Real-time B-mode venous ultrasound

Transcription

1 Real-time B-mode venous ultrasound Eugene D. Sullivan, M.D., David J. Peter, B.S., and John J. Cranley, M.D., Cincinnati, Ohio The ability of real-time B-mode ultrasound to directly visualize arteries and veins and thereby give anatomic rather than physiologic information is unique among the currently available noninvasive methods of vascular evaluation. The usefulness of this technique for examination of the carotid arteries has been well proven. Little attention, however, has been given to its applicability for deep venous evaluation. Over the past 12 months the veins of 108 upper and 215 lower extremities have been studied with high-resolution real-time ultrasound. The technique and interpretive criteria are presented. Thirty extremities underwent confirmatory venography, and in this group the specificity and sensitivity of the ultrasound were 94% and 100%, respectively. In addition, the age of the thrombi detected was accurately predicted in 93% of this group. These preliminary results suggest that real-time B-mode venous ultrasound is an accurate, clinically useful noninvasive technique for the detection of deep venous thrombosis that complements the more widely used physiologic screening tests. (J VAsc SURG 1984; 1: ) Real-time B-mode ultrasound is a well-accepted diagnostic modality in the noninvasive vascular laboratory. This technique is primarily useful in visualizing the extracranial carotid arteries. 1-4 B-mode ultrasound has also been used on a limited basis to assess the common femoral artery bifurcation 5 and subcutaneous arterial grafts. ~ The application of this technique to the detection of deep venous thrombosis is limited to a single preliminary report. 7 We now present our recent experience with real-time B-mode ultrasound for the evaluation of the veins of the upper and lower extremities. MATERIAL AND METHODS Between August 1, 1982, and July 31, 1983, ~36 patients underwent 170 B-mode ultrasound venous examinations; 323 extremities were studied in their entirety. Two hundred fifteen lower extremities were examined in 103 patients, and 108 upper extremities were scanned in 33 patients. Thirty extremities underwent confirmatory venography (seven upper and 23 lower). When venography was obtained, the noninvasive studies were performed and interpreted without prior knowledge of the phlebographic findings. All scans were performed with a commercially available high-resolution real-time B-mode ultra- From the Kachelmacher Memorial Vascular Laboratory and the Department of Surgery, Good Samaritan Hospital. Presented at the Seventh Annual Meeting of the Midwestern Vascular Surgical Society, Chicago, I11., Sept , Reprint requests: John J. Cranley, M.D., Peripheral Vascular Laboratory, Good Samaritan Hospital, Cincinnati, OH sound imager. Two probes were routinely used: an 8 MHz probe with an integrated 5 MHz pulsed Doppler and a 4 MHz probe. This combination of probes allows noninvasive visualization of veins up to 7 or 8 cm deep, as well as determination of their flow characteristics by audible interpretation of their Doppler signals. Both transverse and longitudinal views are possible. Video cassette recordings were obtained of all scans for review and interpretation. The lower is examined with the patient in a reversed Trendelenburg position at 10 to 15 degrees. This allows maximum venous filling and improves visualization. With the patient supine, the common femoral, superficial femoral, greater saphenous, and posterior tibial veins are examined bilaterally. With the patient repositioned in the lateral decubitus or prone position, the distal superficial femoral and popliteal veins may then be visualized. The veins proximal to the inguinal ligament cannot be visualized because of their depth and the overlying bowel gas. Indirect information about the iliac veins, however, can be obtained by assessing the Doppler signals of the common femoral vein for spontaneity, phasicity, and augmentation, as has been = previously described. 8 An entire leg can be evaluated in approximaeely 25 minutes. The upper is examined with the patient supine and either flat or in a slight Trendelenburg position. Views of the jugular, distal subclavian, axillary, and brachial veins are systematically obtained by interchanging the 4 and 8 MHz probes as necessary. The subclavian vein proximal to the cephalic vein junction and the superior vena cava can- 465

2 466 Sullivan, Peter, and Cranley Journal of VASCULAR SURGERY Fig. 1. Acute popliteal vein thrombosis (longitudinal view). PV = popliteal vein; PA = popliteal artery. Table I. Ukrasound criteria of patency 1. Absence of intraluminal thrombus 2. Compressibility of the vein by ukrasound probe 3. Visualization of venous blood flow or valve motion 4. Visualization of changes in vein diameter with quiet respiration or Valsalva maneuver 5. Normal venous Doppler signals (spontaneous, phasic, augmented) not be routinely visualized, but Doppler signals provide indirect information about patency. Complete examination requires approximately 15 minutes. INTERPRETATION Several criteria of patency have been established as a result of our preliminary experience (Table I). Actual visualization of thrombus is a major and the most specific diagnostic criterion (Fig. 1). This ability of high-resolution real-time ultrasound to detect thrombus has been confirmed by others2 '1 Coelho et al. 9 have also shown that acute and chronic thrombi are often distinguishable ultrasonographically, the latter being more brightly echogenic because of fibrous organization and lamination. Fresh thrombus is also somewhat compressible by externally applied pressure by the ultrasound probe, whereas the chronic variety is rigid. Finally, the visualization of free-floating thrombus, without attachment to the vein wall, is presumptive evidence of its acuteness (Fig. 2). We have studied several patients who presented with acute venous thrombosis and whose organizing clot was then followed up ultrasonographically. Usually, acute and chronic thrombi can be differentiated, although the distinction is admittedly somewhat subjective and qualitative. Compressibility of the vein by the examining probe is another sensitive and reproducible indicator of patency. The vein walls should be easily coapted, particularly in the transverse view. The weight of the probe alone often causes complete compression. Venous thrombosis produces either absent or partial compressibility, depending on the completeness of the process. We have observed the progression of compressibility from being absent with acute thrombosis and complete occlusion to being partially compressible with recanalization. A noncorr pressible common femoral vein without visible thrombus suggests venous hypertension due to proximal iliac disease. Compressibility may be difficult to demonstrate when bony structures such as the clavicle or the femoral condyles are in close proximity to the vein. In these instances visualized changes in diameter of the vein with quiet respiration or a Valsalva maneuver ensure patcncy. Blood flow is routinely visualized, especially in the longitudinal view. This is not only quite dramatic when seen but further signifies patency. The echogenicity of flowing blood is thought to represent red blood cell aggregation or rouleau formation. ll Venous valves are also often visualized, and their rhythmic movement further signifies blood flow and therefore venous patency (Fig. 3).

3 Volume 1 Number 3 May 1984 Real-time B-mode venous ultrasound 467 Fig. 2. Chronic recanalized common femoral vein thrombus (longitudinal view). CFA = common femoral artery; CFV = common femoral vein; T = thrombus; L = residual lumen. Fig. 3. Normal internal jugular vein with valve (transverse view). IJV = internal jugular vein; V = valve cusp; CCA = common carotid artery. Finally, the integrated pulsed Doppler in the 8 MHz probe can be used to detect blood flow and assess the spontaneity, phasicity, and augmentation of the venous signal. 8 This well-accepted modality provides indirect information about the iliac veins and the proximal subclavian veins, which cannot be directly visualized because of the technical considerations previously mentioned. Lower results. Two hundred fifteen lower extremities were studied in their entirety in 103 patients. Of these, 23 also underwent confirmatory venography within 48 hours of the B-mode ultrasound study. Venous thrombosis was considered acute only if an actual clot was visualized by venography. As shown in Table II, the specificity and sensitivity of the B-mode ultrasound compared with those of venography were 92% (11 of 12) and 100% (11 of 11), respectively. The one false positive scan occurred early in the experience before the inability of the technique to visualize above the inguinal ligament was fully appreciated. Iliac vein thrombus was thought to be visible, although, in retrospect, fluid-filled small intestine was being seen. Isolated distal posterior tibial vein thrombi were

4 468 Sullivan, Peter, and Cranley Journal of VASCULAR SURGERY Fig. 4. Posterior tibial vein thrombi (transverse view). V = vein; A = artery. Table II. Comparison of B-mode ultrasound of lower and venography (N = 23) Lower B-mode Normal Abnormal Total Normal Abnormal Total Specificity = 11/12 = 92%; sensitivity = 11/11 = 100%; negative predictive value= 11/11= 100%; positive predictive value = 11/12 = 92%. clearly documented by B-mode ultrasound in one additional patient whose venogram was interpreted as normal (Fig. 4). X-ray films of the distal calf and foot, however, were not obtained; therefore this was not considered a false positive B-mode scan. The ability of B-mode ultrasound to distinguish acute and chronic thrombus in contrast to venography is shown in Table III. There was complete nonvisualization of the deep venous system with few collaterals but no visible thrombus in one thought to be acutely thrombosed noninvasively. Otherwise, complete agreement between the two modalities existed in the remaining 10 extremities. An additional patient was thought noninvasively to have free-floating fresh thrombus in the common femoral vein superimposed on chronic occlusion of the superficial femoral vein. Surgical exploration without venography confirmed these noninvasive findings, and thrombectomy of the acutely thrombosed femoral vein with ligation of the chronically Table III. Categorization of thrombus by B-mode ultrasound and venography (11 diseased lower extremities) Lower Acute Chronic Acute and B-mode D VT D VT chronic D VT Equivocal Acute DVT Chronic DVT Acute and chronic DVT DVT = deep venous thrombosis. diseased superficial femoral vein was performed. This potential ability of B-mode ultrasound to distinguish acute and chronic thrombus is intriguing, and of obvious clinical importance. The number of venograms performed in this study population is relatively small because of our extensive reliance on phleborheography (PRG) for clinical decision making.12 One hundred seventy-one lower extremities were studied with both B-mode ultrasound and PRG, without confirmatory venography. The correlations between these two noninvasive modalities are shown in Table IV. Because B-mode ultrasound evaluates anatomy and PRG evaluates venous physiology, discrepancies are inevitable. Meaningful explanations for the 31 mismatches can be offered but remain unproven in the absence ofvenography. Eight of the 10 false positive B-mode ultrasound examinations revealed chronic disease, and a normal PRG in this group implies normal venous return because of recanalization or

5 Volume 1 Number 3 May 1984 Real-time B-mode venous ultrasound 469 CLINICAL EVALUATION PATIENT IS UNCOOPERATIVE, UNABLE TO BE POSITIONED, IN TRACTION, S/P A MEUTA TION EQUIVOCAL OPPLER NEGATIVE UNEQUIVOCALLY POSITIVE / N B-MODE ULTRASOUND NO TREATMENT ANTICOAGULA TE ([ ACUTE, OR? ACUTE VS. CHRONIC I ANTICOA GULATE CHRONIC OR NORMAL I NO ANTICOAGULANTS collateralization, a well-recognized phenomenon. 1~ The remaining two extremities with a normal PRG and an abnormal B-mode scan had ultrasound evidence of acute phlebitis with only partial venous obstruction. Both patients had positive PRG and B-mode scans in the opposite extremities and were, therefore, anticoagulated. In seven of the eight extremities in which PRG was abnormal and ultrasound normal, pelvic disease (hematoma, iliac artery aneurysm, recurrent carcinoma, and so forth), which was suspected to be producing extrinsic compression of the iliac veins, was present. Normal B-mode ultrasound was very useful in this group in at least excluding intraluminal thrombosis distal to the inguinal ligament. Previously we would have insisted on venography to distinguish extrinsic compression from phlebitis. Phle- ~oography was avoided in these cases because of the B-mode ultrasound findings. The one remaining with abnormal PRG and normal B-mode ultrasound had strong clinical evidence of phlebitis, a clearly abnormal PRG on two occasions, and yet a normal B-mode scan. This patient was anticoagulated, and no explanation other than iliac vein thrombosis undetected by the ultrasound can be offered to explain this discrepancy. Technical problems with PRG due to difficulties with patient positioning or apprehension were recognized in 13 extremities while the test was being performed. An interpretation of "equivocal" resulted. A normal B-mode ultrasound in this group was reassuring evidence of normalcy and may have obviated venography or anticoagulation therapy. There was agreement between PRG and B-mode ultrasound in 140 extremities (82%). Seventy-three Fig. 5. Algorithm for suspected phlebitis. Table IV. Comparison of B-mode ultrasound and phleborheography (N = 171) Lower Phleborheography B-mode Normal Abnormal Equivocal Total Normal Acute DVT Chronic DVT Acute and chronic DVT Total ] DVT = deep venous thrombosis. of the extremities with abnormal venous physiology as detected by PRG were further categorized by ultrasound as having acute or chronic thrombosis or both. Because venography was not performed, proof of accuracy is unavailable. Most patients, however, were treated on the basis of the combined findings. In another 21 extremities PRG could not be performed because of lack of patient cooperation. B-mode ultrasound was easily obtained in all instances. Deep venous thrombosis was detected and categorized in 13 extremities, and the patients were treated accordingly. In addition, the common and superficial femoral veins of three extremities with above-knee amputations were studied to exclude postoperative venous thrombosis. This was easily accomplished. Another three extremities with superficial phlebitis of the proximal greater saphenous vein were examined to detect propagation of thrombus into the common femoral vein. Upper results. Of the 108 upper extremities examined in 33 patients, only seven un-

6 470 Sullivan, Peter, and Cranley Journal of VASCULAR SURGERY Table V. Comparison of B-mode ultrasound of upper and venography (N = 7) Upper B-mode Normal Abnormal Total Normal Abnormal Total 7t ~ Specificity = 4/4 = 100%; sensitivity = 3/3 = 100%; negative predictive value = 4/4 = 100%; positive predictive value = 3/3 = 100%. Table VI. Comparison of upper and lower B-mode ultrasound and venography (N = 30) Upper B-mode Normal Abnormal Total Normal Abnormal Total Specificity = 15/16 = 94%; sensitivity = 14/14 = 100%; negarive predictive value= 15/15 = 100%; positive predictive value = 14/15 = 93%. derwent confirmatory venography. As shown in Table V, both the specificity and sensitivity of B-mode ultrasound compared with those of venography were 100%. All three abnormal extremities had both ultrasound and venographic evidence of acute thrombosis. Another three upper extremities revealed acute subclavian and axillary vein thrombosis by ultrasound without venography. Two of these patients had been previously studied and found to be normal but returned with new arm symptoms and ultrasound evidence of acute thrombosis (both patients had subclavian hyperalimentation catheters). Combined results. Thirty extremities (23 lower and seven upper) were studied with both B-mode ultrasound and venography. As shown in Table VI, the composite specificity and sensitivity of the former were 94% (15 of 16) and 100% (14 of 14), respectively. The age of the thrombus in the 14 diseased extremities (11 lower and three upper) was accurately predicted by ultrasound in 93% (13 of 14), as shown in Table VII. One venogram was equivocal, as previously described. DISCUSSION The usefulness of real-time B-mode ultrasound in the vascular laboratory relates to its unique ability Table VII. Categorization of thrombus in all diseased extremities (11 lower and three upper) by B-mode ultrasound and venography B-mode Acute Chronic Acute and ultrasound DVT DVT chronic DVT Equivocal Acute DVT Chronic DVT Acute and chronic DVT DVT = deep venous thrombosis. to noninvasively visualize arteries and veins previously visible only with invasive contrast radiography. Most noninvasive tests use physiologic measurement of pressure or blood flow to obtain indirect information about patency, stenosis, or occlusion (~: veins or arteries. These physiologic tests are generally incapable of differentiating stenosis from occlusion or intraluminal obstruction from extrinsic compression. With the exception of 125I-fibrinogen scanning, all other noninvasive vascular techniques are also incapable of distinguishing acute from chronic thrombosisy Because B-mode ultrasound permits direct visualization of vascular structures in multiple views, it is uniquely capable of differentiating these entities from one another. This is of obvious clinical value, particularly in evaluation of deep venous disease, where it is therapeutically crucial to quickly separate acute from chronic thrombus and intraluminal thrombosis from extrinsic venous compression. Our preliminary experience with B-mode ultrasound of the veins of the upper and lower presented here is quite encouraging. Th(' technique is relatively simple and rapid and requires minimal patient cooperation. It appears to be well tolerated by patients and is without documented ill effects. The expense for the necessary high-resolution real-time ultrasound imager is substantial, but many laboratories are already using the required instrumentation for carotid artery evaluation. It is clear that B-mode ultrasound can accurately detect and localize venous thrombosis. Specificity and sensitivity values of 94% and 100%, respectively, in the 30 limbs with confirmatory venography prove this. Other more conventional noninvasive techniques that rely on physiologic principles, such as PRG or the venous Doppler survey, are capable of similar accuracy, s,12 However, the ability of B-mode ultrasound to visualize the clot and to dis-

7 Volume 1 Number 3 May 1984 Real-time B-mode venous ultrasound 471 tinguish acute from chronic thrombus in 93% of our diseased extremities with confirmatory venograms represents a distinct advantage. Our present diagnostic criteria for distinguishing acute and chronic thrombosis are admittedly subjective and qualitafive. A more quantitative method using the amplitude of the A-mode reflection of the thrombus has been proposed by Sigel et al.,n but this has not been effective in our hands. We are presently exploring other objective methods of determining the age of clots with ultrasound. Certainly the age of the thrombus will at times be difficult to judge, but the fact that we could identify concomitant acute and chronic clots in five extremities (four venographic and one operative correlation) is noteworthy. Equivocal physiologic screening tests are inevitable because of patient discomfort, positioning ar- ~,-~facts, and other factors. 12 The application of B-mode ultrasound to complement screening tests such as PRG when they are equivocally abnormal should improve their overall clinical usefulness. Similarly, in patients with known pelvic disease that could cause extrinsic compression of the iliac veins and therefore an abnormal physiologic test, B-mode ultrasound can exclude intraluminal thrombosis below the inguinal ligament as a cause of the abnormal test. This may be of some clinical value and may decrease the need for venography in such instances. B-mode ultrasound is also a useful method of examining patients unable to be studied noninvasively otherwise because of above-knee amputation or lack of cooperation. Similarly, the proximal extent of thrombus in the greater saphenous vein can be determined, a function not in the province of ~physiologic tests such as PRG. The inability to visualize the iliac and proximal subclavian veins remains a major potential disadvantage. Indirect information is obtained by assessing the common femoral or axillary veins for compressibility, their venous Doppler signals, and the rapidity of visualized blood flow. In this instance the advantages of B-mode ultrasound to distinguish acute from chronic thrombus and intraluminal thrombosis from extrinsic compression are obviously lost. Fortunately the sensitivity of PRG for iliac thrombi approaches 100%, 12 which again emphasizes the desired use of these two techniques in a complementary rather than an exclusive manner. In summary, as a result of this preliminary experience, we believe that real-time B-mode ultrasound examination of the veins of the upper and lower extremities is a useful addition to the noninvasive techniques presently available for venous evaluation. Its capabilities of reliably detecting thrombus and possibly characterizing thrombus age are of obvious clinical usefulness. Critical information about patients unable to be examined with physiologic techniques because of technical factors can also be easily obtained. Rather than replacing the more widespread physiologic methods of evaluation, such as PRG or venous Doppler survey, we propose that B-mode venous ultrasound be used to complement these. As a result of this experience, our present approach to suspected deep venous thrombosis uses both physiologic and anatomic methods, as shown in Fig. 5. should be necessary only rarely. Gratitude to S. R. Talbot, R.V.T., and Clynn R. Ford, M.D., for sharing their technique with us. REFERENCES 1. Comerota AJ, Cranley JJ, Cook SE. Real-time B-mode carotid imaging in diagnosis of cerebrovascular disease. Surgery 1981; 89: Hobson RW, Berry SM, Katocs AS, et al. Comparison of pulsed Doppler and real-time B-mode arteriography for noninvasive imaging of the extracranial carotid arteries. Surgery 1980; 87: Johnson JM, Ansel AL, Morgan S, DeCesare D. Ultrasonographic screening for evaluation and follow-up of carotid artery ulceration. Am J Surg 1982; 144: Johnson JM. Angiography and ultrasound in diagnosis of carotid artery disease: A comparison. Contemp Surg 1982; 20: Hobson RW, Berry SM, Katocs AS, et al. Real-time B-mode ultrasonography of the femoral arteries: Comparison to contrast arteriography. Am Surg 1981; 47: Clifford PC, Skidmore R, Woodcock JP, et al. Arterial grafts imaged using Doppler and real-time ultrasound. Vasc Diagn Ther 1981; 2: Talbot SR. Use of real-time imaging in identifying deep venous obstruction: A primary report. Bruit 1982; 6: Strandness DE, Sumner DS. Ultrasonic velocity detector in the diagnosis of thrombophlebitis. Arch Surg 1972; 104: Coelho JC, Sigel B, Ryva JC, et al. B-mode sonography of blood clots. J Clin Ultrasound 1982; 10: Sigel B, Coelho JC, Spigos DG, et al. Ultrasonography of blood during stasis and coagulation. Invest Radiol 1981; 16: Siget B, Machi J, Beitler JC, et al. Variable ultrasound echogenicity in flowing blood. Science 1982; 218: Comerota AJ, Cranley JJ, Cook SE, Sippel PJ. Phleborbeography--Restdts of a ten-year experience. Surgery 1982; 91: Sauther RD, Larson DE, Bhattacharyya SK, et al. The limited utility of fibrinogen leg scanning. Arch Intern Med 1979; 139: