4. Magnetic Resonance Imaging (MRI)

Home » Radiology » 4. Magnetic Resonance Imaging (MRI)
Headshot of George Keng, MD, Thread Director
George Keng, MD
Thread Director
envelope icon
Headshot of Julie Kaczmark, Clinical Assistant Professor
Julie Kaczmark
Clinical Assistant Professor
envelope icon
Table of Contents

Magnets, computers, and ham radios

MRI, like ultrasound and CT, uses sophisticated computer systems to collect and analyze data, i.e., reflected echos (think fish finder) in ultrasound and X-ray density for CT. MRI measures the mysterious values T1 and T2. T1 and T2 are values that relate to how molecules with a magnetic tendency (dipoles) interact within a magnetic field. A well-known common dipole molecule is water, which fortunately is also the most common molecule in the human body. By manipulating the magnetic environment using static (super-cooled electromagnets) and time varying (radio wave) magnetic fields, T1 and T2 parameters can be detected and used to create an image and characterize tissues.

Any dipole can potentially be imaged, but most have poorer signal strength than water, making usage challenging. Phosphorus is the next likely candidate for medical imaging.

Scans are obtained by placing the patient in a strong, homogenous magnetic field. Within about two seconds, a portion of the water molecules will align with the magnetic field. Then, this steady state is perturbed by beaming in radio waves, causing the aligned water molecules to flip out of alignment with the magnetic field. As they realign and return to the steady state, these molecules give up energy in the form of radio waves, which are detected by antennas. These radio signals are used to make the image, giving information about the tissue composition, namely T1 and T2. T1 and T2 are not to be confused with the strength of the magnetic field, which is termed T for Tesla, a unit of magnetic field strength.

T1 and T2

After the imaging pulse, the flipped water molecules return to the steady state in two ways, simultaneously.

Immediately after the radio wave is absorbed, flipping the molecules out of alignment with the magnet, the susceptible molecules are all aligned (or in-phase) with one another, but no longer aligned with the magnetic field. They rapidly go out of phase with one another, giving off a radio wave. This gives rise to T2 information. T2 relates to the local molecular environment and is the most sensitive parameter for detecting pathology.

As water molecules realign with the main magnetic field, they also generate a radio wave. This gives T1 information. T1 relates to general composition. T1 is useful for tissue characterization and in evaluating for normal anatomy, especially when compared to the T2 image.

Recap
After about two seconds, some of the water molecules are aligned with the machine’s magnetic field, but out of alignment with one another.
A short pulse of specific frequency (the Larmor frequency) radio wave flips the molecules in relation to the machine’s magnetic field but aligns the molecules with one another.
The flipped molecules very rapidly diphase from one another, giving off a detectable radio wave giving T2 information.
The molecules more slowly realign with the machine’s magnetic field, giving off a detectable radio wave giving T1 information.

The generation of T1 and T2 data is manipulated for various purposes by the imager utilizing many complex recipes.

MRI sequences are continually evolving with increasing understanding and advances in machine design. 

Using T1 and T2 information tissue characterization is possible. There are many MRI sequences, but most are various combinations of T1 and T2 information. Images are frequently described as T1 or T2 weighted, meaning containing mainly T1 or T2 information.

A quick way to determine if an image is T1 or T2 weighted is to look at known structures.

Look at water (cerebrospinal fluid, stomach, joint fluid) and fat.

In heavily T1 weighted images, water is dark (low signal), and fat is very bright (gives off a lot of signal). In heavily T2 weighted images, water is bright, and fat is bright but less so than T1.

T1: Water is dark. Fat is very bright.
T2: Water is bright. Fat is bright but less than on T1.

STIR and FATSAT

Imagers commonly superimpose sequences to suppress signal from fat or water. Important sequences are STIR and FATSAT.

STIR imaging
Suppresses signal from free water but not bound water. This is a core sequence in brain imaging so lesions near the ventricles and surface of the brain can be detected.
FATSAT
Suppresses signal from fat. There are myriad sequences, giving an alphabet soup of acronyms. These are detailed in the labeling on each image along with spatial orientation markers. 

Historically, T1 images had much higher signal-to-noise ratio than T2 images, so sections could be very thin, allowing high resolution. T1 imaging also took much less time than T2 imaging. With current advances, T2 images can be obtained rapidly in high-resolution modes. Today, the usual imaging strategy is to use T2 images for lesion detection, and then compare to T1 for characterization.

For each exam, axial, sagittal, and coronal imaging planes are obtained with different T1 and T2 weighting. Each of these is necessarily obtained as separate imaging episodes. Also, each voxel may need to be sampled 2–6 times in each imaging sequence to achieve statistical accuracy. A limited number of voxels can be imaged at a time. Recall that it takes an average of 2 seconds for tissue to return to the steady state after being perturbed by radio waves. It can take a long time to do a complete exam.

If a patient moves during an image acquisition, the entire data set is lost, not just the one image as in CT. There are many clever innovations to keep imaging times down. A typical exam will take 30–45 minutes. 

Most patients can only tolerate 45 minutes in the scanner.

MRI safety

Unlike CT, MRI has no ionizing radiation. No untoward effects have been seen with current MRI imaging parameters. The MRI sequences cause tissue heating. The amount of heating is limited by the FDA and controlled by the imagers. 

The magnetic environment is dangerous around ferromagnetic metals, either implanted or introduced into the MRI room with catastrophic results.

Scanners can alter programs of AICD and pacemaker devices. TENS units are usually problematic. Older heart valves and aneurysm clips may be unsafe. Some tattoos can induce burns, as they may contain metals. The danger increases with increasing magnetic field strength. Original scanners operated at 0.5 Tesla (T); newer scanners operate at up to 3 T. Foreign bodies, such as metallic shrapnel, can shift substantially.

In the MRI imaging suite, the MRI technologist is in charge.

You must receive permission from them to enter the room.

MRI-specific contrast agents may be used; the most common is gadolinium. Allergic-like reactions are unusual and fatal reactions very rare. Patients with renal failure can develop nephrogenic systemic fibrosis. Pregnant women should not be given gadolinium for an MRI exam. Click here for details on contrast from the American College of Radiology.

MRI exams

  • Brain: MRI is 100× more sensitive than CT.
  • Carotid artery: Similar to CT and US but exam failures more common.
  • Circle of Willis: IC aneurysms.
  • Neck: Similar to CT.
  • Spine: Best overall choice.
  • CT has major advantages. Lung is not well imaged with MRI. Cardiac and respiratory motion is problematic.
  • Dedicated cardiac exams are becoming common.
  • Superior for brachial plexus and superior sulcus tumors.
  • Similar to CT. Bowel motion problematic but improved pulse sequences are increasing utility for abdominal imaging.
  • MR Pancreatography and cholangiography.
    • Gold standard for liver metastasis. Use when highest true negative needed (e.g., for liver transplant protocols).
  • Highest detection rate for hepatocellular carcinoma. Used for surveillance in hepatitis C.
  • Emerging bowel applications.
  • Specialized GYN exams.

MR angiography becoming robust and competitive with CT and catheter angiography.

Best choice for:

  • Joint evaluation.
  • Stress fracture.
  • Infection/osteomyelitis.
  • Tumors.
  • Occult hip/foot fractures.
Example accuracy Sensitivity Specificity
Glenoid labrum
80%
90+%
Rotator cuff
95%
90+%
Meniscus
90%
85–93% (varies by portion of meniscus)