Interpretation, Reporting, and Guidelines

Key points

•

Image interpretation is not the same as “pattern recognition” as it requires knowledge of tracer kinetics, instrumentation, protocols, cardiac physiology, etc., skills that are acquired from years of experience and commitment. The board certification in nuclear cardiology, just like other boards, is a test of minimum competency; we should strive to do better, consistently!
•

The reader should recognize normal variations that might be patient, gender, and tracer specific.
•

The reader should be familiar with the capabilities of the software suites illustrated in Chapter 1 .
•

Attention should be paid at every step to quality measures such as indication, acquisition, interpretation, reporting, and personal communication of results. If perfected, the results will enhance patient management. It is the main reason why nuclear cardiology has remained the imaging procedure of choice over almost four decades at our institution.
•

The images should be reviewed before the patient leaves the laboratory and repeated, if necessary, if the quality is suboptimal because of excessive motion or subdiaphragmatic activity. There is a tremendous difference between an equivocal study and a poor-quality study. This is the only way that the entire team will realize that there are no short cuts to quality (and yes, someone has to stay late to repeat the images!). If done a few times, the message will resonate with the individual team members and the department!
•

Interpretation should be systematic and should include a review of the raw images and the use of quantitative techniques and attenuation correction, if available.
•

When reading serial studies, the report should be concise, devoid of jargon, and should be prepared with the primary care provider in mind rather than the imaging specialist. A friendly, and timely, telephone call in selected cases will go a long way toward improving communication.
•

To maintain competence, the reader should be familiar with available guidelines, appropriateness criteria, and new developments in the field.
•

The reader should be cognizant of incidental findings that may have a major impact on patient care.
•

Reports should not be relied on; the current images should be compared side-by-side to prior studies.

Interpretation of the images

On a consistent basis, there is no substitute for good quality images, but quality is more than just pretty images; it includes appropriate indications, acquisition, interpretation, and reporting. We shall address all these issues.

Rotating images: There is, like in many other things, more than one way to do the right thing. What follows is our own approach based on over three decades of experience. A good place to start (always) is to review the rotating images, both the stress and rest (if there is a rest study); we find the gray scale to be most useful. Reset the intensity so the brightness is just right. In some patients the diagnosis is made from these images alone, such as a hiatal hernia or a lung tumor (see Chapter 16 ). The most common things to pay attention to are cardiac motion, breast shadow, arm position, and hot spots.

Motion can be in any of the three axes ( x, y, and z ). The y -axis motion (up and down) is the easiest to recognize and to correct, if not excessive. The z -axis motion (forward-backward) cannot be seen but is inferred when there is evidence of motion on the slices (overlapping walls or hurricane sign; see Chapter 3 ) but no motion in the x – or y -axes (side-by-side). Some programs automatically correct for motion but that is never a substitute for reviewing the raw images. It is important to review the images before the patient leaves the laboratory, because in some patients correction is not feasible and the best way is to repeat the acquisition (as painful as it may be!). One caveat: if the images are normal despite motion, then they are likely to be normal. It is when the images are abnormal that concern arises as to whether the abnormality is real or due to motion. Motion can also affect the LVEF measurement.

The breast shadow is quite variable as it can cover the entire cardiac silhouette, the anterior wall, the lateral wall, the septum, or any combination of the above. A variation in the breast shadow between stress and rest studies can occur. This can have a profound effect on the images, producing both fixed and reversible artifactual defects. When the breast shadow covers the entire cardiac silhouette, an inferior abnormality can be produced because of diaphragmatic attenuation superimposed on uniform attenuation. We have not used prone imaging to differentiate real defects from those due to diaphragmatic attenuation but when used it can prove very useful. Remember, however, that prone imaging might introduce artifacts in the anterior wall and therefore both supine and prone images should be acquired.

The left arm position is usually on the top of the head but in patients who cannot do so, the arm is left at the side of the body, potentially creating lateral wall attenuation. It is important in these patients to have the arm in the same position in both sets of images (stress and rest) ( Figure 2-1 ).

The presence of “hot spots,” either within the heart (such as a prominent papillary muscle) or outside the heart (such as the liver or more likely a bowel loop), can produce downscaling in adjacent areas and also impact the quantitative analysis as the activity is back-projected into the cardiac ROI.

Orthogonal views: Next, the non-gated perfusion slices are reviewed in the SA, HLA, and VLA projections. We use both the “cool” color and gray scales. The color scale makes it easier to recognize milder abnormalities, whereas gray scale makes it easier to recognize reversibility. It is amazing how much more activity can be appreciated on the gray scale compared to the color scale! This is one more reason why quantification is important ( Figure 2-2 ).

The crucial step when looking at the images is to be sure they are properly aligned and scaled. Scaling is the “Achilles’ heel” of proper interpretation and it cannot be learned from books, including this one! It requires experience and knowledge as it is affected by cardiac and extracardiac activity. If the images are too bright, perfusion defects are missed; if the opposite, artifactual defects are created, defects appear larger, or defects are changed from fixed to reversible or vice versa ( Figure 2-3 ). Using the dual control button, adjust each set of images such that the brightest areas on the stress and rest images match in intensity, no more and no less. You may need to use the “expanded mode” in one or both sets if the images are too bright. Adjust by changing the upper scale and leaving the lower scale at 0 all the time!

The four steps for reading images in every patient:

1.

Are the images normal or abnormal?
2.

If abnormal, where is the abnormality (one-, two-, or three-vessel territory)?
3.

What is the size of the abnormality (small, medium, or large)?
4.

What is the nature of the abnormality (reversible, fixed, or mixed)?

Our trainees are drilled on this exercise day in and day out. The practice has paid off as we have had a 100% passing rate on the Nuclear Cardiology Board Certification Examination for the past 10 years (all first attempts).

Site, size, severity, and reversibility: The vascular territory of the LAD includes the anterior wall, septum, and apex (9 segments of the 17-segment model) ( Figure 2-4 ). The RCA territory includes 3 segments (± apex) and the LCX territory includes 5 segments (± apex). There are a few caveats:

1.

The apex is shared by all three vessels; for example, RCA disease causing an inferior defect that can involve the apex. Such a defect should not be called RCA plus LAD disease just because the apex is involved.
2.

The apex is not a point but a segment just like the other segments. But when it is just a dimple, avoid labeling it as a defect as this is due to apical thinning; it is a common source of “false positive” scans.
3.

There is an overlap between the vascular territories. Thus, some LAD defects can extend to the anterolateral wall and some LCX defects can extend to the base of the anterior wall. The same applies for the overlap between the LAD and RCA (the inferior and the inferoseptal segments) and between the LCX and the RCA (the inferior and inferolateral segments). As a general rule, if the overlap involves more than half of a different vascular territory, call it two-vessel disease, otherwise it is an extension.
4.

In LAD disease, the defect is more severe distally than at the base (one exception is in patients with prior CABG; see Chapter 8 ). This is another hint as to whether the LAD is involved in a patient with an LCX abnormality and only a basal anterior wall abnormality.
5.

In patients with disease limited to the diagonal branch of the LAD (not uncommon especially after LAD stenting, if these branches are jailed; see Chapter 8 ), the defect is localized to distal and mid anterolateral segments (between 12 and 2 o’clock). These defects sometimes are seen only in the SA tomograms. Otherwise, in general, we prefer to see the defect in at least two orthogonal planes.
6.

Isolated small septal defects are rare but can be due to a septal branch being jailed after stenting of the LAD or to muscular bridging. If the defect involves most of the septum, however, pay careful attention to the inferior wall. Disease involving both the LAD and RCA tends to produce inferior and septal perfusion abnormalities (i.e., they spare the anterior wall) ( Figure 2-5 ). The reason is that the septum has a dual blood supply from the LAD and RCA and, therefore, is more severely affected than the anterior wall. For the same reason, disease of the LAD and LCX tends to involve the anterior and lateral walls but spare the septum ( Figure 2-6 ). This pattern may potentially represent left main disease as there is no other specific pattern for LM disease. Always be aware of this possibility as ostial LM stenosis could be easily missed on coronary angiography.

Figure 2-5

Perfusion abnormality in a patient with LAD and RCA disease. A, There is a perfusion abnormality involving the septum and inferior wall; the anterior wall abnormality is less conspicuous than the septal abnormality. The septum has a dual blood supply from both these vessels and is therefore more affected than the anterior wall. The TID of 1.3 means the summed LV size is 30% larder on the stress than on the rest images. B, The polar maps are shown.

Figure 2-6

Perfusion abnormality in a patient with LAD and LCX disease. There is a perfusion abnormality involving the anterior wall and lateral wall but sparing the septum. The septum is less affected because the septal perforator branches from the posterior descending branch of the RCA supply the posterior septum.
7.

There is considerable variability in the extent of perfusion abnormalities in patients with CAD that cannot be entirely explained by variations in coronary anatomy ( Figure 2-7 ). It is likely the single most important reason that perfusion imaging provides such powerful prognostic information. In patients undergoing coronary interventions, injection of the tracer during transient coronary occlusion by balloon inflation shows a similar variability in the extent of myocardium at risk ( Figure 2-8 ). It should be noted that although catheter-based methods can be used to assess coronary physiology (whether a stenosis is significant), they do not provide information on the area at risk or the extent of perfusion defects.

Figure 2-7

Variation in the size of the perfusion abnormality during exercise in patients with isolated one-, two-, or three-vessel disease. The defect size was measured by planimetry of planar images.

(Modified from Circulation 1983;67:983.)

Figure 2-8

Variation in the size of a perfusion defect in two patients ( A and B ) with LCX disease. The tracer was injected during transient balloon occlusion (BO) in the cardiac catheterization laboratory during planned angioplasty for severe stenosis. The rest images (R) were normal in both patients. There is a marked difference in the size of the area at risk in these two patients despite comparable coronary angiographic findings.
8.

Attenuation-corrected images should be reviewed side-by-side with the uncorrected images. Obviously, one needs to assure that the quality of the transmission images is acceptable regardless of whether a line source or CT is used for correction. We do not recommend reading only the corrected images at this stage.
9.

The size and severity of a perfusion abnormality can be assessed visually. Size is graded as small, moderate, or large, and severity as mild, moderate, or severe. One needs to adjust for the severity of a defect based on the presence of attenuation artifacts in the same segments. For example, the inferior defect may look more severe in men because the true abnormality is superimposed on an attenuation artifact in the same location. The quantitative methods are more robust, however, than visual methods for assessing size and severity (see below).
10.

The nature of the perfusion abnormality is defined as reversible (ischemia), fixed (scar), or mixed (subendocardial scar). As will be discussed in Chapter 15 , this definition of scar is not precise. The assessment of reversibility, either presence or absence, is more difficult than appreciated, and, in general, reversibility is under-reported. Again some caveats:
- A.
  
  The color scale is not as reliable as the gray scale in detecting reversibility of a mild nature. It requires more training, but if perfected it can be just as useful. So, use the gray scale , at least in the beginning.
- B.
  
  Perfusion defects in some areas that are prone to attenuation artifacts will never appear completely normal on the rest study. Therefore, we tend to label such segments as showing “partial redistribution,” implying there is an element of scarring. As a rule, if the motion on the gated images is normal, call it reversible.
- C.
  
  The reason why segments appear fixed or reversible is not entirely related to the segment in question but rather to the remote and normal zones. If the hyperemic MBF is not as high in the normal area (either because the stress is not adequate, as with submaximal exercise or microvascular disease), then the relative MBF ratio (maximum/rest in the abnormal zone divided by maximum/rest in the normal zone) is not big enough and the defect appears fixed. That is why it is important to pay attention to the type of stress when comparing serial tests in the same patient in which one study shows reversible defects and the second study shows fixed defects.