Prediction of Post-revascularization Ejection Fraction in Patients with CoronoryArtery Disease Using Cavity-to-Myocardial Ratio of Thallium Reinjection Image (Multicenter Trial)

Background: We reported the high correlation between cavity-to-myocardial (C/M) count ratio at stress and rest thallium SPECT, and stress-rest EF calculated by MUGA test, this was confirmed by others. This correlation was explained partially by the functional mass. On the other hand, two important prognostic parameters should be considered before any revascularization technique: (1) Identification of viable myocardium and its amount, (2) Prediction of EF improvement post revascularization. Aim of the study: Correlating EF (C/M) on RD and RI image (EFRD & EFRI) image to actual EF (prevascularization EF1) and 1 year post revascularization EF2. Patients and methods: 78 patients with CAD (68 males and 10 females with mean age of 54.2+9 years) had been subjected to: (1) St-RD-RI thallium SPECT with assessment of reversible or fixed perfusion defects and calculation of C/M and consequently the EFC/M at the three settings. (2) Assessment of EF by MUGA at rest pre and 1 year post revascularization EF1 & EF2 respectively. These patients had been subjected to revascularization either by PTCA and stent (23/78 i.e., 29.5%) or by CABG (55/78, i.e., 70.5%). Results: Out of the 1560 myocardial segments (20 segments × 78 patients), 780 (50%) segments had abnormal resting wall motion. 441/780 (56.5%) of these segments were either of normal thallium uptake or with reversible perfusion defects while the rest (43.5%) showed fixed defects. 233/441 (52.8%) of those normal uptake or reversible segments showed recovery of wall motion post revascularization (PRV) while only 29/339 (15.1%) showed similar improvements. EFRI was found higher than EFRD in 44/78 of patients, no change in 23/78 patients and worsened in 11/78 patients with total agreements of 63/78 (80.8%) with EF2. On the other hand, EFRD was matched with EF1 in 64/78 of patients. 30/64 (46.9%) showed higher EF2, 23/64 (35.9%) showed similar EF2 while 11/64 (17.2%) showed lower EF2. The rest of cases 14/78 showed mismatch between EFRD and EF1 with higher values of EFRD. These patients still had higher values of EFRI and EF2 than EFRD. Conclusion: (1) Mismatch between EFRD and EF1 is an indication of presence of stunning myocardium and of good prognosis. (2) EFRI can be used to predict EF2 and so helps on selecting patients who can benefit from revascularization.


Introduction
In patients with coronary artery disease (CAD), chronically dysfunctional left ventricular myocardium which recovers contractile function after revascularization has been defined as "hibernating" [1].
Presumably, the beneficial effects of revascularization result from restoring blood supply to dysfunctional but viable myocardial regions, with subsequent improvement in regional and global left ventricular function [2]. The proportion of viable myocardium is a critical factor in determining the likelihood of functional recovery after successful revascularization [3]. This fact is a reflection to the previously reported strong correlation between cavity-to-myocardial count ratio (C/M) of stress and redistribution thallium SPECT and stress and rest E.F. as measured by MUGA study [4]. This was explained by the facts that: It represents the actual functional mass [1], and the attenuating power of blood in the cardiac cavity to the cavity radioactivity [2,[4][5].
Accordingly, the aim of the current study is to correlate ejection fraction (EF) as predicted by C/M on redistribution (EF RD ) and reinjection (EF RI ) at rest by MUGA prevascularization (EF 1 ) and postrevascularization (EF 2 ).

Study population
Seventy eight patients with CAD liable to revascularization were subjected to the current study in Cairo university hospital, together with the National Heart Institute (Egypt) and Alhada Armed Force hospital, (Taif, Saudia Arabia). They were 68 males and 10 females (6.8:1) and had 54.2+9.1 years old (Table 1). Patients were eligible for inclusion: If they had severe regional dysfunction in the anatomic distribution of a significantly narrowed or occluded epicardial artery as determined by contrast left ventriculography and if the coronary arteries supplying the area of dysfunction were suitable for coronary artery bypass grafting (CABG) or percutaneous transluminal coronary angioplasty (PTCA) [1,2]. Exclusions criteria included: Unstable angina, or myocardial infarction between revascularization and the scintigraphic studies or after intervention, patients with 3-4+ (moderate to severe) mitral regurgitation [1,2]. Each patient gave informed consent to the study protocol, which was approved by the ethical committees of the participating hospitals.

Methods
All patients underwent exercise-redistribution-reinjection thallium scintigraphy, resting multigated study (MUGA) pre and post revascularization and coronary angiography. Subsequently, 23 patients underwent PTCA while 55 patients underwent CABG. The decision concerning the choice of treatment was left to the referring cardiologists and was not based on the results of the viability studies.

Coronary angiography
Selective coronary angiography and contrast left ventriculography were performed from the femoral approach. Significant coronary artery disease was defined as >70% luminal diameter stenosis in any major coronary branch. Accordingly, 10 patients had one vessel disease, 25 patients had two vessel diseases and 43 patients had three vessel disease.

Thallium SPECT protocol
All patients underwent symptom-limited treadmill exercise testing with a standardized multistage protocol with continuous monitoring of heart rate and rhythm, blood pressure and electrocardiography. Criteria for interrupting the exercise test, included age-predicted maximal heart rate, severe angina, development of marked ST-segment depression, appearance of frequent or complex ventricular arrhythmia, hypotension and exercise-limited dypsnea. Three millicurie (mCi) of thallium was injected IV at peak exercise and continuation of exercise for one minute was done after that. Acquisition of tomographic images was performed within 5 minutes after thallium injection as reported previously [6]. Four hours later redistribution images (RD) were obtained. This was followed by reinjection (RI) of 1.5 mCi of thallium then 1 hour later another SPECT set was done.

Image processing
Comprehensive semiquantitative perfusion defect analysis using 20segment visual analysis has been used. The 20-segment scoring system is based on three-short axis slices (distal, mid, and basal) to represent the entire left ventricle, with the apex represented by two segments visualized in a mid-vertical long axis image. Each of the 20 segments has a distinct name (number) ( Figure 1) [8]. Each segment is visually scored as follows: 0 = normal, 1 = slight reduction of uptake (equivocal); 2 = moderate reduction of uptake (usually implies a significant abnormality); 3 = severe reduction of uptake; 4 = absence of radioactive uptake. Then the summed stress score (SSS) is defined as the sum of the stress scores, the summed rest score (SRS) is the sum of the rest or redistribution scores. The summed difference score (SDS), measuring the degree of reversibility, is defined as the difference between the SSS and the SRS.
For C/M ratio a short axis at the mid-left ventricle (LV) level was used in both initial and delayed images. Two regions of interest had been drawn: one delineating the inner border to represent the cavity©, and the other delineating the outer border to represent myocardium (M). Then, the C/M count ratio will be calculated and E.F. will be derived from the previously mentioned regression equation [5]: E.F. = -4.89 + (C/M×77.4), + 6.9 [5] Multigated study (MUGA) Gated blood pool cardiac scintigraphy was performed 1 week later after myocardial perfusion study to assess left ventricular ejection fraction at rest using red blood cells labeled in vivo with 20-25 mCi 99mTc. Imaging was accomplished using a high sensitivity, parallel hole collimator. Patients were imaged supine in best left anterior oblique view.
The camera was set at 140 KeV photopeak +20%, with 64×64 and 24 frames. Edge detection of left ventricle had been done depending on the phase image. Left ventricular ejection fraction was derived by computer analysis of the scintigraphic data with the lower normal limit of resting EF is 50%.

Statistical analysis
Data were analyzed using SPSS (version 15.0. SPSS Inc., Chicago, IL). Data are presented as mean +standard deviation (SD). Differences in the mean of the different groups were performed using unpaired ttest.
Differences in the frequency of patients of the different groups had been performed using X 2 analysis. Linear regression analysis was done to see if EF RD and EF RI were correlated strongly with EF 1 and EF 2 P < 0.05 was considered significant statistically [9].    Table 3 revealed 63/78 patients agreement, i.e., 80.8% agreement with 37 patients predicted to show this actual improvement, 17 patients expected to show no change in E.F. post-revascularization (EF 2 ) and 9 patients expected to show worsening in E.F. These findings were furtherly analyzed in Table 3 where patients with 1 VD showed the highest agreement 9/10 (90%), patients with 2 VD showed 21/25 (84%) agreement while patients with 3 VD showed lowest agreement 33/44 (75%). Total agreement = 63/78 (80.8%), P < 0.01 Table 3: Degree of agreement between the EF RI and EF 2 . EF RI : Ejection fraction as predicted by calculated by cavity-to-myocardial count ratio in reinjection image; EF 2 : Actual ejection fraction postrevascularization. Figure 2 revealed the mean +SD of patients with improved EF 2 compared to that predicted from EF RD & EF RI . It is significantly higher than that predicted from EF RD (45+10 vs. 34.9+8.6 with P < 0.01) but insignificantly different from that predicted by EF RI (45+10 vs. 47.1+12.1with P > 0.05). Similarly, patients with no change in E.F. where the mean value of EF was the same as predicted from EF RD (36.7+11.8) compared to that predicted from EF RI (36.8+14.5) and the actual postrevascularization EF 2 (37.9+12.3). On the other hand, patients with predicted worsening EF showed significantly higher EF RD (34.7+10.2) than EF RI (30.1+ 8.3) which is marginally higher than that actual postrevascularization EF 2 (28.4+7.4).  Figure 3a and 3b revealed strong correlation between EF 1 and EF RD and between EF 2 and EF RI . Patients had been divided according to the agreement between EF 1 and EF RD    On the other hand, Figure 4 showed flow chart including both matched (M) and mismatched (N) groups and the role of EF RI in prediction of EF 2 : where improved EF 2 as predicted by EF RI in 30/64 (46.9%), no change in EF 2 as predicted by EF RI 23/64 (35.9%) and worsening in EF 2 as predicted in 11/64 (17.2%). On the other hand, in group (N) showed improved EF 2 in all the cases 14/14 (100%). These patients still had higher values of EF RI and EF 2 than EF RD .  Table 4

Discussion
The presents study demonstrates that EF RI is a useful parameter to stratify patients that will be subjected to revascularization and could be used to predict the degree of functional recovery postrevascularization.
In addition, the current study revealed that mismatching between EF 1 and EF RD is an indication of presence of dysfunction viable myocardium i.e., stunning myocardium. While the mismatch between the EF RD and EF RI is an indication of hibernating myocardium.

Pre vascularization EF 1 and predicting EF from C/M ratio
Several reports published a strong correlation between C/M count ratio and actual EF even in poor EF [4,5,[10][11][12][13][14][15] depending on the following rational: The amount of myocardial mass prependicular to the crystal surface of the gamma camera [1] regional thallium accumulation [2], and the effect of tissue attenuation of radiation from the distant myocardial wall by the ventricular blood pool which is decreased in cases of good left ventricular ejection fraction (LVEF) in such way that attenuation is reduced and more radiation is found in the cavity [3]. This was supported by Hachamovich et al. (1998) [16] who stated that SSS is the perfusion analogue of the peak EF and similarly SRS for resting EF. Similarly, Bax et al. (2001) [17] reported strong correlation between the number of viable segments on SPECT and improvement in LVEF after revascularization. Thus, EF RD represents the sum of the actually and potentially functioning mass. This explains the presence of the mismatched group (N group) which included 14 patients where actual EF 1 was 31.4+10 while the predictable EF RD was 36.4+8.3 with P < 0.05. This means that there is potential functioning tissue (stunned) post revascularization.
In patients undergoing biopsy during surgical revascularization, it has been demonstrated that the level of thallium uptake after reinjection closely correlates to the amount of interstitial fibrosis [18]. On the other hand, several authors reported using quantitative analysis of thallium uptake a cutoff uptake value to separate reversibly dysfunctional from irreversibly dysfunctional myocardium with linear correlation between the functional recovery in a dysfunctional segment and thallium uptake. The shape of this correlation is such that it is possible to identify a cutoff point (usually below 30% to 40% of maximal uptake) under which recovery of function almost never occurs, but above which chances of recovery increase in parallel with the amount of tracer uptake [19]. All these lead to the other area of mismatch found between the EF RD and EF RI which could be attributed by the more viable segments that had been detected by RI technique and so more regional thallium accumulation scattering their radioactivity into the cavity to reduce the difference between the count in the cavity and myocardium. This is more apparent by Table 4, which showed that SRS in patients who showed postrevascularization improvement was 0.53+0.09 and also supported by work of Qureshi et al. (1997) who reported high thallium uptake in segments that exhibited ultimate recovery of function (68+12% to 81+11%, n = 42, P < 0.01) than in those that did not (52+19% vs. postrevascularization resting uptake of 55+21%, n = 106; P > 0.05) [20]. This goes on line with Bax et al. [17] report at 2001, they found that a substantial amount of viable myocardium (31% of the left ventricle) needs to be present to result in improvement in LVEF. They again reported division of their patients into three groups according to the amount of viable segments with linear increase in EF with more viable segments.

Degree of agreement between EF 2 and EF RI
The current study could predict the functional outcome postrevascularization accurately in 63/78 patients, i.e., in ~81% of cases. This agreement is more in one and two vessel disease cases but less in three vessel disease cases which could be attributed to the technical problems that would be met during CABG and unpredictable myocardial loss during operations.

Conclusions
Mismatch between EF RD and EF 1 is an indication of presence of stunning myocardium and of good prognosis.
EF RI can be used to predict EF 2 and also helps on selecting patients who can benefit from revascularization.

Study limitations
The current study included small number of patients (78) which on further subdivisions decreases the number for further statistical analysis. In addition, the small number could not separate impact of PTCA versus CABG on post-revascularization ejection fraction (EF 2 ). Further research work with larger number is needed.