Predicting the risk of fetal growth restriction by radiomics analysis of the placenta on T2WI: A retrospective case-control study

A machine learning model based on placental magnetic resonance imaging and clinical factors to predict fetal growth restriction

OBJECTIVES

To create a placental radiomics-clinical machine learning model to predict FGR.

MATERIALS AND METHODS

Retrospectively analyzed placental MRI and clinical data of 110 FGR cases and 158 healthy controls at 28-37 weeks of gestation from two campuses of ZWH. 227 cases from Hubin campus were randomly divided into training (n = 182) and internal testing set (n = 45). 41 cases from Xiaoshan campus were included in external testing set. Placental MRI features were extracted from sagittal T2WI. Mann-Whitney U test, redundancy analysis, and LASSO were used to identify the radiomics signature, and the best-performing radiomics model was constructed by comparing eight machine learning algorithms. Clinical factors determined by univariate and multivariate analyses. A united model and nomogram combining the radiomics Rad-score and clinical factors were established. The performance of the models was assessed by DeLong test, calibration curve and decision curve analysis.

RESULTS

Of 1561 radiomics features, 10 strongly correlated with FGR were selected. The radiomics model using logistic regression performed best compared eight algorithms. 5 important clinical features identified by analysis. The united model demonstrated a good predictive performance in the training, internal testing and external testing sets, with AUC 0.941 (95% CI, 0.0.904-0.977), 0.899 (95% CI, 0.789-1) and 0.861 (95% CI 0.725-0.998), prediction accuracies 0.885, 0.844 and 0.805, precisions 0.871, 0.789 and 0.867, recalls 0.836, 0.833 and 0.684, and F1 scores 0.853, 0.811 and 0.765, respectively. The calibration and decision curves of the united model also showed good performance. Nomogram confirmed clinical applicability of the model.

CONCLUSIONS

The proposed placental radiomics-clinical machine learning model is simple yet effective to predict FGR.

Measuring and interpreting individual differences in fetal, infant, and toddler neurodevelopment

As scientists interested in fetal, infant, and toddler (FIT) neurodevelopment, our research questions often focus on how individual children differ in their neurodevelopment and the predictive value of those individual differences for long-term neural and behavioral outcomes. Measuring and interpreting individual differences in neurodevelopment can present challenges: Is there a "standard" way for the human brain to develop? How do the semantic, practical, or theoretical constraints that we place on studying "development" influence how we measure and interpret individual differences? While it is important to consider these questions across the lifespan, they are particularly relevant for conducting and interpreting research on individual differences in fetal, infant, and toddler neurodevelopment due to the rapid, profound, and heterogeneous changes happening during this period, which may be predictive of long-term outcomes. This article, therefore, has three goals: 1) to provide an overview about how individual differences in neurodevelopment are studied in the field of developmental cognitive neuroscience, 2) to identify challenges and considerations when studying individual differences in neurodevelopment, and 3) to discuss potential implications and solutions moving forward.

Placental quantitative susceptibility mapping and T2* characteristics for predicting birth weight in healthy and high-risk pregnancies

BACKGROUND

The human placenta is critical in supporting fetal development, and placental dysfunction may compromise maternal-fetal health. Early detection of placental dysfunction remains challenging due to the lack of reliable biomarkers. This study compares placental quantitative susceptibility mapping and T2* values between healthy and high-risk pregnancies and investigates their association with maternal and fetal parameters and their ability to predict birth weight (BW).

METHODS

A total of 105 pregnant individuals were included: 68 healthy controls and 37 high-risk due to fetal growth restriction (FGR), chronic or gestational hypertension, and pre-eclampsia. Placental magnetic resonance imaging data were collected using a three-dimensional multi-echo radiofrequency-spoiled gradient-echo, and mean susceptibility and T2* values were calculated. To analyze associations and estimate BW, we employed linear regression and regression forest models.

RESULTS

No significant differences were found in susceptibility between high-risk pregnancies and controls (p = 0.928). T2* values were significantly lower in high-risk pregnancies (p = 0.013), particularly in pre-eclampsia and FGR, emerging as a predictor of BW. The regression forest model showed placental T2* as a promising mode for BW estimation.

CONCLUSION

Our findings underscore the potential of mean placental T2* as a more sensitive marker for detecting placental dysfunction in high-risk pregnancies than mean placental susceptibility. Moreover, the high-risk status emerged as a significant predictor of BW. These results call for further research with larger and more diverse populations to validate these findings and enhance prediction models for improved pregnancy management.

RELEVANCE STATEMENT

This study highlights the potential of placental T2* magnetic resonance imaging measurements as reliable indicators for detecting placental dysfunction in high-risk pregnancies, aiding in improved prenatal care and birth weight prediction.

KEY POINTS

Placental dysfunction in high-risk pregnancies is evaluated using MRI T2* values. Lower T2* values significantly correlate with pre-eclampsia and fetal growth restriction. T2* MRI may predict birth weight, enhancing prenatal care outcomes.

A Hybrid Model for Fetal Growth Restriction Assessment by Automatic Placental Radiomics on T2-Weighted MRI and Multifeature Fusion

BACKGROUND

MRI-based placental analyses have been used to improve fetal growth restriction (FGR) assessment by complementing ultrasound-based measurements. However, these are still limited by time-consuming manual annotation in MRI data and the lack of mother-based information.

PURPOSE

To develop and validate a hybrid model for accurate FGR assessment by automatic placental radiomics on T2-weighted imaging (T2WI) and multifeature fusion.

STUDY TYPE

Retrospective.

POPULATION

274 pregnant women (29.5  ±  4.0 years) from two centers were included and randomly divided into training (N = 119), internal test (N = 40), time-independent validation (N = 43), and external validation (N = 72) sets.

FIELD STRENGTH/SEQUENCE

1.5-T, T2WI half-Fourier acquisition single-shot turbo spin-echo pulse sequence.

ASSESSMENT

First, the placentas on T2WI were manually annotated, and a deep learning model was developed to automatically segment the placentas. Then, the radiomic features were extracted from the placentas and selected by three-step feature selection. In addition, fetus-based measurement features and mother-based clinical features were obtained from ultrasound examinations and medical records, respectively. Finally, a hybrid model based on random forest was constructed by fusing these features, and further compared with models based on other machine learning methods and different feature combinations.

STATISTICAL TESTS

The performances of placenta segmentation and FGR assessment were evaluated by Dice similarity coefficient (DSC) and the area under the receiver operating characteristic curve (AUROC), respectively. A P-value <0.05 was considered statistically significant.

RESULTS

The placentas were automatically segmented with an average DSC of 90.0%. The hybrid model achieved an AUROC of 0.923, 0.931, and 0.880 on the internal test, time-independent validation, and external validation sets, respectively. The mother-based clinical features resulted in significant performance improvements for FGR assessment.

DATA CONCLUSION

The proposed hybrid model may be able to assess FGR with high accuracy. Furthermore, information complementation based on placental, fetal, and maternal features could also lead to better FGR assessment performance.

LEVEL OF EVIDENCE: 3

TECHNICAL EFFICACY

Stage 2.

Prediction of hemorrhage in placenta previa: Radiomics analysis of pelvic MRI images

INTRODUCTION

Prediction of intraoperative massive hemorrhage is still challenging in placenta previa. Radiomics analysis has been investigated as a new evaluation method for analyzing medical images. We used radiomics analysis on placental magnetic resonance imaging (MRI) images to predict intraoperative hemorrhage in placenta previa.

METHODS

We used the sagittal MRI T2-weighted sequence in placenta previa. Using the rectangular region from the uterine os to the anterior wall, we extracted 97 radiomics features. We also collected patient demographics and blood test data as clinical variables. Combining these radiomics features and clinical variables, logistic regression models with a stepwise method were built to predict the risk of hemorrhage, defined as blood loss of > 2000 ml. We evaluated the prediction performance of the models using accuracy and area under the curve (AUC), also analyzing the important variables for the prediction by stepwise methods.

RESULTS

We enrolled a total of 63 placenta previa cases including 30 hemorrhage cases from two institutes. The model combining clinical variables and radiomics features showed the best prediction performance with an accuracy of 0.70 and an AUC of 0.69 in the internal validation data, and accuracy of 0.41 and an AUC of 0.70 in the external validation data, compared with human experts (accuracy of 0.62). Regarding variable selection, two radiomics features. 'original_glrlm_LowGrayLevelRunEmphasis,' and 'diagnostics_Image-original_Minimum,' were important predictors for hemorrhage by the stepwise method.

DISCUSSION

Radiomics features based on MRI could be used as effective predictive variables for hemorrhage in placenta previa. Radiomics analysis of placental imaging could lead to further analysis of quantitative variables related to obstetric diseases.

Fetal MRI: what's new? A short review

Fetal magnetic resonance imaging (fetal MRI) is usually performed as a second-level examination following routine ultrasound examination, generally exploiting morphological and diffusion MRI sequences. The objective of this review is to describe the novelties and new applications of fetal MRI, focusing on three main aspects: the new sequences with their applications, the transition from 1.5-T to 3-T magnetic field, and the new applications of artificial intelligence software. This review was carried out by consulting the MEDLINE references (PubMed) and including only peer-reviewed articles written in English. Among the most important novelties in fetal MRI, we find the intravoxel incoherent motion model which allow to discriminate the diffusion from the perfusion component in fetal and placenta tissues. The transition from 1.5-T to 3-T magnetic field allowed for higher quality images, thanks to the higher signal-to-noise ratio with a trade-off of more frequent artifacts. The application of motion-correction software makes it possible to overcome movement artifacts by obtaining higher quality images and to generate three-dimensional images useful in preoperative planning.Relevance statementThis review shows the latest developments offered by fetal MRI focusing on new sequences, transition from 1.5-T to 3-T magnetic field and the emerging role of AI software that are paving the way for new diagnostic strategies.Key points• Fetal magnetic resonance imaging (MRI) is a second-line imaging after ultrasound.• Diffusion-weighted imaging and intravoxel incoherent motion sequences provide quantitative biomarkers on fetal microstructure and perfusion.• 3-T MRI improves the detection of cerebral malformations.• 3-T MRI is useful for both body and nervous system indications.• Automatic MRI motion tracking overcomes fetal movement artifacts and improve fetal imaging.