Correlation between cardiology fellows’ year of training and HHU interpretation:
To determine whether the number of years in training had any significant influence on the HHU interpretation, all the HHU measurements were stratified for either first year cardiology fellows (fellow in training [FIT] = 1) or those in second year and above (FIT ≥ 2). The correlation coefficient for LVEF between the standard TTE and HHU were 0.64 (95% CI: 0.41 – 0.79) and 0.80 (CI: 0.64 – 0.89) for FIT = 1 and FIT ≥ 2, respectively (Fig 2A-B). A similarly significant increase in the correlation coefficient was found when we compared the WMSI acquired from standard TTE against HHU obtained by cardiology fellows, with r = 0.70 (CI: 0.51 – 0.82) for FIT = 1 and r = 0.87 (CI: 0.76 – 0.93) for FIT ≥ 2 (Fig 2C-D). The higher correlation in FIT ≥ 2 was also reflected in recognizing WMA when comparing each wall segment separately (Table 2).
The influence of number of years in training was also evident in other factors reported by the fellows. These included higher scoring for visualization of segmental endocardial borders (1.45 ± 0.55 in FIT ≥ 2 versus 1.39 ± 0.61 in FIT = 1, p=0.34), higher scores for level of confidence interpreting the studies (1.47 ± 0.60 in FIT ≥ 2 versus 1.13 ± 0.68 in FIT = 1, p<0.01), and a higher percentage of fellows-reported clinical decisions being affected by HHU results: 42% in FIT ≥ 2 group versus 24% in FIT =1, p<0.05) (Table 3).
Discussion :
Among AMI activation cases, there is heterogeneity and uncertainty impacting whether an immediate coronary angiogram and PCI is warranted based on an equivocal ECG and the patient’s symptoms. This is especially challenging in most training centers where cardiology fellows in training are the first contact provider. These may be situations in which HHU can be of significant help in clinical decision making. Such areas where HHU can influence next steps are estimating LVEF, evaluating for pericardial pathology, and identifying the locations of abnormalities in cardiac wall motion, which will help in localizing the location of coronary artery occlusions. These findings may eventually translate into improved diagnosis and better patient care. Thus, it is imperative to determine 1) how reliable the information acquired by HHU is when compared against the gold standard TTE, and 2) whether HHU is associated with a potential change in care in terms of the need for angiography as well as the urgency of angiography, and whether it influences cardiovascular trainees to help make informed decisions.
In this study we found that there is a high correlation for LV function and overall wall-motion abnormality assessment between HHU and standard TTE (concordance correlation coefficient: 0.71 and 0.75) when performed and interpreted by cardiovascular fellows. Furthermore, we found that the absence of WMA on HHU was associated with an ability to defer or delay angiography, which can have important implications for the evaluation and triage of patients presenting with suspected STEMI.
The high degree of correlation between HHU and TTE that we demonstrated in this study can be supported by the fact that cardiology fellows could acquire relatively high-quality images and demonstrate a high level of confidence in their interpretation of the studies. We also found that the accuracy of data acquired by HHU had a positive linear correlation with years of training amongst fellows in detecting wall motion abnormalities, visualizing endocardial borders, and also with their self-reported level of confidence. As a result, second- and third-year fellows relied more heavily on the results of their HHU for making the next clinical decisions in terms of the need for and timing of cardiac catheterization.
Prior work has demonstrated good correlation between HHU and standard TTE for LVEF when experienced users perform the HHU exams (13-20). Lieboet al . in their cross-sectional study of 97 patients concluded that the rapid acquisition of images by skilled ultrasonographers who use pocket mobile echocardiography yields accurate assessments of ejection fraction and some, but not all, cardiac structures in many patients (13). In another similar study, Prinz et al ., using handheld ultrasound, showed that in relation to the basic assessment of cardiac morphology and function, the interpretation by experienced echocardiographers of images obtained using handheld echocardiographic devices demonstrated a moderate to very good correlation with standard echocardiography (r > 0.8, p < 0.01 for wall motion abnormalities, and r > 0.6, p < 0.01 for LVEF assessments) (14).
When compared with existing literature, our study shows comparable findings for correlation between HHU and standard TTE measurements, particularly for LVEF (15-16). Furthermore, there have been similar variable discrepancies in wall motion abnormalities between HHU and TTE in prior work, even in the hands of experienced users (17). This can be explained by lower image resolution and the limited amount of time users often spend optimizing images with HHU. Of note, the mean endocardial visibility score reported in this study (1.41 ± 0.58) was almost similar to what previously reported (1.60 ± 0.50) (18).
In addition to demonstrating that HHU performed by cardiology fellows is efficacious in assessment of suspected STEMI, we highlighted the possible ability of key HHU findings to affect care in this population. Patients with suspected STEMI are a heterogenous group, with a broad differential diagnosis (21). In addition to history, physical exam, and laboratory values, HHU may have the ability to further risk stratify patients. This may occur due to the finding of a competing explanatory diagnosis, such as pericarditis, pericardial effusion, pulmonary embolism, etc. However, even in the absence of alternative findings, lack of WMAs on presentation may not be clinically suggestive of a STEMI (22). We extend similar findings to HHUs performed by cardiology fellows at the point of care in the emergency room. The ability to risk stratify patients is potentially important due to ability to defer invasive angiography, which has been shown to be safe among patients with non-ST elevation MI (23) or consider alternative diagnoses. The ability to safely determine the timing of care may be especially important in the off-hours or in the setting of limited catheterization lab space or requirement for decision making regarding patient transfer to a STEMI center, especially among smaller or community hospitals. The use of HHU for risk stratification for suspected STEMI, as well as other patient populations presenting with suspected AMI, deserves further evaluation in larger studies.
An important consideration in using HHU for decision making in the acute setting is that operators should be familiar with these tools and, more importantly, how to interpret the imaging results. Specific training in this area, therefore, is central to effective use and improved outcomes.