Should lung scan be abandoned for pulmonary embolism diagnosis in the age of multislice spiral CT? No

Abstract

Multidetector computed tomographic angiography (CTA)is increasingly recognized as the first-line imaging tech-nique for suspected pulmonary embolism [1], however,CTA is not always feasible. In the second ProspectiveInvestigation of Pulmonary Embolism Diagnosis (PIOPEDII), over 50% of the 7,284 eligible patients were excludedfor documented contraindications to CTA such as renalfailure, abnormal creatinine levels, allergy to contrastagents, critical illness requiring mechanical ventilation,recent myocardial infarction, or pregnancy [2]. Alternativeimaging protocols are therefore needed.Perfusion scintigraphy was introduced in the 1960s tovisualize the regional distribution of pulmonary blood flow.Some early studies showed that segmental or lobar perfu-sion defects on lung scintigraphy are characteristic of acutepulmonary embolism [3, 4].In the 1970s, ventilation imaging was added to perfusionscintigraphy on the assumption that it could help differ-entiate perfusion defects due to embolism—where venti-lation ought to be relatively preserved—from perfusionabnormalities secondary to disorders of the lung paren-chyma where both ventilation and perfusion are impaired.Contrary to expectations, ventilation-perfusion scintigra-phy often yields inconclusive results, especially whenobsolete ventilation tracers (e.g. 133-Xenon) are used [5].In recent years, the introduction of newer ventilationtracers (e.g. Technegas ) and the implementation of singlephoton emission tomography have led to a substantialimprovement in the diagnostic accuracy [6]. Yet, ventila-tion imaging cannot be obtained in critically ill patients.Moreover, ventilation-perfusion scintigraphy carries ahigher radiation load than perfusion scanning alone, whichmakes it inappropriate in evaluating young subjects orpregnant women with suspected pulmonary embolism.The diagnostic value of perfusion scintigraphy (withoutventilation imaging) was reappraised in the ProspectiveInvestigative Study of Acute Pulmonary Embolism Diag-nosis (PISAPED) published in 1996 [7]. In that study, a lungscan featuring wedge-shaped (segmental) perfusion defectsis diagnostic with a sensitivity of 86% and a specificity of93% against selective or superselective pulmonary angi-ography [7]. These values are in close agreement with thosereported for multidetector CTA in the PIOPED II (sensi-tivity 83%, specificity 96%). The diagnostic performance ofperfusion scintigraphy for lung embolism was confirmedupon examining 889 scans from the PIOPED II [8]. ThePISAPED criteria for interpreting perfusion scans havebeen recently reviewed [9]. An example of a perfusion scansuggestive of acute pulmonary embolism is given in Fig. 1.It appears, therefore, that perfusion scintigraphy is asaccurate as a multidetector CTA in diagnosing or excludinglung embolism. Differently from CTA, it is essentially no-risk, minimally invasive, and requires no iodine injection.In addition, it is less expensive and entails a much lowerradiation burden than CTA.As said, it should be remembered that none of theavailable diagnostic tests for suspected pulmonary embo-lism can, alone, safely confirm or exclude the diagnosiswithout independent assessment of the clinical probabilityof the disease. The latter can be expressed empirically orby means of a prediction rule [10, 11].