A Doppler Echocardiographic Pulmonary

Flow Marker of Massive or Submassive

Acute Pulmonary Embolus
Luis Afonso, MD, Aditya Sood, MD, Emmanuel Akintoye, MD, John Gorcsan, III, MD,
Mobeen Ur Rehman, MD, Kartik Kumar, MD, Arshad Javed, MD, Anupama Kottam, MD, Shaun Cardozo, MD,
Manmohan Singh, MD, Mohan Palla, MD, Tomo Ando, MD, Oluwole Adegbala, MD, Mohamed Shokr, MD,
and Alexandros Briasoulis, MD, PhD, Detroit, Michigan; St. Louis, Missouri; Englewood, New Jersey;
and Iowa City, Iowa

Background: To date, echocardiography has not gained acceptance as an alternative imaging modality for the
detection of massive pulmonary embolism (MPE) or submassive pulmonary embolism (SMPE). The objective
of this study was to explore the clinical utility of early systolic notching (ESN) of the right ventricular outflow
tract (RVOT) pulsed-wave Doppler envelope in the detection of MPE or SMPE.

Methods: Two hundred seventy-seven patients (mean age, 56 6 16 years; 52% women), without known pul-
monary hypertension, who underwent contrast computed tomographic angiography for suspected pulmonary
embolism (PE) and underwent echocardiography were retrospectively studied. Extent of PE was categorized
using standard criteria. ESN identified from pulsed-wave spectral Doppler interrogation of the RVOT was
analyzed, as were other echocardiography parameters such as McConnell’s sign, the ‘‘60/60’’ sign, and ac-
celeration and deceleration times of the RVOT Doppler signal. Analysis was conducted using probability sta-
tistics and receiver operating characteristic curve analysis.

Results: Of the 277 patients studied, 100 (44%) had MPE or SMPE, 87 (38%) had subsegmental PE, and 90
(39%) did not have PE. ESN was observed in 92% of patients with MPE or SMPE, 2% with subsegmental
PE, and in no patients without PE. Interobserver assessment of early systolic notching demonstrated 97%
agreement (k = 0.93, P < .001). Compared with more widely recognized echocardiographic parameters, the
area under the receiver operating characteristic curve (AUC) of 0.96 (95% CI, 0.92–0.98) for ESN was superior
to that for McConnell’s sign (AUC, 0.75; 95% CI, 0.68–0.80), the 60/60 sign (AUC, 0.74; 95% CI, 0.68–0.79),
and RVOT acceleration time # 87 msec (AUC, 0.84; 95% CI, 0.79–0.88), as well as other study Doppler vari-
ables, in patients with computed tomography–confirmed MPE or SMPE.

Conclusions: The pulmonary Doppler flow pattern of ESN appears to be a promising noninvasive sign
observed frequently in patients with MPE or SMPE. Future prospective study to ascertain diagnostic utility
in a broader population is warranted. (J Am Soc Echocardiogr 2019;-:---.)

Keywords: Echocardiography, High-risk pulmonary embolism, Doppler notching, Pulmonary embolism

Pulmonary embolism (PE) is a common and potentially lethal medical people in the United States annually.1,2 The reported in-hospital mor-
condition that accounts for the hospitalization or death of >250,000 tality of patients with massive PE (MPE) varies from 25% to 50%,
whereas mortality for submassive PE (SMPE; defined as the presence
of right ventricular [RV] dysfunction without systemic hypotension)
From the Division of Cardiology, Wayne State University, Detroit Medical Center, ranges from 3% to 15%, and that associated with low-risk or subseg-
Detroit, Michigan (L.A., A.S., M.U.R., K.K., A.J., A.K., S.C., M. Singh, M.P., T.A., mental PE (SSPE) is <5%.3 Diagnostic and treatment algorithms for
M. Shokr); the Division of Cardiology Medicine, Washington University, St. PE have been developed and endorsed by societies such as the
Louis, Missouri (J.G.); the Department of Internal Medicine, Englewood Hospital European Society of Cardiology and the American Heart
and Medical Center, Seton Hall University–Hackensack Meridian School of Association.3,4
Medicine, Englewood, New Jersey (O.A.); and the Division of Cardiovascular Multidetector row computed tomographic angiography (CTA)
Medicine, University of Iowa Hospitals and Clinics, Iowa City, Iowa (E.A., A.B.). and, to a lesser extent, ventilation-perfusion lung scanning are tradi-
Conflicts of Interest: None. tionally the most frequently used noninvasive imaging procedures
Reprint requests: Luis Afonso, MD, Wayne State University, Division of Cardiology, for the diagnosis of acute PE. Given the high mortality associated
Detroit Medical Center, Detroit, MI 48201 (E-mail: lafonso@med.wayne.edu). with undiagnosed PE, it is imperative to risk-stratify acute PE expedi-
0894-7317/$36.00 tiously to ensure prompt initiation of anticoagulant therapy in appro-
Copyright 2019 by the American Society of Echocardiography. priate cases and, conversely, refrain from initiating such therapy when
https://doi.org/10.1016/j.echo.2019.03.004 not indicated.
1
2 Afonso et al Journal of the American Society of Echocardiography
- 2019

Abbreviations
The role of echocardiography the American Heart Association scientific statement, published in
in the workup of suspected PE 2011.3 MPE was defined as acute PE with sustained hypotension (sys-
AT = Acceleration time has historically been supportive tolic blood pressure < 90 mm Hg for $15 min or requiring inotropic
CTA = Computed and primarily reserved for the support, not due to a cause other than PE, such as arrhythmia, hypo-
tomographic angiography detection of RV strain or volemia, sepsis, or left ventricular dysfunction), pulselessness, or
dysfunction in unstable patients. persistent profound bradycardia (heart rate < 40 beats/min with signs
ESN = Early systolic notching In this context, a few echocardio- or symptoms of shock). SMPE was defined as acute PE without sys-
MPE = Massive pulmonary graphic parameters, such as temic hypotension (systolic blood pressure $ 90 mm Hg) but with
embolism McConnell’s sign, the ‘‘60/60’’ RV dilation, defined as RV diameter divided by left ventricular diam-
sign, and RV dysfunction, have eter > 0.9 on CTA (in the four-chamber view). SSPE was defined as
PE = Pulmonary embolism
been studied, but their clinical acute PE and the absence of clinical markers of adverse prognosis
PVR = Pulmonary vascular utility has remained limited.5,6 that define MPE or SMPE.
resistance We recently reported prelimi- Echocardiographic variables for all patients were analyzed from
PW = Pulsed-wave nary pilot data on the potential transthoracic echocardiograms obtained with the Philips iE33, Cx-
value of commonly used as well 50, and EPIQ (Philips Medical Systems, Andover, MA) or the GE
ROC = Receiver operating as less well recognized RVoutflow E9 Ultrasound (GE Medical Systems, Milwaukee, WI) systems. PW
characteristic tract (RVOT) Doppler variables, in Doppler interrogation of the RVOT was performed from the paraster-
RV = Right ventricular patients with MPE or SMPE.7 The nal short-axis view at the level of the aortic valve or from the subcostal
present study expands on these short-axis view, with the sample volume placed approximately 0.5 cm
RVOT = Right ventricular
observations and provides a proximal to the pulmonic valve. The ‘‘early systolic notching’’ (ESN)
outflow tract
head-to-head comparative assess- pattern (spike and dome morphology) was visually assessed and
SMPE = Submassive ment on the performance of deemed present if the Doppler envelope exhibited a narrow peaked
pulmonary embolism several echocardiographic vari- initial wave (spike) with early deceleration of the RVOTenvelope pro-
SSPE = Subsegmental ables in a cohort of patients with ducing a sharp notch within the first half of systole (i.e., notch location
pulmonary embolism MPE or SMPE identified within the initial 50% of ejection, as estimated using the online time-
following CTA for suspected PE. caliper tool), followed by a second Doppler wave (dome) that was
more curvilinear in appearance (Figure 1). Similarly, midsystolic
notching was defined as a distinct notch falling within the second
METHODS half of the systolic ejection period or, if the nadir occurred closer to
the end of ejection, dividing the flow profile into two distinct peaks.
This retrospective investigation was approved by the Wayne State Notch morphology was best appreciated when the Doppler beam
University institutional review board. All cases of contrast CTA per- was aligned parallel to RVOT outflow, with the PW sample volume
formed for suspected PE at our tertiary care institution between placed at the appropriate location at sweep speeds of 50 to
2015 and 2017 were reviewed. For inclusion in this study, all patients 100 mm/sec. (See additional illustrative case examples shown in
were required to have undergone transthoracic echocardiography Figures 1-3.) Care must be exercised not to conflate the ‘‘opening
within 48 hours of computed tomographic angiographic diagnosis click’’ of the pulmonic valve with the systolic notch, a scenario that
of PE. Only patients with complete echocardiographic examinations, can occur if the PW sample gate is placed too close to the
including interpretable pulsed-wave (PW) Doppler signals across the pulmonic valve.
RVOT, measurable tricuspid regurgitation jet Doppler signals, and Ejection time was measured in milliseconds from the beginning to
discernible endocardial border definition of the right ventricle, were the end of the RVOT Doppler envelope. Acceleration time (AT) was
included. Technically suboptimal-quality PW Doppler studies with measured in milliseconds as the time to peak velocity of the RVOT
sample volumes placed at or distal to the pulmonary valve or poorly envelope measured from the beginning of ejection. At least 3 cardiac
aligned to the direction of RVOT flow were excluded. Similarly, pa- cycles were measured and averaged for analysis. Deceleration Time
tients with more than moderate valvular heart disease, known history was measured from the peak Doppler velocity to the end of ejection.
of PE, established chronic thromboembolic pulmonary hypertension, Acceleration and Deceleration slopes were estimated by deploying
and preexisting pulmonary hypertension, regardless of etiology, were the online slope tool using the same time points described for respec-
excluded from this study. tive time measurements. Spike (S) and Dome (D) velocities were
measured as the peak Doppler velocities of these waveforms, respec-
tively. The AT/ejection time ratio was derived from values listed
Imaging above. Notch time represents the duration of the Doppler notch
Contrast-enhanced computed tomographic scans were performed measured in milliseconds, as shown in Figure 1. RVOT Doppler
on multidetector Toshiba 64-slice scanners (Toshiba Medical Systems, velocity-time integral was obtained in the standard manner by tracing
Tokyo, Japan) using a standardized PE timing protocol adopted across the systolic RVOT PW Doppler envelope. RV size was measured as
our hospital systems. Contrast dye used was Isovue 300 or Isovue the basal RV dimension in the four-chamber apical view. The 60/
370 (Bracco, Milan, Italy), tailored to the patient’s renal function. 60 sign was deemed present if RVOT AT was < 60 msec in the pres-
RV strain, dilation, or dysfunction was deemed present if four- ence of a tricuspid peak systolic gradient > 30 mm Hg but < 60 mm
chamber RV diameter divided by left ventricular diameter was Hg.8 Regional pattern of RV dysfunction consistent with McConnell’s
> 0.9 on CTA. The diagnosis of PE was confirmed when thromboem- sign was defined as akinesia of the mid free wall visualized along with
boli were visualized in an at least segmental pulmonary artery on preserved apical contractility.9 All other variables were obtained in
contrast-enhanced multidetector computed tomography. Patients accordance with current American Society of Echocardiography
with PE were stratified according to definitions recommended by guidelines.10
Journal of the American Society of Echocardiography Afonso et al 3
Volume - Number -

HIGHLIGHTS

 ESN reliably identified patients with MPE/SMPE but not those

with SSPE
 ESN demonstrated superior predictive ability with a high nega-
tive predictive value
 Future prospective study to ascertain diagnostic utility in a
broader population is warranted

To assess interobserver variability, two authors (A.S. and Mo.S.)

who were blinded to the PE diagnoses of these patients indepen-
dently reviewed their echocardiograms for the presence of ESN in
a total of 30 randomly selected samples, and agreement between
the two authors was analyzed using the k coefficient.
Figure 1 RVOT Doppler tracing in a patient with MPE, illus-
trating the characteristic ESN pattern and measurement meth-
Statistical Analysis odology used for various other Doppler parameters analyzed
Baseline characteristics were compared among the three groups us- in this study. DT, Deceleration time; D vel, dome velocity; S
ing the c2 test for categorical variables and the Kruskal-Wallis test for vel, spike velocity.
continuous variables. Thereafter, we evaluated the utility of the spike-
and-dome pattern in three steps. First, using CTA-confirmed diagnosis PE who met our selection criteria. In summary, a total of 277 patients
of MPE or SMPE as the gold standard, we evaluated the utility of the (mean age 56 6 16 years; 48% men; 100 [44%] with MPE or SMPE,
spike-and-dome pattern in diagnosing MPE or SMPE using probabil- 87 [38%] with SSPE, and 90 [39%] without PE) were studied.
ity statistics and receiver operating characteristic (ROC) curve anal- Differences in baseline characteristics and hemodynamics among
ysis. Second, we evaluated similar diagnostic utility for prespecified the three groups are presented in Table 1, and details of various echo-
echocardiographic variables, including notch-related parameters, RV cardiographic parameters evaluated in this study are presented in
dilation, and tricuspid regurgitation velocity. Optimal cutoffs for these Table 2.
echocardiographic parameters were derived as the values that mini- ESN was observed in 92% of patients with MPE or SMPE, 2% of
mized the square of the difference between sensitivity and specificity. those with SSPE, and in no patients without PE. There was good
The diagnostic performance of these parameters was then assessed interobserver agreement in the identification of ESN, with 96.7%
using probability statistics and ROC analysis. Third, we identified pa- agreement (k = 0.93, P < .001). Among patients with SSPE, systolic
rameters with potential incremental benefit over the spike-and-dome notching was predominantly midsystolic (19%) and less likely early
pattern using a model-based approach that retained the indicator for systolic (2%). However, among patients with MPE or SMPE, systolic
spike and dome in the model while selecting other variables using a notching was predominantly early systolic (92%) and less likely mid-
backward stepwise selection iteration (P for inclusion = .10, P for systolic (1%). No systolic notching was observed in all control sub-
exclusion = .20). The incremental benefit of combining the finally jects (100%), 79% of patients with SSPE, and 7% of patients with
selected parameters to the spike-and-dome pattern were then evalu- MPE or SMPE. No unique characteristics were identified in the
ated using ROC analysis and probability statistics. seven patients with MPE or SMPE who did not have the early
Secondary analysis for predictive utility of the spike-and-dome notching pattern; of these, five patients had intermittent, poorly
pattern for SSPE was performed after excluding patients with MPE discernible ESN or midsystolic notching, and two patients had a
or SMPE. characteristic triphasic Doppler pattern (see Figures 2B and 3B).
All analyses were performed using Stata version 14 (StataCorp, The precise underlying mechanism for the triphasic pattern is un-
College Station, TX) with a two-tailed a value of 0.05. clear at this time.
Identification of the ESN pattern on echocardiography demon-
strated good to excellent predictive ability for MPE or SMPE, with
RESULTS sensitivity of 92% (95% CI, 84%–97%), specificity of 99% (95%
CI, 96%–100%), positive predictive value of 98% (95% CI, 91%–
A total of 5,152 patients underwent computed tomographic scans for 100%), negative predictive value of 96% (95% CI, 92%–98%), and
suspected PE, of whom 526 (10.2%) had positive results for PE area under the ROC curve of 0.96 (95% CI, 0.92–0.98), which
(including MPE or SMPE and SSPE). After initial screening criteria when compared side by side with the widely recognized
(no echocardiogram available or echocardiogram available but McConnell’s sign suggests superior predictive ability (Table 3).
outside the prespecified 48-hour window, history of embolism, or Notably, McConnell’s sign assessment in this study yielded sensitivity
pulmonary hypertension), a total of 260 patients from this group of 52% (95% CI, 40%–63%), specificity of 97% (95% CI, 94%–
were short-listed for further analysis. Upon further screening and 99%), positive predictive value of 90% (95% CI, 77%–97%), nega-
application of exclusion criteria (limited studies, RVOT PW Doppler tive predictive value of 82% (95% CI, 76%–87%), and area under
data not recorded, or Doppler data recorded but technically inade- the ROC curve of 0.75 (95% CI, 0.68–0.80) while the 60/60 sign
quate for interpretation moderate or greater valvular disease), a total yielded values of 51% (95% CI, 40%–62%), 96% (95% CI, 91%–
of 187 patients with PE qualified and were included in the final 99%), 93% (95% CI, 81%–99%), 70% (95% CI, 61%–77%), and
analysis. The reference group included a total of 90 patients without 0.74 (95% CI, 0.68–0.79), respectively.
4 Afonso et al Journal of the American Society of Echocardiography
- 2019

Figure 2 Illustrative case examples of SMPE showing representative computed tomographic angiographic images and correspond-
ing RVOT Doppler patterns. (A) Saddle embolism with thrombus extension into branch pulmonary arteries and clear early notching
pattern (sweep at 100 mm/sec). (B) Saddle embolism with large thrombus burden; note early notching (white arrow) and intermittent
triphasic notching pattern (white arrowhead). (C) Extensive thrombus in left main pulmonary artery showing early notching (white ar-
rows). Doppler patterns in (B) and (C) were recorded at 50 mm/sec.

Figure 3 Additional case examples. (A) Midsystolic notching in a patient with SSPE and thrombi in branches of the right and left main
pulmonary arteries. (B) Triphasic Doppler pattern in a patient with SMPE and saddle embolism. Arrows denote the locations of
notches along the RVOT Doppler envelope.
Journal of the American Society of Echocardiography Afonso et al 5
Volume - Number -

Table 1 Baseline descriptive demographic, clinical, and RVOT Doppler envelope and transient midsystolic notching of the
hemodynamic characteristics of entire cohort pulmonary valve (M mode), representing systolic flow deceleration
in the RVOT from reflected waves in patients with pulmonary hyper-
No PE SSPE MPE/SMPE tension.15,16 Subsequently, Torbicki et al.17 reported short AT and ESN
Variable (n = 90) (n = 87) (n = 100) P in a patient with acute MPE and appropriately attributed the restora-
Age (y) 56 6 13 56 6 15 55 6 14 .82
tion of flow in midsystole to the distal runoff of blood across the re-
maining patent pulmonary bed. Although there has been renewed
Male 51 36 56 .01
interest in examining the significance of notching patterns in patients
Black race 72 77 76 .06 with chronic pulmonary arterial hypertension,18 expanding previ-
Smoker 20 25 36 .04 ously reported literature on Doppler-hemodynamic correlations in
Hypertension 77 61 57 .01 this population,19-22 data on the utility of RVOT Doppler systolic
Diabetes 36 23 33 .13 notching in the context of acute PE are sparse.20
COPD 16 17 10 .22 The degree and timing of arterial wave reflection are determined in
Malignancy 12 19 16 .50 aggregate by pulmonary vascular resistance (PVR), reflected wave
speed (determined by large and medium vessel stiffness), and the dis-
SBP (mm Hg) 135 6 23 127 6 25 115 6 33 .03
tance of the reflecting sites from the RVOT.18,19,21 Conceptually, the
DBP (mm Hg) 76 6 14 76 6 14 74 6 19 .45 short ATs and ESN pattern characteristic of the ‘‘spike and dome’’
Pulse rate (beats/min) 80 6 19 99 6 18 106 6 19 <.001 observed in our cohort of patients with MPE or SMPE reflect the
COPD, Chronic obstructive pulmonary disease; DBP, diastolic blood proximity of the increased after load to the RVOT and the
pressure; SBP, systolic blood pressure. markedly elevated PVR; this is in contrast to the midsystolic
Data are expressed as mean 6 SD or as percentages. notching or late systolic notching pattern apparent with less severe
grades of chronic precapillary pulmonary hypertension.17,19 Of
Notably, the addition of other echocardiographic parameters note, ‘‘systolic notching’’ is characteristically not observed in the
(Table 2) to the ESN pattern did not show any significant incremental subset patients with postcapillary or left-sided pulmonary venous hy-
benefit. In secondary analysis, the ESN pattern performed poorly for pertension, who by definition have normal PVR.19 In aggregate,
SSPE (Table 4). notching is strongly correlated with PVR (RV afterload) but may be
dissociated from pulmonary artery pressure in patients with acute
PE,20 an observation mirrored in our cohort of patients with MPE
DISCUSSION or SMPE. Our findings demonstrating the ability to identify patients
with extremely elevated PVR (MPE or SMPE) are also thematically
In the present study, an ESN pattern of the RVOT Doppler envelope in agreement with earlier observations suggesting that the combined
was a frequently observed finding in patients with MPE or SSPE, presence of a short RVOT Doppler ATand low tricuspid velocity peak
selected from among a cohort of individuals undergoing CTA for sus- gradient may help discern acute PE from an assorted population of
pected PE. Despite promising diagnostic performance data compared patients with chronic precapillary pulmonary hypertension.19
with existing echocardiographic variables, in a select population of pa- Importantly, we also observed a weak correlation between pulmo-
tients that excluded individuals with preexisting pulmonary arterial nary artery systolic pressure and notch timing in this study. These vari-
hypertension, the diagnostic accuracy of this sign, including its sensi- ables were inversely related but weakly correlated (correlation
tivity and specificity, will need to be validated in a broader population coefficient = 0.22). This observation likely stems from the fact
of patients with less stringent exclusion criteria before it can be recom- that pulmonary artery systolic pressure, in the setting of ‘‘acute’’ hemo-
mended for use in the clinical setting. dynamically significant PE, may be normal or even subnormal, reflect-
Echocardiography has traditionally been relegated to a supportive ing the abrupt, precipitous decline in RV systolic performance (from
role in the evaluation of patients with suspected PE and used primarily RV stunning), although pulmonary artery systolic pressure often rises
to assess RV size and function in hemodynamically unstable patients.11 rapidly in the days following MPE or SMPE.
Several echocardiographic parameters described in the literature, Contrast CTA remains the diagnostic modality of choice for the
including McConnell’s sign (reduced RV free wall contractility with diagnosis of PE, given its high spatial and temporal resolution and
preserved RV apical function), the 60/60 sign (AT < 60 msec in the the ability to visualize segmental and subsegmental arteries.4
presence of a tricuspid peak gradient > 30 mm Hg but < 60 mm Although both tests (CTA and ventilation-perfusion scanning) are reli-
Hg), paradoxical septal motion, and RV dilation can be helpful but able and comparable4 in terms of sensitivity and negative predictive
lack the desired sensitivity, specificity, or negative predictive value (re- value, a time-trend analysis report revealed that the widespread adop-
ported to be about 40%–50%) to be incorporated as first-line options tion of CTA in 1998 for diagnosing PE led to a sharp rise in PE inci-
in evaluation algorithms for suspected PE.4-6,8,9,12,13 Our study dence (an 80% increase from 1998 to 2006), without an
findings on the diagnostic performance of the aforementioned appreciable change in PE-related mortality. Collectively, these obser-
echocardiographic variables are in accord with the published vations, including a declining PE case-fatality rate (in-hospital deaths
literature. Nonetheless, as an exception, it should be acknowledged among patients with diagnoses of PE) seem to reflect an increased
that echocardiographically visualized right atrial thrombi, reported in detection or ‘‘overdiagnosis’’ of clinically insignificant, small or subseg-
<4% of unselected patients with PE, are notoriously associated with mental pulmonary emboli.23,24 Although the lethality of MPE and
hemodynamic compromise and a poor prognosis.14 SMPE is widely recognized,4,25 the clinical significance of isolated
Our observations suggest that the simple visual assessment of SSPE (on the basis of a large meta-analysis), found in roughly 9.4%
RVOT Doppler morphology offers potentially insightful information of individuals with suspected PE, undergoing multidetector CTA is
on the coupling dynamics between the right ventricle and the pulmo- relatively less concerning, presumably because of the low 3-month
nary vasculature. Early studies describe a notching pattern of the thromboembolic risk of approximately 1%.26 Along these lines,
6 Afonso et al Journal of the American Society of Echocardiography
- 2019

Table 2 Echocardiographic parameters in various study groups

Variable No PE (n = 90) SSPE (n = 87) MPE/SMPE (n = 100) P

RVOT ESN 0 2.0 92 <.001

AT (msec) 149 6 45 101 6 38 60 6 19 <.001
DT (msec) 186 6 46 175 6 57 219 6 44 <.001
AT/ejection time ratio 0.46 6 0.11 0.38 6 0.13 0.22 6 0.07 <.001
Acceleration slope (cm/sec2) 574 6 255 945 6 528 1,329 6 545 <.001
Deceleration slope (cm/sec2) 319 6 136 393 6 222 219 6 84 <.001
RV basal diameter (cm) 3.4 6 0.44 3.6 6 0.52 4.4 6 0.74 <.001
Slope ratio (acceleration slope/deceleration slope) 2.0 6 1.0 3.0 6 2.1 6.5 6 2.5 <.001
Notch characteristics*
NT (msec) — 142 6 29 106 6 32 <.001
NT/ejection time — 0.51 6 0.11 0.39 6 0.11 <.001
NT/RVOT VTI — 13 6 5.5 10 6 3.3 <.001
S/D ratio (peak spike-and-dome velocity ratio) — 1.5 6 0.25 1.6 6 0.26 .07
No notching 100 79 7 <.01
TR peak velocity (m/sec) 2.4 6 0.56 2.8 6 0.62 3.1 6 0.60 <.001
60/60 sign 0 7.0 51 <.001
McConnell sign 0 7.1 52 <.001
DT, Deceleration time; NT, notch time; TR, tricuspid regurgitation; VTI, velocity-time integral.
Data are expressed as percentages or as mean 6 SD.
*Among patients with SSPE, systolic notching was predominantly midsystolic (19%) and less likely early systolic (2%). However, among patients
with MPE or SMPE, systolic notching was predominantly early systolic (92%) and less likely midsystolic (1%).

Table 3 Probability statistics of MPE or SMPE in the full cohort

Positive predictive Negative predictive

Variable Sensitivity, % Specificity, % value, % value, % AUROC

60/60 sign* 51 (40–62) 96 (91–99) 93 (81–99) 70 (61–77) 0.74 (0.68–0.79)

McConnell’s sign 52 (40–63) 97 (94–99) 90 (77–97) 82 (76–87) 0.75 (0.68–0.80)
ESN pattern 92 (84–97) 99 (96–100) 98 (91–100) 96 (92–98) 0.96 (0.92–0.98)
AT # 87msec 91 (83–94) 77 (70–83) 60 (52–69) 91 (88–94) 0.84 (0.79–0.88)
AT/ejection time ratio # 0.38 90 (84–94) 59 (51–66) 54 (46–62) 91 (86–95) 0.77 (0.73–0.82)
DT $ 200 msec 64 (53–73) 70 (63–76) 49 (39–58) 81 (74–87) 0.67 (0.61–0.73)
DT/AT ratio $ 2.36 83 (74–90) 79 (71–83) 67 (57–75) 90 (85–93) 0.80 (0.76–0.85)
Acceleration slope $ 810 cm/sec2 89 (80–94) 67 (60–73) 53 (46–63) 91 (87–96) 0.78 (0.73–0.81)
Deceleration slope # 232 cm/sec2 66 (55–76) 74 (67–79) 53 (43–62) 83 (76–88) 0.70 (0.64–0.75)
Slope ratio $ 4 (S/D) 81 (72–89) 83 (79–88) 73 (63–81) 90 (85–93) 0.83 (0.80–0.86)
AUROC, Area under the ROC curve; DT, deceleration time.
*Tricuspid regurgitation velocity < 3.9 m/sec plus pulmonary artery AT < 60 msec.

Table 4 Probability statistics for SSPE

Variable Sensitivity, % Specificity, % Positive predictive value, % Negative predictive value, % AUROC

60/60 sign 7.0 (1.5–19) 94 (90–99) 92 (48–99) 48 (38–70) 0.51 (0.47–0.57)

McConnell sign 7.1 (2.6–13) 95 (90–99) 94 (59–99) 49 (42–56) 0.53 (0.51–0.55)
ESN 2.3 (0.24–7.4) 92 (86–99) 94 (15–99) 50 (43–57) 0.51 (0.49–0.52)
AUROC, Area under the ROC curve.

caution is also advised in interpreting the significance of small vascular large majority of clinicians would likely pursue anticoagulation of
defects on CTA, given the high false-positive rate (secondary to mo- patients with SSPE, it has been reported that the short-term risk of
tion and technical artifacts), the low positive predictive value, and recurrent thromboembolism may be lower than the risk for
concerning levels of interobserver agreement.26-28 Although the adverse events ensuing from anticoagulation in such patients.4,23
Journal of the American Society of Echocardiography Afonso et al 7
Volume - Number -

Accordingly, withholding anticoagulation in selected patients with or SMPE. Future prospective study to better establish its role in the
SSPE may be reasonable given the very low risk for recurrent or management of acute PE is warranted.
fatal PE.27 However, as reiterated in the 2014 European Society of
Cardiology guidelines, it might be prudent to rule out deep vein
thrombosis before opting to withhold anticoagulation in such pa-
REFERENCES
tients.4 In this context, the suboptimal performance and low diag-
nostic accuracy of our study’s echocardiographic variables for SSPE 1. Anderson FA Jr., Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA,
may be viewed as desirable, essentially serving as an inherent filter Jovanovic B, et al. A population-based perspective of the hospital inci-
to reliably and selectively isolate the subset of patients with MPE dence and case-fatality rates of deep vein thrombosis and pulmonary em-
and SMPE from individuals undergoing workup for suspected PE. bolism. The Worcester DVT Study. Arch Intern Med 1991;151:933-8.
A few limitations need to be acknowledged. First, we excluded pa- 2. Beckman MG, Hooper WC, Critchley SE, Ortel TL. Venous thromboem-
tients with known pulmonary hypertension from the study. Hence, bolism: a public health concern. Am J Prev Med 2010;38:S495-501.
our findings may not be applicable to such groups of patients. The cur- 3. Jaff MR, McMurtry MS, Archer SL, Cushman M, Goldenberg N,
rent evaluation was done in a single cohort and reflects a ‘‘best-case’’ Goldhaber SZ, et al. Management of massive and submassive pulmonary
scenario for test performance, and as such, our observations are not embolism, iliofemoral deep vein thrombosis, and chronic thromboem-
bolic pulmonary hypertension: a scientific statement from the American
representative of the entire population of patients presenting with
Heart Association. Circulation 2011;123:1788-830.
symptoms of PE. It is quite likely that our findings would have lower 4. Konstantinides SV, Torbicki A, Agnelli G, Danchin N, Fitzmaurice D,
specificity when applied to a broader, less selected population that Galie N, et al. 2014 ESC guidelines on the diagnosis and management
included patients with pulmonary arterial hypertension, acute hypoxic of acute pulmonary embolism. Eur Heart J 2014;35:3033-69.
respiratory failure and other entities with elevated PVR. Accordingly, 5. Fields JM, Davis J, Girson L, Au A, Potts J, Morgan CJ, et al. Transthoracic
although further prospective study in additional cohorts is warranted, echocardiography for diagnosing pulmonary embolism: a systematic re-
such clinical scenarios emphasize the unique value of CTA as a diag- view and meta-analysis. J Am Soc Echocardiogr 2017;30:714-23.e4.
nostic tool in patients with suspected PE. Our study lacked invasive he- 6. Kurnicka K, Lichodziejewska B, Goliszek S, Dzikowska-Diduch O,
modynamic correlation, but we remain confident in the adoption of Zdonczyk O, Kozlowska M, et al. Echocardiographic pattern of acute pul-
CTA alone as the gold standard for our analysis. Although speculative monary embolism: analysis of 511 consecutive patients. J Am Soc Echocar-
diogr 2016;29:907-13.
and in need of additional validation in a less selected population, early
7. Sood A, Rehman M, Javed AA, Akintoye E, Kumar K, Singh M, et al. Diag-
notching in patients with computed tomography–confirmed PE could nosis and differentiation of acute massive and submassive from subseg-
potentially help identify a subset of patients with more severe pulmo- mental pulmonary embolism using novel echocardiographic methods. J
nary vascular obstruction at risk for greater hemodynamic compro- Am Coll Cardiol 2018;71(11 suppl).
mise. It must be borne in mind that the RVOT Doppler early 8. Kurzyna M, Torbicki A, Pruszczyk P, Burakowska B, Fijalkowska A,
notching pattern is not specific for MPE or SMPE but is a highly specific Kober J, et al. Disturbed right ventricular ejection pattern as a new Doppler
sign of significantly elevated PVR. This hypothesis was elegantly echocardiographic sign of acute pulmonary embolism. Am J Cardiol
demonstrated in experimental dog models following acute proximal 2002;90:507-11.
PE and pulmonary artery banding.29 Not surprisingly, we incidentally 9. McConnell MV, Solomon SD, Rayan ME, Come PC, Goldhaber SZ,
observed (unpublished data) a few instances of the early notching Lee RT. Regional right ventricular dysfunction detected by echocardiogra-
phy in acute pulmonary embolism. Am J Cardiol 1996;78:469-73.
(spike-and-dome) pattern in individuals with very advanced pulmo-
10. Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al.
nary arterial hypertension and in patients with proximal obliteration Recommendations for cardiac chamber quantification by echocardiogra-
or compression of one of the major pulmonary arteries from advanced phy in adults: an update from the American Society of Echocardiography
lung cancer. Mid- or late systolic notching, on the other hand, is and the European Association of Cardiovascular Imaging. J Am Soc Echo-
commonly observed in the vast majority of patients with precapillary cardiogr 2015;28:1-39.e14.
pulmonary arterial hypertension. Regardless of the underlying mech- 11. Roy PM, Colombet I, Durieux P, Chatellier G, Sors H, Meyer G. Systematic
anism, the finding of early notching, when identified for the first time, is review and meta-analysis of strategies for the diagnosis of suspected pul-
abnormal and should prompt a workup for PE, pulmonary artery monary embolism. BMJ 2005;331:259.
compression, or pulmonary arterial hypertension with CTA or a 12. Lodato JA, Ward RP, Lang RM. Echocardiographic predictors of pulmo-
ventilation-perfusion scan and, when appropriate, with right heart nary embolism in patients referred for helical CT. Echocardiography
2008;25:584-90.
catheterization to estimate PVR.
13. Bova C, Greco F, Misuraca G, Serafini O, Crocco F, Greco A, et al. Diag-
Notably, the low positivity rate for CTA, reported as 9% to 14% in nostic utility of echocardiography in patients with suspected pulmonary
the published literature (10.2% in our study),30,31 frames the embolism. Am J Emerg Med 2003;21:180-3.
contemporary use of CTA, as a screening, inasmuch as a valuable 14. Torbicki A, Galie N, Covezzoli A, Rossi E, De Rosa M, Goldhaber SZ. Right
diagnostic tool, while also implicitly underscoring the pressing need heart thrombi in pulmonary embolism: results from the International
to develop more cost-effective screening strategies.32-34 Cooperative Pulmonary Embolism Registry. J Am Coll Cardiol 2003;41:
Until such time as additional validation data are accrued, echocar- 2245-51.
diography (including the presence of ESN) should not be viewed as 15. Kitabatake A, Inoue M, Asao M, Masuyama T, Tanouchi J, Morita T, et al.
the primary screening test or a gatekeeper for CTA in the diagnosis Noninvasive evaluation of pulmonary hypertension by a pulsed Doppler
of acute PE. technique. Circulation 1983;68:302-9.
16. Weyman AE, Dillon JC, Feigenbaum H, Chang S. Echocardiographic pat-
terns of pulmonic valve motion with pulmonary hypertension. Circulation
1974;50:905-10.
CONCLUSION 17. Torbicki A, Tramarin R, Morpurgo M. Role of echo/Doppler in the diag-
nosis of pulmonary embolism. Clin Cardiol 1992;15:805-10.
In patients with suspected acute PE, the pulmonary Doppler flow 18. Takahama H, McCully RB, Frantz RP, Kane GC. Unraveling the
pattern of ESN has potential clinical utility for the detection of MPE RV ejection Doppler envelope: insight into pulmonary artery
8 Afonso et al Journal of the American Society of Echocardiography
- 2019

hemodynamics and disease severity. JACC Cardiovasc Imaging 2017; meta-analysis of the management outcome studies. J Thromb Haemost
10:1268-77. 2010;8:1716-22.
19. Arkles JS, Opotowsky AR, Ojeda J, Rogers F, Liu T, Prassana V, et al. Shape 27. Stein PD, Goodman LR, Hull RD, Dalen JE, Matta F. Diagnosis and man-
of the right ventricular Doppler envelope predicts hemodynamics and agement of isolated subsegmental pulmonary embolism: review and
right heart function in pulmonary hypertension. Am J Respir Crit Care assessment of the options. Clin Appl Thromb Hemost 2012;18:20-6.
Med 2011;183:268-76. 28. Miller WT Jr., Marinari LA, Barbosa E Jr., Litt HI, Schmitt JE, Mahne A, et al.
20. Torbicki A, Kurzyna M, Ciurzynski M, Pruszczyk P, Pacho R, Kuch- Small pulmonary artery defects are not reliable indicators of pulmonary
Wocial A, et al. Proximal pulmonary emboli modify right ventricular ejec- embolism. Ann Am Thorac Soc 2015;12:1022-9.
tion pattern. Eur Respir J 1999;13:616-21. 29. Furuno Y, Nagamoto Y, Fujita M, Kaku T, Sakurai S, Kuroiwa A. Reflec-
21. Turkevich D, Groves BM, Micco A, Trapp JA, Reeves JT. Early partial sys- tion as a cause of mid-systolic deceleration of pulmonary flow wave in
tolic closure of the pulmonic valve relates to severity of pulmonary hyper- dogs with acute pulmonary hypertension: comparison of pulmonary ar-
tension. Am Heart J 1988;115:409-18. tery constriction with pulmonary embolisation. Cardiovasc Res 1991;
22. Lopez-Candales A, Edelman K. Shape of the right ventricular outflow 25:118-24.
Doppler envelope and severity of pulmonary hypertension. Eur Heart J 30. Costantino MM, Randall G, Gosselin M, Brandt M, Spinning K, Vegas CD.
Cardiovasc Imaging 2012;13:309-16. CT angiography in the evaluation of acute pulmonary embolus. AJR Am J
23. Donato AA, Khoche S, Santora J, Wagner B. Clinical outcomes in patients Roentgenol 2008;191:471-4.
with isolated subsegmental pulmonary emboli diagnosed by multidetec- 31. Mamlouk MD, vanSonnenberg E, Gosalia R, Drachman D, Gridley D,
tor CT pulmonary angiography. Thromb Res 2010;126:e266-70. Zamora JG, et al. Pulmonary embolism at CT angiography: implications
24. Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embo- for appropriateness, cost, and radiation exposure in 2003 patients. Radi-
lism in the United States: evidence of overdiagnosis. Arch Intern Med ology 2010;256:625-32.
2011;171:831-7. 32. Paterson DI, Schwartzman K. Strategies incorporating spiral CT for the
25. Torbicki A, Perrier A, Konstantinides S, Agnelli G, Galie N, Pruszczyk P, diagnosis of acute pulmonary embolism: a cost-effectiveness analysis.
et al. Guidelines on the diagnosis and management of acute pulmonary Chest 2001;119:1791-800.
embolism: the Task Force for the Diagnosis and Management of Acute 33. Perrier A, Nendaz MR, Sarasin FP, Howarth N, Bounameaux H. Cost-
Pulmonary Embolism of the European Society of Cardiology (ESC). Eur effectiveness analysis of diagnostic strategies for suspected pulmonary em-
Heart J 2008;29:2276-315. bolism including helical computed tomography. Am J Respir Crit Care
26. Carrier M, Righini M, Wells PS, Perrier A, Anderson DR, Rodger MA, et al. Med 2003;167:39-44.
Subsegmental pulmonary embolism diagnosed by computed tomogra- 34. Fanikos J, Rao A, Seger AC, Carter D, Piazza G, Goldhaber SZ. Hospital
phy: incidence and clinical implications. A systematic review and costs of acute pulmonary embolism. Am J Med 2013;126:127-32.