MR imaging in inferior vena cava thrombosis

EUROPEAN JOURNAL OF RADIOLOGY

ELSEVIER

European Journal of Radiology 19 (1995) 101-107

M R imaging in inferior vena cava thrombosis Rafaela Soler *a, Esther Rodriguez a, Ma Fernanda L6pez b, Milagros Marini a aDepartment of Radiology, Hospital Juan Canalejo, Xubias de Arriba, 84, 15006 La Coruna, Spain bDepartment of Hematology, Hospital Juan Canalejo, Xubias de Arriba, 84, 15006 La Coruna, Spain Received 25 May 1994; revision received 13 October 1994; accepted 28 October 1994

Abstract

Purpose: To prospectively evaluate the efficacy of MR imaging in inferior vena cava (IVC) thrombosis; to differentiate acute from non-acute thrombus, and to identify the presence of changes in the morphology and signal intensity during medical treatment. Methods and materials: Seventeen patients with suspected IVC thrombosis underwent 30 MR examinations. The IVC thromboses were subdivided into two groups (acute and non-acute) according to onset of clinical symptoms. MR imaging of the IVC was analyzed and when an IVC thrombus was identified, a qualitative and quantitative assessment of the thrombus on spin-echo sequences with magnitude and phase reconstruction was performed to evaluate the relationship between signal intensity and the time elapse since the onset of clinical symptoms. Venography and/or CT scan proof was available in all cases. Results: IVC thrombus was correctly identified in 19 MR examinations which showed the size, localization and the degree of lysis during follow-up. No differences were found in the signal intensity of the thrombus related to time. The pattern of the signal intensity was homogeneous in six (86%) acute thrombus and heterogeneous in nine (75%) thrombus of more than 1 week duration. A significant statistical relationship (P < 0.01) existed between the thrombus age and differences in the pattern of signal intensity. Conclusions: MR imaging is accurate to assess the localization and size of IVC thrombus, similar to the imaging techniques of reference. In addition, MR also provides useful information about the age and the morphological variations of thrombus during medical treatment. Keywords: Venae cavae, MR; Vena cavae, abnormalities; Venae cavae, thrombosis; Magnetic resonance (MR), vascular studies

1. Introduction

Inferior vena cava (IVC) thrombosis is a life-threating condition in which early diagnosis has both therapeutic and prognostic importance. Due to the protean clinical presentation of IVC thrombosis, clinical diagnosis is difficult and imaging examination is necessary to confirm or reject the presence of IVC thrombus. Several imaging techniques have been used to evaluate patients with obstruction of the IVC and its tributaries. Although early detection of acute cases is important to establish the appropriate therapy, clear assessment of the thrombus age has not been finally achieved with conventional venography, CT or ultrasound [1,2]. Preliminary reports have described the use of MR imaging for detecting vascular thrombosis. However, the utility of this technique * Corresponding author.

has not been explored in a large number of patients suspected of IVC thrombosis and the ability to differentiate acute or non-acute thrombus has only been described in isolated cases [2-4]. We prospectively analyzed the findings oi 30 MR studies in 17 patients suspected of IVC thrombosis to determine whether MR imaging can be used to predict the age of the thrombus and to follow-up the medical treatment. 2. Materials and methods

2.1. Study group Seventeen patients with documented blood thrombus of the IVC underwent 30 MR examinations. The study group consisted of nine males and eight females with ages ranging from 15 to 82 (mean 58) years. Seven pa-

0720-048X/95/$09.50 © 1995 Elsevier Science Ireland Ltd. All rights reserved SSDI 0720-048X(94)00587-3

102

R. Soler et aL /European Journal of Radiology 19 (1995) 101-107

Table 1 Predisposing factors, clinical presentation and treatment of the study group No. Age (years) Sex

Predisposing factor

Clinical presentation

Treatment

1 33 M 2 57 M 3 80 F 4 72 M 5 68 F 6 41 F 771 F 8 63 F 9 74 M 10 44 M 11 31 F 12 77 F 13 31 M 14 53 M 15 15 F 16 82 M 17 71 M

Tuberculous disease No No Surgery ( < 8 days) Tuberculous disease Ulcerous colitis No No Protein C deficit and surgery Protein S deficit Pregnancy and lupus anticoagulant Cardiopathy and surgery No Hereditary displasminogenemy Surgery No No

Pulmonary thromboembolism Edema of the lower extremities Edema of the left lower extremity Edema of the left lower extremity Edema of the lower extremities Edema of the left lower extremity Edema of the lower extremities Edema of the left lower extremity Edema of the lower extremities Edema of the lower extremities Abdominal pain and fever Edema of the left lower extremity Edema of the right lower extremity Edema of the left lower extremity Edema of the left lower extremity Edema of the left lower extremity Edema of the left lower extremity

Percutaneous urokinase Pereutaneous urokinase Intravenous sodium-heparin None a Intravenous sodium-heparin None a Intravenous sodium-heparin Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant Oral anticoagulant

aTwo patients did not receive treatment due to clinical contraindications.

tients were examined due to acute symptoms related to deep venous thrombosis of less than 1 week duration and phlebocavographic identification of IVC thrombus. The other 10 patients had symptoms of more than 1 week duration or had a previously known IVC thrombosis, and MR examination was included in their follow-up during oral anticoagulant therapy. Nine patients had previous documented history of lower leg thrombosis. The clinical presentation, the predisposing factors and the treatment are summarized in Table 1. Because there is no clear parameter that can serve as reference to define the age of a thrombus, we arbitrarily defined acute thrombosis (group I) when the onset of clinical symptoms was less than 1 week duration; and non-acute (group II) (including those considered by some authors as subacute and chronic) when onset was longer than 1 week.

gap of 0.5-0.8 ram. The field of view was 35 cm in axial and 40 cm in coronal planes with an acquisition matrix of 256 x 204 pixels. Magnitude and phase reconstruction were obtained. Beside the initial MR examination performed in the 17 patients, 10 patients were also evaluated during follow-up. Follow-up MR studies at 1-week interval were performed in six of seven patients with acute IVC thrombosis, at 1-month interval in one patient; and at 1and 2-month intervals in three patients with non-acute IVC thrombus. Thus, a total of 30 MR examinations were analyzed and the MR findings were correlated with venography (n = 30) or enhanced CT scans (n = 10). The interval between the examinations correlated was < 12 h in all cases. All patients provided informed, written consent prior to the MR examination.

2.2. MR imaging technique 2.3. Image analysis MR examinations were performed with a 0.5-T superconductive magnet (Gyroscan T5, Philips, Eindhoven, Netherlands). Axial Tl-weighted spin-echo (SE) images (TR ms/TE ms = 585/20), proton density and T2weighted images (2000/35, 90) with superior and inferior spatial presaturation were obtained. The section thickness was 10-13 mm with 1.0-1.3 mm intervals. Two or four signals were averaged for Tl-weighted images and two signals were averaged for T2-weighted images. Tl-weighted images on coronal planes angulated parallel to the IVC were also obtained in some (n = 18) cases, with a slice thickness of 5-8 mm and an interslice

A qualitative assessment was performed by two radiologists blinded to the clinical, therapeutic status of the patient and the results of corroborative studies. Any discrepancies were resolved by consensus. Analysis of the presence of thrombus on SE sequences with magnitude and phase reconstruction images was based on previously established standard criteria [2,5]. On the magnitude reconstruction images, the signal intensity (SI) was described and compared with that of muscle. On the phase reconstruction images, the thrombus was classified as occlusive when the images showed an area

R. Soler et al. / European Journal of Radiology 19 (1995) 101-107

with SI of soft tissue, or non-occlusive when the clot was surrounded by a different grey level compared with the stationary tissue. The cranial extension of the IVC thrombus was established with respect to the renal veins. The morphological appearance of the IVC was categorized into the following groups: (a) IVC with normal caliber (22-28 mm), (b) IVC with enlarged caliber (> 28 ram), (c) IVC with marked diminution of caliber (< 22 mm) and, (d) presence of collaterals channels. The pattern of signal intensity was established as follows: (I) none (flow void signal), (II) homogeneous, (III) heterogeneous and (IV) no signal analyzable because of the diminution of the IVC caliber. The homogeneous pattern was defined as uniform intermediate, low or hyperintense signal compared with that of muscle, and the heterogeneous pattern was defined as a mixture of intermediate and low or with intermediate and hyperintense signal intensity. A quantitative analysis of the SI of the thrombus was performed with regions of interest (ROIs). The ROIs were placed in the center of the thrombus with an area between 2 to 5 mm 2, and the mean of the thrombus in the SE sequences was performed at 1 week and more than 1 week from the onset of clinical symptoms. The mean of SI of subcutaneous fat by ROIs and the SI ratio

a

b

103

of thrombus to subcutaneous fat (T/F) was calculated and compared with the time elapse from the onset of clinical symptoms.

2.4. Statistical analysis Standard parameters are expressed as mean, and standard error of the mean (S.E.M.) for all patients. The paired two-tailed Fisher's exact test was used to compare the patterns of SI of the thrombus with its age, and a paired two-tailed Mann-Whitney test was used to determine whether statistically significant differences existed between the SI ratio of T/F, and the thrombus age. A probability level <0.05 was considered statistically significant. 3. Results

A total of 30 MR examinations were analyzed and 19 IVC thrombi were identified in the acute period (group I "-7) and the non-acute period (group II = 12). The thrombus showed a SI slightly hyperintense related to paravertebral muscle on both T1- and T2-weighted images, with a decrease in SI from the proton density to the T2-weighted images. On MR, the precise location and

c

Fig. 1. Axial Tl-weighted (TR/TE = 585/20) images show the morphological changes of IVC after fibrinolytic treatment in a patient with occlusive and acute IVC thrombus. (a) MR imaging before therapy shows an enlarged IVC with homogeneous signal intensity (arrows) related to thrombus. (b) Follow-up study at 5 days after the diagnosis of IVC thrombosis, at the end of fibrinolytic therapy shows a marked decrease of IVC caliber and a heterogeneous signal intensity with a hypointense signal in the periphery of IVC (arrow). Collateral paravertebral circulation is seen (arrowheads). (c) Follow-up study during the chronic period shows a patent IVC and flow void signal (arrows) are identified. Paravertebral venous collaterals (arrowheads) are seen.

104

R. Soler et al./ European Journal of Radiology 19 (1995) 101-107

the cephalic extent of all 19 IVC thrombi was similar to that in corroborative imaging techniques. Ascending phlebocavography accurately demonstrated the presence of IVC thrombus in all 19 cases. However, the cephalic extent of the thrombus was not precisely defined with this examination only, and a retrograde cavography was needed in three cases (15%) and a CT study in five (25%). The MR study provided detailed information on the presence of venous collaterals pathways. The Tl-weighted images on axial plane were clear and unequivocally showed patent or thrombosed IVC. The images obtained in coronal plane with an angulation parallel to the IVC and 5 - 8 m m of slice thickness sometimes showed the entire IVC in one image, but partial volume averaging and the out of plane course of the IVC occasionally gave a false appearance of a non-occlusive thrombus in the IVC. MR demonstrated the size of the thrombus and degree of lysis during follow-up examination, similar to that of corroborative studies in all cases (Fig. 1).

In the IVC thrombi of group I (acute period, n = 7), the cephalic extent of the thrombus was infrarenal in all cases, occlusive in four cases and non-occlusive in three. The caliber of IVC was normal in five and enlarged in two. Collateral vessels were detected in three cases. The IVC thrombi of group II (non-acute period, n = 12), were occlusive in four and non-occlusive in eight. The caliber of IVC was normal in five, enlarged in two and decreased in five. Collaterals pathways were identified in five cases. In 11 MR examinations the IVC was patent. In these examinations, the IVC was of normal caliber in three and markedly decreased in eight. Magnitude reconstruction images showed flow void in six examinations and the SI was not analyzable in five due to the very small caliber of the IVC. Phase reconstruction images allowed to identify flow within the IVC in all but one case in which the IVC was atrophic. Collaterals vessels were evident in seven cases. The patterns of SI of the seven IVC thrombi studied

2

3

Fig. 2. Occlusiveand heterogeneous acute thrombus in a patient with duplicated IVC operated on of a rupture aortic abdominal aneurysm. T1weighted (TR/TE = 585/20) coronal image shows a heterogeneous signal intensity in the left IVC with an outer hyperintens¢ (small arrowheads) and a central intermedium signal similar to the right fossa hematoma (large asterisk). Other scans demonstrated its communication with the left common iliac vein and left renal vein. Flow void is identified in the right IVC (arrows). Small asterisks mark the aortic lumen. Fig. 3. Heterogeneouschronic IVC thrombus. Tl-weightext(TR/TE = 585/20) axial image shows an enlarged IVC (arrows) with an intermediate signal intensity mixed with multiple small hypointense signal.

R. Soler et al./ European Journal of Radiology 19 (1995) 101-107 Table 2 Qualitative analysis of the relationship between signal intensity pattern and the onset of clinical thrombosis (P < 0.01) Clinical thrombus

Acute thrombus Non-acute thrombus

Patterns of signal intensity Homogeneous

Heterogeneous

6 (86%) 3 (25%)

1 (14%) 9 (75%)

in the acute period (group I), were homogeneous in six (86%) (Figs. la), and heterogeneous in one (14%) (Fig. 2). The 12 IVC thrombi of more than one week (group II) had a heterogeneous pattern (Fig. 3) of signal intensity in nine (75%) studies and a homogeneous pattern in three. The relationship between the patterns of the SI of the thrombus and the onset of clinical thrombosis are given in Table 2. A good correlation existed between the patterns of SI of the thrombus and the thrombus age (Fisher exact test, two tail: P < 0.01). The sensitivity and specificity of MR imaging to establish the diagnosis of acute thrombus when using the homogeneous pattern was 85% and 75%, respectively; when the heterogeneous pattern is used to diagnose of non-acute thrombus, the sensitivity was 75% and specificity 86%. The SI of the thrombus and the SI ratio of T/F on the quantitative assessment of SE MR images are given in Table 3. No statistical relationship was found between SI of the thrombus or T/F ratios and the age of the thrombus (Mann-Whitney test, two tail: P > 0.05). 4. Discussion

The presence, localization and age of IVC blood thrombus is crucial to the therapeutic decision and has

Table 3 Quantitative analysis of the relationship between the signal intensity of the thrombus and the ratio of signal intensity of the thrombus to that of fat and the onset of clinical thrombus (P > 0.05) Sequence

Thrombus TI DP T2 Thrombus/fat T1 DP T2

Clinical thrombus Acute

Non-acute

773 ± 40.2 1310 4. 50.3 679 4. 83.6

646 4. 38.6 1172 4. 40.1 589 ± 69.7

0.53 ± 0.08 0.78 ± 0.03 1.08 4. 0.12

0.36 4- 0.24 0.79 ± 0.08 0.89 4. 0.09

All values are mean 4. standard error of the mean.

105

prognostic importance. Medical treatment with fibrinolytic or anticoagulant therapy is considered an alternative to surgery in IVC thrombosis. The most important factor in determining the probability of lysis of the thrombus with fibrinolytic therapy is the age of the thrombus. On the other hand, the more chronic and organized the blood thrombus is, the less probability to induce a pulmonary embolism. An imaging technique should have a high sensitivy and specificity to identify the blood thrombus and, ideally, its age. Thus, when we can assess the age of the thrombus, besides its size and localization, a selected therapeutic approach can be better established. Currently used imaging diagnostic modal±ties can identify IVC thrombus but with varying limitations and with little reliable information about the age of the thrombus. Ultrasound is a well-established technique for detecting thrombus in the IVC from the level of the renal veins to the right atrium. However, the inability of ultrasound to clearly show the infrarenal IVC due to overlying bowel gas has limited its use as a reliable method to diagnose or monitoring the evolution of IVC thrombus [61. Venography is considered the diagnostic technique of choice for IVC thrombosis, but is invasive and not without risks. In addition, a single ascendent study is insufficient to delineate the cephalic extent when the thrombus is occlusive, and a complementary imaging technique should be performed. Contrast enhanced CT can provide information about the localization and upper extent of an IVC thrombus. However, a hyperdense recent thrombus can be unappreciated on contrast enhanced CT scans [7,8]. MR images reveal vascular structures in anatomic detail without the use of contrast material or exposing the patient to ionizing radiation; moreover MR imaging can accurately detect intravascular thrombi in many regions of the body. The gradient-echo sequences have been suggested of choice to evaluate vascular structures and to identify thrombosis [9]. On gradient echo sequence the thrombus shows low SI related to the lack of flow or iron-induced signal defasing on these T2*sensitive images. However, turbulent flow and slow flow can be seen as areas of low SI within patent veins. A combination of SE and gradient-echo images are necessary to differentiate ambiguous patterns of intravascular SI from that of thrombus [10]. As one of our aims was to establish in the clinical setting, the possibility of defining the age of the thrombus based in different T1 and T2 relaxation times, we chose to perform the MR examinations with spin-echo sequences with magnitude and phase reconstruction. Spin-echo sequences provide superior contrast of the stationary tissue to evaluate SI of the thrombus on TI, proton density and T2-weighted images. The evaluation of magnitude together with phase reconstruction images is a useful

106

R. Soler et al./European Journal of Radiology 19 (1995) 101-107

technique to differentiate slow flow from thrombus [2,5]. It is assumed that the SI of thrombus is related to age, composition and oxidation state of the hemoglobin iron. However, very few clinical attempts, based on a limited number of patients, have been performed to characterize by MR imaging the age of a thrombus, with unsuccessful results [7,1 l - 13]. In our series the quantitative analysis of thrombus SI showed a wide range of values and we found no statistically significant changes in the SI of the thrombus or the T/F ratio, related to the age of the thrombus. This may be due to the variable contribution of flowing blood that is intimately associated with the organization and recanalization of venous thrombus. On the qualitative analysis of the thrombus we found a correlation between the pattern of the SI (homogeneous vs. heterogeneous) and the age of the thrombus. Thus, a homogeneous pattern was found in 86% of acute thrombus and the only case of acute thrombus with heterogeneous pattern was characterized by an outer hyperintense with a central intermedium signal intensity, whereas a heterogeneous pattern, with hypointense dots within the thrombus was seen only in non-acute thrombus. The heterogeneous pattern that we found in acute thrombus is consistent with a shortened T1 due to methemoglobin formation, similar to the peripheral hyperintense rim on Tl-weighted images described at 3 or 4 days of hematoma evolution [14]. The heterogeneous pattern of non-acute thrombus has been previously reported in an experimental model and was attributed to the recanalization of venous thrombi. Erdman et al. found a less noticeable appearance in clinical thrombi and they suggested that these differences could be related to the greater size and slower resolution of clinical thrombi I1 l]. However in our series, the heterogeneous pattern with hypointense dots within the clot was evident in 75% of the non-acute thrombus and we believe that the medical therapy, in order to accelerate the process of lysis and recanalization of clot, may contribute to the heterogeneous appearance of venous thrombus with time. One of the limitations of this study is the determination of the age of thrombus. We def'me the age of the thrombus clinically by each patient history, so many variations should be considered, but there are no other possible approximations to this topic in a clinical setting. The extent and the morphology patterns of thrombus were best appreciated on Tl-weighted images. As the evaluation of the SI of the thrombus on proton density and T2-weighted images do not give us any additional information, we considered it unnecessary to obtain T2weighted images to study blood thrombus. Patients with documented IVC thrombus can be submitted to medical therapy either for fibrinolytic or anticoagulant therapy. The use of MR imaging in monitoring the effect of these medical therapies has been

previously reported in only one patient with deep venous thrombosis [15], but to our knowledge no study has been performed in a series of patients with IVC thrombosis. The modification in size or complete lysis of IVC thrombus after the anticoagulant or fibrinolytic therapy was correctly identified in all cases of our series. Although the number of patients submitted to the various forms of medical therapy in our series is small, we have found no variation in T1 or T2 relaxation times of the thrombus during treatment that can serve as a parameter to the medical treatment over the thrombus. In conclusion, MR imaging can accurately assess the presence, size and localization of IVC thrombus, similar to the imaging techniques of reference. In addition to the advantages of a noninvasive examination, MR also provides useful information about the morphological variations of the thrombus during medical therapy, as well as the age of the thrombus.

Acknowledgments This study has been supported, in part, by Fondo de Investigaciones Sanitarias de la Seguridad Social (FIS), Instituto Nacional de la Salud (Grant 91/1035). We thank Alvaro Garcla Figueras for his help with statistical analysis.

References [1] Weinreb JC, Mootz A, Cohen JM. MRI evaluation of mediastinal and thoracic inlet venous obstruction. AJR 1986; 146: 679-681. [2] Tavares N J, Auffermann W, Brown J J, Gilbert T J, Sommerhoff C, Higgins ChB. Detection of thrombus by using phase imaging MR scans: ROC curve analysis. AJR 1989; 153: 173-178. [3] Rapoport S, Sostman HD, Pope Ch, Camputaro CM, Holcomb W, Gore JC. Venous clots: evaluation with MR imaging. Radiology 1987; 162: 527-530. [4] Hricak H, Amparo A, Fisher MR, Crooks L, Higgins ChB. Abdominal venous system: assessment using MR. Radiology 1985; 156: 415-422. [5] Von Schulthess GK, Augustiny N. Calculation of T2 values versus phase imaging in the distinction between flow and thrombus in MR imaging. Radiology 1987; 164: 549-554. [6] Park JH, Lee JB, Han MC. Sonographic evaluation of inferior vena cava obstruction. Correlative study with vena cavography. AJR 1985; 145: 752-762. [7] Braun IF, Hoffman JC, Malko JA, Pettigrew RI, Dannels W, Davis PC. Jugular venous thrombosis: MR imaging. Radiology 1985; 157: 357-360. [8] Zerhouni EA, Barth EM, Siegelman SS. Demonstration of venous thrombosis by computed tomography. A JR 1980; 134: 753-758. [9] Rahmouni A, Caseiro-Alves J, Mathieu D, Anglade MC, Golli M, Berger JF, Gheung P, Vasile N. Gradient echo sequences in the diagnosis of inferior vena cava thrombosis. J Radiol 1991; 72: 553-562. [10] Arrivd L, Menu Y, Dessarts I, Bubray B, Vullierme MP, Vilgrain V, Najmark D. Diagnosis of abdominal venous thrombosis by means of spin-echo and gradient-echo MR imaging: analysis with receiver operating characteristic curves. Radiology 1991; 181: 661-668.

R. Soler et al. / European Journal of Radiology 19 (1995) 101-107

[11] Erdman WA, Weinreb JC, Cohen JM, Buja LM, Chancy C, Peshock RM. Venous thrombosis: clinical and experimental MR imaging. Radiology 1986; 161: 233-238. [12] Martin B, Mulopulos GP, Bryan PJ. MRI of puerperal ovarian vein thrombosis. AJR 1986; 147: 291-292. [13] Macchi PJ, Grossman RI, Gomori JM, Goldberg HI, Zimmerman RA, Poilaniuk LT. High field MR imaging of cerebral venous thrombosis. J Comput Assist Tomogr 1986; 10: 10-15.

107

[14] Brooks RA, Di Chiro G, Patronas N. MR imaging of cerebral hematomas at different field strengths: theory and applications. J Comput Assist Tomogr 1989; 13: 194-206. [15] Francis CW, Foster TH, Totterman S, Brenner VJ, Marder VJ, Bryant RG. Monitoring of therapy for deep vein thrombosis using magnetic resonance imaging. Acta Radiol 1989; 30: 445-446.