Can You Tell if a Lump Is Cancerous From an Ultrasound?

Ultrasound can strongly suggest whether a lump is cancerous or benign, but it cannot confirm a diagnosis on its own. What it does extraordinarily well is sort lumps into risk categories based on their shape, borders, internal structure, and blood flow, giving doctors a probability estimate that guides the next step. For some lumps, ultrasound is reassuring enough that you can simply watch and wait. For others, the features are suspicious enough that a tissue sample through biopsy becomes necessary. The gap between “suspicious on ultrasound” and “confirmed cancer” is real and matters more than most people realize.

What Radiologists Actually Look For

When a radiologist examines a lump on ultrasound, they are evaluating a handful of specific visual characteristics. Shape is one of the first things assessed. Oval or round lumps with smooth, well-defined edges tend to be benign, like fibroadenomas or fluid-filled cysts. Irregular shapes with jagged or spiculated borders raise suspicion for cancer. But shape alone is not enough to make a call. Round masses with certain types of irregular margins can still turn out to be malignant, including aggressive forms of breast cancer, mucinous carcinoma, and even lymphoma.¹

Margins matter as much as shape. A circumscribed border, meaning you can draw a clean line around the lump, is a reassuring sign. Microlobulated, angular, or indistinct margins are concerning because cancer tends to infiltrate surrounding tissue rather than pushing it aside neatly. Internal echogenicity also comes into play. A simple cyst filled entirely with fluid appears black on the screen and is almost always benign. A solid mass that is darker than the surrounding tissue (hypoechoic) with shadowing behind it needs closer evaluation.

Orientation relative to the skin matters too. Benign masses often grow parallel to the skin surface, appearing wider than they are tall. Cancers more frequently grow against tissue planes, appearing taller than wide. No single feature is a smoking gun. Radiologists weigh all of these characteristics together, and even then, the picture can be ambiguous.

Standardized Scoring Systems

To make ultrasound interpretation more consistent and less dependent on one radiologist’s gut feeling, the medical community has developed structured classification systems for different body parts. These systems convert visual features into a risk score that corresponds to an estimated probability of cancer.

For breast lumps, the most widely used system is BI-RADS, which stands for Breast Imaging Reporting and Data System. It assigns a category from 0 (needs more information) through 6 (known cancer). The categories that generate the most anxiety for patients are 3, 4, and 5. A BI-RADS 3 lesion is considered probably benign, with a malignancy rate below about 2%, and is usually monitored with follow-up imaging rather than biopsied immediately. Category 4 is subdivided into 4a, 4b, and 4c, reflecting increasing suspicion. In a large reader study, the positive predictive values for malignancy rose from roughly 28% for category 4a to about 50% for 4b, 68% for 4c, and 80% for category 5.² That means even a category 5 lesion, which looks highly suspicious, turns out to be benign about one time in five.

Category 4a is a source of particular frustration. The malignancy rate sits between roughly 2% and 10%, which means most 4a lesions are benign, but the standard recommendation is still to biopsy. A multicenter study found that combining clinical factors with specific ultrasound features like margin irregularity, calcification, and tissue stiffness ratios could identify a low-risk subset within 4a, where the malignancy rate dropped to about 1%, potentially sparing many patients an unnecessary biopsy.³

For thyroid nodules, a parallel system called TIRADS assigns points based on features like composition, echogenicity, shape, margin, and the presence of calcifications. Using a TIRADS score of 4 or higher as a cutoff, one study found that ultrasound could detect malignant thyroid nodules with about 92% sensitivity and 53% specificity.⁴ The high sensitivity means few cancers are missed, but the modest specificity means a fair number of benign nodules still get flagged for biopsy. TIRADS recommendations also take nodule size into account, with smaller suspicious nodules sometimes watched rather than immediately biopsied.⁵

For ovarian and adnexal masses, the O-RADS system stratifies risk using ultrasound features and mathematical models. The most recent version of O-RADS achieved an area under the curve of 0.94 for distinguishing benign from malignant masses, and combining it with a blood marker called HE4 pushed accuracy even higher, to 0.98.⁶ An earlier study comparing two approaches to O-RADS classification found sensitivities above 85% and specificities in the mid-to-high 80s.⁷

Blood Flow, Stiffness, and Contrast Agents

Standard grayscale ultrasound is only one piece of the toolkit. Several add-on techniques give radiologists additional clues about whether a lump is cancerous.

Color and power Doppler ultrasound map blood flow within and around a mass. Cancerous tumors typically recruit new blood vessels to feed their growth, so detecting chaotic or abundant internal blood flow raises suspicion. In one early study of breast cancers, Doppler signals were detected in 96% of malignant tumors, and the flow characteristics carried prognostic value beyond just identifying the tumor as cancer.⁸ The pattern of blood flow matters too. Vessels that radiate from the edges of a mass inward (peripheral vascularity) are more concerning than vessels that follow the normal architecture of the surrounding tissue.⁹

Elastography measures how stiff a lump is by applying gentle pressure or sound waves and tracking how the tissue deforms. Cancers tend to be stiffer than benign tissue because of the dense, fibrous stroma they produce. In breast imaging, hard lesions on elastography have been linked to larger tumor size, lymph node spread, and higher tumor stage.¹⁰ A related technique called acoustic radiation force impulse (ARFI) imaging measures shear wave velocity inside lymph nodes. In one study, a velocity cutoff above 1.9 meters per second identified metastatic lymph nodes with 95% specificity and 81.8% sensitivity.¹¹

Contrast-enhanced ultrasound uses injected microbubbles to highlight blood flow at a very fine level. It is considered the most sensitive real-time method for measuring tissue perfusion noninvasively and can help differentiate benign from malignant tumors in several organs.¹² In fatty tumors, for instance, where the grayscale appearance of a benign lipoma and a low-grade liposarcoma can look frustratingly similar, contrast-enhanced ultrasound revealed dramatically different perfusion levels. A perfusion-based approach correctly identified malignant sarcomas versus benign fatty tumors with a positive predictive value of 93%.¹³

Lymph Nodes and Soft Tissue Lumps

Ultrasound assessment extends well beyond breast, thyroid, and ovarian masses. Swollen lymph nodes in the neck are one of the most common reasons people get an ultrasound, and the same feature-based logic applies. Normal lymph nodes tend to be oval or kidney-bean shaped with a bright fatty center (the hilum). Nodes that become round, lose their hilum, show internal necrosis or calcification, or display chaotic peripheral blood flow patterns are more likely to harbor metastatic cancer or lymphoma.¹⁴ Other red flags include matting, where nodes stick together, and soft-tissue edema around the node.⁹

For soft tissue lumps elsewhere in the body, like those in the arms, legs, or trunk, ultrasound is usually the first imaging step. Tumor size, margin clarity, and internal blood flow are the three strongest predictors of malignancy. One study developed a scoring system from these three parameters that achieved high accuracy in separating benign from malignant soft tissue tumors.¹⁵ Another study comparing lipomas to well-differentiated liposarcomas found that malignant fatty tumors were more often deep-seated, irregularly shaped, larger, brighter on ultrasound than surrounding tissue, and showed Doppler blood flow signals.¹⁶

Why Biopsy Still Matters

Even the best ultrasound assessment leaves a gap that only a tissue sample can close. Ultrasound tells you what a lump looks like. Biopsy tells you what its cells actually are. When a radiologist inserts a needle into a suspicious lump under ultrasound guidance, the ultrasound is doing double duty: it flagged the concern and it is now steering the needle to the right spot in real time.

Ultrasound-guided core needle biopsy has become the standard first step for tissue diagnosis of many lumps. In a large analysis of soft tissue tumors, core needle biopsy produced a conclusive histological diagnosis in about 89% of cases. When the biopsy result could be compared to the final surgical pathology, the two agreed on whether the tumor was benign or malignant 97% of the time and on the specific tumor type 93% of the time.¹⁷ Similar accuracy has been reported in gynecologic oncology, where ultrasound-guided core biopsy established a diagnosis in about 90% of cases and agreed with surgical pathology about 96% of the time.¹⁸

The point worth emphasizing is that the ultrasound finding alone, no matter how suspicious, is not the diagnosis. It is a probability estimate that determines the urgency and type of next step. A BI-RADS 5 breast mass warrants an immediate biopsy. A BI-RADS 3 mass warrants a follow-up ultrasound in six months. Both decisions hinge on the ultrasound, but neither replaces the microscope.

How Ultrasound Compares to MRI and Mammography

For breast lumps specifically, patients sometimes wonder whether MRI would be more definitive than ultrasound. Each modality has strengths in different situations. In one study comparing all three, mammography had 56% sensitivity but 88% specificity, ultrasound had 75% sensitivity but only 19% specificity, and MRI had 100% sensitivity and 50% specificity.¹⁹ In plain terms, MRI in that study caught every cancer but flagged a lot of benign lumps as suspicious. Ultrasound caught most cancers but also flagged many benign lumps. Mammography missed more cancers but made fewer false alarms.

The tradeoffs shift in dense breast tissue. A systematic review found that MRI’s precision was particularly strong in women with dense breasts, though ultrasound performed well at differentiating cancers with about 84% specificity in some studies within that review.²⁰ In practice, ultrasound and mammography are often used together as complementary tools, with MRI reserved for high-risk patients or situations where the other modalities give conflicting results. Ultrasound has practical advantages: no radiation, lower cost, wider availability, and the ability to perform a biopsy in the same appointment if something looks concerning.

Artificial Intelligence on the Horizon

Machine learning models are increasingly being trained to read ultrasound images and flag suspicious features. One deep learning model achieved about 94% accuracy in classifying breast ultrasound images as benign or malignant.²¹ These systems are designed to be more consistent than human readers, who can vary in how they apply the BI-RADS criteria and other scoring systems. A review of AI applications in breast ultrasound described how both traditional machine learning and newer deep learning approaches can extract and quantify features from images automatically, potentially reducing reader variability and helping less-experienced practitioners.²²

A particularly interesting application is in settings with limited access to specialist radiologists. In a pilot study in Mexico, minimally trained operators acquired ultrasound images of palpable breast lumps and fed them to a computer-aided diagnosis system. The software classified all cancers as suspicious and all benign lesions as benign, matching the accuracy of specialist radiologists reviewing the same images.²³ That study was tiny, with only 32 patients, so the numbers should be read cautiously. But it illustrates the direction the field is heading: ultrasound as a triage tool that does not necessarily require a radiologist on-site.

The Follow-Up Problem

When ultrasound classifies a lump as probably benign (BI-RADS 3, for example), the standard recommendation is to come back for follow-up imaging, usually at six months and then annually for a couple of years. The idea is sound: if the lump does not change over time, you can feel confident it is benign without needing a biopsy. But compliance is a real problem. In one large cohort, only about 17% of patients with a BI-RADS 3 lesion completed all recommended follow-up visits. About 31% never returned for a single follow-up.²⁴

The reasons are familiar: anxiety, cost, inconvenience, and the assumption that a “probably benign” result means everything is fine. But the surveillance protocol exists precisely because a small percentage of BI-RADS 3 lesions do turn out to be cancer. Dropping out of follow-up defeats the purpose of the cautious approach the radiologist intended. The anxiety associated with being told to “watch and wait” is a well-documented side effect of this strategy, and some researchers have noted it as a barrier to compliance.²⁵ If you are in this category and find the waiting difficult, talk to your doctor about whether a one-time biopsy might be a reasonable alternative for your specific situation.

When the Ultrasound Appearance Fools Everyone

Even experienced radiologists get surprised. Some cancers are masters of disguise. Round breast cancers with smooth borders can look identical to fibroadenomas on ultrasound. At least seven distinct types of cancer can present as round, well-circumscribed masses, including aggressive high-grade ductal carcinomas and colloid carcinomas.¹ Conversely, some benign conditions like sclerosing adenosis and radial scars look alarmingly irregular on ultrasound and are frequently mistaken for cancer.

Thyroid nodules present a similar challenge. A highly suspicious TIRADS 5 nodule might still be benign, and a low-suspicion nodule can occasionally harbor a small cancer, particularly papillary microcarcinomas that grow slowly enough that many would never cause harm during a patient’s lifetime. The imperfect specificity of ultrasound across every organ system is the reason tissue sampling exists as the gold standard. Ultrasound is a powerful filter, not a final verdict.

Non-mass lesions, meaning areas of abnormality that do not form a discrete lump, present their own difficulties. In one study of breast non-mass lesions classified using BI-RADS, the sensitivity was about 83% but specificity was only about 42% when using category 4B as the cutoff for suspicion.²⁶ These are the cases that tend to generate the most diagnostic uncertainty, because the classic rules about shape and margins simply do not apply to something that does not have a clear shape or margin.

The bottom line for anyone staring at an ultrasound report is this: the imaging can tell you a great deal about what a lump probably is, and the probability estimates from structured scoring systems are remarkably well calibrated. But “probably” and “definitely” are different words for a reason, and the biopsy exists to bridge that gap.