Deep venous thrombosis (DVT) is the presence of coagulated blood, a thrombus, in one of the deep venous conduits that return blood to the heart. The clinical conundrum is that symptoms (pain and swelling) are often nonspecific or absent. However, if left untreated, the thrombus may become fragmented or dislodged and migrate to obstruct the arterial supply to the lung, causing a potentially life-threatening pulmonary embolus (PE).
In terms of incidence, lower-extremity DVT is the most common venous thrombosis, with a prevalence of 1 case per 1000 population. In addition, it is the underlying source of 90% of acute PEs, which cause 25,000 deaths per year in the United States (National Center for Health Statistics [NCHS], 2006). Other than the immediate threat of PE, the risk of long-term major disability from postthrombotic syndrome (PTS) is high. [1, 2, 3, 4, 5]
In 1856, Virchow described the classic triad of predisposing factors for DVT, namely, venous stasis, injury of the vascular wall, and a hypercoagulable state.  Events or conditions that alter the equilibrium of one or more of these factors may produce DVT.
Anatomic variants that result in diminution or absence of the IVC or iliac veins may contribute to venous stasis. In iliocaval thromboses, an underlying anatomic contributor is identified in 60-80% of patients. The best-known anomaly is compression of left common iliac vein at the anatomic crossing of the right common iliac artery. The vein normally passes under the right common iliac artery during its normal course (see the image below).
Conclusive diagnosis historically required invasive and expensive venography, which is still considered the criterion standard. Since 1990, the diagnosis has been obtained noninvasively by means of sonographic exam. The recent validation of the simpler and cheaper D-dimer test as an initial screening test permits a rapid, widely applicable screening that may reduce the rate of missed diagnoses. Algorithms are based on pretest probabilities and D-dimer results. As many of 40% of patients with a low clinical suspicion and a negative D-dimer result require no further evaluation. [14, 15]
Access to the popliteal vein is usually obtained with ultrasonographic guidance, though the common femoral, tibial, or internal jugular veins are also used. When thrombolysis is planned, use of ultrasonography and a micropuncture 21-gauge needle are recommended to minimize bleeding risk. Diagnostic venography is used to identify the extent of DVT, and fluoroscopic guidance is the most accurate and straightforward means of placing infusion catheters or devices. A sheath is placed, and a multiple–side-hole catheter or wire is used to deliver the drug and maximize exposure of the lytic to the surface area of the thrombus.
The CT finding of intraluminal thrombus is documented as a filling defect on a delayed contrast-enhanced scan.
Spiral multidetector-row CT venography (CTV) from the popliteal fossa to the pelvis offers good diagnostic accuracy and correlation with sonographic findings. The radiation dose, cost, and scanning time, as well as the recent explosion in the number of requests for CT scanning at most hospitals, have made it prohibitive to use CT to evaluate extremity DVT alone. [16, 17, 18]
Studies of multidetector-row CTV showed that that venous-phase scanning after arterial-phase scanning is feasible and possibly cost-effective. In practice, adding indirect CTV to the now relatively standard CT pulmonary arteriography for suspected PE leads to additional diagnoses of thrombotic disease in only a few patients. However, this is an incremental increase of 15-38% of diagnoses of venous thromboembolism. Among patients in emergency departments, the yield is relatively low, and management is unlikely to be changed because deep venous thrombosis is rarely identified in the absence of PE. In the converse, in an oncology population, the addition of CTV to CT pulmonary angiography (CTPA) resulted in a 20% increase in detection of thrombotic disease. CTV showed DVT isolated to iliac or pelvic veins in 4.5%. [17, 18, 19, 20]
CTV is useful for evaluating for DVT versus other causes of leg swelling in patients with equivocal or negative Doppler sonographic results and for obtaining additional information in patients with known DVT before endovascular treatment. CTV reliably depicts the extent of the thrombi and underlying anatomic abnormalities, and it may help in defining the chronicity of the lesion. Increased attenuation of the thrombus (>60 HU) and an increased diameter of the vessel (>150% the diameter of the contralateral normal vein) are correlated with acuity, and they are predictors of successful CDT.
CT requires the use of iodinated contrast agent, to which some patients are allergic. In addition, renal insufficiency is a contraindication because of the large dose of contrast agent needed.
The radiation dose for bilateral lower-extremity CTV is 3-8 mSv (less than that of abdominal CT).
Claustrophobia, extreme patient girth, certain metallic implants, or inability to remain immobile can produce nondiagnostic studies, though these factors are generally less important with CT than with MRI.
In preliminary studies, CTV findings that were used to exclude iliofemoral thrombi had a sensitivity equivalent to that of ultrasonography. Studies in which indirect CTV was compared with venography showed 100% sensitivity and 96-97% specificity. Contrast enhancement of vessels to greater than 60 HU is desired. In the studies, CTV required 80% less contrast agent than venography.
In an ICU setting, combined CTPA and CTV yielded a 25% incidence of nondiagnostic DVT studies because of inadequate contrast opacification or because of artifacts due to metallic hardware. CT scans do not help in differentiating chronic from acute DVT.
False-positive findings include tumor thrombus and/or invasion (pelvis and cava), compression by extrinsic mass (usually detected), inflow defects from unopacified blood (usually seen at the iliac confluence to the IVC, brachiocephalic confluence, or inflow at the renal vein), and poorly timed CT scanning with indeterminate findings. False-negative findings include small thrombi (< 1 cm) when CT is performed at large gaps or intervals (ie, 5 mm of every 20 mm scanned) to reduce the radiation dose. This technique reduces sensitivity as well.
Magnetic Resonance Imaging
Findings on magnetic resonance venography (MRV) depend on the sequence used. If nonenhanced (flow, bright blood) or contrast-enhanced (gadolinium-enhanced) images are obtained, they demonstrate a bright rim around a dark intraluminal filling defect. Although MRI is highly sensitive and relatively specific, the cost of the examination, the technical complexity, and the lack of general availability limit the use of MRV as a screening tool. Specific indications for MRV are primarily as an alternative to CT (particularly in patients with an allergy to contrast material, in those with renal failure, and those in whom an evaluation of the iliocaval veins are required for questionable sonographic findings) or for a preinterventional evaluation of the extent of a thrombus. [21, 22, 23, 24]
MRI cannot be used in patients with ferromagnetic implants or in those who depend on metallic devices that cannot be placed in the imaging unit.
Claustrophobia, extreme patient girth, certain metallic implants, or an inability to remain immobile can produce nondiagnostic studies.
MRV is effective and accurate, with a sensitivity and specificity for iliac and femoral DVT approaching 100% compared with venography and a 92% sensitivity in detecting isolated calf-vein thrombus. In addition, pelvic veins that are nearly impossible to visualize on sonography and difficult to view by other means are consistently imaged well with MRV.
In general, MRI findings are subject to many artifacts that simulate vascular disease. Adjacent metallic objects, inadequate contrast enhancement, turbulent or sluggish venous blood flow, inflow from another vein into a vessel filled with contrast agent, and reflux (reversal) of venous blood flow may affect the signal received, depending on the machine and protocol chosen. False-positive findings may result from slow or turbulent flow, an adjacent pulsatile structure, or hypointense inflow defects.
Gadolinium-based contrast agents (gadopentetate dimeglumine [Magnevist], gadobenate dimeglumine [MultiHance], gadodiamide [Omniscan], gadoversetamide [OptiMARK], gadoteridol [ProHance]) have been linked to the development of nephrogenic systemic fibrosis (NSF) or nephrogenic fibrosing dermopathy (NFD). The disease has occurred in patients with moderate to end-stage renal disease after being given a gadolinium-based contrast agent to enhance MRI or MRA scans. NSF/NFD is a debilitating and sometimes fatal disease. Characteristics include red or dark patches on the skin; burning, itching, swelling, hardening, and tightening of the skin; yellow spots on the whites of the eyes; joint stiffness with trouble moving or straightening the arms, hands, legs, or feet; pain deep in the hip bones or ribs; and muscle weakness.
Ultrasonography is the current first-line imaging examination for deep venous thrombosis (DVT) because of its relative ease of use, absence of irradiation or contrast material, and high sensitivity and specificity in institutions with experienced sonographers. [25, 26, 27] Compression ultrasonography entails imaging the calf to the groin in the axial plane with a 5- to 10-MHz transducer. Compression is intermittently applied to induce complete coaptation of the walls of the patent vein. If the vein does not compress, it is occluded. Attempts to visualize iliac and pelvic veins are made.  See the images below.
Regarding clinical outcomes, the negative predictive value at 3 months after compression ultrasonography yields normal results is 97-98%, and it is greater than 99% with serial ultrasonography. A more comprehensive study includes color Doppler imaging. In addition to compressibility, the evaluation includes an assessment for incomplete color filling, flow augmentation (vein patency peripheral to the transducer), and respiratory variation (patency central to the transducer). A negative single, complete duplex color sonogram of the entire lower extremity obtained to assess suspected DVT has a negative predictive value of 99.5%.
Specific findings include the following:
Incompressibility: A thrombosed vein does not compress.
Loss of augmentation: Loss of appropriate increased flow when the lower extremity is compressed implies an obstruction (clot) between the transducer and the compressed area.
Visualization: DVT may be directly visualized as moderately echoic to hyperechoic masses separate from anechoic fluid.
Doppler flow: Doppler color-flow imaging can depict absent or abnormal flow in an area where isoechoic clot might not be visible.
Below-the-knee thrombus: A clot below the popliteal vein level remains an elusive area in duplex scanning. It is challenging to detect, and detection is operator dependent.
Patient size limits the use of sonography because large patients are difficult to scan with accuracy. Good-quality sonograms depend on the experience of the technologist performing the examination. The iliac and pelvic veins are not imaged consistently with sonography.
In patients with clinically suspected disease, compression ultrasonography is 95-99% sensitive for proximal venous thrombus compared with contrast venography. For isolated calf-vein thrombus, the sensitivity decreases below 50%. The high accuracy of ultrasonography versus venography for the diagnosis of proximal DVT has been demonstrated in asymptomatic patients. In clinical evaluations in which anticoagulation was withheld on the basis of negative serial compression sonograms, the incidence of thromboembolic complications was 0.07-1.5% at 6-month follow-up. [14, 15]
Because of the limitation of diagnostic study in the proximal veins, serial scans are required to ensure that calf-vein DVT is not progressing. A few become positive over 7-day follow-up. However, because of the cost and patient-compliance issues with follow-up testing, investigators evaluated the usefulness of single, complete lower-extremity compression sonography. The technical-failure rate was 1.5%; these cases required additional study. The outcome evaluated was thromboembolic complication at 3 months, which occurred in 0.2-0.8% of studies.
Visualization of iliac or proximal thrombus is often difficult. In the presence of thigh swelling or an abnormal common femoral vein, the central iliocaval veins warrant evaluation. Interposed bowel gas may compromise duplex ultrasonography, and CTV or MRV have been useful adjuncts. Visualization at the adductor canal is similarly difficult, and a focal thrombus may not be identified; however, this has not compromised the clinical relevance of a negative study. If indeterminate findings occur, extremity venography remains the diagnostic criterion standard.
False-positive findings may result from a technical error in scanning or from interpreting chronic DVT as acute DVT. However, the use of compression ultrasonography with a consideration of venous diameter is highly sensitive in identifying recurrence. [29, 30] False-negative findings may result from inadequate scanning due to the size of the patient's leg; edema; or inexperience of the technologist, who must carefully scan each segment. In addition, iliac or pelvic DVT may be missed because of overlying bowel gas, which is the major limitation of duplex scanning in patients with DVT. In most patients, an iliac or pelvic DVT cannot be completely excluded.
Femoral vein duplication is a congenital variant that poses a pitfall in diagnosis. If a patent femoral vein is identified, an occluded duplicated vein may be missed if the anomaly is not recognized.
Radiolabeled peptides that bind to various components of a thrombus have been investigated. Apcitide, a technetium-labeled platelet glycoprotein IIb/IIIa receptor antagonist, is approved for diagnostic studies of DVT. Other peptides in development include fragments of fibronectin with a distinct fibrin-binding domain and analogs of laminin and thrombospondin, which bind to platelet receptors. [31, 32]
The cost of the tests and the inability to visualize the anatomy of the area of involvement (which many clinicians prefer) has led to the underuse of scintigraphy. The radiation dose is 6.8 mSv, equivalent to lower-extremity CTV.
Foci of increased activity indicate an acute thrombus in that location. This scanning technique is used in institutions where practitioners have experience and confidence in the technique.
In a multicenter evaluation of technetium-99m-apcitide scintigraphy compared with contrast venography in 243 symptomatic or high-risk patients, 99mTc-apcitide had a sensitivity of 75.5% and a specificity of 72.8%. However, after patients with a history of DVT or PE were excluded, the sensitivity and specificity were 90.6% and 83.9%, respectively, for 99mTc-apcitide. 
The classic finding of acute thrombus is an intraluminal filling defect in the contrast opacified vein. Lack of opacification of a vein or venous segment indicates occlusion. Occlusion is consistent with an acute or chronic thrombus. Findings of intraluminal septation, webs, or stenoses are consistent with a healed or remote deep venous thrombosis. In chronic DVT, recanalization can result in a linear filling defect in the vein, sometimes termed the tram-track pattern. The vein appears as if it were 2 small, paired veins.
Until the 1980s, venography was the criterion standard examination for DVT. This procedure is now uncommonly performed because of the patient's discomfort from needle puncture, the potential for infiltration of contrast agent at the injection site or allergy to the agent, and the cost in time and infrastructure necessary to perform the examination. The development of highly sensitive, noninvasive ultrasonography and impedance plethysmography protocols for DVT has relegated the use of venography to specific indications.
Venography remains the examination of choice when absolute determination of the presence and extent of thrombus is needed. This study is often required in obese patients, in patients with severe leg edema, or in patients in whom results of noninvasive tests are equivocal or negative in the setting of high clinical suspicion. If technically adequate, the study offers a high degree of confidence. Technical limitations include poor IV access in the foot, poor contrast opacification of the deep veins (contrast material shunted to superficial veins, injection too slow, poor tourniquet compression), motion artifact, and excessive muscular contractions or spasms.
An IV line is placed in a dorsolateral foot vein, and several tourniquets (placed at the ankle and below and above the knee) or the reverse Trendelenburg positioning is used to shunt contrast material into the deep venous system. The pelvis is imaged by compressing the femoral vein while the leg is elevated or while the table is moved from the reverse Trendelenburg to the Trendelenburg (head-down) position. Compression is then released while the external iliac vein is rapidly imaged.
Images are obtained from the foot to the pelvis, and detailed images of the entire deep venous system, including the paired tibial veins, iliacs, and IVC can be obtained. The internal iliac vein in the pelvis is not imaged, and a clot in this area cannot be excluded. The mean radiation dose for a single extremity is 6 mSv.