Beyond the Scan: Why Evidence and Human Judgment Matter in Breast Imaging

Imaging Does Not Provide Certainty

One of the most important principles in breast imaging is that imaging is not equivalent to histological diagnosis.

A mammogram can demonstrate suspicious calcifications. Ultrasound can show an irregular hypoechoic mass. MRI can reveal an enhancing lesion with concerning kinetics and morphology. These findings may strongly suggest malignancy, but they do not independently establish what the tissue is. Conversely, an imaging appearance that seems reassuring does not create absolute certainty that malignancy is absent.

Radiologists therefore have to manage probabilities. This is reflected in systems such as BI-RADS, which translate imaging appearances into levels of suspicion and recommended management. The purpose is not simply to attach a number to a finding. It is to connect observation with an evidence-based next step.

In that sense, breast imaging is not merely the recognition of abnormalities. It is risk estimation followed by clinical decision-making. The same image can mean different things in different patients.

A small circumscribed mass in a 25-year-old woman would be approached very differently if it were present in a 70-year-old woman with a new palpable abnormality.

Likewise, an area of subtle architectural distortion may carry different implications depending on whether it is new, whether it has a sonographic correlate, whether the patient has had previous surgery, and whether prior examinations demonstrate stability.

Clinical history matters, as does age, symptoms, breast density, previous imaging, family history, personal cancer history and known genetic or pathological risk factors.

The scan provides visual information, and the radiologist must decide how much that information means in that particular patient. This is why good breast imaging cannot be reduced to image pattern recognition alone. Evidence guides us — but does not remove judgment.

Modern breast imaging is strongly evidence-based. Screening programmes, recall thresholds, surveillance intervals, biopsy recommendations and the use of supplemental imaging are all informed by clinical trials, observational studies, consensus statements and professional guidelines.

Evidence is indispensable because it prevents practice from becoming dependent solely on individual habit or intuition. But evidence rarely dictates every clinical decision with mathematical precision. Studies describe populations, but each patient is an individual with their own unique features.

Guidelines may recommend an approach for an average-risk woman, yet the patient in front of us may have unusually dense breasts, a strong family history, an unexplained palpable abnormality or discordant imaging findings. Evidence defines the boundaries of reasonable practice, but in the end it is clinical judgment that determines how that evidence should be applied.

Comparison with Previous Imaging Is a Form of Evidence

One of the most valuable tools available to a breast radiologist is often not a new imaging modality, but an old examination. Stability over time can fundamentally change the interpretation of a finding.

A focal asymmetry that appears concerning on a single examination may prove unchanged over many years. Conversely, a subtle area that might initially appear innocuous can become important when comparison demonstrates that it is genuinely new or evolving.

This is particularly relevant in breast imaging because many findings are subtle and because normal breast tissue varies considerably between individuals. Prior imaging provides a patient-specific baseline and helps the patient to avoid unnecessary biopsies.

For this reason, retrieving and comparing previous mammograms, ultrasounds and MRIs should not be regarded as an administrative detail. It is a crucial part of the diagnostic process itself.

Concordance Matters More Than Any Single Test

An important concept in breast imaging is concordance. In essence, it is about ensuring that all the available information — clinical findings, imaging findings and pathological findings — makes sense.

Consider a woman with a clearly palpable hard breast mass.

If mammography is negative but the palpable abnormality persists, the negative mammogram does not automatically end the investigation. Targeted ultrasound, clinical reassessment or tissue sampling will still be required.

Similarly, if imaging suggests a highly suspicious lesion but a biopsy returns a benign result that does not adequately explain the imaging appearance, the radiologist will call this discordant and suggest further tests.

A benign biopsy is reassuring only when the tissue diagnosis plausibly accounts for what was seen on imaging. This is known as radiological-pathological correlation, and is one of the most important safeguards in breast imaging.

The main principles are that no individual result should be interpreted without asking whether it fits the rest of the clinical picture, and that more imaging is not always better imaging.

There is an understandable tendency to assume that additional imaging must always improve diagnostic certainty. However, every investigation has strengths, limitations and potential consequences. Supplemental ultrasound may detect additional cancers in women with dense breasts, but it may also identify benign abnormalities that lead to additional follow-up or biopsy. MRI is highly sensitive, but that sensitivity can generate findings of uncertain significance that require further assessment.

Contrast-enhanced mammography, tomosynthesis, abbreviated MRI and emerging technologies all offer important advantages in specific situations, but none eliminates the need to select investigations appropriately. As radiologists, we have the duty to decide whether an additional test will meaningfully change diagnosis or management of the patient.

Technology Is Changing the Radiologist's Role

Breast imaging has always been technologically driven with multiple imaging modalities available and being developed. Digital mammography, tomosynthesis, advanced ultrasound, contrast-enhanced mammography and MRI have progressively increased the amount and complexity of information available to radiologists.

With the advent of Artificial Intelligence, the possibilities are endless. AI systems can assist with cancer detection, lesion prioritisation, breast density assessment, risk prediction and workflow optimisation. However, AI also reinforces an important distinction between detection and decision-making.

AI may identify a suspicious region, but it has no clinical awareness. It does not know whether the patient can feel a lump there, whether the finding was present five years ago, whether the biopsy result is discordant, whether the enhancement pattern has a benign explanation, or whether an additional investigation is likely to benefit the patient.

These tasks require integration of evidence, which is where the radiologist comes in. The radiologist's role is therefore unlikely to become less important as imaging technology advances. It may become more interpretative. The challenge will increasingly be not simply to identify abnormalities, but to determine which abnormalities matter.

Expertise Includes Knowing When Not to Intervene

Clinical expertise is related not only to detecting disease but also to the ability to recognise when an abnormality should be left alone.

Breast imaging contains a large spectrum of benign findings: cysts, fibroadenomas, postsurgical changes, vascular calcifications, fat necrosis and numerous normal anatomical variants, just to name a few. Over-investigation can be harmful in many ways by creating anxiety due to unnecessary recall, exposing patients to discomfort, cost and procedural risk after an unnecessary biopsy and prolonged uncertainty from short-term repeat imaging. Therefore, much of the breast radiologist's work is about balancing the risk of missing clinically significant disease against over-investigation.

Communication Is Part of Diagnostic Quality

Completion of the radiology report does not mean the breast radiologist's work is done; it is only the beginning of the patient’s journey, and the radiologist needs to remain involved. The report must communicate what has been found, how concerning it is and what should happen next. The language of the report has to be clear, as ambiguity can create unnecessary uncertainty. A technically accurate report that fails to provide clear management guidance may still be clinically inadequate.

In breast imaging, the management recommendation is often as important as the descriptive finding itself. This is why structured reporting systems such as BI-RADS were created to allow a common language between radiologists, surgeons, oncologists, primary-care physicians and patients. This highlights how important communication is in radiology, and often just the report alone is insufficient. Suspicious or unexpected findings may require direct discussion with the referring clinician or patient. Multidisciplinary review may be necessary when imaging, pathology and clinical findings disagree.

Judgment Improves Through Experience — But Experience Must Remain Accountable to Evidence

After seeing thousands of examinations, certain appearances become immediately familiar and intuition develops naturally. Subtle asymmetries, distortion or patterns of enhancement may trigger concern before the observer can fully articulate why. This is part of expertise.

But intuition is not immune to scrutiny and in many instances experienced radiologists can still be wrong. Familiarity may produce cognitive shortcuts and previous diagnoses can anchor interpretation. A reassuring history may lead to premature closure, while an alarming clinical history can increase the tendency to overcall equivocal findings. This is why it is important to have a strong evidence-based framework to provide an important counterbalance.

The strongest and safest clinical practice combines experience with discipline in a structured manner to recognise the patterns, examine the evidence, consider the alternatives, test for concordance and then decide on the next step.

Breast Imaging Is Ultimately a Decision Science

So much focus is placed on the sophisticated technologies of modern breast imaging. However, the radiologist's most important contribution is not the image but the collective decision that results from it. The scanner merely supplies information. Evidence tells us what that information has meant in populations and what will benefit the patient. Experience helps us recognise patterns but human judgment brings these elements together for the individual patient.

That is why excellent breast imaging extends beyond finding abnormalities on a scan. It requires knowing what they mean, how certain we can be, what evidence supports the next step and when the information in front of us does not yet tell the whole story.

That is the work beyond the scan.