**5. Image interpretation**

In a normal patient, perfusion and ventilation are both homogeneous (figure1). There is usually a ventilation and perfusion gradient which increases from the anterosuperior to the posteroinferior region of the lungs. There is often a thin band of hyperventilation located in the lower two thirds of the posterior aspect of both lungs. The normal indentations of the mediastinum should be recognized.

Fig. 1. Normal perfusion and ventilation coronal SPECT slices.

Non-segmental partial mismatches (preserved ventilation with abnormal perfusion) can occur in normal subjects. Physiologically, they are explained by the fact that ventilation with Technegas is usually evaluated with deep breathing. Such deep breathing can temporarily overcome partially compressed lung and result in hyperventilation. Those lung sections are usually poorly ventilated with tidal breathing. Since perfusion will be physiologically matched to tidal breathing, this explains the potential for partial mismatches. Perfusion is usually maintained to some degree (figure 2). The most common areas are the medial postero-inferior regions of both lungs because of compression by the mediastinum in the supine position. Both the inferior and posterior costal phrenic angles are also often subject to this phenomenon (figure 3). Also, the superior portion of the large fissure may be the site of a small mismatch. None of these anomalies follow the topography of normal lung vessels.

Image display should strive to match precisely each ventilation slice with the corresponding perfusion slice in all three planes (transverse, coronal, sagittal). This can be easily done either by not moving the patient or bed position between ventilation and perfusion or, alternatively,

In a normal patient, perfusion and ventilation are both homogeneous (figure1). There is usually a ventilation and perfusion gradient which increases from the anterosuperior to the posteroinferior region of the lungs. There is often a thin band of hyperventilation located in the lower two thirds of the posterior aspect of both lungs. The normal indentations of the

by using commercially available software which will co-register each set of images.

**5. Image interpretation** 

mediastinum should be recognized.

Fig. 1. Normal perfusion and ventilation coronal SPECT slices.

Non-segmental partial mismatches (preserved ventilation with abnormal perfusion) can occur in normal subjects. Physiologically, they are explained by the fact that ventilation with Technegas is usually evaluated with deep breathing. Such deep breathing can temporarily overcome partially compressed lung and result in hyperventilation. Those lung sections are usually poorly ventilated with tidal breathing. Since perfusion will be physiologically matched to tidal breathing, this explains the potential for partial mismatches. Perfusion is usually maintained to some degree (figure 2). The most common areas are the medial postero-inferior regions of both lungs because of compression by the mediastinum in the supine position. Both the inferior and posterior costal phrenic angles are also often subject to this phenomenon (figure 3). Also, the superior portion of the large fissure may be the site of a small mismatch. None of these anomalies follow the topography of normal lung vessels.

Fig. 2. Passive atelectasis (coronal and sagittal slice triangulation). Note non-segmental partial mismatch.

Fig. 3. Sagittal SPECT slices showing non-embolic mismatch at posterior costophrenic angle caused by lung compression.

Ventilation Perfusion Single Photon Emission

occasionally be seen in this setting.

Tomography (V/Q SPECT) in the Diagnosis of Pulmonary Embolism 147

The preceding discussion applies only to acute pulmonary embolism. In the sub-acute or chronic phases, when partial reperfusion has occurred, the aspect can vary considerably and interpretation can be less straightforward because strange shaped partial mismatches can

Fig. 5. Typical wedge-shaped emboli in two different patients (coronal and transverse slices)

Fig. 4. Gravity dependant atelectasis. (A) Coronal slices centered on posterior surface of the lungs show diffuse shallow non segmental partial mismatches. (B) Excerpt from a 3D reconstruction of perfusion in oblique view showing a topography totally incompatible with a vascular origin. (C) CT of another patient showing the typical pattern of gravity dependent atelectasis on posterior surfaces.

In some patients, gravity dependent atelectasis can result in widespread partial mismatches on the posterior surface of both lungs. This pattern is usually easily recognized by the occurrence of multiple shallow perfusion defects which are often in a linear pattern (figure 4). A 3D display will often best demonstrate the topography.

PE is diagnosed when there is a severe and well demarcated perfusion defect which is pleural-based and clearly larger at the periphery (typically wedge-shaped, triangular or half-oval). Small size partial defects that are not well defined are much less specific and should be ignored in most acute settings, even if they are partially mismatched. The defect should clearly follow an orientation compatible with known pulmonary vascular anatomy. Ventilation should be normal or at least much better preserved than perfusion. One such large sub-segmental defect is sufficient for the diagnosis (figure 5 & 6). There are however multiple mismatched regions in most cases (figures 7, 8 & 12). For smaller sub-segmental defects, at least two are required for a confident diagnosis (figure 9 & 11). Distal PE is usually totally occlusive. However, more proximal PE (i.e.: lobar) can be partially occlusive and the perfusion defect may at times be moderate. It should be remembered that an isolated whole lung mismatch is usually not caused by PE but rather by compression of the main pulmonary artery by a mediastinal or hilar lesion.

Fig. 4. Gravity dependant atelectasis. (A) Coronal slices centered on posterior surface of the lungs show diffuse shallow non segmental partial mismatches. (B) Excerpt from a 3D reconstruction of perfusion in oblique view showing a topography totally incompatible with

In some patients, gravity dependent atelectasis can result in widespread partial mismatches on the posterior surface of both lungs. This pattern is usually easily recognized by the occurrence of multiple shallow perfusion defects which are often in a linear pattern (figure

PE is diagnosed when there is a severe and well demarcated perfusion defect which is pleural-based and clearly larger at the periphery (typically wedge-shaped, triangular or half-oval). Small size partial defects that are not well defined are much less specific and should be ignored in most acute settings, even if they are partially mismatched. The defect should clearly follow an orientation compatible with known pulmonary vascular anatomy. Ventilation should be normal or at least much better preserved than perfusion. One such large sub-segmental defect is sufficient for the diagnosis (figure 5 & 6). There are however multiple mismatched regions in most cases (figures 7, 8 & 12). For smaller sub-segmental defects, at least two are required for a confident diagnosis (figure 9 & 11). Distal PE is usually totally occlusive. However, more proximal PE (i.e.: lobar) can be partially occlusive and the perfusion defect may at times be moderate. It should be remembered that an isolated whole lung mismatch is usually not caused by PE but rather by compression of the

a vascular origin. (C) CT of another patient showing the typical pattern of gravity

dependent atelectasis on posterior surfaces.

4). A 3D display will often best demonstrate the topography.

main pulmonary artery by a mediastinal or hilar lesion.

The preceding discussion applies only to acute pulmonary embolism. In the sub-acute or chronic phases, when partial reperfusion has occurred, the aspect can vary considerably and interpretation can be less straightforward because strange shaped partial mismatches can occasionally be seen in this setting.

Fig. 5. Typical wedge-shaped emboli in two different patients (coronal and transverse slices)

Ventilation Perfusion Single Photon Emission

amputation with preserved ventilation.

registered coronal slices)

Tomography (V/Q SPECT) in the Diagnosis of Pulmonary Embolism 149

Fig. 8. Massive embolism (co-registered coronal slices). Ventilation slices are on 1st and 3rd row. Corresponding perfusion slices on 2nd and 4th row show massive areas of vascular

Fig. 9. Small left sub-segmental embolus. Pleural effusion is noted on the right side (co-

Fig. 6. Typical embolus, coronal, sagittal and transverse slice triangulation.

Fig. 7. Multiple emboli.

Fig. 6. Typical embolus, coronal, sagittal and transverse slice triangulation.

Fig. 7. Multiple emboli.

Fig. 8. Massive embolism (co-registered coronal slices). Ventilation slices are on 1st and 3rd row. Corresponding perfusion slices on 2nd and 4th row show massive areas of vascular amputation with preserved ventilation.

Fig. 9. Small left sub-segmental embolus. Pleural effusion is noted on the right side (coregistered coronal slices)

Ventilation Perfusion Single Photon Emission

Tomography (V/Q SPECT) in the Diagnosis of Pulmonary Embolism 151

Fig. 12. Transverse slices showing multiple right sided emboli. Pleural effusion on left.

narrow indeterminate category can be acceptable for highly atypical cases.

hours of aerosol therapy.

Probabilistic interpretation with classification of cases into "normal", "low", "indeterminate" or "high" probability which is still in use with planar scintigraphy is absolutely not warranted with V/Q SPECT. Indeed, there has been no study to specifically address the validity of probabilistic interpretation with V/Q SPECT. Therefore, using a probabilistic interpretation in that context has no scientific basis. All authors that have published on the subject have used a straightforward positive-negative approach and it is the only one that is acceptable. It makes no sense to use a mode of interpretation that has been motivated by the use of an inferior ventilation agent (xenon-133) in a single planar view, associated with planar perfusion imaging. As for any other type of imaging, a very

Causes of non-embolic pathological mismatches are well known and are essentially the same with V/Q SPECT as they are with the conventional planar V/Q scintigraphy. They may be the source of false-positive readings. Septic, fat or amniotic fluid embolization may occasionally occur in specific settings. The mismatches are usually small. Intravenous illicit drug use may occasionally result in small sized mismatches, although larger mismatches may occur. Vasculitis can be considered when clinically appropriate. A compression of a segmental or sub-segmental branch of the pulmonary artery by a lung nodule can be rarely seen. Much more frequent is a very large mismatch caused by a compression of hilar or mediastinal origin. Small partial mismatches can also occur in emphysematous bullae (Figure 13) because of occasional penetration of Technegas. However, the mismatch does not usually have a vascular pattern and is typically located at the apex. Lung scarring or fibrosis can also cause small or partial mismatches. Rarely, asthma can present with some strange looking small but multiple partial mismatches, presumably because there can be a lag time between restoration of regional ventilation and adjustment of the physiologically matched perfusion anomaly (figure 14). In such cases, V/Q SPECT will normalise after 24

Fig. 10. Typical wedge-shaped embolus.

Fig. 11. Example of segmental (long arrow) and subsegmental (short arrow) emboli.

Fig. 11. Example of segmental (long arrow) and subsegmental (short arrow) emboli.

Fig. 10. Typical wedge-shaped embolus.

Fig. 12. Transverse slices showing multiple right sided emboli. Pleural effusion on left.

Probabilistic interpretation with classification of cases into "normal", "low", "indeterminate" or "high" probability which is still in use with planar scintigraphy is absolutely not warranted with V/Q SPECT. Indeed, there has been no study to specifically address the validity of probabilistic interpretation with V/Q SPECT. Therefore, using a probabilistic interpretation in that context has no scientific basis. All authors that have published on the subject have used a straightforward positive-negative approach and it is the only one that is acceptable. It makes no sense to use a mode of interpretation that has been motivated by the use of an inferior ventilation agent (xenon-133) in a single planar view, associated with planar perfusion imaging. As for any other type of imaging, a very narrow indeterminate category can be acceptable for highly atypical cases.

Causes of non-embolic pathological mismatches are well known and are essentially the same with V/Q SPECT as they are with the conventional planar V/Q scintigraphy. They may be the source of false-positive readings. Septic, fat or amniotic fluid embolization may occasionally occur in specific settings. The mismatches are usually small. Intravenous illicit drug use may occasionally result in small sized mismatches, although larger mismatches may occur. Vasculitis can be considered when clinically appropriate. A compression of a segmental or sub-segmental branch of the pulmonary artery by a lung nodule can be rarely seen. Much more frequent is a very large mismatch caused by a compression of hilar or mediastinal origin. Small partial mismatches can also occur in emphysematous bullae (Figure 13) because of occasional penetration of Technegas. However, the mismatch does not usually have a vascular pattern and is typically located at the apex. Lung scarring or fibrosis can also cause small or partial mismatches. Rarely, asthma can present with some strange looking small but multiple partial mismatches, presumably because there can be a lag time between restoration of regional ventilation and adjustment of the physiologically matched perfusion anomaly (figure 14). In such cases, V/Q SPECT will normalise after 24 hours of aerosol therapy.

Ventilation Perfusion Single Photon Emission

Tomography (V/Q SPECT) in the Diagnosis of Pulmonary Embolism 153

Fig. 14. Asthma, acute attack (retrospective diagnosis). Aerosol therapy was begun before

are mainly seen on the ventilation part of this study. Distribution is heterogeneous and, in the more severe cases, there may be focal deposition of Technegas in Airways. With advanced disease, there may be widespread focal deposition. On the other hand, perfusion is usually better preserved. With a pattern of relatively pure and advanced emphysema, perfusion and ventilation are more matched, reflecting mainly focal architectural pulmonary changes. It has been demonstrated that the degree of heterogeneity on the ventilation study, as well as the degree of heterogeneity of perfusion and ventilation matching, are both proportional to the severity of COPD. In fact, these measures appear to be more sensitive to the presence of COPD than high-resolution CT which, despite its higher resolution, has a limited capacity for the detection of airway closure. However, heterogeneous distribution of ventilation and perfusion can also be found in pulmonary oedema, lung fibrosis and

Although pulmonary oedema is usually well demonstrated on a chest x-ray, in the early stages of volume overload the only sign will be vascular redistribution to the upper lung zone. On a V/Q SPECT study, this is very easily appreciated. Typically, the examination being performed in a supine position, redistribution will be most marked anteriorly and superiorly and will usually be much more apparent on perfusion then on ventilation (figure 18). In the earliest stages, the ventilation gradient will be totally preserved which produces a rather large scale partial mismatch. It is important that this pattern be recognized and not confused with bilateral partially occluding inferior lobar PE. It should be noted that cardiac failure is not the only cause of vascular redistribution. Volume overload, whether iatrogenic

V/Q SPECT. Notice multiple non-vascular looking partial mismatches.

infectious or non-infectious diffuse lung inflammation.

or caused by hepatic of renal failure may produce the same images.

**6.2 Cardiac failure and volume overload** 

It is therefore highly recommended that image interpretation is made with full knowledge of the clinical data and that correlation should be made with a recent chest x-ray. Correlation with existing anterior thoracic CT may be helpful in selected pathological cases. In this manner, high specificity can be achieved. Also, equivocal cases should be interpreted in light of the pre-test probability and knowledge of prior pulmonary pathology. Evidently, knowledge of prior PE or venous thrombotic disease is essential for the correct interpretation of positive cases. This type of interpretation ("holistic" or "Gestalt") is now considered as standard in most parts of the world and has been officially endorsed by the guidelines of the European Association of nuclear medicine (Bajc, Neilly et al. 2009).

Fig. 13. Small apical mismatch caused by a bullae. This should be suspected in all nonsegmental apical mismatch, although scarring may also cause a similar image.
