Clinical value of deep vein thrombosis density on pre-contrast and post-contrast lower-extremity CT for prediction of pulmonary thromboembolism

Abstract

Background

There was a lack of studies assessing the relationship between deep vein thrombosis (DVT) Hounsfield unit (HU) density and pulmonary thromboembolism (PTE).

Purpose

To evaluate the clinical value of DVT density measured on pre- and post-contrast lower-extremity computed tomography (CT) for the prediction of PTE.

Material and Methods

From 2017 to 2021, patients who underwent pulmonary CT angiography within one week after diagnosis of DVT on lower-extremity CT were included in this retrospective study. Then, the patients without PTE were included in “DVT group” and those with both DVT and PTE were included in the “DVT-PTE group.” The DVT HU density was measured by drawing free-hand region of interests (ROIs) within the thrombus at the most proximal filling defect level. A receiver operating characteristic (ROC) analysis was used to evaluate the predictive value of DVT density for the risk of PTE.

Results

This study included a total of 94 patients (DVT group: n=56; DVT-PTE group: m=38). DVT density was significantly higher in the DVT-PTE group than the DVT group in both pre-contrast (53.5 ± 6.2 HU vs. 44.1 ± 7.9 HU; P < 0.001) and post-contrast CT (67.0 ± 8.6 HU vs. 57.1 ± 10.6 HU; P < 0.001). ROC analysis revealed that the area under curve, sensitivity, and specificity for predicting the risk of PTE were 0.739, 71.1%, and 64.2%, respectively, at a DVT density cutoff of 48.2 HU on pre-contrast CT and were 0.779, 73.7%, and 69.6% at a DVT density cutoff of 61.8 HU on post-contrast CT.

Conclusion

The DVT density on both pre- and post-contrast CT could be a predictive factor of PTE.

Keywords

Deep vein thrombosis pulmonary thromboembolism Hounsfield unit density

Introduction

Venous thromboembolism is the third most common cardiovascular disease after heart attack and stroke (1,2). As part of the clinical diagnosis, deep vein thrombosis (DVT) is a major clinical manifestation of venous thromboembolism and pulmonary thromboembolism (PTE) is a major complication of DVT (3 5). DVT and PTE are important causes of morbidity and mortality; therefore, early diagnosis is needed to treat patients more effectively and to reduce mortality (6,7).

For the diagnosis of PTE, pulmonary computed tomography (CT) angiography is a sensitive imaging modality. However, there was concern for overuse of pulmonary CT angiography due to the low positivity rate and unnecessary exposure to radiation (8,9). Therefore, to evaluate the clinical probability for PTE, previous studies suggested several prediction models such as Wells score, Geneva/modified Geneva score, and pulmonary embolism rule-out criteria (10 12). These prediction models are based on the patients’ symptoms and signs; however, patients exhibit a wide range of non-specific signs and symptoms including dyspnea, chest pain, or tachycardia (13). Therefore, the diagnosis of PTE could be delayed at initial presentation.

For the diagnosis of DVT, contrast-enhanced lower-extremity CT is a rapid and accurate diagnostic modality (14). A previous study suggested that DVT density on contrast-enhanced CT venography could be a reliable predictor for acute PTE (15). To minimize the radiation dose, only the venous phase images were usually obtained at contrast-enhanced CT for the diagnosis of DVT. Since not only normal vein density but thrombus density can be also affected by contrast agent (16,17), pre-contrast unenhanced CT may be more suitable for measuring DVT density; however, there was a lack of studies assessing the clinical value of DVT density on pre-contrast CT.

Therefore, the aim of the present study was to evaluate the clinical value of DVT density on both pre- and post-contrast lower-extremity CT and to elucidate the relationship between PTE and DVT density.

Material and Methods

Study population

This single-center retrospective study was approved by our institutional review board and requirements for informed consent were waived.

From January 2017 to December 2021, patients who underwent pulmonary CT angiography within one week after the diagnosis of DVT (involving inferior vena cava, common/external iliac vein, common/superficial femoral vein, or popliteal vein) on two-phase (pre- and post-contrast) lower-extremity CT were eligible for inclusion in this retrospective study. Among them, patients with recurrent DVT, prior history of PTE, iliac vein compression syndrome, metallic implants in lower limbs, malignancy diagnosed within one year, and too small DVT to measure the density of thrombus were excluded. Finally, a total of 94 patients with DVT were included in this study. Then, the patients without PTE were classified into the “DVT group” and those with both DVT and PTE were classified into the “DVT-PTE group” (Fig. 1).

CT protocol

The pre- and post-contrast lower-extremity CT of the study population was performed using a 128-slice CT scanner (Somatom Definition Flash; Siemens Healthcare, Forchheim, Germany). On post-contrast CT, using bolus tracking with a region of interest (ROI) on the ascending aorta after administration of 150 mL of non-ionic contrast media (Omnipaque 300 [Iohexol]; Nycomed Amersham, GE Healthcare, Milwaukee, WI, USA) intravenously at a rate of 5.0 mL/s, the venous phase images were obtained at 150 s after being triggered when the CT attenuation exceeded 100 Hounsfield units (HU). The CT parameters were as follows: tube voltage = 80 kV; tube current = 150–248 mAs (automatically adjusted to each patient's body size); detector collimator = 128 × 0.6 mm; and slice thickness = 2 mm.

Image analysis

The pre- and post-contrast venous phase images of lower-extremity CT of the study population were retrospectively reviewed. The DVT HU density was measured from each pre- and post-contrast axial CT image by the free-hand drawing of ROIs with areas >10 mm² within the thrombus at the most proximal filling defect level. For accurate analysis and comparison, the measurement of DVT density on pre-contrast CT was assisted by obtaining the CT image number of the thrombus on post-contrast CT. The reference unenhanced normal vein HU density was also measured at the contralateral vein on the same pre-contrast axial image where the DVT density was measured (Fig. 2). If the most proximal DVT involved the inferior vena cava, the normal vein density was measured at the inferior vena cava, 8–10 mm superior to the thrombus.

The measurement of DVT density and normal vein density was independently performed by three radiologists (a fourth-year radiology resident, a board-certified musculoskeletal radiologist with seven years of experience, and a board-certified intervention radiologist with 13 years of experience), respectively, who were blinded to any clinical information of patients.

The radiation dose of pre- and post-contrast lower-extremity CT was obtained based on the dose-length product (DLP) recorded in CT images.

Clinical records

Clinical and laboratory data including age, sex, body mass index (BMI), medical history, duration of symptoms due to DVT, and the values of white blood cells, hemoglobin, hematocrit, platelets, neutrophil, C-reactive protein, and D-dimer were retrieved from the patients’ electronic medical charts.

Statistical analysis

Continuous variables are expressed as mean ± standard deviation (SD). Continuous variables were compared using independent t-tests, while categorical variables were compared using chi-square or Fisher’s exact tests.

Inter-observer agreements for normal vein density and DVT density on both pre- and post-contrast CT were evaluated by intraclass correlation coefficients (ICCs). ICC results were interpreted according to the following criteria: ICC < 0.50 = poor; 0.50 < ICC < 0.75 = moderate; 0.75 < ICC < 0.90 = good; and ICC > 0.90 = excellent (18). A receiver operating characteristic (ROC) analysis was conducted to assess the value of DVT density for predicting the risk of PTE, based on the sensitivity, specificity, and area under curve (AUC). The optimal cutoff value was determined to maximize the sum of sensitivity and specificity.

Significant differences between AUCs were assessed by the method described by Hanley and McNeil (19). The AUCs were compared using MedCalc Statistical Software version 19.1.2 (MedCalc Software, Ostend, Belgium). All the other statistical analyses were performed with SPSS 25.0 for Windows (IBM Corp., Armonk, NY, USA). P values <0.05 were considered statistically significant.

Results

Baseline characteristics of study population

The baseline characteristics of the study population are summarized in Table 1. The present study included a total of 94 patients (mean age = 41.4 ± 10.2 years) and the mean DLP of pre- and post-contrast lower-extremity CT examinations was 834.1 ± 138.5 mGy × cm (range = 522.0–1093.0 mGy × cm). Of these 94 patients, 56 patients without PTE were classified into the “DVT group” (mean age = 41.2 ± 9.9 years) and 38 patients with PTE were classified into the “DVT-PTE group” (mean age = 41.7 ± 10.7 years) (Figs. 2 and 3). There were no significant differences in age, sex, BMI, medical history, symptom duration, and laboratory data including D-dimer concentration between the two groups.

Table 1.

Baseline characteristics of the study population.

Characteristic	DVT group	DVT-PTE group	P value	Total
No. of patients	56	38		94
Age (years)	41.2 ± 9.9	41.7 ± 10.7	0.833	41.4 ± 10.2
Sex			0.484
Male	42 (75.0)	26 (68.4)		68 (72.3)
Female	14 (25.0)	12 (31.6)		26 (27.7)
BMI (kg/m²)	25.8 ± 2.8	25.6 ± 2.6	0.807	25.7 ± 2.7
Medical history
Hypertension	10 (17.9)	8 (21.1)	0.699	18 (19.1)
Diabetes mellitus	8 (14.3)	6 (15.8)	0.841	14 (14.9)
Dyslipidemia	7 (12.5)	2 (5.3)	0.242	9 (9.6)
Smoking	39 (69.6)	22 (57.9)	0.242	61 (64.9)
Chronic renal disease	1 (1.8)	1 (2.6)	0.780	2 (2.1)
Pneumonia	3 (5.4)	6 (15.8)	0.092	9 (9.6)
Coronary artery disease	2 (3.6)	2 (5.3)	0.690	4 (4.3)
Long bone fracture	12 (21.4)	9 (23.7)	0.797	21 (22.3)
Malignancy*	1 (1.8)	0 (0.0)	0.408	1 (1.1)
Laboratory data
WBC (×10³/μL)	8.0 ± 1.9	7.9 ± 1.6	0.816	7.9 ± 1.8
Hemoglobin (g/dL)	14.0 ± 2.3	14.0 ± 2.2	0.986	14.0 ± 2.2
Hematocrit (%)	40.3 ± 6.2	40.2 ± 6.0	0.933	40.2 ± 6.1
Platelets (×10³/μL)	260.3 ± 111.9	232.4 ± 75.6	0.154	249.0 ± 99.3
Neutrophil (%)	61.0 ± 10.2	64.9 ± 19.7	0.267	62.6 ± 14.8
CRP (mg/dL)	2.2 ± 2.9	1.8 ± 2.8	0.547	2.0 ± 2.8
D-dimer (μg/mL)	9.9 ± 1.4	10.3 ± 2.2	0.340	10.1 ± 1.8
Duration of symptom (days)	7.5 ± 5.1	7.7 ± 3.7	0.855	7.6 ± 5.0

Values are given as n (%) or mean ± SD.

*Patients with any cancer that has been diagnosed >1 year previously and successfully treated.

BMI, body mass index; CRP, C-reactive protein; DVT, deep vein thrombosis; PTE, pulmonary thromboembolism; SD, standard deviation; WBC, white blood cell.

Comparison of DVT density on pre- and post-contrast CT between the two groups

The inter-observer agreements for DVT density on pre- and post-contrast CT of the three readers in all participants were 0.830 (95% confidence interval [CI] = 0.788–0.866) and 0.921 (95% CI = 0.899–0.939), respectively. The DVT density on pre-contrast CT was significantly higher in patients in the DVT-PTE group than in those in the DVT group (53.5 ± 6.2 HU vs. 44.1 ± 7.9 HU; P < 0.001). In addition, the DVT density on post-contrast CT was also significantly higher in patients in the DVT-PTE group than in those in the DVT group (67.0 ± 8.6 HU vs. 57.1 ± 10.6 HU; P < 0.001) (Fig. 4). However, there was no significant difference in reference normal vein density between the two groups (36.1 ± 7.8 HU vs. 36.6 ± 8.2 HU; P = 0.768). The inter-observer agreement for reference normal vein density on pre-contrast CT was 0.908 (95% CI = 0.881–0.929). Furthermore, there was no significant difference in the ratio of limbs and veins involved by DVT between the two groups (Table 2).

Table 2.

CT findings of the study population and comparison between two groups.

Characteristic	DVT group	DVT-PTE group	P value	Total
No. of patients	56	38		94
Involved limb*			0.505
Right	16 (28.6)	8 (21.1)		24 (25.5)
Left	36 (64.3)	25 (65.8)		61 (64.9)
Bilateral	4 (7.1)	5 (13.2)		9 (9.6)
Involved vein*			0.763
Inferior vena cava	7 (12.5)	3 (7.9)		10 (10.6)
Common or external iliac vein	23 (41.1)	18 (47.4)		41 (43.6)
Common or superficial femoral vein	18 (32.1)	10 (26.3)		28 (29.8)
Popliteal vein	8 (14.3)	7 (18.4)		15 (16.0)
Pre-contrast DVT density (HU)	44.1 ± 7.9	53.5 ± 6.2	<0.001	47.9 ± 8.6
Post-contrast DVT density (HU)	57.1 ± 10.6	67.0 ± 8.6	<0.001	61.1 ± 11.0
Normal vein density (HU)	36.1 ± 7.8	36.6 ± 8.2	0.768	36.3 ± 8.0

Values are given as n (%) or mean ± SD.

*Involved vein was determined by the location of the most proximal DVT. If the most proximal DVT involved the inferior vena cava, the side of the involved limb was categorized by the location of the distal DVT.

CT, computed tomography; DVT, deep vein thrombosis; HU, Hounsfield unit; PTE, pulmonary thromboembolism; SD, standard deviation.

Predictive value of DVT density for pulmonary thromboembolism

A ROC analysis revealed that a DVT density >48.2 HU on pre-contrast lower-extremity CT represented a potential cutoff value for the prediction of PTE, with a sensitivity of 71.1% and a specificity of 64.2% (AUC = 0.739, 95% CI = 0.644–0.834; P < 0.001). Furthermore, DVT density >61.8 HU on post-contrast CT represented a cutoff value for the prediction of PTE, with a sensitivity of 73.7% and a specificity of 69.6% (AUC = 0.779, 95% CI = 0.696–0.862; P < 0.001) (Fig. 5). However, there was no statistically significant difference in AUCs between pre-contrast and post-contrast DVT density (P = 0.167).

Discussion

In the present study, the difference of DVT HU density on pre- and post-contrast lower-extremity CT between the DVT group and DVT-PTE group was evaluated. The results of the present study demonstrated that the DVT-PTE group showed a significantly higher DVT density on both pre- and post-contrast CT than the DVT group, which was in line with the results of a previous study (15). On ROC analysis, the DVT cutoff value for predicting PTE was calculated as 48.2 HU on pre-contrast CT and 61.8 HU on post-contrast CT. This showed that the assessment of DVT density could be an imaging predictor for acute PTE. Furthermore, AUC of ROC analysis revealed post-contrast DVT density (AUC = 0.779, 95% CI = 0.696–0.862) may be more discriminative than pre-contrast DVT density (AUC = 0.739, 95% CI = 0.644–0.834) for the presence of PTE.

Most of the studies regarding the clinical importance of thrombus density were conducted for arterial and venous thromboses of brain CT. These previous studies were focused on the treatment results that a higher thrombus HU could predict successful recanalization in patients with acute ischemic stroke (20,21). In addition, studies evaluating the diagnostic value of thrombus density were mostly performed based on non-contrast brain CT (22 24), while there were few studies evaluating the relationship between DVT density and acute pulmonary embolism. A previous study assessed the relationship between DVT density and acute PTE, but the DVT density was measured only at contrast-enhanced lower-extremity CT (15). The density of a thrombus can be affected by multiple factors including contrast-agent as well as protein, hematocrit contents, and acuity of clots. Furthermore, the fibrin content in a thrombus was associated with contrast uptake (25 27). Therefore, considering the results of previous studies suggesting that higher thrombus density was associated with a higher thrombolysis success rate, we hypothesized that a thrombus with a higher HU was more likely to migrate into the pulmonary circulation, resulting in PTE. Furthermore, considering that thrombus density could be affected by the contrast agent, we hypothesized that measuring DVT density on pre-contrast CT may be more accurate for the determination of thrombus HU density than measuring density on post-contrast CT.

As with a previous study (15), the mean DVT density was significantly higher in the DVT-PTE group than the DVT group on both pre-contrast (53.5 ± 6.2 HU vs. 44.1 ± 7.9 HU) and post-contrast lower-extremity CT (67.0 ± 8.6 HU vs. 57.1 ± 10.6 HU). Though there was no statistically significant difference in AUCs between pre- and post-contrast DVT density (P = 0.167), contrary to our expectation, the DVT density on post-contrast CT showed a superior predictive value for PTE than the DVT density on pre-contrast CT (AUC = 0.779 vs. 0.739). We thought that this result was due to the inability to accurately recognize DVT on pre-contrast CT. On pre-contrast CT, the mean DVT density (47.9 ± 8.6 HU) was higher than normal vein density (36.6 ± 8.0 HU), but the difference was only about 11.3 HU. Though the measurement of DVT density on pre-contrast CT was assisted by reference post-contrast CT images, the accuracy of drawing ROIs on the most proximal DVT was likely to be inferior on pre-contrast CT than post-contrast CT, while DVT could be accurately recognized and measured on post-contrast CT. In addition, the inter-observer agreement was good for pre-contrast DVT density (ICC = 0.830), while the agreements were excellent for post-contrast DVT density (ICC = 0.921) and reference normal vein density (ICC = 0.908)

The present study has some limitations. The first was its retrospective design and inherent selection bias, which could affect the results. Second, the sample size was relatively small in each group. Third, we did not assess the density of DVT in the calf veins. Fourth, DVT HU density was assessed based on free-drawing ROIs manually at selected sites. The manual drawing of ROIs could be inaccurate by operator bias or beam-hardening artifact due to adjacent soft tissue and contrast agent. Fifth, there are many factors including tube voltage of CT, patients’ BMI, heart function, or amount, concentration, and flow rate of contrast agents that can affect the thrombus HU density. Though some previous studies reported that DVT density was not significantly different according to the various CT tube voltages (28,29), the validation of cutoff values in this study needs prospective cohort studies. Finally, although the results of our study demonstrated that post-contrast DVT showed superior predictive value for acute PTE than pre-contrast DVT density, further studies using deep learning are required to improve the accuracy of detecting DVT and measuring DVT density on lower-extremity CT.

In conclusion, DVT HU density on both pre- and post-contrast lower-extremity CT could be a predictive factor for PTE. Furthermore, though the results of our study should be considered within the limitations, post-contrast DVT density showed superior predictive value for PTE than pre-contrast DVT density.

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.

ORCID iDs

Jae Hyeop Jung

Dong Kyu Kim

References

Icli

Aksoy

Turker

, et al. Relationship between mean platelet volume and pulmonary embolism in patients with deep vein thrombosis. Heart Lung Circ 2015;24:1081–1086.

Glynn

Rosner

. Comparison of risk factors for the competing risks of coronary heart disease, stroke, and venous thromboembolism. Am J Epidemiol 2005;162:975–982.

Dickson

. Virchow‘s triad. Br J Haematol 2009;145:433–433.

Huisman

Klok

. Current challenges in diagnostic imaging of venous thromboembolism. Blood 2015;126:2376–2382.

Eberhardt

Raffetto

. Chronic venous insufficiency. Circulation 2005;111:2398–2409.

Giordano

Jansson

Young

, et al. Epidemiology, pathophysiology, stratification, and natural history of pulmonary embolism. Tech Vasc Interv Radiol 2017;20:135–140.

Liu

Xie

Gao

, et al. Predictive value of circulating microRNA-134 levels for early diagnosis of acute pulmonary embolism: meta-analysis. J Cardiovasc Transl Res 2021;14:744–753.

Venkatesh

Kline

Courtney

, et al. Evaluation of pulmonary embolism in the emergency department and consistency with a national quality measure: quantifying the opportunity for improvement. Arch Intern Med 2012;172:1028–1032.

Meng

, et al. Development and validation of a prediction model to estimate risk of acute pulmonary embolism in deep vein thrombosis patients. Sci Rep 2022;12:649.

10.

Wells

Anderson

Rodger

, et al. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thromb Haemost 2000;83:416–420.

11.

Le Gal

Righini

Roy

, et al. Prediction of pulmonary embolism in the emergency department: the revised Geneva score. Ann Intern Med 2006;144:165–171.

12.

Penaloza

Soulié

Moumneh

, et al. Pulmonary embolism rule-out criteria (PERC) rule in European patients with low implicit clinical probability (PERCEPIC): a multicentre, prospective, observational study. Lancet Haematol 2017;4:e615–e621.

13.

Agnelli

Becattini

. Acute pulmonary embolism. N Engl J Med 2010;363:266–274.

14.

Cham

Yankelevitz

Shaham

, et al. Deep venous thrombosis: detection by using indirect CT venography. Radiology 2000;216:744–751.

15.

Jeong

Kwon

Noh

, et al. Relationship of lower-extremity deep venous thrombosis density at CT venography to acute pulmonary embolism and the risk of postthrombotic syndrome. Radiology 2019;293:687–694.

16.

Tumma

Feitell

Han

, et al. Thrombus can enhance on delayed enhancement imaging. J Cardiovasc Magn Reson 2014;16:P97.

17.

Donato

Ganame

Bogaert

. Contrast-enhancing left ventricular apical thrombus. Eur Heart J 2009;30:1977.

18.

Bland

Altman

. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;327:307–310.

19.

Hanley

McNeil

. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839–843.

20.

Mokin

Morr

Natarajan

, et al. Thrombus density predicts successful recanalization with Solitaire stent retriever thrombectomy in acute ischemic stroke. J Neurointerv Surg 2015;7:104–107.

21.

Puig

Pedraza

Demchuk

, et al. Quantification of thrombus Hounsfield units on noncontrast CT predicts stroke subtype and early recanalization after intravenous recombinant tissue plasminogen activator. AJNR Am J Neuroradiol 2012;33:90–96.

22.

Tayyebi

Akhavan

Shams

, et al. Diagnostic value of non-contrast brain computed tomography in the evaluation of acute cerebral venous thrombosis. Sci Rep 2020;10:883.

23.

Alharbi

Alahmadi

. The diagnostic utility of unenhanced computed tomography of the brain and D-dimer levels in acute cerebral venous sinus thrombosis: A quantitative study. J Clin Imaging Sci 2022;12:15.

24.

Buyck

De Keyzer

Vanneste

, et al. CT density measurement and H: h ratio are useful in diagnosing acute cerebral venous sinus thrombosis. AJNR Am J Neuroradiol 2013;34:1568–1572.

25.

Kirchhof

Welzel

Mecke

, et al. Differentiation of white, mixed, and red thrombi: value of CT in estimation of the prognosis of thrombolysis—phantom study. Radiology 2003;228:126–130.

26.

Borggrefe

Kottlors

Mirza

, et al. Differentiation of clot composition using conventional and dual-energy computed tomography. Clin Neuroradiol 2018;28:515–522.

27.

Qazi

Wilson

, et al. Identifying thrombus on non-contrast CT in patients with acute ischemic stroke. Diagnostics 2021;11:1919.

28.

Cho

Chung

Kim

, et al. CT venography for deep vein thrombosis using a low tube voltage (100 kVp) setting could increase venous enhancement and reduce the amount of administered iodine. Korean J Radiol 2013;14:183–193.

29.

Sasaki

Fujimoto

Ishitoya

, et al. Improved detectability of thromboses of the lower limb using low kilovoltage computed tomography. Medicine (Baltimore) 2018;97:e9775.