Prediction of venous malformations with localized intravascular coagulopathy with diffusion-weighted magnetic resonance imaging

Abstract

Background

Venous malformations may be complicated by localized intravascular coagulopathy which is a serious condition with hematological sequel. Prediction of localized intravascular coagulopathy is mandatory for prompt anticoagulation therapy. Laboratory and routine magnetic resonance imaging can predict localized intravascular coagulopathy in venous malformations; however, the results are variable.

Purpose

To predict venous malformations with localized intravascular coagulopathy with diffusion-weighted magnetic resonance imaging.

Material and methods

A retrospective analysis was performed on 55 patients (34 male, 21 female aged 14–64 years: mean 39 years) with venous malformations that underwent diffusion-weighted magnetic resonance imaging. The apparent diffusion coefficient value of venous malformations was calculated.

Results

The mean apparent diffusion coefficient value of venous malformations with localized intravascular coagulopathy (n = 26) (1.28 ± 0.18 × 10⁻³mm²/s) was significantly different (P = 0.001) from venous malformations without localized intravascular coagulopathy (n = 29) (1.60 ± 0.18 × 10⁻³mm²/s). When apparent diffusion coefficient value of 1.454 × 10⁻³mm²/s was used as a threshold value for the prediction of venous malformations with localized intravascular coagulopathy, the best result was obtained with an accuracy of 83.6%, sensitivity of 84.6%, specificity of 82.8%, and area under the curve of 0.895. The apparent diffusion coefficient value of venous malformations was correlated with D-dimer level (r = −0.59, P = 0.006) and fibrinogen level (r = 0.73, P = 0.001).

Conclusion

The apparent diffusion coefficient value is a non-invasive imaging parameter that can be used to predict venous malformations with localized intravascular coagulopathy.

Keywords

Diffusion venous malformations localized intravascular coagulopathy

Introduction

Venous malformations (VMs) are the most common type of congenital vascular malformation with an incidence of 1 to 2 in 10,000 and commonly seen in the head, neck, limbs, and trunk.^1–7 Routine T2-weighted magnetic resonance (MR) imaging can determine the extent of VMs and to detect the relation with the surrounding structures. The use of gadolinium contrast enables the differentiation of VMs from lymphatic malformations and arterio-venous malformations.^8,9 Contrast MR Angiography is used for the differentiation of VMs from high flow vascular malformations and classification of VMs.^10–13 The CT angiography has limited role in evaluation of VMs and associated with radiation exposure.^14,15

Localized intravascular coagulopathy (LIC) is a common finding (42–88%) in large VMs that activate the coagulation leading to a consumption of clotting factors. It may be associated with bleeding, deep venous thrombosis, and pulmonary embolism. Patients with LIC clinically manifest with thrombotic and bleeding problems responsible for swelling, severe pain, functional impairment, and marked consumption of platelets, coagulation factors and fibrinogen. Early diagnosis of LIC and prompt anticoagulation treatment with adjunctive therapies has been used to prevent the progression of LIC to disseminated intravascular coagulation in case of trauma or surgery and thromboembolic disease resulting in death in some cases. The diagnosis of LIC is made by a combination of high fibrin degradation products, such as D-dimers and decreased fibrinogen levels. The sensitivity of laboratory tests is considerably lower (43%) as smaller or solitary VMs do not result in enough clotting to become evident as systemic elevations in D-dimer levels.^5–7 Routine MR imaging may suggest presence of LIC in VMs in presence of large lesion size, truncal location, intramuscular lesions, lesions involving the bone and viscera, spongiform morphology, and the presence of phleboliths; however, these findings at routine MR are subjective, qualitative, overlapping, and varies between different radiologists.^8,9 Therefore, there is an urgent need for quantitative non-invasive parameter that can suspect LIC in VMs.

Diffusion-weighted MR imaging is based on the free diffusion of water molecules due to the Brownian movement, which defines that any molecules in liquid or gaseous compounds tend to move in a random and probabilistic way in any space direction. Diffusion-weighted echo-planar MR imaging can provide better characterization of soft tissues because it reflects the random motion of water protons, which is disturbed by intracellular organelles and macromolecules located in the tissues.^16–20 The extent of translational diffusion of molecules measured in the human body is referred to as the apparent diffusion coefficient. The apparent diffusion coefficient is expected to vary according to the cellular density of the lesion. Diffusion-weighted MR imaging has been widely used in many clinical settings; for evaluation of head and neck cancer, thyroiditis, nasal tumors,^18–20 soft tissue masses and pediatric lymphatic, and VMs.^21–25 Recently, diffusion-weighted MR imaging used for differentiation orbital VM from simulating lesions in the orbit.²⁵ To our knowledge, the relationship between apparent diffusion coefficient of VMs and presence of LIC has not been established.

We hypothesized that VMs with LIC will behave differently on diffusion-weighted MR imaging from those without LIC. The aim of this work is to predict VMs with LIC with diffusion-weighted MR imaging.

Material and methods

Patients

The study was approved by institutional review board and informed consent from patients was waived because this a retrospective study. Retrospective analysis of 62 patients with VMs that underwent routine and diffusion-weighted MR imaging in the period from March 2005 to August 2017 were performed. We excluded 7 patients that undergone therapeutic intervention and 2 patients with bad image quality with susceptibility artifacts in a small lesion less than 1 cm. The study involved 55 patients (34 males and 21 females aged from 12 years to 64 years, mean 39 years) that presented with chronic pain (n = 37), cosmetic disfigurement (n = 30), and bluish discoloration (n = 25). Laboratory tests included D-dimer and fibrinogen level was collected. The VMs with LIC was diagnosed when D-dimer exceeded 1000 ng/mL and/or fibrinogen was less than 200 mg/dL.

MR imaging

MR imaging was performed for all patients on a 1.5 T MR machine (symphony; Siemens Medical systems, Erlangen, Germany) using a head circular polarization surface coil. All patients underwent T1-weighted images (TR/TE of 800/15 ms) and T2-weighted images (TR/TE of 4500/80 ms) with a section thickness of 5 mm, an interslice gap of 1–2 mm, a field of view (FOV) of 20 × 25 cm, and an acquisition matrix of 256 × 256. Diffusion-weighted MR images were obtained using a multislice spin echo single shot echo planar imaging sequence. Imaging parameters were; TR/TE of 10,000/108 ms, FOV of 20 × 25 cm, an acquisition matrix of 256 × 128, and section thickness of 5 mm with an interslice gap of 1–2 mm. Diffusion-weighted MR images were acquired with diffusion weighted factor, factor b of 0, 500, and 1000 s/mm² and an apparent diffusion coefficient maps were generated for all images. The data acquisition time for the diffusion-weighted images was 1 min. Finally enhanced T1-weighted images (TR/TE of 800/15 ms) were obtained after an intravenous bolus injection of 0.2 ml/kg of body weight of gadopentate, dimeglumine.

Image analysis

Analysis of MR images was performed by one radiologist expert in MR imaging since 15 years (AG) who was blinded to the patient data and the final diagnosis. On apparent diffusion coefficient maps, a region of interest (ROI) was placed within the inner margin of the VMs using an electronic cursor. A quantitative analysis of the apparent diffusion coefficient map was made and the apparent diffusion coefficient values were calculated.

Statistical analysis

The statistical analysis of data was done by using Excel program and SPSS program (Statistical Package for Social Science version 22). The analysis of data was done to test statistical significant difference. To compare between two groups student t-test was used. The receiver operating characteristic (ROC) curve was done to determine the cutoff point of apparent diffusion coefficient used to differentiate VMs with LIC and those without LIC. Pearson correlation was used to correlate the apparent diffusion coefficient with laboratory tests. The P value was considered significant if ≤ 0.05 at confidence interval 95%.

Results

The VMs were classified according to level of D-dimer and fibrinogen levels into VMs without LIC (n = 29) and VMs with LIC (n = 26). The site of VMs was head and neck (n = 25), extremity (n = 23), and trunk (n = 8).

The mean apparent diffusion coefficient value of VMs was 1.45 ± 0.24 (1.02–1.82) ×10⁻³mm²/s. The mean apparent diffusion coefficient value of VMs with LIC (Figure 1) was 1.28 ± 0.18 (1.02–1.60) ×10⁻³mm²/s and of VMs without LIC was 1.60 ± 0.18 (1.06–1.82) ×10⁻³mm²/s. There was significant difference in the apparent diffusion coefficient values of both groups (P = 0.001) (Table 1). When apparent diffusion coefficient of 1.45 ×10⁻³mm²/s was used to differentiate VMs without LIC from VMs with LIC revealed an accuracy of 83.6%, sensitivity of 84.6%, specificity of 82.8%, positive predictive value of 81.5%, negative predictive value of 85.6%, and an area under the curve of 0.895 (Figure 2).

Figure 1.

Venous malformation with LIC: (a) axial T2-weighted image shows intramuscular hyperintense lesion is seen in the right masseter muscle with signal void regions of phleboliths (arrow). (b) Axial contrast T1-weighted image shows intense enhancement of the lesion with signal void region of phleboliths (arrow). (c) Apparent diffusion coefficient map shows the lesion with low apparent diffusion coefficient value (1.21 × 10⁻³mm²/s).

Table 1.

Mean, standard deviation, minimum, and maximum apparent diffusion coefficient (×10⁻³mm²/s) of VMs with and without LIC.

	VM without LIC (n = 29)	VM with LIC (n = 26)	P value
Apparent diffusion coefficient	1.60 ± 0.18 (1.06–1.82)	1.28 ± 0.18 (1.02–1.6)	0.001

VM: venous malformation; LIC: localized intravascular coagulopathy.

Figure 2.

Receiver operating characteristic (ROC) curve: the vertical axis represents the sensitivity and the horizontal axis represents the specificity, the blue line represents the diagnostic performance of the selected cut off apparent diffusion coefficient value while the green line represents reference line i.e. zero sensitivity and zero specificity. Selection of 1.45 × 10⁻³mm²/s as a threshold value that used for prediction of VMs with LIC revealed an accuracy of 83.6%, sensitivity of 84.6%, specificity of 82.8%, positive predictive value of 81.5%, negative predictive value of 85.6% and an under the curve of 0.895.

The mean values of D-dimer and fibrinogen levels in VMs with LIC were 1973 (1080–3130) ng/mL and 154.5 (69–190) mg/dL and in VMs without LIC were 593 (105–954) ng/mL and 347 (234–564) mg/dL respectively (Table 2). There was significant difference in D-dimer and fibrinogen levels between VMs with LIC and VMs without LIC (P = 0.001 both). There was significant negative correlation of the apparent diffusion coefficient value and D dimer level (r = −0.59, P = 0.006) and positive correlation of apparent diffusion coefficient value and fibrinogen level (r = 0.73, P = 0.001) (Table 3).

Table 2.

Median, minimum, and maximum fibrinogen and D-dimer level in VMs with LIC and without LIC.

	VM without LIC (n = 29)	VM with LIC (n = 26)	P value
Fibrinogen level	347 (234–564)	154.5 (69–190)	0.001
D-dimer level	593 (105–954)	1973 (1080–3130)	0.001

VM: venous malformation; LIC: localized intravascular coagulopathy.

Table 3.

Correlation of apparent diffusion coefficient value with D-dimer and fibrinogen levels.

	Apparent diffusion coefficient
	R	P value
Fibrinogen	0.733	0.001
D-dimer	−0.59	0.006

Discussion

The main finding of this study is that apparent diffusion is an non-invasive quantitative imaging parameter that can be incorporated to the routine pre- and post contrast MR imaging of VMs. The addition of diffusion-weighted MR imaging into routine MR imaging in the evaluation of VMs may provide new information about bleeding tendency of VMs and may predict bleeding tendency with LIC without contrast medium injection within short examination time.

Diffusion-weighted MR imaging adds another new dimension to the conventional MR imaging sequences owing to its ability to act as a potential marker for tissue component. Combining this signal intensity measurement with differing b-values of diffusion-weighted MR imaging, the apparent diffusion coefficient is calculated. The apparent diffusion coefficient value is able to look beyond the anatomy and give physiological parameter of the lesion. The apparent diffusion coefficient value estimates the amount and speed of proton movement within the tissue. Diffusivity is affected by cell size, density, and integrity.

In this study, the apparent diffusion coefficient value of VMs is 1.45 ± 0.24×10⁻³mm²/s. One study reported that the apparent diffusion coefficient value of VMs was 1.63 ± 0.03 × 10⁻³mm²/s²² and another study added that the mean apparent diffusion coefficient of distensible orbital VMs was 2.80 ± 0.48 ×10⁻³mm²/s.²⁵ The difference in apparent diffusion coefficient value between different studies may be attributed to different applied B value, different parameters of the diffusion-weighted MR imaging as well as different composition and regions of VMs. One study reported that vascular lesions such as hemangiomas and VMs revealed high apparent diffusion coefficient value than other benign solid tumors such as neurofibromas.²³ This may be attributed to VMs being composed of dilated and dysfunctional venous channels that are deficient in smooth muscle cells. These slow-flow venous sacs progressively expand with stagnation of venous blood.

In this work, the mean apparent diffusion coefficient value of VMs with LIC is significantly lower (P = 0.001) than that of VMs without LIC lesions and the apparent diffusion coefficient value can predict VMs with LIC. This is explained by the difference in pathologic features of both types of VMs. The difference in the apparent diffusion coefficient in the VMs with LIC may be attributed to VMs with LIC having slow–flow circulation, micro-thrombi, phleboliths, and hemorrhage within the lesions that reduce the diffusion space of water protons in the extracellular and intracellular dimensions with a resultant decrease in apparent diffusion coefficient.⁴

In this work, there is an overlap of the apparent diffusion coefficient values between VMs with LIC and VMs without LIC. One patient with VM without LIC exhibited low apparent diffusion coefficient value on apparent diffusion coefficient map. This is attributed to excess fibrous tissue within the lesion with subsequent restriction of diffusion. Two patients with VMs with LIC exhibited high apparent diffusion coefficient value due to small amount of fibrous tissue and absence of phleboliths within the lesion.

In this study, there is a significant difference in the D-dimer and fibrinogen levels between VMs with LIC and those without LIC and the apparent diffusion coefficient value of VMs well correlated with the D-dimer and fibrinogen level. Elevated D-dimer levels are the hallmark of hypercoagulable/prothrombotic state of VMs due to the venous stasis in numerous dilated abnormal venous channels. The elevated D-dimer level is due to activation of coagulation with blood stagnation in the enlarged venous channels. Patients with LIC are associated with generation of large amount of thrombin and monomers of fibrin, consumption of fibrinogen, and some clotting factors and increased risk of bleeding.^5–7

The advantages of diffusion-weighted MR imaging are: it is non-invasive, rapid, and the duration of scan is short without radiation exposure or injection of intravenous contrast material. This can add new physiological information rendering it easy to be incorporated into a standard routine T2-weighted MR images and contrast study imaging protocol of VMs without extra-cost. Therefore, both conventional morphologic and physiological assessment of VMs can be done during the same examination.^16–21

There are few limitations of this study. First, this study is limited due to the small number of patients. Further multicenter studies upon large number of patients are recommended. Second, this study applied diffusion-weighted MR imaging. Further studies with application of multi-parametric MR imaging of diffusion tensor imaging,²⁶ dynamic contrast MR imaging,²⁷ and proton MR spectroscopy²⁸ at higher 3 T scanner^29–31 will improve the results in the future.

Conclusion

We concluded that the apparent diffusion coefficient value is a non-invasive imaging parameter that can be used to predict VMs with LIC.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

Obtained

Guarantor

Abdel Razek A

Contributorship

N/A

Acknowledgements

N/A

References

Ali

Mitchell

. Outcomes of venous malformation sclerotherapy: a review of study methodology and long-term results. Semin Interv Radiol 2017; 34: 288–293.

Colletti

Ierardi

. Understanding venous malformations of the head and neck: a comprehensive insight. Med Oncol 2017; 34: 42–42.

Razek

. Vascular neurocutaneous disorders: neurospinal and craniofacial imaging findings. Jpn J Radiol 2014; 32: 519–528.

Dompmartin

Vikkula

Boon

. Venous malformation: update on aetiopathogenesis, diagnosis and management. Phlebology 2010; 25: 224–235.

Yasumoto

Ishiura

Narushima

et al.

Successful treatment with dabigatran for consumptive coagulopathy associated with extensive vascular malformations. Blood Coagul Fibrinolysis 2017; 28: 670–674.

Zhuo

Russell

Wargon

et al.

Localized intravascular coagulation complicating venous malformations in children: associations and therapeutic options. J Paediatr Child Health 2017; 53: 737–741.

Hung

Leung

Liu

et al.

Venous malformation and localized intravascular coagulopathy in children. Eur J Pediatr Surg 2017; 27: 181–184.

Koo

Dowd

Mathes

et al.

MRI phenotypes of localized intravascular coagulopathy in venous malformations. Pediatr Radiol 2015; 45: 1690–1695.

Dompmartin

Acher

Thibon

et al.

Association of localized intravascular coagulopathy with venous malformations. Arch Dermatol 2008; 144: 873–877.

10.

Abdel Razek

Ashmalla

Samir

. Clinical value of classification of venous malformations with contrast enhanced MR angiography. Phlebology 2017; 32: 628–633.

11.

Razek

Gaballa

Megahed

et al.

Time resolved imaging of contrast kinetics (TRICKS) MR angiography of arteriovenous malformations of head and neck. Eur J Radiol 2013; 82: 1885–1891.

12.

Lee

Suh

. Visualization of soft tissue venous malformations of head and neck with 4D flow magnetic resonance imaging. Neurointervention 2017; 12: 110–115.

13.

Razek

Saad

Soliman

et al.

Assessment of vascular disorders of the upper extremity with contrast-enhanced magnetic resonance angiography: pictorial review. Jpn J Radiol 2010; 28: 87–94.

14.

Henzler

Vogler

Lange

et al.

Low dose time-resolved CT-angiography in pediatric patients with venous malformations using 3rd generation dual-source CT: initial experience. Eur J Radiol Open 2016; 3: 216–222.

15.

Abdel Razek

Denewer

Hegazy

et al.

Role of computed tomography angiography in the diagnosis of vascular stenosis in head and neck microvascular free flap reconstruction. Int J Oral Maxillofac Surg 2014; 43: 811–815.

16.

Abdel Razek

Gaballa

Denewer

et al.

Diffusion weighted MR imaging of the breast. Acad Radiol 2010; 17: 382–386.

17.

Connolly

Srinivasan

. Diffusion-weighted imaging in head and neck cancer technique, limitations, and applications. Magn Reson Imaging Clin North Am 2018; 26: 121–133.

18.

Razek

Sieza

Maha

. Assessment of nasal and paranasal sinus masses by diffusion-weighted MR imaging. J Neuroradiol 2009; 36: 206–211.

19.

Abdel Razek

Mossad

Ghonim

. Role of diffusion-weighted MR imaging in assessing malignant versus benign skull-base lesions. Radiol Med 2011; 116: 125–132.

20.

Abdel Razek

Samir

El-Said

. Role of diffusion-weighted MR imaging in differentiation of Graves’ disease from painless thyroiditis. Polish J Radiol 2017; 28: 536–541.

21.

Surov

Nagata

Razek

et al.

Comparison of ADC values in different malignancies of the skeletal musculature: a multicentric analysis. Skeletal Radiol 2015; 44: 995–1000.

22.

Razek

Nada

Ghaniem

et al.

Assessment of soft tissue tumours of the extremities with diffusion echoplanar MR imaging. Radiol Med 2012; 117: 96–101.

23.

Abdel Razek

Gaballa

Elhawarey

et al.

Characterization of pediatric head and neck masses with diffusion-weighted MR imaging. Eur Radiol 2009; 19: 201–208.

24.

Sauer

Müller

Pabst

et al.

Diffusion weighted MRI for diagnosis and differential diagnosis of lymphatic malformations in paediatric patients. Cancer Res Frontiers 2016; 2: 300–310.

25.

Kalin-Hajdu

Colby

Idowu

et al.

Diagnosing distensible venous malformations of the orbit with diffusion-weighted magnetic resonance imaging. Am J Ophthalmol 2018. DOI: 10.1016/j.ajo.2018.02.005.

26.

Razek

Shabana

El Saied

et al.

Diffusion tensor imaging of mild-moderate carpal tunnel syndrome: correlation with nerve conduction study and clinical tests. Clin Rheumatol 2017; 36: 2319–2324.

27.

Abdel Razek

Gaballa

. Role of perfusion magnetic resonance imaging in cervical lymphadenopathy. J Comput Assist Tomogr 2011; 35: 21–25.

28.

Razek

Nada

. Correlation of choline/creatine and apparent diffusion coefficient values with the prognostic parameters of head and neck squamous cell carcinoma. NMR Biomed 2016; 29: 483–489.

29.

Abdel Razek

Elkhamary

Al-Mesfer

et al.

Correlation of apparent diffusion coefficient at 3T with prognostic parameters of retinoblastoma. AJNR Am J Neuroradiol 2012; 33: 944–948.

30.

Razek

AAKA

Ashmalla

. Assessment of paraspinal neurogenic tumors with diffusion-weighted MR imaging. Eur Spine J 2017. DOI: 10.1007/s00586-017-5265-6.

31.

Razek

AAKA

. Assessment of masses of the external ear with diffusion-weighted MR imaging. Otol Neurotol 2018; 39: 227–231.