**7.2 Computer modeling**

The superiority of V/Q SPECT (sensitivity 97%) to planar imaging (sensitivity 77%) was tested in a computer model which clearly predicted much better accuracy with V/Q SPECT (Magnussen, Chicco et al. 1999). It is very interesting to note that this model accurately predicted the results of clinical studies which would appear several years later.

#### **7.3 Animal studies**

The technique was also tested in a pig model with artificially produced small size emboli (Bajc, Bitzen et al. 2002). Again, the superiority of V/Q SPECT to planar imaging was clearly demonstrated, as sensitivity for V/Q SPECT was 91% while it was 64% for planar imaging for the same emboli. Specificity was also better for SPECT (87% vs 79%). It should be mentioned that a similar model had been employed earlier for validation of invasive pulmonary angiography and that the performance numbers in terms of

Ventilation Perfusion Single Photon Emission

**8.2 Radiation dose** 

1,5 mSv.

organ system.

breasts.

to exclude) for PE (Leblanc, Leveillee et al. 2007).

**8.3 Contraindications and technical success rate**

failure requiring dialysis is extremely high.

Tomography (V/Q SPECT) in the Diagnosis of Pulmonary Embolism 161

been discussed earlier. Causes of false-positive filling defect on CTPA do exist and have been discussed elsewhere (Kuriakose and Patel 2010). Indeterminate interpretation is usually low for CTPA and is generally related to technical factors. It is also very low for V/Q SPECT, occurring in less than 5% of cases in all published studies. Finally, different studies have proven that both techniques have a high negative predictive value (capacity

Estimated radiation dose is a complex subject and a detailed discussion is beyond the scope of this chapter and can be found elsewhere (Schembri, Miller et al. 2010). However, there is little doubt that when comparing state-of-the-art technology for both modalities, incurred radiation dose is much higher with CTPA and this is particularly true of the female breast. Depending on the exact protocol that is employed, total radiation dose for CTPA is in the range of 8-20 milliSievert (mSv) while it is 2,0 - 3,5 mSv for V/Q SPECT. The dose to the female breast varies between 10 and 70 mSv for CTPA (equivalent to 10-25 mammograms or 100-400 chest x-rays) while the corresponding breast dose for V/Q SPECT is less than

There are no contraindications to V/Q SPECT imaging. Some degree of prudence is required in cases of severe pulmonary hypertension (the number of particle injected should be limited) and reasonable efforts should be made to limit the dose during pregnancy, but any patient may undergo V/Q SPECT as long as he can tolerate a supine position for 20 min. Allergies are virtually nonexistent. There are no known deleterious effects on any

On the other hand, CTPA has specific contraindications. Allergies are relatively frequent and, depending on the severity, constitute an absolute or relative contraindication. If the decision is made to proceed with the study, patients must be prepared appropriately. Also, because of the injection of contrast, renal failure is a possible complication especially in patients with established underlying renal disease. In some subgroups, the risk of renal

Performance in pregnancy has also been reviewed (Ridge, McDermott et al. 2009). The radiation dose to the foetus is very low for both techniques (< 1mSv) although it is significantly lower for CTPA. However, at this level, there is no increased risk for either technique. On the other hand, it is clear that the technical performance of CTPA in pregnancy is poor, with as many as one third of the studies being technically inadequate. This is probably due to increased pressure in the inferior vena cava during pregnancy with aspiration of large amounts of non-opacified blood during inspiration that interferes with optimal mixing of contrast coming from the superior vena cava. There are no technical limits to V/Q SPECT during pregnancy. Therefore, V/Q SPECT should be the preferred modality in this situation, especially considering the very high breast radiation dose given with CTPA in young patients with actively proliferating

High-quality imaging with CTPA requires accurate timing for contrast injection. In most studies, the technical failure rate hovers between 5 and 10%. The technical success rate for

sensitivity and specificity were no better than that obtained with V/Q SPECT (Baile, King et al. 2000).

#### **7.4 Validation with alternative perfusion techniques**

New generation techniques to measure regional perfusion independently (by opposition to simply visualizing the artery and the thrombus) by means of perfusion thoracic CT (Wildberger, Klotz et al. 2005) or nuclear magnetic resonance (NMR) (Kluge, Gerriets et al. 2006) have correlated very well with V/Q SPECT measures of regional perfusion. Indeed, work with those techniques suggests that embolism is almost always totally occlusive and that inclusion of perfusion data enhances sensitivity over an approach based purely on luminology.

#### **7.5 Clinical studies**

It is beyond the scope of this chapter to review individually all studies. However, even if proof of superiority of V/Q SPECT to planar imaging seems redundant, it has been proven clinically, with V/Q SPECT having an edge in sensitivity of more than 20% while still maintaining better specificity (Gutte, Mortensen et al. 2010; Bajc, Olsson et al. 2004). The negative predictive value of V/Q SPECT (the ability to reliably exclude PE) has been validated and is excellent, in the order of 98-99%, even in the presence of abnormal perfusion with a nonvascular pattern (Leblanc, Leveillee et al. 2007). Sensitivity is in the range of 96-99% and specificity hovers between 85% and 98%, depending on the study. The rate of non-diagnostic studies is 1-3%. Comparison to CTPA is unfortunately limited because of the lack of a large-scale prospective study comparing both techniques. A detailed discussion on each of the available studies can be found elsewhere (Leblanc and Paul 2010). However, all published studies have demonstrated that V/Q SPECT performs at least as well as CTPA for the diagnosis of pulmonary embolism (Reinartz, Wildberger et al. 2004; Suga, Yasuhiko et al. 2008; Gutte, Mortensen et al. 2009; Miles, Rogers et al. 2009).

#### **8. V/Q SPECT and CTPA: Relative advantages and limitations**

#### **8.1 Diagnostic performance**

Although a large randomized prospective study comparing the two techniques is not available, the pooled published results suggest at this point that V/Q SPECT may have an edge in sensitivity while CTPA may have an edge on specificity. Better sensitivity of V/Q SPECT can be attributed essentially to sub-segmental embolism. Indeed, to visualize directly a thrombus in a sub-segmental vessel is difficult even with the latest CTPA technology. Also, existing literature suggest that inter-observer agreement is very low for sub-segmental embolism with CTPA (Ghanima, Nielssen et al. 2007). Since V/Q SPECT visualizes the resulting perfusion defect, it has a clear advantage. Indeed, even for a small sub-segmental defect implicating 25% of a segment, the pleural base of the defect will have at least 3 cm, a dimension easily resolved by the SPECT technique. On the other hand, since CTPA directly visualizes a filling defect, it is less prone to false positive studies since most (but not all) filling defects will represent embolus. This may not be true of sub-segmental emboli because the poor inter-observer agreement in this setting suggests limited specificity. Causes of false-positive mismatches on V/Q SPECT have been discussed earlier. Causes of false-positive filling defect on CTPA do exist and have been discussed elsewhere (Kuriakose and Patel 2010). Indeterminate interpretation is usually low for CTPA and is generally related to technical factors. It is also very low for V/Q SPECT, occurring in less than 5% of cases in all published studies. Finally, different studies have proven that both techniques have a high negative predictive value (capacity to exclude) for PE (Leblanc, Leveillee et al. 2007).
