Objective
The objective of the study was to investigate whether vector velocity imaging (VVI), a non-Doppler speckle tracking ultrasound technology, is feasible in twin pregnancies and can aid management of twin-twin transfusion syndrome (TTTS).
Study Design
Twenty-seven women pregnant with monochorionic diamniotic twins affected by TTTS and 28 monochorionic pregnancies that did not develop TTTS were included in a prospective case-control study at a fetal medicine center. Fetal echocardiograms were recorded with dummy electrocardiography to retain original frame rates when exported for offline speckle tracking analysis using Syngo-VVI software (Siemens Corp, Munich, Germany). Right and left ventricular (LV) free wall Lagrangian strain was measured from the original coordinates. Within-twin pair ventricular strain differences including relationship to Quintero staging and response to laser therapy for TTTS were analyzed by Wilcoxon signed-rank test.
Results
The VVI strain measurements could be analyzed in 182 of 200 TTTS and 96 of 112 non-TTTS control ventricles. Within-pair strain was concordant in non-TTTS controls. Recipient LV strain was reduced at all Quintero stages compared with donors ( P < .01). Recipient right ventricular strain was reduced only in stages 3 and 4 ( P < .01). Strain improved at a median of 2 weeks following successful laser therapy. Intertwin differences in strain were independent of weight discordance.
Conclusion
Recipient LV strain is reduced in stages 1 and 2 TTTS. Within-pair strain discordance may distinguish early TTTS from growth discordance and guide timing of and management following treatment.
Prediction of development and progression of twin-twin transfusion syndrome (TTTS) in genetically identical monochorionic diamniotic (MCDA) twins remains difficult. Current measurements cannot reliably identify pregnancies that will progress rapidly to advanced disease, or even fetal demise, within days.
Attempts to find improved methods of characterizing and risk stratifying in TTTS have turned to measures of cardiac dysfunction. The Tei or Myocardial Performance Index (MPI) index compares the time the heart spends in relaxation and in contraction from simultaneous Doppler assessment of mitral and aortic flow profiles. A prolonged isovolumic relaxation time has been described in recipients with early-stage disease and the resultant increased MPI proposed as one measurement distinguishing early TTTS from discordant growth and an aid to assessment in TTTS. However, the Tei index is not a measure of intrinsic myocardial function but alters in response to acute changes in load and distal impedance ; it therefore reflects the pathophysiology faced by the recipient early in the disease process and probably before the effects of unbalanced inter-twin transfusion satisfy the Quintero staging. To date the ability of cardiovascular parameters to predict the need for, or response to, treatment is unsatisfactory, with only a weak correlation with Quintero staging and wide variation at each stage.
Vector velocity imaging (VVI) is a non-Doppler ultrasound technology that tracks myocardial speckles frame by frame to produce measures of myocardial deformation, including strain. Strain signifies the fractional change in ventricular wall length and allows quantification of myocardial performance offline. Although strain, like MPI, is load dependent, the assessment of cardiac wall deformation provides a more direct reflection of intrinsic cardiomyocyte function.
The feasibility and pitfalls of speckle tracking in normal singleton fetuses has been described. Many studies of VVI speckle tracking in the fetus have been limited by measurement at low frame rates, resulting in falsely low measures of strain, particularly in the left ventricle, or the use of software calculated global or natural strain rather than Lagrangian strain. Global strain is the software-averaged sum of short segmental strain values, whereas Lagrangian strain is manually calculated from stored coordinates of the entire length of the ventricular wall and produces values with less variation and therefore greater reliability. These technical factors may limit the ability of VVI to measure strain and have resulted in an inaccurate measure of strain, masking gestational changes and reporting interventricular differences in strain that are not present when high frame rate clips are measured.
In this study, we assessed the feasibility of using VVI speckle tracking to measure Lagrangian strain in twin pregnancies including those complicated by TTTS. We investigated whether ventricular strain measurement might aid risk stratification of TTTS, specifically whether within-pair differences in ventricular strain exist in early TTTS. Finally, we assessed whether VVI measurements can reflect functional recovery following fetal laser therapy for TTTS.
Materials and Methods
This was a prospective case-control study enrolling consecutive women with MCDA pregnancies referred for ultrasound assessment of their pregnancy from local maternity units. Fetal echocardiograms were performed using dummy electrocardiogram gating (described in the following text) to ensure optimal transfer of frame rates to the Syngo software (Siemens Corp, Munich, Germany) to perform offline speckle tracking. The institutional review board deemed ethical approval was unnecessary because the evaluation was an integral part of routine clinical visits in which the mother had consented to the examination. Twenty-seven pregnancies with TTTS (as defined by standard Quintero staging criteria ) and 28 controls who did not develop TTTS were studied. Single scans were performed for each control. Eighteen of the 27 TTTS pregnancies underwent serial scans and data from a total of 50 TTTS scans (200 ventricles) were available for analysis.
Sonographic data were collected prospectively by several of the coauthors to a standard protocol using Acuson Sequoia ultrasound systems (Siemens Corp) with a 6-2 multi-Hertz curvilinear probe (Acuson, Mountain View, CA). The 4-chamber image was optimized to achieve high contrast between the myocardial walls and cavity, and the clearest loop of each image acquired during the clinical examination was stored digitally in a DICOM format. As previously described, we attached a commercially available metronome (BOSS DB-60; Roland Corp, Hamamatsu, Japan), to the line input of the ultrasound system with the metronome’s dummy spike signals, set to 60 beats per minute, to permit the storage of 2 virtual cycles (2 seconds) at their original frame rate. The identity of the clips were blinded, assigned a code so that twin pairs could not be identified, and imported in batches retrospectively by one examiner (H.M.) into Syngo Work Station (Siemens Corp) using Syngo VVI version 1 for offline analysis of ventricular strain.
The fetal heart rate was determined by mitral valve closure and a single heartbeat selected for analysis. The endocardial border was tracked manually in end-diastole and tracking curves created automatically in subsequent frames. The 2 orthogonal X,Y tracking coordinates stored by the software’s database were used to calculate peak negative systolic strain along the whole length of the right and left free wall (Lagrangian strain ). Strain is defined as a relative change in length and is represented in this paper as a positive percentage change. We allowed 10 minutes for tracking and considered recordings taking longer unsuitable for analysis.
Outcome measures
- 1.
Within-twin pair differences in left (LV) and right (RV) ventricular strain between heavier vs lighter non-TTTS control pairs and recipient vs donor TTTS pairs.
- 2.
Relationship of Quintero staging to strain: within-twin pair differences in LV strain and RV strain between recipient vs donor TTTS in early (prestage, stages 1 and 2) and advanced (stages 3 and 4) TTTS.
- 3.
Effect of laser therapy for TTTS on ventricular strain: within-twin pair differences in LV strain and RV strain between recipient vs donor TTTS before and after treatment.
Statistical analysis
Outcomes 1 and 2 were analyzed using Wilcoxon signed ranks test and significance of 2-tailed test taken as P < .05. Because some twins had serial measurements, a mixed-effects linear regression model was used to analyze the effect of staging and laser intervention, with patients entered as a random effect. All analyses were performed using Stata 12.1 (StataCorp, College Station, TX).
Results
VVI tracking was suitable for analysis in 96 of 112 (86%) non-TTTS control scans and in 182 of 200 (91%) TTTS scans; RV strain and LV strain for each twin were tracked with similar success in both cohorts. Reproducibility was good with an intraclass correlation for RV and LV strain (0.894 and 0.928, respectively).
Median gestational age at examination was 21 ± 3 (range, 16 ± 3 to 32 ± 5) weeks and was similar between non-TTTS controls and TTTS cases (22 ± 4 vs 21 ± 2; P = .18). A trend toward decreasing ventricular strain with gestational age was observed in the non-TTTS group, which did not reach statistical significance (LV, P = .052; RV, P = .18). No congenital or chromosomal abnormalities were identified in either group.
To enable comparisons controlling for differences in fetal weight, non-TTTS control twins were described as heavy or light controls. The mean estimated fetal weight at examination is shown in Table 1 . Recipient and donor twins were generally lighter than heavy and light twins, respectively. A trend toward decreasing ventricular strain with increasing estimated fetal weight was observed in the non-TTTS group, which did not reach statistical significance (LV, P = .054; RV, P = .2). The mean estimated fetal weight discordance was greater in the TTTS group compared with the non-TTTS group as expected; however, Kendall’s tau showed no significant relationship between estimated weight discordance and strain discordance for either left or right ventricles (LV: tau, b = 0.14; 95% confidence interval [CI], −0.07 to 0.35; P = .13; RV: tau, b = −0.08; 95% CI, −0.27 to 0.11; P = .40).
| Characteristic | Heavy | Light | Recipient | Donor |
|---|---|---|---|---|
| EFW, g | 527 (60.5) | 458 (51.6) | 462 (36.4) | 343 (28.6) |
| Gestational age, wks | 22.6 (20–25) | 21.3 (18–24) | ||
| Maternal age, y | 32.5 (30–35) | 33.4 (24–37) | ||
| EFW discordance, % | 13 (2.0) | 23 (2.1) | ||
The Quintero stages of TTTS at presentation were as follows: 1 (n = 2); 2 (n = 3); 3 (n = 19); 4 (n = 1). Two pregnancies were studied initially at prestage disease and later developed TTTS. All underwent laser therapy for TTTS at a median gestational age of 20 ± 5 (range, 16 ± 3 to 25 ± 5) weeks. Sixteen women had serial VVI assessment before and following treatment. Outcomes for all pregnancies are shown in Table 2 .
| Outcome | TTTS (n = 27) | Non-TTTS (n = 28) |
|---|---|---|
| Live birth | 22 | 27 |
| Single fetal demise | 3 | 1 |
| Double fetal demise | 1 | 0 |
| Fetocide | 1 | 0 |
The mean ventricular strain values for heavy and light non-TTTS control twins and recipient and donor TTTS twins are shown in Table 3 . In TTTS, within-pair differences were significant for both the LV strain and RV strain, with lower values seen in recipients compared with donors. Interventricular strain differences were significant in recipients but not donors or controls.
| Variable | Heavy | Light | Recipient | Donor | ||
|---|---|---|---|---|---|---|
| LV | 22.07 ± 4.4 | 22.21 ± 3.0 | P = .11 | 18.28 ± 2.7 | 22.90 ± 2.2 | P < .001 |
| RV | 22.41 ± 3.3 | 22.53 ± 4.0 | P = .93 | 19.60 ± 3.4 | 22.59 ± 3.4 | P < .01 |
| P = .77 | P = .6 | P = .013 | P = .77 | |||
Differences observed in the TTTS cohort were investigated further by analyzing the effect of Quintero stage on ventricular strain; the mean values are shown in Table 4 . Within-twin pair comparison of strain was significant for the left ventricle but not the right at Quintero stage 2 or less and significant for both ventricles at higher Quintero stages. The Figure represents mean ventricular strain values for non-TTTS controls and TTTS cases at early and later Quintero stages with significant within-twin pair differences indicated.
| Variable | Quintero stage 2 or less | Quintero stages 3 and 4 | ||||
|---|---|---|---|---|---|---|
| Recipient | Donor | P value | Recipient | Donor | P value | |
| LV | 20.34 (10.1) | 23.0 (7.7) | .018 | 17.8 (8.8) | 22.6 (8.8) | < .001 |
| RV | 22.3 (11.4) | 21.8 (14.1) | > .9 | 19.8 (12.0) | 22.6 (10.1) | < .001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
