The accuracy of V/Q SPECT in the diagnosis of pulmonary embolism: a meta-analysis

Abstract

Background

Ventilation perfusion single photon emission computed tomography (V/Q SPECT) and CT pulmonary angiography have all been used in the diagnosis of acute PE. Previous studies have shown higher sensitivity and specificity and a marked decrease in the non-diagnostic rate of V/Q SPECT than planar scan.

Purpose

To systematically review and perform a meta-analysis of published data on the performance of V/Q SPECT in the diagnosis of acute PE.

Material and Methods

A comprehensive computer search was conducted on literature published through 31 December 2013 in an effort to find relevant articles on the diagnostic performance of V/Q SPECT in the diagnosis of PE patients. Pooled sensitivity, specificity, negative likelihood ratio (LR), and positive LR, the area under the receiver-operating characteristic (ROC) curve of V/Q SPECT in the diagnosis of PE patients were calculated.

Results

Nine studies, comprising a total sample size of 3454 patients, were included in our meta-analysis. The pooled sensitivity, specificity of V/Q SPECT in the diagnosis of acute PE patients, calculated on a per-patient-based analysis, was 96% (95% confidence interval [CI], 95–97%), 97% (95% CI, 96–98%). The pooled negative LR, positive LR of V/Q SPECT in acute PE patients was 0.06 (range, 0.02–0.19) and 16.64 (range, 9.78–31.54). The area under the ROC curve of V/Q SPECT in the diagnosis of acute PE patients was 0.99 on a per-patient-based analysis.

Conclusion

V/Q SPECT is an accurate method in acute PE patients with high sensitivity and high specificity in the diagnosis of PE.

Keywords

Pulmonary embolism ventilation perfusion SPECT

Introduction

Acute pulmonary embolism (PE) is a common disease. The incidence of venous thromboembolism is still high, with an estimated 250,000 patients diagnosed each year in the Americas (1). Acute PE is often undiagnosed and is potentially a fatal disease. The mortality of acute PE is between 10% and 30%, if not treated in a timely manner. However, the mortality can be reduced to 2–8% with appropriate treatment (2). Although various clinical symptoms, signs, laboratory tests, predisposing factors can strengthen or weaken the clinical possibility of PE, the diagnosis is challenging even when PE is highly suspected. Ventilation perfusion (V/Q) planar scan, single photon emission computed tomography (SPECT) tomographic scan, and CT pulmonary angiography (CTPA) have been used in the diagnosis of acute PE (3). There is still a debate as to which test is the most accurate or appropriate in this setting. V/Q SPECT can provide cross-sectional images, which overcome the superimposition of lung with normal perfusion, which may conceal PE. Previous studies have showed higher sensitivity, specificity, and marked decrease in the non-diagnostic rate of V/Q SPECT compared to planar scan (4 –7). However, these studies had relative small sample size and limited power for any single study (4 –12). The purpose of this study was to systematically review and perform a meta-analysis of published data on the performance of V/Q SPECT in the diagnosis of acute PE.

Material and Methods

Search strategy

A comprehensive computer literature search of the Pubmed, Embase, and Scopus databases was conducted to find relevant published articles on the diagnostic performance of V/Q SPECT in the diagnosis of acute PE. We used a search algorithm that was based on a combination of the terms: (i) “V/Q” or “ventilation and perfusion scan”; (ii) “SPECT”; and (iii) “pulmonary embolism”. No beginning date limitation was used. The last date of search was 31 December 2013. The search was limited to English language journals. To maximize our search result, references of the retrieved articles were also scrutinized for additional papers.

Study selection

Papers investigating diagnostic performance of V/Q SPECT tomographic scan in the diagnosis of acute PE patients were eligible for inclusion. The exclusion criteria were: (i) case reports or very small case series; (ii) same patient data (such as duplicate publication); (iii) review articles, editorials, letters, author replies, comments, erratum, conference proceedings; (iv) insufficient data to reassess sensitivity or specificity from individual studies; and (v) articles not within the field of our study. CTPA before and after V/Q SPECT scan was not used as an inclusion or exclusion criteria. Three researchers (Y, LL, and JG) independently reviewed the titles and abstracts of the retrieved papers, according to the inclusion and exclusion criteria described above. The same three researchers then independently reviewed the full papers to decide their eligibility for inclusion. Differences in opinions about the papers were resolved in a consensus meeting.

Data extraction from these studies

For every included study, data were collected including basic information (authors), patient information, PE prevalence of the included patients, prospective or retrospective study, sensitivity, specificity, and reference standard. Based on these studies, the number of true positive, true negative, false positive, false negative findings for V/Q SPECT in the diagnosis of acute PE patients were recorded on a per-patient-based analysis.

Quality assessment

Three independent reviewers (Y, LL, and JG) evaluated the methodology of the included papers using the quality assessment of diagnostic accuracy studies (QUADAS) criteria (13,14).

Statistical analysis

Sensitivity and specificity of diagnostic performance of V/Q SPECT in the diagnosis of acute PE patients were obtained from individual studies on a per-patient-based analysis. Pooled sensitivity, specificity, negative likelihood ratio (LR), and positive LR of V/Q SPECT based on the pooled number of true positive, false positive, false negative and true negative. Pooled data were presented with 95% confidence intervals (CI). An I-square statistic was also performed to test for heterogeneity between studies (15).The area under the receiver-operating characteristic (ROC) curve was calculated to measure the accuracy of V/Q SPECT in the diagnosis of acute PE patients. Statistical analyses were performed using Meta-DiSc statistical software version 1.4 (16).

Results

Literature search

After the comprehensive computer literature search from Pubmed, Embase, and Scopus databases, 93 papers were included after the initial analysis. Upon review, 62 papers were excluded from the initial analysis: 21 were reviews, letters, or editorials, and 41 were not within the field of interest of this meta-analysis. Thirty-one papers were selected and retrieved in full text version. No additional papers were found by screening the references of the 31 papers. After reviewing the full text papers, 22 papers were excluded as case reports, duplicate data, or insufficient data from the 31 papers. Finally, nine studies (4 –12) , comprising a total sample size of 3454 patients, met all inclusion and exclusion criteria and were included in our meta-analysis (Fig. 1). All patients were suspicious of PE based on clinical symptoms and signs. V/Q SPECT scan was performed in all patients. The characteristics of the included studies on diagnostic performance of V/Q SPECT in the diagnosis of acute PE patients are presented in Table 1.

Fig. 1.

Flow chart of the search for eligible studies on the diagnostic performance of V/Q SPECT in the diagnosis of PE patients.

Table 1.

Basic study, patient characteristics, and technical aspect of diagnostic performance of V/Q SPECT in the diagnosis of PE patients.

Author	nPs	Prev	Design	Reference	Conclusion	TP	FP	FN	TN
Bajc M	53	25	Prosp	V/P scan, radiography, CTPA, follow-up	No difference	8	3	0	42
Bajc M	1785	34	Retrosp	SPECT, CTPA, follow-up	High diagnostic value	601	24	6	1154
Gutte H	36	31	Prosp	SPECT/CT, CTPA, follow-up	SPECT with higher sen and spe	8	4	0	24
Miles S	87	26	Prosp	V/P scan, radiography, CTPA, follow-up	95% agreement between SPECT and CTPA	19	1	4	63
Le Duc-Pennec A	243	20	Prosp	V/P planar, CTPA, ultrasound	V/P SPECT was diagnostic	45	4	3	191
Weinmann P	95	21	Prosp	CTPA, limb ultrasound	V/P SPECT was diagnostic	56	4	15	20
Reinartz P	83	45	Retrosp	SPECT, CTPA	No significant difference	40	4	1	38
Gutte H	81	38	Prosp	SPECT/CT, CTPA, follow-up	No significant difference	26	6	1	48
Lemb M,	991	ND	Retrosp	V/P SPECT, radiography, follow-up	V/P SPECT was diagnostic	171	7	24	789

ND, non-description; Prev, Prevalence; Prosp, Prospective; Retrosp, Retrospective; Sen, Sensitivity; Spe, Specificity; TP, true positive; FP, false positive; FN, false negative; TN, true negative.

Quality assessment

Using QUADAS criteria, studies were scored between 7 and 13 with a median score of 11. Two out of nine (22.2%) studies scored between 8 and 9 while 77.8% or seven out of nine studies scored 10 or more. Although none of the studies achieved an A rating, four (44.4%) received a B rating, three (33.3%) received a C rating, and two (22.2%) received a D rating. Overall, the methodological quality of the included studies was medium–high.

Diagnostic performance

The performance of V/Q SPECT in the diagnosis of acute PE in the nine included studies are presented in Figs. 2 and 3. The pooled sensitivity of V/Q SPECT in the diagnosis of acute PE patients, calculated on a per-patient-based analysis, was 96% (95% CI, 95–97%). The included studies were statistically heterogeneous in their estimates of sensitivity on a per-patient-based analysis (I-square, 85.8%) (Fig. 2). The pooled specificity of V/Q SPECT in the diagnosis of acute PE patients was 97% (95% CI, 96–98%). The included studies of V/Q SPECT in the diagnosis of acute PE patients were statistically heterogeneous in their estimates of specificity (I-square: 75.4%) (Fig. 3). The pooled negative LR of V/Q SPECT in the diagnosis of PE patients was 0.06 (range, 0.02–0.19). The included studies of V/Q SPECT in the diagnosis of acute PE patients were statistically heterogeneous in their estimates of negative LR (I-square, 91.6%) (Fig. 4). The pooled positive LR of V/Q SPECT in the diagnosis of acute PE patients was 16.64 (range, 9.78–31.54). The included studies of V/Q SPECT in the diagnosis of acute PE patients were statistically heterogeneous in their estimates of positive LR (I-square, 85.1%) (Fig. 5). The area under the ROC curve of V/Q SPECT in the diagnosis of acute PE patients was 0.99 on a per-patient-based analysis (Fig. 6).

Fig. 2.

Plot of individual studies and pooled sensitivity of V/Q SPECT in the diagnosis of PE patients, including 95% CI. The size of circles indicates the weight of each study. The included studies were statistically heterogeneous in their estimates of sensitivity (I-square, 85.8%).

Fig. 3.

Plot of individual studies and pooled specificity of V/Q SPECT in the diagnosis of PE patients, including 95% CI. The size of circles indicates the weight of each study. The included studies were statistically heterogeneous in their estimates of specificity of V/Q SPECT (I-square, 75.4%).

Fig. 4.

The pooled negative LR of V/Q SPECT in PE patients, including 95% CI. The included studies of V/Q SPECT in the diagnosis of PE patients were statistically heterogeneous in their estimates of negative LR (I-square, 91.6%).

Fig. 5.

The pooled positive LR of V/Q SPECT in the diagnosis of acute PE patients, including 95% CI. The included studies of V/Q SPECT in the diagnosis of acute PE patients were statistically heterogeneous in their estimates of negative LR (I-square, 85.1%).

Fig. 6.

Summary ROC curves of diagnostic accuracy of V/Q SPECT in the diagnosis of PE patients, including 95% CI. The area under the ROC curve was 0.99, demonstrating that V/Q SPECT is an accurate method in the diagnosis of PE patients.

Discussion

To the best of our knowledge, this meta-analysis is the first to evaluate the diagnostic performance of V/Q SPECT in the diagnosis of acute PE. Several single-center studies have used V/Q SPECT in the diagnosis of acute PE. However, many of these studies had limited statistical power as only small numbers of patients were included. In order to derive more robust estimates of the diagnostic performance of V/Q SPECT in the diagnosis of acute PE patients, we pooled the published studies. A systematic review process was adopted in ascertaining studies, thereby avoiding selection bias (17).

V/Q scan became the preferred imaging method for the diagnosis of PE between the late 1960s and the early 1990s. However, the rate of low or intermediate probability in PIOPED I of these early V/Q scans was 65% (18). All these factors limited the application of V/Q scan in the diagnosis of PE. From the 1990s, when CTPA became available, clear, anatomic details could be revealed. CTPA has been accepted as the first-line examination in suspected PE patients (19). Therefore, CTPA application in PE increased quickly, especially in the last 20 years due to faster scan times and most important availability 24 hours over 7 days (20). However, the breast exposure from CTPA was in the range of 20–60 mSv (21), which is significantly higher than that of the V/Q scan of 0.28–0.9 mSv (21,22). The radiation dose should be taken into account by the physician, especially in young women and pregnant patients. Stein et al. reported that the number of CT examinations dropped from 62.4% in 2006 to 42% in 2007 of the total imaging studies performed for PE due to the radiation dose (23). CTPA does have contraindications, and cannot be performed in some inpatients (20). Onyedika et al. reported that V/Q scan has re-emerged in the diagnosis of PE due to the above factors and reasons (22).

Although V/Q planar scan has been used in the diagnosis when there is a suspicion of PE in patients, this method has some limitations (24). For example, V/Q planar scans are two dimensional image scans (2D). There is significant overlap of the anatomical segments in 2D images. Embolic defects may not be demonstrated if there is “shine-through” occurring from underlying lung segments with normal perfusion (25). SPECT is widely used in many nuclear medicine scans because of its ability to image in three dimensions (3D). In contrast to V/Q planar imaging, V/Q SPECT is able to image all segments of the lungs and more accurately define the size and location of perfusion defects (25). For these reasons, it would be expected that V/Q SPECT should be superior to planar imaging (24). V/Q SPECT should be the primary imaging test for most patients with suspected PE in clinical practices today (26).

The pooled sensitivity and specificity of V/Q SPECT in the diagnosis of acute PE patients in this meta-analysis, calculated on a per-patient-based analysis, was 96% (95% CI, 95–97%) and 97% (95% CI, 96–98%). The area under the ROC curve of V/Q SPECT in the diagnosis of acute PE patients was 0.99 on a per-patient-based analysis. In head-to-head comparisons of V/P SPECT and planar scan, more and smaller defects were shown on SPECT. The median differences between SPECT and planar V/P scans in sensitivity, specificity, and accuracy were 21%, 6%, and 7%, respectively (4 –6). On the other hand, V/P SPECT is diagnostic in more than 95% of the patients. A substantial part of V/P planar scans are indeterminate (up to 73%) (9). Some authors reported that V/Q SPECT demonstrated less intra-observer (94% vs. 91%) inter-observer (88% vs. 79%) variation compared with V/Q planar scan (4). Considering these factors, V/Q SPECT had a greater sensitivity and specificity, and improved reproducibility compared with planar imaging (24).

The pooled negative LR and positive LR of V/Q SPECT in the diagnosis of acute PE patients in this meta-analysis was 0.06 (range, 0.02–0.19) and 16.64 (range, 9.78–31.54), respectively. The PE prevalence of all these studies in this meta-analysis was in the range of 20–45%. The PE prevalence may correlate with the negative LR and positive LR of V/Q SPECT in the diagnosis of PE. The included studies were statistically heterogeneous in their estimates of sensitivity, specificity, negative LR, and positive LR on a per-patient-based analysis (I-square, 85.8%, 75.4%, 91.6%, and 85.1%, respectively). Many influencing factors may result in heterogeneity of these included studies. Different diagnostic criteria had been used in these studies, such as Le Duc-Pennec et al. who applied PIOPED I as the diagnostic criterium (27), while Weinmann et al. (7) and Miles et al. (11) applied PIOPED II. Le Roux et al. assessed the performance of V/Q SPECT using various criteria for interpretation. This study reported that the best performance was achieved using a diagnostic cutoff of at least one segmental or two subsegmental mismatches of all the tested criteria, with sensitivity and specificity of 0.92 and 0.91, respectively. With a negative V/Q SPECT result, the post-test probability of PE was 0.010, 0.037, and 0.119 for a low, intermediate, and high clinical probability, respectively. With a positive V/Q SPECT result, the posttest probability of PE was 0.531, 0.814, and 0.939 for a low, intermediate, and high probability, respectively. This study concluded that a diagnostic cutoff of one segmental or two subsegmental mismatches was best for confirming or excluding acute PE for V/Q SPECT interpretation. Some studies combined pretest clinical probability with V/Q SPECT. Prevalence of PE was 0 of 41 (0%; 95% CI, 0–9%), six of 134 (4%; 95% CI, 2–9%), 15 of 36 (42%; 95% CI, 27–58%), and 28 of 32 (88%; 95% CI, 72–95%) in the normal, low, intermediate, and high V/Q SPECT probability groups, respectively. Overall, the combination of V/Q SPECT and clinical probability was diagnostic in 220 patients (88.4%): 177 patients (71.1%) had either a normal V/Q SPECT, or the combination of an intermediate V/Q SPECT with a low clinical probability, 174 (98.3%) of whom had no PE. Conversely, 43 patients had either the combination of a high probability V/Q SPECT with an intermediate V/Q high clinical probability, or the combination of an intermediate / SPECT with a high clinical probability, 40 (93.0%) of whom had PE (8). Most studies in this meta-analysis did not combine pretest clinical probability with V/Q SPECT. Some studies in this meta-analysis had a prospective design (4,5,11), while others were retrospective (6,9). The reference standard in all studies included in this meta-analysis was not the same, as, some reference standards included lower limb compression ultrasonography (7,8), while others did not. The V/Q SPECT acquisition protocol of these studies in this meta-analysis was different, such as, a simultaneous V/P SPECT protocol using 81mKr / 99mTc-MAA (10), a sequential V/P SPECT protocol using 99mTc-DTPA radioaerosol / 99mTc-MAA (9), a sequential V/P SPECT protocol using 99mTc-Technegas / 99mTc-MAA (6). In these different protocols, the main difference was the tracer of the ventilation scan, including 81mKr, 99mTc-DTPA radioaerosol, 99mTc-Technegas. 99mTc-DTPA radioaerosol and 99mTc-Technegas are much more widely used because of their greater availability, low cost, and good image quality (24). 99mTc-Technegas is an ideal agent for ventilation SPECT because of its small particle size (30–60 nm), resulting in greater alveolar penetration and less central deposition than 99mTc-DTPA (24).

Finally, based on high sensitivity and high specificity, V/Q SPECT may be considered as suitable in the diagnostic workup of acute PE patients. The Society of Nuclear Medicine guidelines state that the use of SPECT to obtain a three-dimensional evaluation of the lungs is recommended by some investigators (28), whereas the European Association of Nuclear Medicine guidelines state that V/Q SPECT should be the preferred modality whenever possible (29).

In conclusion, V/Q SPECT is an accurate method in the diagnosis of acute PE with high sensitivity and high specificity. Ultimately future prospective studies are needed to assess the accuracy of this technique compared to other imaging techniques, CTPA, SPECT/CT, and pulmonary angiography.

Footnotes

Funding

Jigang Yang was partially supported by Natural Science Foundation of China (No. 81101069), Beijing Health System High Level Technical Personnel (No. 2013-3-066), Beijing Science New Star Cooperation Project (No. xxhz201306).

References

Silverstein

Heit

Mohr

. Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study. Arch Intern Med 1998; 158: 585–593.

Sandler

Martin

. Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep vein thrombosis? J R Soc Med 1989; 82: 203–205.

Soler

Kerr

Marsh

. Pilot study comparing SPECT perfusion scintigraphy with CT pulmonary angiography in chronic thromboembolic pulmonary hypertension. Respirology 2012; 17: 180–184.

Bajc

Olsson

. Diagnostic evaluation of planar and tomographic ventilation/perfusion lung images in patients with suspected pulmonary emboli. Clin Physiol Funct Imaging 2004; 24: 249–256.

Gutte

Mortensen

Jensen

. Comparison of V/Q SPECT and planar V/Q lung scintigraphy in diagnosing acute pulmonary embolism. Nucl Med Commun 2010; 31: 82–86.

Reinartz

Wildberger

Schaefer

. Tomographic imaging in the diagnosis of pulmonary embolism: a comparison between V/Q lung scintigraphy in SPECT technique and multislice spiral CT. J Nucl Med 2004; 45: 1501–1508.

Weinmann

Moretti

J-L

Brauner

. Usefulness of tomographic versus planar lung scintigraphy in suspected pulmonary embolism in a daily practice. Open Med Imaging J 2008; 2: 49–55.

Le Duc-Pennec

Le Roux

Cornily

. Diagnostic accuracy of single-photon emission tomography ventilation/perfusion lung scan in the diagnosis of pulmonary embolism. Chest 2012; 141: 381–387.

Bajc

Olsson

Palmer

. Ventilation/Perfusion SPECT for diagnostics of pulmonary embolism in clinical practice. J Intern Med 2008; 264: 379–387.

10.

Gutte

Mortensen

Jensen

. Detection of pulmonary embolism with combined ventilation-perfusion SPECT and low-dose CT: head-to-head comparison with multidetector CT angiography. J Nucl Med 2009; 50: 1987–1992.

11.

Miles

Rogers

Thomas

. A comparison of single-photon emission CT lung scintigraphy and CT pulmonary angiography for the diagnosis of pulmonary embolism. Chest 2009; 136: 1546–1553.

12.

Lemb

Pohlabeln

. Pulmonary thromboembolism: a retrospective study on the examination of 991 patients by ventilation/perfusion SPECT using Technegas. Nuklearmedizin 2001; 40: 179–186.

13.

Yang

Hao

Zhu

. Diagnostic role of 18F-dihydroxyphenylalanine positron emission tomography in patients with congenital hyperinsulinism: a meta-analysis. Nucl Med Communs 2013; 34: 347–353.

14.

Hao

Yuan

Kan

. Diagnostic performance of 18F-FDG PET/CT in patients with fever of unknown origin: a meta-analysis. Nucl Med Commun 2013; 34: 682–688.

15.

Higgins

Thompson

Deeks

. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560.

16.

Zamora

Abraira

Muriel

. Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6: 31–31.

17.

Bailey

Hutton

Walker

. Improved SPECT using simultaneous emission and transmission tomography. J Nucl Med 1987; 28: 844–851.

18.

PIOPED Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism. Results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). JAMA 1990; 263: 2753–2759.

19.

Perrier

Howarth

Didier

. Performance of helical computed tomography in unselected outpatients with suspected pulmonary embolism. Ann Intern Med 2001; 135: 88–97.

20.

Stein

Fowler

Goodman

. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med 2006; 354: 2317–2327.

21.

Hurwitz

Reiman

Yoshizumi

. Radiation dose from contemporary cardiothoracic multidetector CT protocols with an anthropomorphic female phantom: implications for cancer induction. Radiology 2007; 245: 742–750.

22.

Onyedika

Glaser

Freeman

. Pulmonary embolism: role of ventilation-perfusion scintigraphy. Semin Nucl Med 2013; 43: 82–87.

23.

Stein

Haramati

Chamarthy

. Success of a safe and simple algorithm to reduce use of CT pulmonary angiography in the emergency department. Am J Roentgenol 2010; 194: 392–397.

24.

Roach

Bailey

Harris

. Enhancing lung scintigraphy with single-photon emission computed tomography. Semin Nucl Med 2008; 38: 441–449.

25.

Meignan

. Lung ventilation/perfusion SPECT: the right technique for hard times. J Nucl Med 2002; 43: 648–651.

26.

Roach

. V/Q SPECT: the primary imaging test for suspected pulmonary embolism. BMJ 2013; 347: f5119–f5119.

27.

Gottschalk

Sostman

Coleman

. Ventilation-perfusion scintigraphy in the PIOPED study. Part II. Evaluation of the scintigraphic criteria and interpretations. J Nucl Med 1993; 34: 1119–1126.

28.

Parker

Coleman

Grady

. SNM practice guideline for lung scintigraphy 4.0. J Nucl Med Technol 2012; 40: 57–65.

29.

Bajc

Neilly

Miniati

. EANM guidelines for ventilation/perfusion scintigraphy: Part 1. Pulmonary imaging with ventilation/perfusion single photon emission tomography. Eur J Nucl Med Mol Imaging 2009; 36: 1356–1370.