Evaluation of Brain Microstructure Changes in Surviving Fetus of Monochorionic Twin Pregnancies With Demise of One Fetus Using Apparent Diffusion Coecient

Objective To compare the differences of (ADC) in twin controls and single controls using and present study investigated the differences of ADC values of brain regions among surviving fetuses, twin controls and single fetus controls. Our study showed lower ADC values of frontal, parietal and occipital lobe in surviving fetuses compared with twin controls and single fetus controls. Furthermore, our ndings showed the negative correlations between ADC values in several brain regions and gestational age.


Background
Monochorionic twin pregnancies have some complications [1][2][3][4][5][6][7] . Demise of one fetus, as one of the main complications, occurs up to 20% [8] , which usually has an effect on the surviving fetus, and accounts for increasing in morbidity and mortality [9][10] . Placental sharing and vascular communications in monochorionic twins might result in clinical symptom. A clinical study including 49 monoamniotic pregnancies with single fetal demise has suggested that severe cerebral injury occurred in 26% of the survivors [11] . How to detect these abnormalities early needs to be explored.
In recent years, fetal MRI, as a valuable complementary technique, is being increasingly used to evaluate fetal brain, provides improved anatomic detail and physiologic information than ultrasound [12][13][14][15][16] . For surviving fetus with demise of one fetus, severe brain damage might be detected by ultrasound and prenatal MR imaging [17] . In some cases, no abnormal echo or signal of brain in the surviving fetus could be found on ultrasound or conventional MRI sequences after the death of one fetus. However, brain abnormalities were found in subsequent prenatal examination or postnatal follow-up. In the circumstances, it is necessary to identify the potential anomalies of brain in surviving fetuses by other methods. Diffusion weighted imaging (DWI) has been applied to fetal MRI to evaluate the microstructure and biophysical status of tissues and intracranial lesions [18][19] , which contributes to identify abnormal ischemic tissue at an early stage. Hoffmann et al [20] reported that acute ischemic lesions can be detected on DWI within one week after the death of one fetus, which can't be detected on ultrasound or conventional MRI. Apparent diffusion coe cient (ADC) is derived from DWI sequence, which can quantitatively detect the degree of diffusion of water molecules. The decrease of ADC value may be an index of brain injuries and an early ischemic manifestation of fetal brain, which can be found before conventional MRI [21][22] . ADC value of fetus brain has been studied in many studies, about fetal maturity and fetal brain injury [23][24][25][26] .
This study is to compare the differences of brain ADC values in surviving fetuses, twin controls and single fetus controls, and to perform follow-up study to nd out whether abnormal ADC values exist or not in surviving fetus, in order to reveal the underlying cerebral microstructure changes.

Subjects
This prospective study enrolled eighty-three pregnant women that were admitted to the department of obstetrics and gynecology in Peking University Third Hospital from January 2018 to March 2020.
Pregnant women who were diagnosed as demise of one fetus were chosen as surviving fetus group. The inclusion criteria were as follows: 1) gestational age ≥ 20 weeks; 2) Twin pregnancy was diagnosed and one fetus died in utero which con rmed by ultrasound.3) No abnormal echo of the brain was found by ultrasound in the surviving fetus.
Healthy twins were chosen as twin control group, and healthy single fetuses were chosen as one fetus control group. The exclusion criteria were as follows: (1) gestational age < 20 weeks; (2)brain abnormalities diagnosed by ultrasound; and (3) any contraindications to MRI examinations.
Both the maternal age and gestational age were recorded. The current study was approved by the medical ethical committee of Peking University Third Hospital. All methods of the study were performed in accordance with the guildlines and regulations of Peking University Third Hospital ethical committee.
Participants under 18 years old were not involved in this study. All the participants of the study signed written informed consents.

The followed-up clinical outcomes
The clinical outcomes were followed up after at least one year, including terminate pregnancy, died or various complication after birth. A trained researcher who was blinded to the MR images did the follow-up via telephone call after at least one year.

MR imaging protocol
The fetal MR image acquisition was performed on a 1.

Image processing
ADC maps were obtained automatically on post-processing software (W4.7, GE Healthcare, Waukesha, Wisconsin, USA) after DWI sequence scanning. Two radiologists with more than 2 years' experience in fetal MR imaging independently drew the regions of interests (ROIs) at each axial slice. ROIs were placed bilaterally over the desired anatomical areas, including white matter of frontal lobes, parietal lobes, temporal lobes and occipital lobes, basal ganglia, thalamus and cerebellum. Each ROI was measured twice by radiologist and the average value was chosen as the nal result. As large a ROI as possible was obtained, avoiding adjacent structures such as cerebrospinal uid spaces. The areas of ROIs were 29.3 mm 2 .

Statistical analysis
Statistical analyses were performed with SPSS software (version 19.0, SPSS, Inc., an IBM Company). One-way ANOVA was used to compare ADC across surviving fetuses, twin controls and single fetus controls in the same ROIs. ADC values in left hemisphere was compared with those in right hemisphere of surviving fetuses by using paired t test. Correlations between ADC values in surviving fetus group and time of one fetus demise, gestational age, mean maternal age were calculated by Pearson or Spearman correlation coe cients. Receiver-operating-characteristic-curve (ROC) was generated and the area-underthe-curve (AUC) of ADC values were calculated in discriminating surviving fetuses and twin controls. P < 0.05 (two-tailed) was considered as statistically signi cant.

Demographic characteristics of the study population
A total of 83 subjects were included in this study. Six patients were excluded from analysis due to poor MR image quality with motion artifacts. There were no signi cant differences of gestational ages and maternal ages among three groups (P > 0.05) ( Table 1). Of 32 demise of one twin, 27 (84.4%) had determined time of one fetus death. No signi cant structural and signal abnormalities in surviving fetus were found on SSFSE and FIESTA sequences ( Fig. 1A-B). No signi cant signal abnormalities were found on DWI sequence (Fig. 1C).

ADC value differences
The mean ADC values of surviving fetuses were shown in Table 2 Table 2). ADC values were lower in bilateral frontal, parietal and occipital lobes of surviving fetuses compared with that of twin controls and single fetus controls (

Follow-up results
Among the 32 surviving fetuses, one was terminated, and one was died after birth. Of the remaining 30 fetuses (mean gestational age, 33.4 ± 2.5 weeks; birth weight, 2739.0 ± 619.0 g), 9 (28%) were born prematurely (1 of which was complicated with congenital pneumonia and cryptorchidism). 4 (13.3%) had lower Apgar scores at birth and about 3-4 points in the rst minute, 7 (23.3%) were followed up by ultrasound and MRI (3 of which had brain abnormalities, 1 of which had enhanced echoes of white matter in the right paraventricular with effusion in the posterior fossa, and 1 had enlarged supratentorial ventricles after birth by MRI) and 1 (3.3%) had obvious growth retardation.

Discussion
The present study investigated the differences of ADC values of brain regions among surviving fetuses, twin controls and single fetus controls. Our study showed lower ADC values of frontal, parietal and occipital lobe in surviving fetuses compared with twin controls and single fetus controls. Furthermore, our ndings showed the negative correlations between ADC values in several brain regions and gestational age.
Meta-analysis showed that the risk of neuro developmental morbidity in monochorionic twins was about 5 times higher compared with that in dichorionic twins after a single fetal death [3] . Previous study reported that cerebral hypoxic-ischemic injury might occur in the surviving fetus after demise of one fetus due to the sharing of one placenta. The surviving fetus might present with hypotension and hypoperfusion due to the loss of circulatory equilibrium and the shunting of blood ow. The reduced cerebral blood ow might be accepted as causative factors for cerebral damage in surviving fetus. In this study, compared with twin and single fetus controls, ADC values of bilateral frontal lobes, parietal lobe and occipital lobes were lower in surviving fetuses. The reduction of ADC values might be indicative of parenchymal damage and metabolic compromise while no abnormalities were detected on ultrasound and MR conventional sequences. Decrease of ADC values might re ect the intracellular/extracellular water compartmentalization [27] , especially the decrease in the extracellular water content. ADC value might be more sensitive to detect the subtle anomalies, even changes of signal were not shown on DWI images. Our ndings further provided evidence of the possibility of detecting potential brain damage by measuring ADC values in surviving fetuses after demise of one fetus. There were no signi cant differences of ADC values in all ROIs between left and right hemisphere in surviving fetuses. The characteristics of symmetry indicated that decreases of ADC values might be caused by cerebral hypoperfusion, but not a single vessel supply.
Our present study showed the negative correlations between ADC values of basal ganglia, thalamus and cerebellum and gestational age in control groups which were consistent with previous study. Decreases of ADC values in the majority of brain during fetal development were reported in previous publication [28] . Signi cant decreases of ADC values were detected in thalamus, basal ganglia, pons and cerebellum with gestational age, but the decrease was not detected in frontal white matter [29][30] . However, the conclusions were controversial. Hoffmann et al [20] found that a weak trend for regional ADC decline was shown in all regions which didn't reach statistical signi cance with brain development. In the present study, ADC values in bilateral frontal, temporal lobes were not correlated with gestational age in two control groups. The above results implied that ADC values in these regions were relatively stable and might be served as developmental indicator. However, ADC values of bilateral thalamus and cerebella, bilateral frontal lobe, parietal lobes were negatively correlated with gestational age in surviving fetuses, which indicated the existed potential damage other than the effects of gestational age in these regions. Our ndings suggested the possibility of ADC values within frontal lobes in distinguishing the potential lesions of surviving fetuses.
In this study, we also found that average ADC values of frontal white matter in single fetus group was lower than that in previous study [28] . The possible reason was that the average gestational age in our study was larger than that in other studies. No signi cant differences of ADC values were found between twin controls and single fetus controls, which suggested that single fetus might serves as control group if there was no suitable twin control group in future study. The follow-up results showed that 3 fetuses subsequently developed brain abnormality by ultrasound or MRI. It indicated the underlying changes of brain might exist even through no abnormal signals on conventional and DWI sequence. The measurement of ADC values in surviving fetuses might help to detect the potential subtle anomalies earlier. Although changes of ADC values were shown in surviving fetuses, making crucial decisions (such as pregnancy termination) on the basis of DWI alone also could be very di cult. It might be more reasonable to discuss the possibility of termination of pregnancy in cases with large cerebral lesion on MR and DWI. For surviving fetuses with ADC values changes alone, the follow-up is necessary. What's more, as fast sequence, DWI has potential value of clinical application due to limitation of many sequences. It could be effective supplement to conventional fetal MRI examination and may detect the underlying changes earlier. Furthermore, the present study showed the good predictive value of ADC values in single lobe including frontal and parietal lobe in discriminating surviving fetuses and twin controls. The predictive value of combination of ADC value of frontal lobes, parietal lobes and gestational age was stronger than that of single lobe alone. The study demonstrated that ADC values might be effective indicators of subtle anomalies in surviving fetuses in future.
There were several limitations in the current study. Firstly, the sample size of the present study is small, and a larger sample group may contribute to more accurate conclusion. Secondly, there might be manual errors as ROI is selected. And thirdly, well-controlled and long-term studies are needed to reveal the relation between reduction of the ADC values and postnatal outcome.

Conclusion
In conclusion, decreases of ADC values were detected in surviving fetus of one fetus demise, when no visible abnormalities were detected on conventional MR. DWI, especially ADC value, is a very useful sequence for detecting underlying changes. ADC values should also be evaluated in larger clinical studies with ongoing pregnancies before adopting it as a formal work-up in cases of one fetus demise.

Declarations
Ethics approval and consent to participate This study was approved by the medical ethics committees of our hospital (IRB00006761-M2017316). Written informed consent was obtained from all subjects in this study.

Consent for publication
Not applicable.

Availability of data and material
All data generated or analysis during this study are included in this published article.

Competing interests
The authors of this manuscript declare no any nancial and non-nancial competing interests.

Funding
This study is funded by Key Clinical Projects of Peking University Third Hospital No. BYSY2017021, which provided enough testing and processing cost for completing the present study.

Authors' contributions
All the authors made equal contributions including conception and study design (L.Y. and H.R.), data collection (L.Y., H.R. and Z.Q.), statistical analysis (L.Y., H.R. and W.Z.), interpretation of results (L.Y. and Y.H.), drafting the manuscript the work or revising it critically for important intellectual content (L.Y., H.R. W.Y. and Y.H.) and approval of nal version to be published and agreement to be accountable for the integrity and accuracy of all aspects of the work (All authors).