What Does an Ultrasound Show and What It Misses

Ultrasound excels at imaging soft tissues, fluid-filled structures, and moving anatomy in real time, all without radiation. It can show a beating fetal heart, a gallstone, a blood clot, or a torn tendon with impressive clarity. Where it struggles is equally consistent: bone, air, deep tissue in larger patients, and anything that demands the kind of fine-grained structural detail that CT or MRI provides. The gap between what ultrasound reveals and what it leaves hidden is shaped by real physics, and understanding that gap matters whether you are headed for a prenatal scan, an emergency room visit, or a check on a suspicious lump.

What Ultrasound Shows Best

Ultrasound’s greatest strength is distinguishing fluid from solid tissue. A simple cyst filled with clear fluid looks completely different from a solid mass on a sonogram, and the technology is remarkably good at making that call. Research on breast imaging has achieved near-perfect separation between fluid-filled and solid masses using coherence-based metrics, with one study reporting ideal sensitivity and specificity of 1.00 for that distinction.¹ That ability to tell “liquid bag” from “solid lump” in a fraction of a second, without cutting anyone open, is the reason ultrasound is often the first imaging test ordered for thyroid nodules, ovarian cysts, liver masses, and breast lumps.

Once a cyst is identified, the picture gets more nuanced. Not all cystic lesions are harmless. Studies classifying breast cysts by internal features found that simple and mildly complex cysts were uniformly benign, but cystic masses with solid components had malignancy rates climbing above 60 percent in some categories.² This is exactly the kind of sorting that ultrasound does well: it cannot always tell you what something is, but it can frequently tell you whether to worry about it.

Ultrasound also captures motion. Because it produces images in real time, it is the go-to tool for watching things that move: a fetal heartbeat, a valve opening and closing, blood flowing through an artery. Echocardiography, the cardiac version, provides detailed information about heart function and blood flow patterns and is the most commonly used cardiovascular imaging test after an electrocardiogram and a chest X-ray.³ For a patient arriving at an emergency department with chest pain, echocardiography can simultaneously check for wall motion abnormalities suggestive of a heart attack, pericardial effusion, and aortic dissection, all in one exam at the bedside.

Blood Flow and Doppler

Standard ultrasound shows anatomy. Doppler ultrasound adds a layer by detecting the speed and direction of blood flow. The technique works by measuring how sound waves shift in frequency when they bounce off moving red blood cells. Clinicians routinely use Doppler to evaluate blood flow in both large arteries and tiny vessels throughout the body, and the resulting waveforms can distinguish normal flow from narrowed, blocked, or reversed patterns.⁴

This is how doctors check for deep vein thrombosis, assess whether a carotid artery is dangerously narrowed, or confirm blood flow to a transplanted organ. Newer flexible Doppler devices can even continuously monitor blood flow velocity in deeply embedded arteries, a capability being explored for long-term wearable monitoring.⁵ At the research frontier, techniques are emerging that can map microvascular flow without injecting contrast agents, reaching vessel networks too small for conventional Doppler to resolve.⁶

Muscles, Tendons, and Joints

Musculoskeletal ultrasound has become a workhorse for evaluating soft-tissue injuries around joints. It provides detailed views of tendons, ligaments, and nerves, which makes it especially useful for diagnosing rotator cuff tears, tennis elbow, carpal tunnel syndrome, and similar problems.⁷ A pilot study comparing musculoskeletal ultrasound with MRI for hand tendon evaluation found ultrasound to be very accurate for that specific task.⁸

The practical advantages are significant: the exam is done in real time, so the clinician can ask you to flex or extend a joint and watch the tendon move under stress. No other imaging modality offers that dynamic view. It is also fast, portable, and far less expensive than an MRI. The catch, as with much of ultrasound, is that the quality of the result depends heavily on who is holding the probe.

The Depth-Versus-Detail Trade-Off

Every ultrasound exam involves an inherent compromise between how deep the sound waves can reach and how sharp the resulting image will be. Higher-frequency probes produce finer detail but cannot penetrate as deeply, while lower-frequency probes reach deeper structures but sacrifice resolution.⁹ This is not a flaw in the equipment; it is a physical law that cannot be engineered away.

In practice, this means a high-frequency probe aimed at a thyroid gland or a finger tendon just beneath the skin produces remarkably detailed images. But those same frequencies cannot reach the liver or kidneys of a larger patient. For deep abdominal organs, the sonographer switches to a lower-frequency probe that penetrates further but produces a grainier picture. Some structures, particularly those deep in the pelvis or behind ribs, sit in a zone where neither option is ideal.

What Ultrasound Commonly Misses

Sound waves travel poorly through two things: bone and air. This single physical fact accounts for most of ultrasound’s blind spots. The skull blocks ultrasound from the brain in adults (though a small acoustic window above the ear allows limited transcranial Doppler studies). Ribs partially obstruct views of the lungs and parts of the heart. Air-filled loops of bowel scatter sound waves and create a noisy mess, which is why ultrasound is often unhelpful for detailed evaluation of the stomach, intestines, and much of the colon. The pancreas, tucked behind the stomach, is famously difficult to see on ultrasound and its diagnostic performance varies greatly by operator experience and patient habitus.

Body size has a major impact. As BMI increases, the quality of liver and kidney imaging deteriorates regardless of which ultrasound probe is used.¹⁰ Subcutaneous fat absorbs and scatters sound energy before it reaches the target organ, and the result is often a hazy, inconclusive image. For these patients, CT or MRI may be necessary to get the same information that ultrasound provides easily in a thinner person.

Ultrasound also struggles with very small structures. A sub-centimeter lymph node deep in the abdomen, a tiny early-stage tumor, or a hairline fracture in a bone cortex are all likely to be missed. The technology is best at catching things that are moderate-sized, superficial, and sitting in soft tissue, and least reliable for things that are small, deep, and surrounded by bone or gas.

The Operator Problem

Unlike CT or MRI, where the machine produces a standardized dataset that any radiologist can review, ultrasound is uniquely dependent on the person scanning. The operator chooses the probe, sets the angle, applies the right pressure, and decides in the moment which structures to examine closely. Ultrasound is widely recognized as highly operator-dependent and difficult to perform, which limits its broader use.¹¹

This has been quantified. A study tracking how surgeon accuracy improved with practice found that the ability to detect full-thickness rotator cuff tears kept getting better up to about 100 ultrasound examinations, at which point accuracy plateaued.¹² That is a steep learning curve. In an emergency setting, where ultrasound is often performed by physicians with varying levels of training rather than dedicated sonographers, accuracy depends heavily on the operator’s technical skill and the specific conditions of each patient.¹³

What this means for you: the same exam on the same day might produce a confident diagnosis from one sonographer and an inconclusive result from another. If your ultrasound report comes back unclear, it is not always because the pathology is ambiguous. Sometimes the exam just needs to be repeated by someone with more experience or a different approach.

Prenatal Ultrasound and Its Limits

Pregnancy is where most people first encounter ultrasound, and it is worth understanding what those scans can and cannot find. A large Cochrane systematic review established the expected performance: a first-trimester scan correctly identifies about 91 percent of lethal fetal anomalies but only about 37.5 percent of all structural anomalies. Adding a second-trimester scan raises overall detection to roughly 84 percent, with a very low false-positive rate of about 0.1 percent.¹⁴

Those numbers highlight a meaningful gap. Even with two rounds of scanning, roughly one in six structural anomalies goes undetected before 24 weeks. Early scans are good at catching severe brain defects and major heart problems but can miss subtler issues like minor kidney abnormalities, some bowel defects, and certain limb anomalies that only become visible later in pregnancy. One study found that while operators could visualize 90 percent of mandatory anatomical structures in the first trimester using transvaginal scanning, nuchal translucency, a key marker for chromosomal abnormalities, was successfully measured in only about 58 percent of cases.¹⁵

The technology you use matters somewhat as well. A comparative study of two-dimensional versus three-dimensional ultrasound found that standard 2D imaging successfully visualized all 17 assessed fetal structures in about 65 percent of cases, compared to 35 percent for 3D, with statistically significant differences for several structures including the cardiac axis, posterior fossa, and orbits.¹⁶ Three-dimensional imaging is often marketed as superior, but at least in first-trimester screening, the conventional approach performed better for basic anatomical assessment.

How Ultrasound Stacks Up Against CT, MRI, and PET

Each imaging modality has a niche, and understanding where ultrasound sits relative to the alternatives helps explain when doctors choose it and when they escalate to something else. A study comparing all four methods for detecting cervical lymph node metastases in oral cancer found that ultrasound, CT, and MRI all achieved accuracy above 80 percent on both a regional and per-node basis. PET/CT had higher sensitivity but lower specificity, meaning it caught more positive nodes but also flagged more false alarms. Ultrasound and CT showed high positive predictive values, likely because their superior spatial resolution allows more accurate detection of structural changes inside the node.¹⁷

In general terms, CT is better for bone detail, lung imaging, and emergency whole-body assessment. MRI is better for soft-tissue contrast in the brain, spinal cord, and joints. PET is better for detecting cancer that has spread to unexpected sites. Ultrasound’s advantages are portability, speed, cost, real-time imaging, and no ionizing radiation. For many questions, ultrasound is the right first test even when it cannot give the final answer, because it narrows the possibilities cheaply and safely before a more expensive or invasive test is needed.

Artifacts That Can Fool You

Ultrasound images are constructed from reflected sound waves, and the process is not perfect. Artifacts, which are features that appear on the image but do not correspond to real anatomy, are common and can mislead both the operator and the patient. Some artifacts are actually useful: acoustic shadowing behind a gallstone confirms the stone is dense, and posterior enhancement behind a cyst confirms it contains fluid. But the same physics can display echoes incorrectly within shadow regions or shift the apparent position of real structures.¹⁸

Other common artifacts include reverberation (repeated echoes that create false horizontal lines), mirror artifacts (where a structure appears duplicated on the other side of a strong reflector like the diaphragm), and side-lobe artifacts that place echoes in the wrong location. These are not rare edge cases; they appear in routine exams. An experienced sonographer recognizes them instantly and adjusts. An inexperienced operator or an automated system may not. This is another reason why ultrasound reports sometimes hedge with language like “limited study” or “technically difficult exam,” phrases that essentially mean the images were not clean enough for a confident answer.

Safety Profile

Diagnostic ultrasound has been officially declared safe, and no harmful biological effects in humans have been demonstrated with current imaging modes.¹⁹ Sound waves interact with tissue through two mechanisms: thermal (heating) and mechanical (pressure-related). Safety indices called the thermal index and mechanical index are displayed on the machine screen during every scan, giving the operator a real-time estimate of the potential for biological effects. In practice, a standard diagnostic exam keeps both well within safe limits.

That said, newer techniques like shear wave elastography and contrast-enhanced ultrasound operate with acoustic output parameters that may approach or exceed traditional diagnostic limits, which is why guidelines recommend physicians remain informed about potential risks and follow the ALARA principle: keep exposure as low as reasonably achievable.²⁰ This mostly matters for prolonged research scans or specialized therapeutic applications, not for the quick diagnostic exam most patients experience.

Preparation Tips That Actually Matter

You have probably been told to drink water and hold your bladder before a pelvic ultrasound. That instruction dates from a time when transabdominal scanning was the only option, and a full bladder was needed to push bowel gas out of the way and create an acoustic window to the uterus and ovaries. With transvaginal probes now standard, the evidence shows that transvaginal scanning with an empty-bladder transabdominal supplement can replace the full-bladder technique for routine pelvic exams.²¹ Only about 1.5 percent of patients in one study needed a full bladder to visualize a normal ovary that the other approaches missed. If your clinic still routinely requires a full bladder, it may be following older protocols.

For abdominal ultrasound, fasting for 8 to 12 hours before the exam is commonly requested. The reason is practical: eating causes the gallbladder to contract and empties it of the bile that makes it easy to see, and food in the stomach generates gas that degrades the image. For thyroid, breast, or musculoskeletal exams, no preparation is typically needed.

Where the Technology Is Headed

Several developments are expanding what ultrasound can detect. Contrast-enhanced ultrasound uses microbubbles injected into the bloodstream to light up blood vessels in ways standard ultrasound cannot. Studies have shown it is diagnostically superior to conventional ultrasound for identifying vessel irregularities and measuring new blood-vessel growth, which helps assess things like plaque vulnerability in carotid arteries.²²

Super-resolution ultrasound imaging, still largely experimental, pushes the resolution of ultrasound down to the micron level by tracking individual microbubbles as they flow through tiny vessels. Animal studies have demonstrated imaging depths greater than 35 millimeters at resolutions far beyond what conventional ultrasound achieves.²³ If this translates to routine clinical use, it could open up ultrasound assessment of microvascular disease in organs that currently require more invasive evaluation.

Artificial intelligence is addressing the operator-dependence problem from two angles. On the detection side, AI models applied to point-of-care ultrasound for abdominal hemorrhage have shown sensitivity ranging from 88 to 98 percent and specificity from 68 to 99 percent in systematic review.²⁴ On the acquisition side, AI guidance systems are being tested that allow people without ultrasound training to perform compression scans for deep vein thrombosis, with about 87 percent of AI-guided scans achieving diagnostic image quality and a median scan-plus-review time under 8 minutes.²⁵ If that approach scales, it could make basic ultrasound screening available in clinics and rural settings that currently have no access to trained sonographers, chipping away at the biggest single limitation of the technology: the need for a skilled human behind the probe.