Mammogram vs. Ultrasound vs. Breast MRI

It is tempting to sort these three from worst to best. They do not sort that way. Each answers a different question, and the right test depends on which question is being asked: are we looking for cancer in someone with no symptoms, are we characterising something we can already feel or see, or are we mapping a cancer we already know about?

A mammogram is a low-dose X-ray of compressed breast tissue. It is the only breast imaging test shown in randomised trials to reduce deaths from breast cancer, which is why every major guideline builds screening around it.

• Screening mammogram — no symptoms, standard views, sometimes read later the same day.
• Diagnostic mammogram — done for a symptom or a screening callback. More views from more angles, a higher radiation dose, and usually read while you wait.
• 3-D mammography (digital breast tomosynthesis) — takes images in slices. It finds more tumours than 2-D alone and tends to produce fewer callbacks, though whether it reduces deaths more than 2-D is still unknown.

Its main limitation: dense tissue and tumour both look white, so mammography misses more cancers in dense breasts.

Ultrasound uses sound waves, no radiation, and works in real time. It is what a radiologist reaches for to answer a specific, local question:

• Is this lump a simple cyst (fluid) or a solid mass?
• What is actually going on at the exact spot the mammogram flagged?
• What is this lump in someone who is young, pregnant or breastfeeding, where mammography is less informative?
• Where should this biopsy needle go?

Ultrasound is excellent at those jobs. As a stand-alone screening test it does not replace mammography: it finds a small number of additional cancers in dense breasts and generates a large number of false positives.

MRI uses a magnetic field and an injected contrast agent (gadolinium) to show how tissue takes up blood supply. It is the most sensitive of the three, and it is used where sensitivity is the point:

• Annual screening alongside mammography for people at high risk — a known BRCA1 or BRCA2 variant, certain genetic syndromes, chest radiation before age 30, or an estimated lifetime risk of roughly 20% to 25% or more. ACS recommends starting around age 30 in that group.
• Mapping a known cancer before surgery, or assessing response to treatment given before surgery.
• Resolving a question that mammography and ultrasound have left open.

Its costs: more false positives and more benign biopsies, a contrast injection, a longer scan lying face down, higher cost and patchier access. It is not suitable for everyone — some implanted devices, some kidney problems and severe claustrophobia all get in the way. Abbreviated or "fast" MRI protocols are being studied to reduce time and cost.

In practice these are layered, not swapped. A typical callback path runs screening mammogram → diagnostic mammogram → targeted ultrasound → biopsy if needed. MRI enters when risk is high, when a diagnosis is already made, or when the first two tests leave a question open.

Two things worth holding onto. A negative MRI does not cancel out a suspicious mammogram finding. And a lump you can feel gets evaluated regardless of what the imaging says — a normal mammogram does not close that question.

Before accepting or declining any of these, it helps to know what question the test is being asked to answer and what happens with each possible answer. "What would a normal result here change, and what would an abnormal one change?" is a fair question, and the answer usually makes clear whether the test is worth doing.

• NCI — Breast Cancer Screening (PDQ®) Patient Version
• American Cancer Society — Recommendations for the Early Detection of Breast Cancer
• USPSTF — Breast Cancer: Screening (2024)
• NCI — Dense Breasts: Answers to Commonly Asked Questions