Chapter 2: Imaging Modalities

2.1 X-ray Radiography

X-ray radiography remains the most widely used imaging modality, providing fast, low-cost imaging of bones and soft tissues.

Physical Principles

X-rays are electromagnetic radiation produced when electrons strike a target material. Different tissues absorb X-rays at different rates based on their density and atomic composition. Dense structures like bone absorb more X-rays, appearing white on images, while air-filled structures appear black.

Clinical Applications

Application	Common Indications	Key Findings
Chest X-ray	Pneumonia, heart failure, lung cancer	Infiltrates, cardiomegaly, masses
Musculoskeletal	Fractures, arthritis, bone tumors	Fracture lines, joint space narrowing
Abdominal	Bowel obstruction, kidney stones	Air-fluid levels, calcifications
Dental	Caries, periodontal disease	Decay, bone loss, impacted teeth

2.2 Computed Tomography (CT)

CT scanning uses multiple X-ray projections and computer reconstruction to create detailed cross-sectional images, providing superior anatomical detail compared to conventional radiography.

Technical Components

• X-ray Tube: Rotating source producing collimated beam
• Detector Array: Multiple rows detecting transmitted radiation
• Gantry: Rotating frame containing tube and detectors
• Computer System: Image reconstruction algorithms

CT Generations and Technologies

Technology	Configuration	Advantages
Single-slice CT	One detector row	Basic applications, lower cost
Multi-detector CT (MDCT)	64-320 detector rows	Fast scans, cardiac imaging
Dual-energy CT	Two X-ray energies	Material characterization
Photon-counting CT	Energy-resolving detectors	Reduced dose, spectral imaging

2.3 Magnetic Resonance Imaging (MRI)

MRI uses strong magnetic fields and radio waves to generate detailed images of soft tissues without ionizing radiation. It provides superior soft tissue contrast compared to CT.

MRI Sequences

Sequence	Characteristics	Clinical Use
T1-weighted	Fat bright, fluid dark	Anatomy, post-contrast enhancement
T2-weighted	Fluid bright, muscle dark	Pathology detection, edema
FLAIR	T2 with CSF suppression	Brain lesions near ventricles
DWI	Water molecule motion	Stroke, tumor cellularity
MRA	Blood flow imaging	Vascular anatomy, aneurysms

2.4 Ultrasound

Ultrasound uses high-frequency sound waves to produce real-time images. It is portable, radiation-free, and relatively inexpensive, making it ideal for bedside imaging.

Ultrasound Principles

• Transducer: Emits sound waves and receives echoes
• Frequency: Higher = better resolution, lower penetration (2-18 MHz)
• Reflection: Echoes return from tissue interfaces
• Doppler: Detects blood flow direction and velocity

Clinical Applications

Application	Organs/Structures	Key Capabilities
Obstetrics	Fetus, placenta, uterus	Real-time fetal imaging, anomaly detection
Cardiac (Echo)	Heart chambers, valves	Function assessment, valve disease
Abdominal	Liver, gallbladder, kidneys	Masses, stones, cysts
Vascular	Carotid, peripheral vessels	Stenosis, DVT, aneurysms
Point-of-Care	Multiple organs	Rapid bedside assessment

2.5 Nuclear Medicine

Nuclear medicine uses radioactive tracers to image physiological processes rather than just anatomy. This functional imaging reveals metabolism, blood flow, and receptor binding.

PET Imaging

Positron Emission Tomography (PET) detects gamma rays from positron-emitting radiotracers. The most common tracer, 18F-FDG, accumulates in metabolically active tissues, making PET invaluable for oncology.

SPECT Imaging

Single Photon Emission Computed Tomography (SPECT) uses gamma-emitting tracers and rotating cameras. Common applications include cardiac perfusion imaging and brain blood flow studies.

2.6 Mammography

Mammography is specialized low-dose X-ray imaging optimized for breast tissue. It is the primary screening tool for breast cancer detection.

Mammography Technologies

Technology	Description	Advantages
2D Digital Mammography	Standard digital X-ray	Established screening standard
3D Tomosynthesis (DBT)	Multiple angles, reconstructed slices	Reduced overlapping tissue, better detection
Contrast-Enhanced	Iodine contrast injection	Tumor vascularity assessment
Breast MRI	MRI with contrast	High-risk screening, staging

2.7 Emerging Modalities

New imaging technologies continue to expand diagnostic capabilities:

Advanced and Hybrid Imaging

• PET/CT: Combined metabolic and anatomical imaging
• PET/MRI: Soft tissue detail with functional imaging
• Optical Coherence Tomography: Microscopic imaging (ophthalmology, cardiology)
• Photoacoustic Imaging: Combines light absorption and ultrasound
• Molecular Imaging: Targeted tracers for specific disease markers

2.8 Simulator ENUM and Threshold Mapping (Format Domain)

The diverse medical imaging formats surveyed in this chapter map directly onto the simulator's FORMAT ENUM. Each ENUM standardises a serialisation strategy and a corresponding header structure, ensuring both semantic and syntactic interoperability between systems with different native storage choices.

Table 2.8.1 — Format ENUM to Data Model Mapping

ENUM	Extensions	Primary Use	Byte Order
DICOM_3_0	.dcm	Clinical imaging	Little-Endian (default)
DICOM_PS3	.dcm	DICOM Part 3 IODs	Little-Endian
NIFTI	.nii / .nii.gz	Neuroimaging research	Little-Endian
NIFTI_2	.nii / .nii.gz	NIfTI-2 with 64-bit indexing	Little-Endian
NRRD	.nrrd / .nhdr	3D medical research	Endian explicit
METAIMAGE	.mha / .mhd	ITK image processing	Endian explicit
MINC	.mnc	Montreal Neurological Institute	HDF5-based
ANALYZE_7_5	.hdr + .img	Legacy neuroimaging	Endian explicit
HDF5	.h5 / .hdf5	Large-scale scientific data	Platform independent
OME_TIFF	.ome.tif	Microscopy and pathology	TIFF-based

Each ENUM is selectable from the simulator's Panel 1 (Algorithm) format dropdown. Upon selection the simulator surfaces the header layout, pixel data alignment, and spatial transform matrices (qform/sform in NIfTI, IJK to LPS in DICOM, space directions in NRRD) so that learners can compare structural differences side by side.

NIfTI Header Essentials

The NIfTI format has become the de facto research standard for neuroimaging. The header is a fixed 348 bytes followed by an optional extension section and the pixel data block. Key fields include dim[8] (dimensions), pixdim[8] (voxel spacing), datatype (8/16/32/64-bit numeric types), qform_code and sform_code (spatial transform codes), quatern_b/c/d (quaternion rotation parameters), and srow_x/y/z (affine transformation rows).

NIfTI-2, standardised in 2011, preserves the semantic structure of NIfTI-1 while expanding indexing from 16-bit to 64-bit. This permits single-file representation of ultra-high-resolution imaging — for example, 7-Tesla MRI acquisitions at 0.5 mm isotropic resolution covering the entire brain. The simulator's NIFTI_2 ENUM is exposed separately from NIFTI to prevent the common implementation error of conflating the two specifications.

Table 2.8.2 — DICOM versus NIfTI Core Differences

Aspect	DICOM	NIfTI
Primary purpose	Clinical imaging	Research neuroimaging
File granularity	One file per slice	One file per volume
Metadata count	4000+ tags	~40 fields
Patient information	Rich (PHI)	Minimal (anonymised)
Spatial representation	Direction cosines	Affine matrix and quaternions
Standards body	NEMA	NIfTI Data Format Working Group
Conformance declaration	Mandatory	Informal

2.9 Format Conversion and Data Integrity

Format conversion is a routine activity in medical imaging pipelines but it presents non-trivial data integrity risks. The dominant conversion paths are clinical DICOM to research NIfTI (typically via dcm2niix), DICOM to NRRD (typically via ITK or 3D Slicer), and Analyze 7.5 to NIfTI (modernisation of legacy datasets).

Information at risk of loss or distortion during conversion includes patient identifiers (which research formats deliberately strip), pixel units (Hounsfield Units, Standardised Uptake Values, or arbitrary modality units), slice ordering (cross-referencing DICOM (0020,0013) Instance Number and (0020,1041) Slice Location), spatial coordinate systems (LPS in DICOM versus RAS in NIfTI), and the consolidation of multi-frame or multi-echo structures into single volumes.

"Conversion from DICOM to NIfTI is in principle non-destructive, yet without rigorous verification of slice ordering, coordinate system transformation, and metadata preservation it can introduce systematic errors that propagate silently into downstream analyses." — dcm2niix documentation, 2024.

BIDS — A Structural Standard for Neuroimaging

The Brain Imaging Data Structure (BIDS) specification, first released in 2016, standardises the directory layout and naming conventions for NIfTI-based neuroimaging datasets. BIDS does not redefine the file format; instead, it defines how files are organised, named, and accompanied by JSON sidecars carrying the clinical metadata that DICOM tags would otherwise express. Major datasets including OpenNeuro, the Human Connectome Project, and UK Biobank have adopted BIDS as their distribution format.

OME-TIFF — Microscopy and Digital Pathology

The Open Microscopy Environment TIFF (OME-TIFF) format has become the de facto standard for digital pathology and biological microscopy. It combines a multi-page TIFF container with an OME-XML metadata block, allowing multi-dimensional (channel, time, Z-stack) imagery and rich experimental metadata to coexist in a single file. The simulator's OME_TIFF ENUM is reused in Chapter 5 to drive the whole-slide image workflow, automatically surfacing the pyramidal tile structure.

2.10 Korean Edition Note (한국어판 안내)

The Korean edition (한국어판) extends this chapter with section §2.A on Korean medical imaging standardisation infrastructure, covering the Korean Ministry of Food and Drug Safety (MFDS) Guidance on Medical Imaging Standardisation (2024), the Korean Society of Medical Informatics (KOSMI) DICOM Standardisation Committee, the Electronics and Telecommunications Research Institute (ETRI) medical imaging R&D programmes, and dominant domestic vendors Infinitt Healthcare and Vieworks. Readers concerned with Korean regulatory positioning are referred to the Korean edition.

2.11 Beyond Pixel Data — Auxiliary Format Concerns

Medical imaging formats are not limited to pixel containers. Increasingly important auxiliary formats include DICOM Structured Reports (DICOM SR) for AI inference results, DICOM Segmentation Objects for tissue masks, DICOM Surface Segmentation for 3D mesh storage, and DICOM Encapsulated PDFs for legacy report carriage. Each of these auxiliary objects participates in the same Patient–Study–Series–Instance hierarchy and is governed by the same Conformance Statement matrix as conventional image objects.

The growth of AI-driven imaging analytics has elevated DICOM SR to a position of strategic importance. Modern AI inference pipelines produce probabilistic findings (e.g., lung nodule, 87 per cent confidence, RUL apex), heat maps, and bounding boxes; persisting these structured outputs in DICOM SR (with templates governed by DICOM PS3.16) ensures that downstream PACS, radiology workstations, and electronic health records can render the findings using vendor-neutral viewers.