**4. Sensitivity and specificity of V/P SPECT and other methodological considerations**

In a clinical study, 53 % more mismatch points were identified with V/P SPECT compared to planar technique (Bajc et al., 2004). Similar results have been found by others (Gutte et al., 2010; Reinartz et al., 2001). SPECT eliminates superimposed structures, clarifying segmental and sub-segmental nature of perfusion defects caused by PE.

The value of V/P SPECT is further confirmed in clinical studies (Bajc et al., 2008; Gutte et al., 2009; Leblanc et al., 2007; Lemb & Pohlabeln, 2001). V/P SPECT is today the recommended method for clinical diagnosis, follow up and research (Bajc et al., 2009b).

Powell reported that sensitivity and specificity of CT for central PE are about 90% (Powell & Muller, 2003). Perrier et al. found in a broad clinical material that CT had a sensitivity of 70% for PE and a specificity of 91 % (Perrier et al., 2001). They concluded: "clinical CT should not

Quantitative Ventilation/Perfusion Tomography:

**6. Additional findings** 

diagnoses than PE.

**6.1 Chronic obstructive pulmonary disease** 

on the basis of V/P SPECT and relevant clinical information.

The Foremost Technique for Pulmonary Embolism Diagnosis 195

and quantified in terms of mismatch points. In a prospective study Olsson et al. studied 102 out-patients with a moderate degree of PE (up to 40%) (Olsson et al., 2006). After only 5 days, embolism diminished by 44 % on average. There was no tromboembolic mortality in the trial. Later follow up indicated that PE had not recurred in patients showing resolution after 5 days. Since 2004, more than 1000 patients, of Skåne University Hospital, Lund, with up to 50% extension of PE have been treated as out-patients. After this positive experience, out-patient treatment is now perceived as a safe routine in patients who have been selected

PE is a condition known for its non-specific symptoms. Medical imaging, such as V/P SPECT, is therefore necessary to confirm or exclude the diagnosis among patients with suspected PE. The majority of patients that are examined with V/P SPECT, due to the initial assumption of PE, will not have PE. It is therefore important that any alternative diagnoses, which could explain the patients' symptoms, are identified and provided to the referring physician. Possible alternative diagnoses include pneumonia, heart failure, pleural fluid, malignancy and chronic obstructive pulmonary disease (COPD) (Richman et al., 2004). Another important aspect is that these conditions sometimes coexist and that they also elevate the risk of PE (Elliott et al., 2000). V/P SPECT can be employed to identify other

COPD is a major cause of both morbidity and mortality globally (Mannino et al., 2006). It is one of the few diseases that continues to rise in numbers in many countries. COPD is an inflammatory disease characterized by airflow limitation that is not fully reversible (Celli & MacNee, 2004). The airflow limitation is caused by a combination of airway obstruction and parenchymal destruction (emphysema). The pulmonary changes in COPD lead to inhomogeneous regional ventilation. V/P SPECT is sufficiently sensitive as a method to identify the functional changes in COPD. In comparison with DTPA aerosols, Technegas penetrates the lung periphery better (Jögi et al., 2010), which is especially important in COPD (exemplified in Fig 3). Technegas ventilation imaging has been shown to visualise the early changes of COPD before they can be observed with high resolution CT (HRCT) (Yokoe et al., 2006). As COPD initially affects the airways, the ventilation defects are commonly more prominent than those of perfusion. Perfusion within the lungs also becomes abnormal as the lungs attempt to adapt the regional blood flow to ventilation to preserve an efficient gas exchange. Often this adaptation is incomplete and perfused but non-ventilated areas (low V/P ratio) occur, i.e. reverse mismatch (Gottschalk et al., 1993). With progressive disease, concurrent destruction of airways and blood vessels takes place and matched defects with absence of both ventilation and perfusion are seen. Vascular remodeling in COPD may lead to regions with elevated V/P ratios. Garg et al. (1983) found that the degree of abnormality on aerosol ventilation images significantly correlated to pulmonary function tests. In a recent paper, which evaluated the role of V/P SPECT in patients with COPD, it was shown that V/P SPECT correlated significantly both to traditional lung function tests as well as the extent of emphysema as measured with HRCT (Jögi et al., 2011). It was also shown that V/P SPECT could be used to characterize the severity of COPD. Pulmonary embolism and heart failure are common comorbidities with overlapping symptoms that

be used alone for suspected PE but could replace angiography in combined strategies that include ultrasonography and lung scanning". Likewise, van Strijen et al. found in a multicentre prospective study that sensitivity of CT was 69% while specificity was 84% and "concluded that the overall sensitivity of spiral CT is too low to endorse its use as the sole test to exclude PE" and that "this holds true even if one limits the discussion to patients with larger PE in segmental or larger pulmonary artery branches" (Van Strijen et al., 2005). Our experience supports this view. Also, CT as a second procedure following scintigraphy has limited value (van Strijen et al., 2003). Multislice CT seems to improve resolution but sensitivity for small PE appears not to be improved (Stein et al., 2006).

A problem associated with limited sensitivity of CT and incomplete coverage of the total lung is that the degree of embolism and lung function deficiency cannot be quantified. Quantification is important for treatment selection.

In spite of excellent diagnostic qualities of V/P SPECT and documented low sensitivity of CT, the latter method is often recommended. A high number of non-diagnostic scintigraphies were reported in the PIOPED study (65%) (1990). This is still used as an argument against lung scintigraphy. In PIOPED, scintigraphy was performed with inferior technique and inflexible sub-optimal interpretation criteria. Even with planar scintigraphy, a reduction in the number of non-diagnostic reports to 10% can be achieved with adequate acquisition and a holistic interpretation strategy (Bajc et al., 2002a). With V/P SPECT, this number is further reduced to between 1 and 4%, as found in several studies (Bajc et al., 2008; Leblanc et al., 2007; Lemb & Pohlabeln, 2001).

Some practitioners hold that sub-segmental emboli are of little importance for otherwise healthy people and may be left untreated and, as a consequence, are prepared to accept less sensitive methods for PE diagnosis. However, small emboli are important because they 1) may be a first and only sign of silent deep venous thrombosis, 2) may precede larger emboli 3) if not diagnosed and/or untreated, further episodes may lead to chronic PE and pulmonary hypertension (Fig. 6). 4) form a threat to patients with limited cardio-pulmonary reserve, 5) are clinically essential for quantification, which is necessary to scientifically establish appropriate treatment protocols. Thus, each embolus is relevant, irrespective of size. Sub-segmental emboli should not be left untreated without further scientific evidence.

#### **5. Selection of therapeutic strategy**

Management of PE was previously confined to in-hospital therapy, using anticoagulation, heparin injections followed by oral anticoagulants for extended periods of time.

About 20 % to 55 % of patients with deep venous thrombosis have concomitant PE, which is usually not diagnosed because symptoms of PE are absent. Outpatient treatment of patients with deep venous thrombosis, which is perceived as a safe routine, implies that many patients with PE are treated at home. Home treatment of patients with diagnosed PE has been suggested (Kovacs et al., 2000) but in order to determine the appropriateness of the treatment on an individual basis, the extension of PE obviously need to be estimated. Whilst patients with limited extension of PE may be treated at home, intermediate cases and patients with co-morbidity may need in-hospital treatment. Those with very extensive PE may require thrombolysis, necessitating inpatient treatment.

Obviously, quantification requires studies of the whole lung with methods allowing identification of large and small emboli. V/P SPECT is the ideal method for this purpose as segmental and sub-segmental emboli can be both detected with a high degree of sensitivity and quantified in terms of mismatch points. In a prospective study Olsson et al. studied 102 out-patients with a moderate degree of PE (up to 40%) (Olsson et al., 2006). After only 5 days, embolism diminished by 44 % on average. There was no tromboembolic mortality in the trial. Later follow up indicated that PE had not recurred in patients showing resolution after 5 days. Since 2004, more than 1000 patients, of Skåne University Hospital, Lund, with up to 50% extension of PE have been treated as out-patients. After this positive experience, out-patient treatment is now perceived as a safe routine in patients who have been selected on the basis of V/P SPECT and relevant clinical information.
