Project Summary/Abstract
Twin-to-Twin Transfusion Syndrome (TTTS) is a serious complication of twin pregnancies, affecting up to 15%
of twins sharing the same placenta (monochorionic twins). TTTS occurs when monochorionic twins develop
abnormal connections (anastomosis) in placental vessels resulting in imbalanced blood flow: one twin (the
recipient) receives more blood at the expense of the other (donor) twin with both at increased risk for demise
and compromised neurodevelopmental outcomes. The natural history of TTTS is dismal with perinatal losses up
to 70-100%. Although fetoscopic laser photocoagulation of the placental anastomoses has improved outcomes,
losses remain as high as 34% with cerebral palsy and neurodevelopmental impairment occurring in 10-18%.
Thus, margins for improvement remain and controversy exists over optimal type of laser surgery, appropriate
definition of TTTS and when to intervene. Surgical planning requires accurate localization of the umbilical cord
insertions and placenta surface vessels, which remains challenging with 2D magnetic resonance (MR) or
ultrasound (US) images. Diagnosis of TTTS and response to laser surgery is based on user dependent, indirect
measures of placental function (e.g. umbilical arterial and venous Doppler assessment, maximum vertical pocket
of amniotic fluid, fetal bladder size). In addition, fetal growth discordance, a critical reflection of TTTS severity, is
also based on user dependent 2D US techniques. Improved preoperative placental vascular visualization, direct
measures of placental function and user independent measures of fetal size are needed. We propose to build
on our prior work developing CNN to perform rapid 3D uterine reconstructions with 2D HASTE images with
flattening of placental images to improve visualization of the placental surface vessels. To directly assess
placental function, we will build upon our prior work where we found oxygen transport correlated with discordance
and predicted birth weight in monochorionic twins. Instead of monitoring relative T2* changes, we propose to
quantify T2* at baseline and with maternal oxygen. We also use intravoxel incoherent motion (IVIM) with and
without flow compensation (FC) to explore pulsatile flow in the fetal villous tree. Finally, we will extend our prior
work that identified key points for fetal pose providing limb length and motion statistics in singletons, to twins and
add estimation of fetal intracranial, bladder and body volumes. In summary, to address the unmet needs in
monochorionic twins, we propose the following aims: Aim 1) Provide rapid visualization of the placental surface
vessels and umbilical cord insertion in native and flattened space; Aim 2) Characterize local placental function
pre and post laser ablation; and Aim 3) Calculate discordance of intracranial, bladder and body volumes, fetal
limb biometrics and motion statistics pre and 4-6 weeks post treatment. If successful, we will be poised to perform
a clinical study assessing the impact of these advancements on surgical planning and outcomes and determine
application to earlier gestational ages.