Get Started with Medical Imaging

Medical imaging refers to the non-invasive acquisition and processing of images of the human body for clinical applications. Medical imaging often involves collaboration between clinicians, radiologists, and image processing experts. You can use medical image processing for applications including clinical diagnosis, surgical planning, and research.

This list provides some sample applications you can explore in Medical Imaging Toolbox™.

• Diagnostic Systems: Detect tumors in brain MRI scans or breast ultrasound images, COVID-19 in CT scans, or pneumonia in chest X-rays.

• Biomedical Engineering: Generate 3-D models of bones for finite element modeling, 3-D printing, or orthopedic implant design.

• Functional Analysis: Analyze brain function from functional MRI, or estimate respiratory motion from CT scans.

• Pharmaceutical Research: Measure drug efficacy and clearance time.

• Device Design: Develop next generation imaging devices and algorithms.

Medical Imaging Modalities

A medical imaging modality refers to the technology used to capture the image, such as X-ray, magnetic resonance imaging (MRI), or computed tomography (CT). Medical Imaging Toolbox supports most 2-D and 3-D modalities including, but not limited to, CT, MRI, positron emission tomography (PET), single-photon emission CT (SPECT), ultrasound, X-ray, X-ray angiography, mammography, and cine-MRI. You can also use Medical Imaging Toolbox to explore multimodal data sets, which combine multiple modalities. Additionally, the Medical Imaging Toolbox Interface for Cellpose Library support package enables cell segmentation in microscopy images.

Note

Medical Imaging Toolbox functionality generally works with all images in supported medical file formats, regardless of modality.

This table lists a subset of common imaging modalities. Each entry provides basic background information, a representative image, tips for getting started, and links to examples that use the corresponding modality.

Modality	Image	Get Started	See Also
Computed tomography (CT) — Visualize 3-D patient anatomy using X-rays.  More About CT works by rotating an X-ray tube about the patient, applying a narrow X-ray beam at various angles. The scanner measures X-ray attenuation, which depends on tissue composition in the path of the beam. CT is most suitable for bones, but can also visualize soft tissue with or without contrast agents. CT scans are versatile, and faster and more cost-effective than MRI. Mathematically, a CT scanner reconstructs a 3-D image from the raw X-ray attenuation data. There are different options for reconstruction, but generally, an algorithm first applies the Radon transform to obtain projections, represented as a sinogram. Then, the algorithm applies techniques such as the inverse Radon transform or other iterative methods to generate a 3-D image. CT can be prone to poor contrast in soft tissue, and artifacts from patient motion, metal objects, or physical limitations of the X-ray detector.	For an example that uses this image [1], see Display Labeled Medical Image Volume in Patient Coordinate System.	Interactively explore and label CT volumes in a Medical Image Labeler 3-D volume session. Programmatically load CT volumes as a medicalVolume object. Programmatically display CT volumes using volshow, sliceViewer, or montage.	Read, Process, and Write 3-D Medical Images Label 3-D Medical Image Using Medical Image Labeler Segment CT Scan Using MONAI Label Segment Lungs from CT Scan Using Pretrained Neural Network
Magnetic resonance imaging (MRI) — Visualize 3-D patient anatomy using a powerful magnetic field and radio waves.  More About MRI works by applying a strong magnetic field and radio frequency energy that disturb hydrogen nuclei within water molecules. The scanner then measures how long the hydrogen nuclei take to return to equilibrium. This relaxation time varies between different tissues, which helps to create an image. MRI is most suitable for imaging soft tissues such as the brain, spinal cord, organs, and muscle. Different MRI sequences such as T1-weighted, T-2 weighted, Flair, diffusion, and functional MRI images vary the magnetic and RF energies to focus on different tissues. Mathematically, an MRI scanner generates 2-D slice images in the Fourier domain, or k-space. The scanner transforms the k-space images to the spatial domain to get a 3-D image volume. MRI can be prone to artifacts from acquisition noise, undersampling, and patient motion.	For an example that uses this image [2][3], see Brain MRI Segmentation Using Pretrained 3-D U-Net Network.	Interactively explore and label MRI volumes in a Medical Image Labeler 3-D volume session. Programmatically load MRI volumes as a medicalVolume object. Programmatically display MRI volumes using volshow, sliceViewer, or montage.	Read, Process, and Write 3-D Medical Images Brain MRI Segmentation Using Pretrained 3-D U-Net Network Align Brain MRI and Dynamic PET Data
Positron emission tomography (PET) — Visualize metabolic or functional activity in 3-D using radioactive tracers.  More About PET is a type of nuclear medicine imaging that works by tracking radioactive tracers, or radiotracers, in the body. The patient inhales, ingests, or receives an injection of the radiotracer before the imaging procedure. PET imaging can be used to diagnose cancer, using radiotracers that accumulate near metabolically overactive tumors. You can also use PET with inhaled radiotracers to measure lung function, which is useful for radiation treatment planning in lung cancer patients. Mathematically, PET imaging is a 3-D reconstruction of energy emitted from the radiotracer. As the radiotracer decays, it releases positrons that collide with electrons. Each collision releases two gamma rays, which move in opposite directions. The PET scanner detects pairs of gamma rays, and uses algorithms such as filtered back projection and iterative reconstruction to calculate a 3-D image of the radiotracer distribution. PET imaging has low spatial resolution, and scanners often combine PET with a more detailed modality like CT or MRI.	For an example that uses this image [4], see Display Multimodal Medical Image Data from PET and CT.	Interactively explore and label PET volumes in a Medical Image Labeler 3-D volume session. Programmatically load CT volumes as a medicalVolume object. Programmatically display PET volumes using volshow, sliceViewer, or montage. Display multimodal PET/CT data by using the OverlayData argument of the volshow function.	Display Multimodal Medical Image Data from PET and CT Align Brain MRI and Dynamic PET Data
Ultrasound — Visualize 2-D patient anatomy using sound waves.  More About Ultrasound works using a probe that emits ultrasound waves. The probe then measures the strength of reflected echo waves, which depends on tissue composition. Ultrasound is suitable for imaging soft tissue near the surface of the body. Often, clinicians use ultrasound to monitor fetal development during pregnancy, or to image the heart, breasts, and abdomen. You can also use Doppler ultrasound to measure blood flow through the heart and blood vessels. Mathematically, an ultrasound scanner derives a 2-D image or video sequence from the reflected ultrasound waves. Ultrasound images are prone to speckle noise, and can have a limited field of view and resolution.	For an example that uses this image, see Read, Process, and View Ultrasound Data.	Interactively explore and label ultrasound series in a Medical Image Labeler 2-D image session. Programmatically load ultrasound series as a medicalImage object. Programmatically display ultrasound series using implay or montage. Extract and display individual frames by using extractFrame and imshow.	Read, Process, and View Ultrasound Data Label 2-D Ultrasound Series Using Medical Image Labeler Breast Tumor Segmentation from Ultrasound Using Deep Learning
X-ray — Visualize 2-D anatomy using X-rays.  More About X-ray imaging works by measuring X-ray attenuation on a 2-D sensor array or a radiographic film. The patient lies between the X-ray source and the detector. X-ray imaging is fast and cost-effective compared to other modalities, and is a popular option for preliminary clinical diagnosis.	For an example that uses this image, see Create Medical Image Object from Filename.	Interactively explore and label X-ray images in a Medical Image Labeler 2-D image session. Programmatically load X-ray images as a medicalImage object. Programmatically display X-ray images using imshow.	medicalImage
Microscopy — Visualize tissue samples or cells at the microscopic level.  More About Microscopy uses microscopes, which magnify objects that cannot be seen with the naked eye. There are different types of microscopes, such as light, electron scanning, and fluorescence microscopes, that apply different techniques to visualize samples. For example, light microscopes use visible light and lenses to illuminate and magnify a sample, while fluorescent microscopes use specialized stains and fluorescent light sources to highlight specific molecules or cellular structures. Some microscopes capture 2-D images, while others capture 3-D stacks of image slices through a tissue. Microscopy is useful for detecting cancer cells, infections, or other pathogens in tissue samples from patients.	For an example that uses this image, see Getting Started with Cellpose.	Load microscopy images using a function that supports your file type, such as imread or tiffreadVolume. Segment cells using the cellpose object and its object functions.	Choose Pretrained Cellpose Model for Cell Segmentation Refine Cellpose Segmentation by Tuning Model Parameters

Typical Workflow for Medical Image Analysis

Import and Spatial Referencing

A typical medical imaging workflow begins with importing data into the workspace. Medical images are available in file formats such as NIfTI, DICOM, NRRD, Analyze7.5, and Interfile. Medical Imaging Toolbox provides general 2-D and 3-D medical image objects, as well as file-format-specific functions, that you can use to import and export medical images. For more information, see Read, Process, and Write 3-D Medical Images.

Display, Volume Rendering, and Surfaces

Once you have imported a medical image into the workspace, you can view and inspect the image to plan your workflow. Medical Imaging Toolbox provides tools for 2-D and 3-D display, and for generating surfaces for 3-D printing or modeling applications. For more information, see Choose Approach for Medical Image Visualization.

Preprocessing and Augmentation

Image preprocessing and augmentation prepare images for advanced workflows. Use image preprocessing to reduce artifacts such as noise, or to register multiple images. For deep learning workflows, you can increase the size of your training data set using augmentation, which applies random transformations to simulate variations in image acquisition and patient anatomy. For more information, see Medical Image Preprocessing.

Ground Truth Labeling, Segmentation, and Analysis

The toolbox supports interactive and deep learning-based image segmentation and pixel labeling using the Medical Image Labeler app. Label large ground truth data sets for training deep learning networks, or segment individual images to isolate regions of interest for analysis workflows. You can analyze segmented images using standardized radiomics features, or develop custom analysis workflows using image processing functions. For more information, see Get Started with Medical Image Labeler and Analysis and Applications.

References

[1] Medical Segmentation Decathlon. "Lung." Tasks. Accessed May 10, 2018. http://medicaldecathlon.com/. The Medical Segmentation Decathlon data set is provided under the CC-BY-SA 4.0 license. All warranties and representations are disclaimed. See the license for details.

[2] “NITRC: CANDI Share: Schizophrenia Bulletin 2008: Tool/Resource Info.” Accessed October 17, 2022. https://www.nitrc.org/projects/cs_schizbull08/.

[3] Frazier, J. A., S. M. Hodge, J. L. Breeze, A. J. Giuliano, J. E. Terry, C. M. Moore, D. N. Kennedy, et al. “Diagnostic and Sex Effects on Limbic Volumes in Early-Onset Bipolar Disorder and Schizophrenia.” Schizophrenia Bulletin 34, no. 1 (October 27, 2007): 37–46. https://doi.org/10.1093/schbul/sbm120.

[4] Eslick, Enid M., John Kipritidis, Denis Gradinscak, Mark J. Stevens, Dale L. Bailey, Benjamin Harris, Jeremy T. Booth, and Paul J. Keall. “CT Ventilation as a Functional Imaging Modality for Lung Cancer Radiotherapy (CT-vs-PET-Ventilation-Imaging).” The Cancer Imaging Archive, 2022. https://doi.org/10.7937/3PPX-7S22.