1072
The Utility of MRI for Measuring Hematocrit in Fetal Anemia1Department of Physiology, University of Toronto, Toronto, ON, Canada, 2Institute of Medical Science, University of Toronto, Toronto, ON, Canada, 3Division of Cardiology, Hospital for Sick Children, Toronto, ON, Canada, 4Division of Maternal-Fetal Medicine, Mount Sinai Hospital, Toronto, ON, Canada, 5Division of Translational Medicine, Hospital for Sick Children, Toronto, ON, Canada, 6Mouse Imaging Centre, Hospital for Sick Children, Toronto, ON, Canada, 7Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada
Synopsis
To determine the accuracy of MRI for estimating hematocrit in anemic fetuses, the current study compared MRI-estimated hematocrit against the gold-standard fetal blood sampling and Doppler assessment of the middle cerebral artery for peak systolic velocity (MCA-PSV). MRI-estimated hematocrit was calculated from T1 and T2 measurements of intrahepatic umbilical vein blood. MRI results correlated well with fetal blood sampling and had a higher specificity for predicting anemia than Doppler MCA-PSV. In conclusion, MRI is feasible and accurate for the detection of fetal anemia and may be used to confirm the need for blood transfusion when MCA-PSV is suggestive of anemia.
Background
Fetal anemia is evaluated noninvasively by Doppler ultrasonographic measurements of the peak systolic velocity of the middle cerebral artery (MCA-PSV), which has higher likelihood of being falsely-positive after the fetus has received intrauterine blood transfusions and in late gestation.1–4 A previous in vitro study showed that hematocrit can be estimated using MRI T1 and T2 relaxation times.5 This current study aimed to assess the accuracy of MRI for diagnosing fetal anemia in vivo.Method
This study was designed as a prospective observational study. Fetuses undergoing fetal blood sampling or intrauterine blood transfusion were scanned at 1.5 Tesla (Siemens Avanto, Erlangen, Germany). A modified Look-Locker inversion pulse sequence and a T2 preparation sequence were applied for T1 and T2 mapping of the umbilical vein. Estimated fetal hematocrit was calculated using a combination of T1 and T2 values using an in vitro calibration previously reported using human fetal blood5. MRI estimates of hematocrit was then compared with measurements from fetal blood sampling and Doppler MCA-PSV. The cut-offs of fetal anemia for Doppler MCA-PSV and hematocrit were adopted from Mari et al.6. The analysis was adjusted for repeated measures.Results
23 fetuses were assessed with 33 MRI scans. The gestational age of these fetuses ranged from 19 to 36.6 weeks. The mean difference between laboratory and MRI hematocrit was 6% ± 5% with a significant correlation of 0.76 as determined by generalized estimation equation (p < 0.001) (Figure 1a). Bland-Altman analysis revealed a systematic bias of -3% between MRI and laboratory measurements (Figure 1b). The area under the curve for the MRI ROC curve (0.96) was significantly higher (p value = 0.03) than that of Doppler (0.69) to predict fetal anemia (Figure 2). MRI and Doppler had similar sensitivity at around 90% in predicting moderate to severe anemia (Table 1). However, MRI had a specificity of 93% (95% CI: 66-100%), which was higher than the specificity of Doppler (71%, 95% CI: 42-92%).Discussion
After adjusting for the repeated measurements of the same fetuses and the scan type (i.e. pre- or post-transfusion scan), MRI estimates of hematocrit showed good correlation and excellent agreement with fetal blood sampling data, even as early as 19 weeks of gestation, suggesting that MRI may be more accurate for estimating fetal hematocrit in vivo across wide gestational age ranges. In addition, MRI had higher specificity than Doppler ultrasound, potentially supporting the use of MRI to confirm the need for intra-uterine transfusion in fetuses with elevated MCA-PSV.
Fetal MRI can be challenging, especially in early gestation due to the small vessels and the extensive fetal motion. The feasibility of T2 mapping in human fetal vessels has been demonstrated previously.7,8 Our experience confirms the feasibility of T1 mapping for clinical practice as we were able to obtain diagnostic images as early as 19 weeks of gestation. Our technique is particularly suited to estimate hematocrit during the third trimester, when there is more uncertainty about how to interpret abnormal MCA-PSV findings.
Conclusion
Moderate to severe anemia can be detected by MRI non-invasively with high accuracy and specificity. Our results indicate a potential clinical application for fetal MRI in determining which patients should undergo intrauterine blood transfusion, particularly following previous transfusions and in late gestation.Acknowledgements
This study was supported by the Health Research Grant of Physicians' Services Incorporated Foundation.References
1. Mari G, Norton ME, Stone J, et al. Society for Maternal-Fetal Medicine (SMFM) Clinical Guideline #8: The fetus at risk for anemia-diagnosis and management. Am J Obstet Gynecol [Internet] 2015;212(6):697–710. Available from: http://dx.doi.org/10.1016/j.ajog.2015.01.059
2. Scheier M, Hernandez-Andrade E, Fonseca EB, Nicolaides KH. Prediction of severe fetal anemia in red blood cell alloimmunization after previous intrauterine transfusions. Am J Obstet Gynecol 2006;195(6):1550–6.
3. Friszer S, Maisonneuve E, Macé G, et al. Determination of optimal timing of serial in-utero transfusions in red-cell alloimmunization. Ultrasound Obstet Gynecol 2015;46(5):600–5.
4. Zimmermann R, Durig P, Carpenter RJ, Mari G. Longitudinal measurement of peak systolic velocity in the fetal middle cerebral artery for monitoring pregnancies complicated by red cell alloimmunisation: A prospective multicentre trial with intention-to-treat. BJOG An Int J Obstet Gynaecol 2002;109(7):746–52.
5. Portnoy S, Seed M, Sled JG, Macgowan CK. Non-invasive evaluation of blood oxygen saturation and hematocrit from T 1 and T 2 relaxation times: In-vitro validation in fetal blood. Magn Reson Med [Internet] 2017;00(July 2016):1–8. Available from: http://doi.wiley.com/10.1002/mrm.26599
6. Mari G, Deter RL, Carpenter RL, et al. Noninvasive Diagnosis by Doppler Ultrasonography of Fetal Anemia Due to Maternal Red-Cell Alloimmunization. N Engl J Med [Internet] 2000;342(1):9–14. Available from: http://www.nejm.org/doi/abs/10.1056/NEJM200001063420102
7. Zhu MY, Milligan N, Keating S, et al. The hemodynamics of late-onset intrauterine growth restriction by MRI. Am J Obstet Gynecol [Internet] 2016;214(3):367.e1-367.e17. Available from: http://dx.doi.org/10.1016/j.ajog.2015.10.004
8. Sun L, Macgowan C, Sled J, et al. Reduced Fetal Cerebral Oxygen Consumption is Associated With Smaller Brain Size in Fetuses With Congenital Heart Disease. Circulation 2015;131(15):1313–23.
Figures
Figure 1. Comparison of MRI and laboratory measurements obtained from fetal blood sampling from 23 patients (n=33).
a) A correlation (0.76, p < 0.001) was found between MRI estimates of hematocrit and laboratory hematocrit using generalized estimating equation. Dashed line represents the identity line. b) Bland-Altman analysis showed a bias of -0.03 between the two methods. Dashed line represents the 95% limits of agreements.
Figure 2. Receiver operating characteristic curves for Doppler ultrasonography and MRI pre-transfusion measurements for the prediction of moderate to severe fetal anemia.
Cut off of moderate to severe fetal anemia: 0.65 multiple of median for MRI, 1.50 multiple of median for Doppler.6 Data were compared against the gold-standard fetal blood sampling measurements. In brackets: sensitivity (%) and specificity (%) of the cut points. Areas under curve (AUC) and the p values are shown. * represents significant results.
