DVT Ultrasound: How Compression Scans Detect Clots

Compression ultrasound is the first-line imaging test for diagnosing deep vein thrombosis, and it has held that role for decades. The basic idea is surprisingly simple: a technician or physician presses the ultrasound probe against a vein, and a healthy vein collapses flat under the pressure while a vein containing a clot does not. That single finding, non-compressibility, is the core diagnostic sign. But the technique has evolved well beyond that one maneuver, and its accuracy varies depending on where the clot sits, which scanning protocol is used, and who is holding the probe.

How Compression Ultrasound Detects a Clot

The concept dates back to the late 1980s, when researchers demonstrated that applying gentle transducer pressure to leg veins could reliably distinguish clotted veins from open ones. In an early validation study, ultrasound identified clots in 25 patients whose veins could not be compressed, while correctly clearing 23 patients whose veins compressed normally, yielding a sensitivity of 89% and specificity of 100%.¹ Before this, diagnosing DVT typically required contrast venography, an invasive procedure that involved injecting dye into a foot vein and taking X-rays as it traveled upward. Ultrasound replaced that approach because it is painless, repeatable, and uses no radiation.

Modern DVT ultrasound usually combines the compression maneuver with Doppler assessment, which is why you will sometimes hear it called “duplex ultrasound.” The Doppler component evaluates blood flow within the vein. In a normal leg vein, flow speeds up when you breathe out and slows when you breathe in, and it also pulses slightly with the heartbeat. When a clot partially or fully blocks the vein, those flow patterns change or disappear.² This matters most when a vein is hard to compress for anatomical reasons rather than because of a clot. Abnormal Doppler signals can also hint at a blockage further upstream that the probe cannot directly see.

How Accurate Is It, and Where Does It Struggle

For clots in the large veins of the upper thigh and behind the knee (called proximal DVT), ultrasound performs well. A large meta-analysis combining many studies found that sensitivity for proximal DVT was about 94%, with specificity around 94%.³ Those numbers mean it catches the vast majority of large proximal clots and rarely calls one present when it is not.

The picture is less reassuring for clots confined to the calf veins below the knee, so-called distal DVT. The same meta-analysis found sensitivity dropped to roughly 64% for distal clots.⁴ Calf veins are smaller, deeper, and more numerous, which makes compression harder to perform and interpret. Whether missing some of these smaller clots matters clinically is debated: many isolated calf clots resolve on their own, but a fraction will extend upward into the proximal veins, where they pose a greater risk of traveling to the lungs. This is one reason some protocols call for a repeat scan about a week later if the initial exam is negative but suspicion remains.

Scanning Protocols and What They Cover

Not every DVT ultrasound covers the same territory. Three main approaches are in common use, and understanding the differences helps make sense of what your scan did and did not evaluate.

Two-point compression: The probe checks only two locations, the common femoral vein at the groin and the popliteal vein behind the knee. This is the quickest approach and catches most dangerous proximal clots, but it skips the femoral vein in the mid-thigh and the entire calf.
Three-point compression: Adds the femoral vein in the mid-thigh to the two sites above. This addresses a blind spot, since roughly 6% of new DVTs in one study involved clots isolated to the femoral vein that a two-point scan would have missed.⁵
Whole-leg ultrasound: Scans from the groin all the way down through the calf veins in a single session. This picks up distal clots on the first visit and avoids the need for a repeat scan a week later.

A randomized trial directly comparing the two-point approach (with a repeat scan if initially negative) against whole-leg ultrasound in a single visit found that the rate of subsequent symptomatic clots was similar in both groups, about 1% over three months.⁶ A later meta-analysis echoed this, finding no significant difference in failure rates among the three strategies.⁷ The practical trade-off is time and convenience: whole-leg scanning takes longer but resolves ambiguity in one visit, while limited protocols are faster but may require a return trip.

The Role of D-Dimer Testing Before the Scan

In many hospitals, ultrasound is not the very first step when someone shows up with a swollen leg. Clinicians typically start by estimating how likely DVT is based on symptoms and risk factors, then check a blood marker called D-dimer. D-dimer is a protein fragment produced when blood clots break down. A negative result in a patient whose clinical picture makes DVT unlikely is often enough to rule it out without any imaging at all.⁸ This combination of clinical assessment and blood testing keeps a large number of patients from needing an ultrasound, reducing wait times and costs.

The catch is that D-dimer goes up in many situations besides DVT: after surgery, during pregnancy, with infections, and in cancer patients. A positive D-dimer does not mean you have a clot; it just means ultrasound is the logical next step. And in patients where clinical suspicion is high, most guidelines recommend going straight to ultrasound regardless of D-dimer.

Point-of-Care Ultrasound in the Emergency Department

One of the biggest shifts in DVT imaging over the past decade is the spread of bedside ultrasound performed by emergency physicians rather than dedicated vascular labs. The appeal is speed: waiting for a formal study ordered from radiology can take hours, while a bedside scan takes minutes. A 2024 systematic review found that the time from triage to diagnosis was significantly shorter when emergency physicians performed the scan themselves.⁹

The accuracy numbers are encouraging. Pooled data from that same review showed that the two-point bedside technique had sensitivity around 92% and specificity around 97%, while three-point bedside scanning came in at about 89% sensitivity and 93% specificity.¹⁰ An earlier meta-analysis found very similar performance between the two approaches, with false-negative rates of about 4% for both.¹¹ Training and experience mattered: studies where attending emergency physicians performed the scan tended to show higher accuracy than those where only residents did.¹²

When bedside operators performed complete compression ultrasound of the whole leg, the pooled sensitivity in the 2024 review reached 100% with specificity around 97%.¹³ That said, whole-leg scanning demands more skill and time, so in a busy emergency department the abbreviated protocols remain the norm for an initial look.

Where the Probe Cannot Reach

Ultrasound has genuine anatomical blind spots. The iliac veins deep in the pelvis sit behind bowel gas and bone, making direct compression impossible. This is a particular concern in pregnant patients, who are at elevated risk for DVT and whose clots are disproportionately located in these pelvic veins.¹⁴ Standard compression ultrasound simply cannot visualize clots above the inguinal ligament. Clinicians working up a pregnant patient with high suspicion but a negative leg ultrasound will often look for indirect signs on Doppler, such as absent respiratory variation in the common femoral vein or continuous reflux in the saphenous vein, which can suggest an upstream obstruction.¹⁵ One case report described exactly this scenario: a patient’s leg ultrasound appeared normal, but unilateral saphenous vein reflux prompted further investigation with venography, which confirmed a completely blocked iliac vein.¹⁶ When pelvic clots are suspected, CT or MRI venography typically takes over as the imaging modality.

The upper extremity presents its own challenge. Arm DVTs have become more common with the widespread use of central venous catheters and port devices. Ultrasound works reasonably well for the jugular, axillary, and brachial veins, but the subclavian vein passes beneath the collarbone, where the probe cannot compress it. This is the main reason ultrasound sensitivity for upper-extremity DVT is lower than for the legs, and it is worse on the left side, where the subclavian vein takes a longer course behind the clavicle.¹⁷

Conditions That Mimic DVT on Ultrasound

A swollen, painful leg is not always a blood clot, and one of the underappreciated roles of ultrasound is ruling DVT out while sometimes revealing the actual cause. The differential diagnosis for leg symptoms that prompt a DVT scan is broad. At the groin, enlarged lymph nodes, soft-tissue tumors, and hernias can cause swelling. In the thigh, muscle infections or abscesses may be responsible. Behind the knee, a ruptured popliteal (Baker’s) cyst is a classic DVT mimic, often producing sudden calf swelling and pain. In the lower leg, muscle tears, cellulitis, and fluid retention from heart or kidney problems can all look similar to DVT from the outside.¹⁸ A skilled sonographer will often spot these alternative diagnoses during the same examination.

Monitoring After Treatment and Detecting Recurrence

DVT ultrasound is not a one-time event for many patients. After treatment with blood thinners, a follow-up scan is often performed to establish a new baseline. Why? Because after a first clot, the vein may not return to completely normal. Residual thrombus, scar tissue, or a vein that remains partially non-compressible can persist for months or years. If a patient later develops new symptoms in the same leg, the physician needs to know whether what the ultrasound shows is old or new.

The standard approach compares the current vein diameter under compression with the measurement from the earlier baseline scan. An increase of two millimeters or more, or non-compressibility in a segment that had previously normalized, is considered diagnostic of recurrent DVT.¹⁹ The reliability of this comparison is imperfect, though. Interobserver agreement on residual vein diameter is moderate, with paired measurements sometimes differing by several millimeters.²⁰ To manage this, safety data on withholding anticoagulation when vein diameters are stable or improved have been studied, allowing clinicians to use a combination of ultrasound measurements and clinical judgment rather than relying on the scan alone.²¹

Superficial Clots and the Deep System

Patients sometimes undergo a DVT ultrasound because a superficial vein clot has been found and the question is whether it has extended deeper. This happens more often than you might expect. In one study of patients with superficial vein thrombosis, the most common path into the deep system was extension from the great saphenous vein in the thigh into the common femoral vein, accounting for 70% of cases with deep vein involvement. Many of these extensions were non-occlusive and had a free-floating component, which raises concern for embolization to the lungs.²² Superficial clots in the trunk of the great saphenous vein and those located close to junctions with the deep veins carry the highest risk of concurrent DVT.²³ This is why guidelines increasingly recommend a full duplex ultrasound whenever a superficial clot is found near the groin or behind the knee.

AI-Guided Ultrasound for Non-Specialists

One of the most interesting frontiers is using artificial intelligence to let people with little or no ultrasound training perform a usable DVT scan. The motivation is practical: many hospitals and clinics do not have a vascular lab available around the clock, and trained sonographers are in short supply. Several systems now use AI software that guides the operator through probe positioning in real time, telling them where to place it and when the image is adequate for interpretation.

In one approach, nurses with no prior ultrasound experience used an AI-guided system to scan 381 patients with suspected DVT. Each patient also received a standard scan from an expert sonographer for comparison. The study concluded that diagnosing lower-extremity DVT through AI-guided image acquisition with remote clinician review was feasible.²⁴ A commentary on this work emphasized that such tools could fill important gaps in healthcare delivery, particularly for timely DVT diagnosis in settings where ultrasound technicians are unavailable.²⁵ Another system, designed for a similar workflow, has been investigated for its potential to reduce the volume of formal vascular laboratory testing by triaging straightforward negative cases before they reach the specialist queue.²⁶

These systems are still being validated, and most current implementations involve a remote physician reviewing the acquired images rather than fully automated interpretation. But if the technology proves robust in larger trials, it could reshape DVT diagnostics in rural emergency departments, urgent care clinics, and overnight hospital shifts where a sonographer is simply not on site.