INTRODUCTIONThe diagnosis of pulmonary thromboembolism remains a fre-

quent clinical dilemma although this is a common and potentially

life-threatening condition. Ventilation–perfusion scintigraphy

(VQS) is the most frequently used imaging investigation but it

affords only a probability of embolism. The Prospective

Investigation of Pulmonary Embolism Diagnosis (PIOPED)

investigators showed that a normal VQS study excluded

pulmonary embolism in 96% of cases, although unfortunately

only 25% of patients fall into this category. Approximately 50% of

VQS studies yield a result of low or intermediate probability for

embolus. Patients with intermediate-probability VQ scintigrams

who proceed to pulmonary angiography have a positive

diagnosis of embolism in 30% of cases;1 this figure increases to

66% when only those with a high pretest clinical probability of

embolism are considered. Ferretti et al. have demonstrated

DiagnosticRadiology

Subsequent investigation and management ofpatients with intermediate-category and -probability ventilation–perfusion scintigraphyGeraldine Walsh and D Neil JonesDivision of Medical Imaging, Flinders Medical Centre, Bedford Park, South Australia, Australia

SUMMARY

The authors wished to determine the proportion of patients with intermediate-category and intermediate-probabilityventilation–perfusion scintigraphy (IVQS) who proceed to further imaging for investigation of thromboembolism, toidentify the defining clinical parameters and to determine the proportion of patients who have a definite imagingdiagnosis of thromboembolism prior to discharge from hospital on anticoagulation therapy. One hundred and twelveVQS studies performed at the Flinders Medical Centre over a 9-month period were reported as having intermediatecategory and probability for pulmonary embolism. Medical case notes were available for review in 99 of these patientsand from these the pretest clinical probability, subsequent patient progress and treatment were recorded. Eight caseswere excluded because they were already receiving anticoagulation therapy. In the remaining 91 patients the pretestclinical probability was considered to be low in 25; intermediate in 30; and high in 36 cases. In total, 51.6% (n = 47) ofthese patients (8% (n = 2) with low, 66% (n = 20) with intermediate, and 69.4% (n = 25) with high pretest probability)proceeded to CT pulmonary angiography (CTPA) and/or lower limb duplex Doppler ultrasound (DUS) evaluation. Ofthe patients with IVQS results, 30.7% (n = 28) were evaluated with CTPA. No patient with a low, all patients with ahigh and 46% of patients with an intermediate pretest probability initially received anticoagulation therapy. This wasdiscontinued in three patients with high and in 12 patients with intermediate clinical probability prior to dischargefrom hospital. Overall, 40% of patients discharged on anticoagulation therapy (including 39% of those with a highpretest probability) had a positive imaging diagnosis of thromboembolism The results suggest that, although themajority of patients with intermediate-to-high pretest probability and IVQS proceed to further imaging investigation,CTPA is relatively underused in this group. Most patients with a high pretest clinical probability receiveanticoagulation therapy irrespective of imaging findings, and less than half of all patients discharged from hospital onanticoagulation therapy have a positive imaging diagnosis of thromboembolism.

Key words: intermediate; investigation; management; ventilation–perfusion scintigraphy.

G Walsh FRCR; DN Jones FRANZCR.

Correspondence: Dr G Walsh, Department of Radiology, Sunnybanks Private Hospital, Level 1, 245 McCullough Street, Sunnybank, Brisbane, Qld

4109, Australia.

Submitted 12 April 1999; resubmitted 16 November 1999; accepted 8 March 2000.

Australasian Radiology (2000) 44, 424–427

g gy

recently that patients with intermediate-probability VQS and

normal lower limb duplex Doppler ultrasound studies (DUS)

have emboli diagnosed on CT pulmonary angiography (CTPA)

in approximately 24% of cases.2

A retrospective study by Schulgar et al. in 1988 showed

physician behaviour at their institution to be at variance with the

recommended practice of proceeding to pulmonary angio-

graphy in patients with intermediate-probability VQ scinti-

graphy.3 They found that this group of patients was treated on

the basis of clinical probability most probably because of

physician concerns about the risks of contrast reactions.

A further review at the same institution in 1991, in the wake of

the PIOPED study (which demonstrated the relative safety of

conventional pulmonary angiography), showed an actual

decrease in physician use of angiography but an increase in the

use of further non-invasive studies, particularly duplex Doppler

evaluation of lower limb veins.4

Khorasani et al. recently reported that most patients at their

institution with intermediate-probability VQS in 1991–1992 did

not proceed to further imaging and were treated without a

definitive imaging diagnosis.5 Schulgar et al., Henschke et al. and

Khorasani et al. reported physician practice at major US teaching

institutions in the late 1980s/early 1990s at which conventional

pulmonary angiography was readily available.3–5 In the past 5

years much progress has been made in non-invasive direct

visualization of pulmonary embolus, most notably with the advent

of CTPA,6,7 but also with MR pulmonary angiography (MRPA).8

We wished to determine present practice at Flinders Medical

Centre, South Australia, where conventional angiography is

rarely requested but where CTPA is readily available and its

appropriate use was brought to the attention of referring

clinicians throughout the period of the present review.

PurposeThe purpose of the study was to determine the proportion of

patients with intermediate-probability VQS who proceed to

further imaging investigations, to identify which clinical

parameters are most important in determining the need for

further investigation and/or treatment and to assess the

percentage who have a definite imaging diagnosis of thrombo-

embolism prior to discharge from hospital on anticoagulation

therapy.

METHODSBetween July 1997 and March 1998, 380 patients had VQS

performed at Flinders Medical Centre, South Australia for

suspected acute pulmonary embolism. One hundred and

twelve of these studies were reported as intermediate in

probability for pulmonary embolism by one of four nuclear

medicine readers based on revised PIOPED criteria.9 The

clinical case notes were available for review in 99 of these

patients and, from these, the pretest clinical probability together

with patient progress and treatment following VQS studies were

recorded. Plasma D dimer measurements were not performed

routinely.10 Pretest probability was usually indicated in the

clinical admission summary and was interpreted as low when

the diagnosis of pulmonary embolism was considered unlikely

but required exclusion; as intermediate when both pulmonary

embolism and an alternative diagnosis were considered to be

equally likely based on clinical criteria; and as high when

pulmonary embolism was the most probable clinical diagnosis.

A more structured approach as recently described by Scott et

aI.11 was not employed during the period of this review.

Ventilation–perfusion scintigraphyVentilation perfusion scintigraphy was performed following

administration of 40 MBq of technetium-99m Technegas (Tetley

Technologies/Vitamedical, Sydney, NSW, Australia) and

200 MBq of technetium-99m-labelled macroaggregates of

albumin (MAA; Mallinckrodt, St Louis, MO, USA) intravenously.

The ventilation images were acquired first, followed by the

perfusion study. Scintigraphic images (n = 6) were acquired in

the anterior, posterior, lateral and posterior oblique projections.

Computed tomography pulmonary angiographyComputed tomography pulmonary angiography was per-

formed using a Hi Speed Advantage system (GE Medical

Systems, Milwaukee, WI, USA). Data were acquired com-

mencing from a level 2 cm cranial to the more inferior

hemidiaphragm to the level of the aortic arch; collimation was

3 mm and a pitch of 2.0 was used during intravenous injection

of 150 mL Ultravist 240 mg/mL by a pump injector, Biotel PJ3

(Biotel, Sydney, New South Wales, Australia); rate of injection

was 3.5 mL/s, and the delay (from commencement of injection

to imaging) was 20 s.

Doppler sonographyLower limb venous sonography (DUS) was performed using a

linear array 7-MHz transducer and employed a standard

protocol. Lack of venous compressibility and alteration in

venous flow signals on Doppler and colour flow Doppler

imaging were the criteria employed for diagnosis of venous

thrombosis.

RESULTSClinical data were available in 99 patients with intermediate-

probability VQS studies. Eight patients were excluded because

they were already receiving anticoagulation therapy at the time

of presentation. In the remaining 91 patients, pretest clinical

probability was considered to be low in 25, intermediate in 30

and high in 36 cases. In total 51.6% (n = 47) of these patients

(80% (n = 2) with low, 66% (n = 20) with intermediate and

69.4% (n = 25) with high pretest probability) proceeded to

further imaging.

425INVESTIGATION/MANAGEMENT OF INTERMEDIATE VQS

Further imagingDoppler ultrasound only was performed in a total of 19 patients

(8% (n = 2) with low, 23% (n = 7) with intermediate and 27%

(n = 10) with high pretest probability). Computed tomography

pulmonary angiography only was performed in 17 patients

(26% (n = 8) with intermediate and 25% (n = 9) with high pretest

probability). A total of 16.6% (n = 5) of patients with intermediate

and 16.6% (n = 6) of patients with high clinical probability

proceeded to both DUS and CTPA (Table 1). Therefore, CTPA

was performed in a total of 30.7% (n = 28) of patients with

intermediate VQS studies (Table 1).

Patient managementNo patients with low pretest probability were treated with

anticoagulation therapy. A total of 46% (n = 14) of patients with

intermediate pretest probability initially received anti-

coagulation therapy; further studies were positive in just one of

these cases. Twelve of these patients had their anticoagulation

therapy discontinued when neither CTPA nor DUS showed

evidence of thromboembolism or/and when clinical progress

while in hospital suggested this diagnosis to be unlikely.

All patients with a high pretest probability in the intermediate

VQS category initially received anticoagulation therapy. Of

these patients 69% (n = 25) proceeded to further imaging and

CTPA was requested in 41.6% (n = 15). Three patients had their

anticoagulation therapy discontinued when these further

studies proved to be negative (Table 2). A total of 40% (n = 14)

of all patients (including 39% (n = 13) with high pretest

probability who were discharged on anticoagulation therapy)

had a positive imaging diagnosis of thromboembolic disease

(pulmonary embolus, n = 7; venous thrombus, n = 6; pulmonary

embolus, n = 1; and venous thrombus, n = 1).

DISCUSSIONCorrect diagnosis of pulmonary thromboembolism presents a

frequent clinical dilemma although this is both a common and

life-threatening condition having a reported mortality for

untreated embolism of 30–50%.12,13 Lung scintigraphy has been

well established as the primary non-invasive imaging modality

for patients with suspected embolism, but its lack of specificity

has led to continuing controversy over the further investigation

and management of patients with intermediate, indeterminate

and inconclusive VQS results.14 The PIOPED study showed

that approximately 50% of studies yield a result of low or

intermediate probability for embolism; 30% of patients with

intermediate-probability VQS and 14% of those with low-

probability VQS subsequently had emboli diagnosed at

pulmonary angiography.1 This group therefore merits further

investigation prior to treatment because there is a significant

mortality associated with untreated embolism, and empirical

anticoagulation also has potential risks.15

We have found that the most important factor determining

the subsequent course of patients with intermediate VQS

appears to be the pretest clinical probability. Only 8% of

patients with a low pretest probability of embolism proceeded to

further investigation and none received anticoagulation

therapy; the majority were treated for pulmonary infection or

oedema. As the likelihood ratio (post-test probability/pretest

probability) of an intermediate VQS result approaches 1,14 the

contribution of this result to the diagnostic pathway in this group

does not appear significant. By PIOPED results, pulmonary

embolism is present in 16% of these cases but with the caveat

that any ventilation–perfusion abnormality is less specific in the

presence of prior cardiorespiratory disease.1 A total of 66% of

patients with an intermediate pretest probability proceeded to

426 G WALSH AND DN JONES

Table 1. Further imaging evaluation of patients with intermediate-probability ventilation perfusion scintigraphy

Pretest probability Further imaging studies DUS CTPA DUS + CTPA

(n) (n) (n) (n) (n)

51.6% (47)

Low (25) 8% (2) 8% (2)

Intermediate (30) 66% (20) 23% (7) 26% (8) 16.6% (5)

High (36) 69.4% (25) 27% (10) 25% (9) 16.6% (6)

DUS, Doppler ultrasound; CTPA, CT pulmonary angiography.

Table 2. Subsequent management of patients with intermediate-probability ventilation perfusion scintigraphy

Pretest probability Initial anticoagulation therapy Results of further imaging Discharged on anticoagulation therapy

(n) (n) +ve –ve

(n)

Low (24) 0

Intermediate (30) 66% (14) 3% (1) 43% (13) 7% (2)

High (36) 100% (36) 36% (13) 33% (12) 92% (33)

427INVESTIGATION/MANAGEMENT OF INTERMEDIATE VQS

further imaging; most of those with negative results on these

studies had anticoagulation therapy discontinued prior to

discharge, suggesting that, in this group, clinical management

is influenced to some degree by negative DUS/CTPA results.

By contrast, 69.4% of patients with a high pretest clinical

probability proceeded to further investigation; these were

negative in 48% of cases but all of these patients subsequently

received anticoagulation therapy. An intermediate-probability

VQS together with a high pretest clinical probability was

associated with a 66% probability of pulmonary embolism in the

PIOPED study.1 When these are combined with both negative

CTPA and lower limb DUS, the probability of embolism is

unknown. But because only 5.6% of PIOPED patients had

evidence of isolated subsegmental embolus at pulmonary

angiography, the probability of pulmonary embolism in patients

with high clinical probability, intermediate VQS, normal DUS

and CTPA might be inferred to be quite low.

In patients in the intermediate and high pretest probability

groups who proceeded to further imaging investigations, only

65% of those with intermediate and 60% of those with high

probability proceeded to CTPA. Overall, only 37.5% of those

who were discharged from hospital on anticoagulant therapy

underwent CT evaluation and only 40% had a positive imaging

diagnosis of thromboembolic disease. These latter findings are

not significantly different from those reported by Khorasani et al.

for conventional pulmonary angiography,5 and those reported

more recently by Murchison et al. for periods prior to the clinical

adaptation of CTPA.16 The relative hesitancy of clinicians to

request a non-invasive modality such as CTPA may relate to

their past dependence on clinical judgement for patients with

intermediate-category VQS studies in institutions where

conventional angiography was not frequently employed. This

may also reflect the present lack of data on clinical outcome in

patients with negative CTPA studies who do not receive

anticoagulation therapy. If, in time, these become available,

then use of CTPA in this group of patients may increase and its

impact on their subsequent clinical management may become

more significant.

Our findings also suggest that anticoagulation therapy may,

at present, be overutilized in patients with intermediate-

category VQS studies; this is undesirable because of the

significant associated risk of bleeding complications. In

addition, the cost of long-term therapy with its inherent require-

ment for regular clinical and haematological review might be

expected to exceed considerably that of further imaging

studies; this factor is likely to become increasingly significant in

the present economic climate of managed health care.

Our results indicate that although patients with intermediate

and high pretest clinical probability and intermediate VQS

proceed to further imaging, CTPA, even when freely available,

is underused in this group. Most patients with high pretest

probability receive anticoagulation therapy irrespective of

imaging findings, but negative CTPA and DUS studies appear to

have more influence on the subsequent management of patients

with intermediate pretest probability. Overall, less than half of

patients discharged from hospital on anticoagulation therapy

have a positive imaging diagnosis of thromboembolic disease.

REFERENCES1. Prospective Investigation of Pulmonary Embolism Diagnosis

(PIOPED). Value of the ventilation/perfusion scan in acute

pulmonary embolism results of the prospective investigation of

pulmonary embolism diagnosis. JAMA 1990; 263: 2753–9.

2. Ferretti GR, Bosson JL, Buffaz PD et al. Acute pulmonary

embolism. Role of helical CT in 164 patients with intermediate

probability at ventilation–perfusion scintigraphy and normal results

at duplex US of the legs. Radiology 1997; 205: 453–8.

3. Schulgar N, Henschke CI, King T et al. Academia versus clinic:

Practice patterns in the diagnosis of pulmonary embolism at a

large teaching hospital. J Thorac Imaging 1994; 9: 180–4.

4. Henschke CI, Mateescu I, Yarkelevitz DF. Changing practice

patterns in the workup of pulmonary embolism. Chest 1995; 107:

940–5.

5. Khorasani R, Gudas TF, Nikpoor N, Polak JF. Treatment of

patients with suspected pulmonary embolism and intermediate-

probability lung scans: Is diagnostic imaging underused? AJR1997; 169: 1355–7.

6. Remy-Jardin MJ, Remy J, Wattinne L, Giraud F. Central

pulmonary thromboembolism: Diagnosis with spiral volumetric CT

with the single-breath-hold technique: Comparison with pulmonary

angiography. Radiology 1992; 185: 381–7.

7. Goodman LR, Curtin JJ, Mewissen MW et al. Detection of

pulmonary embolism in patients with unresolved clinical and

scintigraphic diagnosis: Helical CT versus angiography. AJR 1995;

164: 1369–74.

8. Meaney JFK, Weg JG, Chenevert TL, Stafford-Johnson D, Hamilton

BK, Prince MR. Diagnosis of pulmonary embolism with magnetic

resonance angiography. N Engl J Med 1997; 336: 1422–7.

9. Gottschalk A, Sostman HD, Coleman RE. Ventilation–perfusion

scintigraphy in the MOPED study. Part 11. Evaluation of the

scintigraphic criteria and interpretations. J Nucl Med 1993; 34:

1119–26.

10. Flores J, Lancha C, Perez Rodriguez EP, Avello AG, Bollo E, Frade

LJG. Efficacy of D-Dimer and total fibrin degradation products

evaluation in suspected pulmonary embolism. Respiration 1995;

62: 258–62.

11. Scott IM, Gillen GJ, Shand J, Bryden F, Milroy R. A structural

approach to the interpretation and reporting of ventilation/

perfusion scans. Nucl Med Commun 1998; 19: 107–12.

12. Lilienfeld DE, Chan E, Ehland J et al. Mortality from pulmonary

embolism in the United States: 1962–84. Chest 1990; 98: 1067–72.

13. Anderson FA Jr, Wheeler HB, Goldberg RJ et al. A population-

based perspective of the hospital incidence and case-fatality rates

of deep venous thrombosis and pulmonary embolism: The

Worcester DVT study. Arch Intern Med 1991; 151: 933–8.

14. Robinson PJA. Ventilation–perfusion lung scanning and spiral

computed tomography of the lungs: Competing or complementary

modalities? Eur J Nucl Med 1996; 23: 1547–53.

15. Goldhaber SZ, Morpurgo M. Diagnosis, treatment and prevention

of pulmonary embolism. JAMA 1992; 268: 1727–33.

16. Murchison JT, Gavan DR, Reid JH. Clinical utilization of the non-

diagnostic lung scintigram. Clin Radiol 1997; 52: 295–8.