Lung scintigraphy--doubt and certainty in the diagnosis of pulmonary embolism

Do preliminary chest X-ray findings define the optimum role of pulmonary scintigraphy in suspected pulmonary embolism?

AIM

To investigate if preliminary chest radiograph (CXR) findings can define the optimum role of lung scintigraphy in subjects investigated for pulmonary embolism (PE).

MATERIALS AND METHODS

The CXR and scintigraphy findings from 613 consecutive subjects investigated for suspected PE were retrieved from a radiological database. Of 393 patients with abnormal CXRs, a subgroup of 238 was examined and individual radiographic abnormalities were characterized. CXR findings were related to the scintigraphy result.

RESULTS

Scintigraphy was normal in 286 subjects (47%), non-diagnostic in 207 (34%) and high probability for PE in 120 (20%). In 393 subjects (64%) the preliminary CXR was abnormal and 188 (48%) of scintigrams in this group were non-diagnostic. Individual radiographic abnormalities were not associated with significantly different scintigraphic outcomes. If the preliminary CXR was normal (36%), the proportion of non-diagnostic scintigrams decreased to 9% (19 of 220 subjects) (P < 0.05).

CONCLUSION

In subjects investigated for PE, an abnormal CXR increases the prevalence of non-diagnostic scintigrams. A normal pre-test CXR is more often associated with a definitive (normal or high probability) scintigram result. The chest radiograph may be useful in deciding the optimum sequence of investigations.Forbes, K. P. N., Reid, J. H., Murchison, J. T.(2001). Clinical Radiology56, 397-400.

A comparison of radiopharmaceutical agents used for the diagnosis of pulmonary embolism

Radioactive gas or technetium-99m aerosols are used to perform pulmonary ventilation scintigraphy. The aim of this study was to compare three radiopharmaceuticals, Kryptoscan, Technegas and Venticis II, in terms of their costs and user preferences rather than on the basis of diagnostic efficacy. For each radiopharmaceutical agent, an analysis questionnaire was sent to nuclear medicine departments setting out the criteria (and subcriteria) to be assessed: diagnosis quality: imaging quality, distribution homogeneity, examination procedures and capacity to examine particular patients (e.g. smokers); safety: for patient, paramedical and medical staff and the environment; use: availability in cases of emergency, ergonomics of the apparatus, simplicity and time of preparation. A score, ranging from 0 to 5, and a weighting (importance of one criterion with regard to the others) were assigned to each criterion. The direct cost of a ventilation (drugs, generator systems, disposable materials) was calculated for each radiopharmaceutical agent according to the number of patients examined per day (1-6) and the number of examination days per week (2-5). Fourteen questionnaires concerning at least two of the products were returned out of the 30 mailed. A 'preference score' was calculated using Pharma Decision software. The mean score of Kryptoscan was significantly higher than that of Venticis II (444 vs. 286, P < 0.001) and higher than the mean score of Technegas (444 vs. 344, P < 0.01). For Venticis II and Technegas, the changes in patient direct costs were minor and depended on the number of patients per day and the number of examination days per week. Respectively, they were: $US 117.66 (5 patients.day-1; 5 days.week-1) to $US 147.74 (2 patients.day-1; 2 days.week-1) and $US 56.60 (6 patients.day-1; 5 days.week-1) to $US 132.08 (2 patients.day-1; 2 days.week-1). The direct cost of ventilation using Kryptoscan varied only according to the number of patients examined per day: $US 104.66 (6 patients.day-1) to $US 266.47 (2 patients.day-1). This study shows that Kryptoscan appears to be preferable for ventilation scintigraphy whenever at least four patients are examined daily.

Communication of doubt and certainty in radiological reports

We have investigated the reliability of communication of uncertainty in radiological reports. The 18 most commonly used verbal expressions of probability were identified from a series of radiological reports. 11 clinicians (three radiologists, three rheumatologists and five orthopaedic surgeons) recorded the probability that they ascribed to each of the 18 expressions using visual analogue scales. Each subject was re-tested on four occasions at least 1 week apart. The results were analysed to assess reproducibility within and between individuals. We found considerable variation in the probabilities assigned to many commonly used expressions between subjects, and between repeated testing of the same subject. Some expressions were rated much more consistently than others. "Absent", "excludes", "unlikely", "probable", "certain" and "definite" were the most consistently rated expressions. We have identified a potential source of misunderstanding in radiological reports owing to differences in interpretation of expressions used by radiologists and referring clinicians.

Clinical utilization of the non-diagnostic lung scintigram

By virtue of the poor specificity of ventilation perfusion lung scintigraphy, a significant number of examinations for suspected pulmonary embolism (PE) result in a report which is neither normal nor high probability. These are unhelpful in establishing a firm clinical diagnosis. Patients with an indeterminate report should therefore undergo further investigation to establish the diagnosis particularly when treatment with anticoagulants is proposed. All lung scintigram reports issued over a 2-year period were reviewed and 102 indeterminate lung scintigram reports were identified. The case notes of 94 of these patients were examined and details of further investigation and management recorded. Fifty-one patients (55%) had no further radiological investigations and 19 (37%) of these were unequivocally categorized as having had pulmonary embolism by the referring clinician. Eighteen of these were treated with anticoagulation therapy. When patients proceeded to further radiological investigation then the result usually influenced the final clinical diagnosis. Clinicians frequently treat an intermediate report as the end point in investigation of suspected PE and consequently some patients may receive suboptimal management. By implementing a hospital wide policy of further investigation of non-diagnostic lung scintigrams, using a standard protocol, patient management could be improved.

Ventilation-perfusion lung scanning and spiral computed tomography of the lungs: competing or complementary modalities?

The recently developed technique of spiral computed tomographic angiography (CTA) is being used for the detection of pulmonary emboli (PE), and several studies have assessed its accuracy using pulmonary angiography as the gold standard. CTA shows a high level of accuracy in the detection of pulmonary emboli in segmental or larger central vessels. The specificity is high enough to eliminate the requirement for angiography if a positive CTA result is found. The main factor limiting the sensitivity of CTA is the frequency of peripheral emboli in the vessels outside the central chest field covered by CTA. The incidence of such peripheral emboli varies in different reports from 0% to 36%, and their significance remains arguable. Interpretative criteria for V./Q.- lung scintigraphy have been refined as a result of the lessons learned from the PIOPED study. Using these modified criteria, and taking into account the prior probability of PE based on the presence or absence of clinical risk factors, treatment decisions can be reasonably made in patients in the following categories: those with normal lung scans, those with high probability scans and high prior probability of PE, and those with low probability scans and low clinical suspicion. Patients with intermediate probability or indeterminate scans, and those in whom the scan results conflict with the clinical expectation, will need further tests. Ultrasound examination of the leg veins, if positive, will select a further subgroup of patients for active treatment. Patients with a negative or inconclusive ultrasound result, who previously have been candidates for pulmonary angiography, can now go on to CTA. The advantages in specificity which CTA offers will make it an important part of the diagnostic workup for selected patients, but in view of its increased cost and high radiation dose compared with V./Q. scintigraphy, the argument that CTA should completely replace lung scintigraphy is currently unsupportable.

Pulmonary embolism in children

Pulmonary embolism is an uncommon diagnosis in the pediatric patient, often discovered on autopsy. The mortality rate of untreated pulmonary embolism approaches 30%, and the best way to prevent death is by swift diagnosis and initiation of therapy. This article reviews the risk factors, clinical presentation, pathophysiology, diagnostic modalities, and treatment of pulmonary embolism in children.

Pulmonary scintigraphy in suspected pulmonary embolism: how often does it change clinical management?

The effect of pulmonary scintigraphy on patient management in suspected pulmonary embolism has been assessed in 158 patients. A change in management occurred in 30% following scintigraphy. The major effect was a reduction in the total number of patients who were anticoagulated. Heparin was stopped in 42 of the 74 patients to whom it had been administered prior to the scintigram.

Observer variation in the interpretation of ventilation-perfusion lung scintigraphy

The diagnosis of pulmonary embolism (PE) remains one of the most difficult in clinical medicine, and the diagnostic value of lung scintigraphy has been questioned. To evaluate the observer variation in the interpretation of ventilation-perfusion lung scanning in the diagnosis of PE, 87 lung scintigrams from consecutive patients with phlebography-proven deep venous thrombosis and without clinical signs of PE were randomly mixed with 50 reference lung scintigrams from patients with PE symptoms. The scintigrams were reevaluated blind by two experienced clinical physiologists. Each observer evaluated each lung scintigram twice and recorded whether the lung scan was normal or abnormal. If it was abnormal, the location and number of segment defects were registered. The intraobserver agreement, including number and location of segments, ranged from 0.77 to 0.85 and for the diagnosis of PE from 0.88 to 0.92 with a kappa of 0.80-0.84. The values for the interobserver agreement for the diagnosis of PE were 0.73-0.80 with a kappa of 0.56-0.67. It is concluded that in the interpretation of ventilation-perfusion lung scintigraphy the use of a simple scheme-deciding whether there is segmental ventilation-perfusion mismatch or not-has a good reproducibility with a high kappa for inter- and intraobserver variation and can serve as a simple routine method for diagnosing PE.

A comparison of SPECT and planar ventilation perfusion lung scanning

In order to define the role of Single Photon Emission Computed Tomography (SPECT) in the diagnosis of pulmonary embolus; SPECT and Planar ventilation and perfusion lung studies were performed consecutively on eleven patients referred with suspected embolus. Three patients were shown to have 'high probability' ventilation perfusion mismatches. SPECT imaging allowed segmental localisation of the perfusion defect and revealed additional defects not seen on planar scans. SPECT lung study was performed with minimal technical difficulty and was well tolerated by all patients studied. SPECT is likely to become the method of choice for investigating patients referred with suspected pulmonary embolus.

Lung scanning in the diagnosis of pulmonary embolism: the emperor redressed

Diagnosis and management of the patient with pulmonary embolism remains a vexing clinical problem. Emboli are often unsuspected clinically. Ventilation-perfusion (VQ) scintigraphy is known to miss some emboli found on pulmonary angiography. Some physicians recommend diagnostic approaches in which the lung scan plays a relatively minor role, and angiography is required for many patients. Major prospective studies recently have made available objective data for formulation and evaluation of diagnostic and therapeutic strategies. These data suggest that the lung scan is a better predictor of patient outcome than has been previously appreciated. The goal of V/Q scanning is not detection of pulmonary emboli per se, but rather the identification of patients at a high or low risk for future embolic events if they are not anticoagulated. Several studies have shown that some emboli found on angiography are clinically benign and, in the absence of persistent thrombosis of the lower extremities, do not require anticoagulation. Treatment trials and clinical follow-up studies have shown that although the V/Q scan is not always predictive of angiogram results, it is a reliable predictor of patient outcome. In the absence of deep vein thrombosis, the low-probability scan identifies a patient population not requiring anticoagulation. When combined with objective studies of the venous system, the ventilation-perfusion lung scan provides a guide to management in the great majority of patients. Pulmonary angiography is required only occasionally.

Survey on the use of pulmonary scintigraphy and angiography for suspected pulmonary thromboembolism in the UK

A survey of UK radiologists in 360 acute hospitals was undertaken to assess the current use of pulmonary angiography and radionuclide lung scanning in the investigation of suspected pulmonary thromboembolism (PTE). Replies were received from 340 (94%) hospitals, of which 48 (out of 50) were teaching hospitals or centres which perform cardiothoracic surgery. Lung scintigraphy was provided by 200 (59%) hospitals and angiography by 120 (35%), with 99 (29%) providing both. Twenty-two hospitals which could undertake angiography had not performed more than one angiogram for suspected PTE in the preceding 3 years. Average numbers of lung scans performed in each hospital were 21.8 per month, in contrast with an average of 4.1 angiograms (including digital subtraction angiography) per year performed for suspected PTE over the previous 3 years. The total number of V/Q lung scans performed for the diagnosis of PTE per year was approximately 47,000 compared with 490 pulmonary angiograms.