Improving the diagnostic performance of lung scintigraphy in suspected pulmonary embolic disease

Aimto determine the effectiveness of a new imaging algorithm in the investigation of suspected pulmonary embolism (PE).Materials and methodsA new imaging algorithm for suspected PE was introduced following the installation of a multisection computed tomography (CT) machine at our institution. Before its installation, patients with suspected PE were evaluated with ventilation/perfusion (V/Q) scintigraphy. Subsequently, patients were triaged according to chest radiography (CR) and respiratory history to either lung scintigraphy or CT pulmonary angiography (CTPA). Patients with a normal CR and no history of lung disease were evaluated using perfusion (Q) scintigraphy [ventilation (V) scintigraphy was no longer performed]. Patients with an abnormal CR, asthma or chronic lung disease were evaluated using CTPA. All V/Q images in a continuous 3-year period before the introduction of the new imaging algorithm and all Q images performed in a 3-year period after its introduction were retrospectively reviewed. Imaging reports were categorized into normal, non-diagnostic (low or intermediate probability) or high probability for PE. Patients in the later group who subsequently underwent CTPA, were also reviewed.ResultsAfter the policy change the percentage of normal scintigrams significantly increased (39 to 60%; p < 0.001). There was a non-significant increase in the percentage of high probability scintigrams (15 to 18%; p = 0.716). Overall the diagnostic yield of lung scintigraphy improved significantly (54 to 78%; p < 0.001).Conclusionthe diagnostic performance of lung scintigraphy can be improved by careful triage of patients to either Q scintigraphy or CTPA based on clinical history and CR findings. Q scintigraphy remains a valuable diagnostic test in the investigation of suspected PE in carefully selected patients.