Application: material decomposition (separate calcium from iodine), virtual non-contrast, gout crystal detection. Advantage: additional tissue characterization beyond standard CT.
Cardiac CT
ECG-gating: synchronize acquisition with heart rhythm. Temporal resolution: 66-175 ms (must freeze cardiac motion). Application: coronary artery disease, calcium scoring, structural heart assessment. Requirement: low heart rate (<65 bpm) for best quality (beta-blocker pre-medication).
Contrast Agents
Iodinated Contrast
Mechanism: iodine absorbs X-rays strongly (increases attenuation). Administration: IV injection (1-2 mL/kg, power injector). Phases: arterial (25-30 sec), portal venous (60-70 sec), delayed (3-5 min). Application: vascular imaging, tumor enhancement, organ perfusion. Cost: $50-200 per dose.
Contrast Reactions
Mild: nausea, warmth, metallic taste (common, self-limited). Moderate: urticaria, vomiting (treat with antihistamine). Severe: anaphylactoid (bronchospasm, hypotension, laryngeal edema), rare (~0.01%). Treatment: epinephrine, IV fluids, airway management. Pre-medication: steroids + antihistamine if prior reaction.
Contrast-Induced Nephropathy
Risk: patients with renal insufficiency (eGFR <30 mL/min/1.73m²). Mechanism: direct tubular toxicity + renal vasoconstriction. Prevention: hydration (IV normal saline), minimize contrast volume. Monitoring: creatinine 48-72 hours post-contrast. Incidence: decreasing with modern iso-osmolar contrast agents.
Oral Contrast
Material: dilute barium or iodinated water-soluble contrast. Application: bowel opacification (distinguish bowel from other structures). Timing: 1-2 hours before scan (transit time). Alternative: water as negative contrast (CT enterography).
Special Techniques
CT Angiography (CTA)
Technique: bolus-tracking IV contrast, arterial phase acquisition. Application: pulmonary embolism, aortic dissection, stroke (intracranial vessels). Speed: entire chest in 2-5 seconds (motion-free). Post-processing: MIP, VRT, CPR (curved planar reformation). Advantage: fast, widely available, excellent for emergencies.
CT Perfusion
Technique: repeated scans during contrast passage (time-attenuation curves). Parameters: cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT). Application: acute stroke (identify ischemic penumbra), tumor vascularity. Dose: higher than standard CT (repeated scanning). Processing: deconvolution algorithms.
Virtual Colonoscopy (CT Colonography)
Preparation: bowel cleanse + CO2 insufflation. Scanning: prone and supine positions. Processing: 3D fly-through visualization (virtual endoscopy). Application: colorectal cancer screening (alternative to optical colonoscopy). Sensitivity: >90% for polyps >10 mm. Limitation: requires bowel prep, no biopsy capability.
High-Resolution CT (HRCT)
Technique: thin slices (0.5-1 mm), bone algorithm, no contrast. Application: lung parenchymal disease (interstitial lung disease, bronchiectasis). Detail: visualize individual secondary pulmonary lobules. Advantage: exquisite lung detail. Standard: essential for ILD diagnosis and monitoring.
Radiation Dose
Dose Metrics
CTDIvol: dose to phantom per rotation (mGy). DLP: dose × length (mGy·cm). Effective dose: whole-body risk equivalent (mSv). Conversion: effective dose ≈ DLP × conversion factor (organ-specific). Background radiation: ~3 mSv/year (natural).
Typical Doses
| Examination | Effective Dose (mSv) | Equivalent Background |
|---|---|---|
| Head CT | 2 | 8 months |
| Chest CT | 7 | 2.3 years |
| Abdomen/Pelvis CT | 10 | 3.3 years |
| Coronary CTA | 5-15 | 1.7-5 years |
Dose Reduction Strategies
Tube current modulation: automatically adjusts mA based on patient thickness. Iterative reconstruction: reduces noise at lower dose. Low kVp: for smaller patients and contrast-enhanced studies. Organ shielding: bismuth shields for breast, thyroid (controversial). Justification: every scan should have clear clinical indication (ALARA principle).
Cancer Risk
Estimated risk: ~0.05% lifetime cancer risk per 10 mSv CT (uncertain at low doses). Linear no-threshold (LNT) model: assumes any radiation dose carries some risk. Controversy: LNT may overestimate risk at diagnostic levels. Pediatric: higher concern (longer life to develop cancer, more radiosensitive). Principle: benefit must outweigh risk for every scan.
Image Artifacts
Metal Artifact
Cause: high-density metal objects (joint replacements, dental fillings). Effect: streaking, beam hardening (dark/bright bands). Reduction: higher kVp, thinner slices, metal artifact reduction algorithms (MAR). Clinical impact: limits evaluation near prostheses.
Motion Artifact
Cause: patient movement during scan. Effect: blurring, double contours, streaking. Prevention: fast scan (helical MDCT), breath-hold instructions, sedation if needed. Cardiac: ECG-gating eliminates cardiac motion.
Beam Hardening
Cause: polychromatic X-ray beam (low-energy photons absorbed preferentially). Effect: cupping artifact (edges brighter than center), dark bands between dense structures. Correction: beam hardening correction algorithms, filtration. Common: between petrous bones (posterior fossa artifact).
Partial Volume Effect
Cause: voxel contains multiple tissue types. Effect: averaged HU value (neither tissue accurately represented). Reduction: thinner slices. Clinical: small lesions may be missed or mischaracterized.
Clinical Applications
Emergency Medicine
Trauma: whole-body CT (head, C-spine, chest, abdomen, pelvis). Stroke: non-contrast head CT (exclude hemorrhage), CTA (vessel occlusion), perfusion (penumbra). Pulmonary embolism: CTA chest (gold standard). Aortic emergency: dissection, aneurysm rupture. Speed: entire trauma CT in <60 seconds.
Oncology
Staging: tumor size, lymph nodes, metastases (TNM staging). Monitoring: response to treatment (RECIST criteria). Screening: low-dose lung CT (lung cancer screening program). Biopsy guidance: CT-guided needle biopsy (precise targeting).
Pulmonary
Lung disease: HRCT for interstitial lung disease, emphysema, bronchiectasis. Pulmonary embolism: CTA with high sensitivity/specificity. COVID-19: ground-glass opacities (characteristic pattern). Airways: virtual bronchoscopy, airway measurements.
Musculoskeletal
Fractures: complex fractures (pelvis, spine, calcaneus). Pre-surgical planning: 3D reconstructions for orthopedic surgery. Bone tumors: cortical destruction, periosteal reaction. Advantage: superior bone detail compared to MRI.
Emerging CT Technologies
Photon-Counting CT
Detector: directly converts X-ray photons to electrical signals (no scintillator). Advantage: higher resolution, lower noise, energy discrimination. Multi-energy: simultaneous spectral information (material decomposition). Dose: potential for significant dose reduction. Status: first clinical systems installed (Siemens NAEOTOM Alpha, 2021).
AI in CT
Reconstruction: deep learning reduces noise, improves image quality. Detection: AI detects pulmonary nodules, hemorrhage, fractures. Quantification: automated measurements (tumor volume, calcium scoring). Workflow: prioritize urgent cases (triage). Evidence: FDA-cleared products available for multiple applications.
Spectral CT
Material decomposition: separate iodine, calcium, water. Virtual non-contrast: subtract iodine from enhanced images (eliminate non-contrast scan). Gout: identify uric acid crystals (specific diagnosis). Kidney stones: differentiate stone composition. Application: reducing radiation dose, improving tissue characterization.
Ultra-High Resolution CT
Detector element: 0.25 mm (vs. standard 0.5 mm). Resolution: up to 0.15 mm spatial resolution. Application: temporal bone, coronary stents, lung detail. Trade-off: higher dose for equivalent noise. Status: clinical systems available (Canon Aquilion Precision).
References
- Bushberg, J. T., Seibert, J. A., Leidholdt, E. M., and Boone, J. M. "The Essential Physics of Medical Imaging." Lippincott Williams & Wilkins, 4th ed., 2020.
- Hounsfield, G. N. "Computerized Transverse Axial Scanning (Tomography)." British Journal of Radiology, vol. 46, no. 552, 1973, pp. 1016-1022.
- McCollough, C. H., Leng, S., Yu, L., and Fletcher, J. G. "Dual- and Multi-Energy CT: Principles, Technical Approaches, and Clinical Applications." Radiology, vol. 276, no. 3, 2015, pp. 637-653.
- Brenner, D. J., and Hall, E. J. "Computed Tomography,An Increasing Source of Radiation Exposure." New England Journal of Medicine, vol. 357, no. 22, 2007, pp. 2277-2284.
- Willemink, M. J., and Noel, P. B. "The Evolution of Image Reconstruction for CT,From Filtered Back Projection to Artificial Intelligence." European Radiology, vol. 29, no. 5, 2019, pp. 2185-2195.