Assessment of fetal lung maturity and prediction of neonatal respiratory distress syndrome by MV-Flow imaging

Article information

Ultrasonography. 2025;44(6):470-482

Publication date (electronic) : 2025 August 4

doi : https://doi.org/10.14366/usg.24200

Qi Wang^,*, Yuzhuo Chen^,*, Shijian Xu, Yuhong Lin

Department of Ultrasound, The Fifth Affiliated Hospital of Sun Yat-sen University, Zhuhai, China

Correspondence to: Yuhong Lin, MD, The Fifth Affiliated Hospital of Sun Yat-sen University, No. 52, Meihua East Road, Zhuhai 519000, China Tel. +86-13726219345 Fax. +86-0756-2528888 E-mail: linyh97@mail.sysu.edu.cn

^*

These authors contributed equally to this work.

Received 2024 October 29; Revised 2025 August 3; Accepted 2025 August 4.

Abstract

Graphic Abstract

Introduction

Neonatal respiratory distress syndrome (NRDS) is a leading cause of neonatal morbidity and mortality, particularly among premature infants [1]. The condition arises from pulmonary immaturity and insufficient surfactant, resulting in alveolar collapse and life-threatening breathing difficulties immediately after birth [2]. Although advances in perinatal care, such as maternal corticosteroid administration and neonatal surfactant therapy, have improved outcomes, NRDS continues to present significant risks in preterm neonates, especially those with very low birth weight [3]. Therefore, early identification of fetal lung immaturity is clinically important to guide obstetric decision-making and interventions aimed at preventing NRDS.

For decades, obstetricians have sought reliable prenatal indicators of fetal lung maturity to anticipate the risk of NRDS. Historically, the gold standard involved analyzing amniotic fluid obtained via amniocentesis for biochemical markers of surfactant production, including the lecithin/sphingomyelin ratio and lamellar body count [4], as well as more recent markers such as the lecithin/surfactant ratio [5], phosphatidylglycerol levels [6], and surfactant proteins A and B [7]. Although informative, amniocentesis is invasive and associated with procedure-related risks, making it less common in routine practice. Magnetic resonance imaging (MRI), which assesses the fetal lung-to-liver signal intensity ratio, demonstrates 86% sensitivity and 72% specificity for distinguishing mature from immature fetal lungs [8]. However, MRI is costly, time-consuming, and not feasible for routine use in late pregnancy. In ultrasound diagnostics, researchers have attempted to correlate various prenatal ultrasound findings (e.g., fetal biparietal diameter, liver or diaphragm echo patterns, bone ossification, placental grading, and fetal colonic gas) with fetal lung maturity [9]. Nonetheless, these methods also lack sufficient accuracy. More recently, image-based technologies like QuantusFLM have been developed, enabling quantitative analysis of fetal lung structures to predict NRDS risk [10]. Multicenter studies of QuantusFLM have reported high sensitivity (~90%) and an excellent negative predictive value (NPV; ~98%), making it reliable for identifying truly mature lungs [10-12]. Nevertheless, its positive predictive value (PPV) is only ~50%, indicating many false positives for immaturity, and the method requires strict image-acquisition protocols and proprietary software, limiting broad clinical adoption.

Microvascular flow imaging (MV-Flow) is an emerging third-generation Doppler ultrasound technology that provides high-resolution, non-invasive visualization of slow blood flow and microvascular structures [13]. By using advanced algorithms and high sampling frequencies, it enhances sensitivity and enables detection of small vessels and low-velocity flow that conventional Doppler imaging often misses [14]. MV-Flow has shown promise across multiple specialties. In reproductive medicine, it improves assessment of endometrial receptivity by quantifying microvascular perfusion [15]. In gynecology, it increases the sensitivity and specificity of ovarian tumor diagnosis, particularly in distinguishing benign from malignant masses [16]. In obstetrics, it can evaluate placental microcirculation and detect fetal growth restriction, and its ability to visualize fetal brain microvasculature offers new insights into early neurovascular development and vascular abnormalities [17]. As a non-invasive imaging technique, MV-Flow holds promise for prenatal diagnostics by providing reliable assessments of fetal organ development.

The fetal lung is among the most physiologically active organs in terms of microcirculation, yet real-time in vivo imaging of fetal pulmonary microcirculation remains largely unexplored. This study aims to evaluate, for the first time, the ability of MV-Flow to visualize fetal lung microcirculation and to assess its feasibility for quantitative analysis of fetal pulmonary maturity. The findings are expected to fill a gap in fetal lung microcirculation imaging by providing a non-invasive method for prenatal evaluation of lung development and maturity, thereby offering imaging-based evidence to help reduce the risk of neonatal respiratory complications.

Materials and Methods

Study Sample

This single-center prospective study was conducted at the Fifth Affiliated Hospital of Sun Yat-sen University from May 2023 to August 2024. The study comprised two parts, and the participant flow is shown in Fig. 1. In part 1, a cross-sectional observational study, 167 eligible pregnant women were enrolled. Inclusion criteria were: singleton pregnancy; gestational age confirmed by the last menstrual period and consistent with first-trimester average ultrasound age; and no obvious fetal structural abnormalities on routine prenatal ultrasound. Exclusion criteria were: major maternal comorbidities during pregnancy (e.g., gestational hypertension, diabetes, anemia, hyperthyroidism, autoimmune disease); and fetal structural malformations, chromosomal abnormalities, or other congenital anomalies. Part 2 was a prospective cohort study involving 42 pregnant women who received routine prenatal care and subsequently delivered at the authors’ affiliated hospital. Most participants presented with premature rupture of membranes or had other clinical indications for elective cesarean delivery. Fetuses with structural lung anomalies or abnormal prenatal imaging findings were excluded. All participants underwent fetal lung ultrasound and measurement of the vascular intensity of microvascular volume (VI^MV) within 72 hours before delivery. Neonatal outcomes were recorded, and newborns were categorized into an NRDS group (n=17) or a control group (n=25) based on postnatal diagnosis. Diagnostic criteria for NRDS followed the 2022 European Consensus Guidelines on the Management of Neonatal Respiratory Distress Syndrome [18]. Complete clinical data collection and standardized ultrasound image acquisition were performed for all participants.

Fig. 1.

Flowchart of participant enrollment and grouping in the reference and prediction cohorts.

The study consisted of two parts. Part 1 was a cross-sectional study establishing normal fetal lung VI^MV values, with 167 participants included after exclusion due to poor image quality or unmet inclusion criteria. Part 2 was a prospective cohort study assessing the predictive value of VI^MV for NRDS. After excluding three cases due to incomplete questionnaires, 42 participants remained and were categorized into the NRDS group (n=17) and the non-NRDS control group (NC; n=25). VI^MV, vascular intensity of microvascular volume; NRDS, neonatal respiratory distress syndrome.

Results

Part 1. Evaluation of MV-Flow for Peripheral Pulmonary Vascular Imaging

Baseline characteristics of participants

A total of 167 pregnant women at the Fifth Affiliated Hospital of Sun Yat-sen University underwent fetal lung MV-Flow examinations. Maternal age ranged from 24 to 41 years (mean, 30.3±4.0 years). Body mass index (BMI) ranged from 15.4 to 29.2 kg/m² (mean, 20.7±3.1 kg/m²). Detailed demographic and obstetric characteristics are provided in Table 1. Overall, the cohort represented a healthy pregnant population spanning all trimesters.

Table 1.

Maternal demographic and obstetric characteristics of the study cohort (n=167)

Comparison of vascular visualization by CDFI, PDI, and MV-Flow

Fetal pulmonary vascular imaging was performed using CDFI, PDI, and MV-Flow (Fig. 2). In early gestation (12-15⁺⁶ weeks), MV-Flow showed high sensitivity to slow blood flow, depicting sparse vascular branches in the central lung field. With advancing gestation, MV-Flow visualized increasingly dense and continuous peripheral vascular networks that extended toward the subpleural region. By contrast, CDFI and PDI consistently failed to depict distal vasculature throughout gestation. These observations underscore the technical superiority of MV-Flow for assessing fetal pulmonary microvascular development, particularly peripheral-lung perfusion.

Fig. 2.

Comparison of fetal lung vascular visualization using different imaging modalities across gestation.

Representative images demonstrating peripheral pulmonary vasculature in fetuses at different gestational ages using color Doppler flow imaging (CDFI), power Doppler imaging (PDI), and microvascular flow imaging (MV-Flow). MV-Flow reveals finer vascular detail in peripheral lung regions compared to conventional Doppler techniques.

For larger proximal pulmonary vessels near the hilum, MV-Flow, CDFI, and PDI performed comparably during early to mid-gestation (12-35⁺⁶ weeks). After 36 weeks, MV-Flow image quality in the hilar region was occasionally degraded by signal overflow from cardiac and great-vessel pulsations; however, appropriate adjustment of gain and wall-filter settings effectively mitigated interference and preserved acceptable visualization of proximal vessels.

Consistency and reproducibility of VI^MV under different ROI strategies

Four ROI configurations were evaluated for VI^MV measurement—rectangle, ellipse, whole lung, and peripheral lung—across gestational ages (Fig. 3). Intra- and inter-observer reproducibility was assessed using ICCs and CVs (Fig. 3, Supplementary Table 1). The whole-lung ROI achieved the highest reproducibility, with an intra-observer ICC of 0.966 (95% CI, 0.914 to 0.990), an inter-observer ICC of 0.991 (95% CI, 0.977 to 0.998), and corresponding CVs of 1.04% and 1.96%. The peripheral-lung ROI also showed excellent agreement (intra-observer ICC, 0.925; 95% CI, 0.818 to 0.978; CV, 1.17%; inter-observer ICC, 0.980; 95% CI, 0.947 to 0.994; CV, 1.65%). By comparison, the ellipse ROI had lower repeatability (intra-observer ICC, 0.607; CV, 3.38%), and the rectangle ROI performed worst (intra-observer ICC, 0.575; CV, 2.68%). These findings indicate that whole-lung and peripheral-lung ROIs yield stable, reliable measurements, whereas rectangular and elliptical ROIs may introduce greater operator-dependent variability.

Fig. 3.

Comparison of ROI strategies for fetal lung VI^MV measurement and reproducibility.

A. Representative MV-Flow images show the fetal lungs at 24 and 33 weeks of gestation using four ROI strategies: rectangle, ellipse, whole lung, and peripheral lung. ROI boundaries are indicated in yellow. B. In a reproducibility analysis of the four ROI strategies, the left and middle panels show inter- and intra-observer ICC (95% CI); the right panel displays intra- and inter-observer coefficient of variation (CV). Whole-lung and peripheral-lung ROIs demonstrated higher reproducibility with lower measurement variability compared to rectangle and ellipse ROIs. ROI, region of interest; VI^MV, vascular intensity of microvascular volume; MV-Flow, microvascular flow imaging; ICC, intraclass correlation coefficient; CI, confidence interval.

Gestational trends and statistical comparison of fetal lung VI^MV

To examine the relationship between gestational age and fetal pulmonary vascular development, VI^MV values were analyzed by gestational-age segment in 167 fetuses using both whole-lung and peripheral-lung ROIs (Table 2, Fig. 4, Supplementary Table 2). Regardless of ROI, VI^MV increased progressively with gestation. Mean whole-lung VI^MV rose from 62.0%±11.0% at 12-14⁺⁶ weeks to 97.9%±1.3% at 39-42 weeks, while peripheral-lung VI^MV increased from 42.6%±13.7% to 96.2%±2.7%. Differences across gestational segments were statistically significant for both ROI types (P<0.001). Throughout gestation, whole-lung VI^MV values were consistently higher than peripheral-lung values (P<0.001); however, the increase in peripheral-lung VI^MV was steeper—especially in mid and late gestation—narrowing the gap between measures. At 36-38⁺⁶ weeks, the difference between whole-lung and peripheral-lung VI^MV became statistically non-significant (P=0.130), suggesting nearequal perfusion.

Table 2.

Comparison of fetal lung VI^MV values across gestational age segments

Fig. 4.

Gestational trends and variability in fetal lung VI^MV measured by MV-Flow.

A. Box-and-whisker plots show whole-lung and peripheral-lung VI^MV (%) across gestational ages (12-42 weeks). B. Plot showing variation in fetal lung VI^MV measurements (expressed as CV) across different gestational age segments. VI^MV, vascular intensity of microvascular volume; MV-Flow, microvascular flow imaging; CV, coefficients of variation.

Measurement variability declined with advancing gestation (Fig. 4B, Supplementary Table 3). At 12-14⁺⁶ weeks, the CVs for whole-lung and peripheral-lung VI^MV were 31.48% and 21.58%, respectively; these decreased to 12.72% and 7.16% at 16-24⁺⁶ weeks, and to 6.69% and 4.07% at 25-34⁺⁶ weeks, indicating greater perfusion stability as the lungs matured.

Spearman correlation analysis demonstrated strong positive associations between gestational age and both whole-lung VI^MV (r=0.821, P<0.001) and peripheral-lung VI^MV (r=0.860, P<0.001), highlighting gestational age as a key determinant of fetal pulmonary microvascular development.

Part 2. VI^MV-Based Fetal Lung Maturity Assessment and NRDS Prediction

Maternal and neonatal clinical characteristics

Forty-two pregnant women and their neonates were included in the prospective cohort. Maternal and neonatal characteristics are summarized in Table 3. Maternal age (30.5±3.4 vs. 31.7±5.1 years, P=0.396) and BMI (19.44±2.07 vs. 20.81±3.14 kg/m², P=0.123) were comparable between the NRDS and control groups. The NRDS group had a higher proportion of nulliparous women (88.2% vs. 60.0%, P=0.047), pregnancy complications (58.5% vs. 24.0%, P=0.023), and antenatal corticosteroid use (29.4% vs. 4.0%, P=0.021).

Table 3.

Comparison of maternal and neonatal characteristics between NRDS and control groups

Neonatal outcomes also differed. Birth weight was lower in the NRDS group (2,811±471 g vs. 3,102±316 g, P=0.021), and 5-minute Apgar scores <7 were more frequent (23.5% vs. 0%, P=0.011). All NRDS cases required neonatal intensive care unit (NICU) admission (100% vs. 8.0%, P<0.001), with longer NICU stays (8.4 vs. 0.7 days, P=0.001). These findings reflect the severity of NRDS and its association with adverse neonatal outcomes.

Comparison of fetal lung VI^MV between NRDS and control group

At equivalent gestational-age segments, fetal lung VI^MV was significantly lower in the NRDS group than in controls (Fig. 5, Supplementary Table 4). At 34-38⁺⁶ weeks, whole-lung VI^MV was 96.13%±1.90% in the NRDS group versus 98.12%±1.05% in controls (P=0.008), and peripheral-lung VI^MV was 93.71%±2.77% versus 96.49%±1.61% (P=0.013). At 39-42 weeks, absolute differences narrowed but statistical significance increased, likely due to reduced variability at term: whole-lung VI^MV was 96.76%±0.89% in NRDS versus 98.6%±0.56% in controls (P<0.001), and peripheral-lung VI^MV was 93.4%±1.71% versus 97.54%±0.89% (P<0.001). These results suggest reduced prenatal pulmonary microvascular perfusion in NRDS, independent of gestational age.

Fig. 5.

Comparison of fetal lung VI^MV values between NRDS and control groups across gestational age segments.

Whole-lung VI^MV (A) and peripheral-lung VI^MV (B) values are plotted by gestational age (left) for NRDS (red squares/triangles) and control (blue circles) group. The VI^MV values grouped by gestational age segments (right) showed significant differences between NRDS and control group (^*P<0.05, ^**P<0.01, ^***P<0.001). VI^MV, vascular intensity of microvascular volume; NRDS, neonatal respiratory distress syndrome; NC, normal control group.

Logistic regression analysis of VI^MV and NRDS risk

Two multivariable logistic regression models evaluated whether fetal lung VI^MV was independently associated with NRDS, adjusting for gestational age, fetal sex, birth weight, and maternal age (Table 4). In model 1 (whole-lung VI^MV), each 1% increase was associated with a 73% reduction in NRDS risk (OR, 0.270; 95% CI, 0.115 to 0.633; P=0.003). In Model 2 (peripheral-lung VI^MV), each 1% increase reduced risk by 65% (OR, 0.354; 95% CI, 0.184 to 0.682; P=0.002). None of the other covariates were significantly associated with NRDS (all P>0.1). These findings indicate that lower VI^MV is an independent and robust predictor of NRDS.

Table 4.

Multivariable logistic regression analysis of factors associated with NRDS

Diagnostic performance of fetal lung VI^MV for NRDS prediction

Using gestational age–adjusted reference thresholds, diagnostic performance was assessed for both ROI strategies (Table 5). Model 2 (peripheral-lung VI^MV) demonstrated better overall performance, with a sensitivity of 82.4% (14/17), a specificity of 84.0% (21/25), a PPV of 77.8% (14/18), and an NPV of 87.5% (21/24). The overall accuracy was 83.3%.

Table 5.

Diagnostic performance of VI^MV measured by MV-Flow for predicting NRDS

In contrast, model 1 (whole-lung VI^MV) demonstrated limited diagnostic value, with a sensitivity of 11.8% (2/17), specificity of 100% (25/25), PPV of 100% (2/2), and NPV of 62.5% (25/40). Although no false positives were detected, the very low sensitivity reflects poor case detection and limited clinical utility for screening.

These findings suggest that model 2 (peripheral-lung VI^MV) offers a more balanced and clinically useful measure for NRDS risk prediction than model 1 (whole-lung VI^MV).

Discussion

This study demonstrates the feasibility and clinical value of using MV-Flow to assess fetal pulmonary microvascular perfusion and to predict NRDS. The principal findings are as follows: (1) MV-Flow enables non-invasive, real-time visualization of the fetal lung microvasculature; (2) VI^MV increases with gestational age, providing a quantitative index of fetal pulmonary vascular development and maturity; (3) fetuses who later developed NRDS consistently exhibited lower VI^MV values than gestational age–matched controls; and (4) peripheral-lung VI^MV showed superior diagnostic performance for identifying fetuses at high risk of NRDS. Collectively, these results suggest that impaired prenatal pulmonary microvascular development may contribute to neonatal respiratory morbidity and that MV-Flow offers a novel, non-invasive imaging approach to assess fetal lung maturity and predict NRDS risk.

The analysis of fetal pulmonary microvascular development showed a gestational-age–dependent rise in VI^MV, with the most rapid increase between 16 and 24 weeks. After 25 weeks, the rate slows, and by 35 weeks VI^MV stabilizes, with average values reaching 93% for the whole lung and 96% for the peripheral lung. This trajectory aligns with established stages of fetal lung development. In the pseudoglandular phase (5-16 weeks), rapid bronchial branching and pulmonary artery formation occur, but terminal respiratory units have not yet formed, yielding sparse microvasculature in the peripheral lung parenchyma [23]; correspondingly, VI^MV values are low and variable in early gestation (e.g., 62% for the whole lung and 42% for the peripheral lung at 12-14 weeks). During the canalicular phase (16-25 weeks), the emergence of terminal bronchioles and primitive alveolar sacs coincides with brisk capillary proliferation [23], which accounts for the observed increase in VI^MV. In the saccular and alveolar stages (after 25 weeks), the expansion of alveolar sacs and ongoing capillary growth significantly enhance the alveolar gas exchange capacity, marking the maturation of the fetal lung [23]. By 32 weeks, surfactant secretion by type II alveolar cells reaches functional levels, and although alveolar numbers continue to rise, capillary perfusion density stabilizes [23], resulting in a plateau of VI^MV. At term (≥36 weeks), VI^MV values for the whole lung and peripheral lung exceed 97% and 95%, respectively, indicating a mature and well-perfused pulmonary microvascular network that underpins the structural and functional capacity required for independent breathing after birth. This pattern aligns with clinical observations, as term fetuses generally exhibit lung maturity sufficient for spontaneous respiration and have a significantly reduced risk of NRDS.

Statistical analysis confirmed strong correlations between VI^MV and gestational age (whole lung r=0.821, peripheral lung r=0.860; both P<0.001), reinforcing VI^MV as a robust marker of fetal lung maturity. Furthermore, both whole-lung and peripheral-lung ROI strategies demonstrated excellent reproducibility (ICC, 0.966-0.991 and 0.925-0.980, respectively), supporting their suitability as standard ROI configurations for future research.

Regarding the relationship between VI^MV and NRDS, fetuses who developed NRDS exhibited VI^MV values that were 2%-4% lower than age-matched controls. At 39-42 weeks, for example, whole-lung VI^MV in the NRDS group was 96.8%±0.9% versus 98.6%±0.6% in controls (P<0.001), and peripheral-lung VI^MV was 93.4%±1.7% versus 97.5%±0.9% (P<0.001). Although numerically small, these differences were statistically significant and likely reflect subtle deficits in pulmonary microvascular perfusion. The persistence of VI^MV disparities even at term indicates that chronological age does not always equate to functional maturity, consistent with prior observations that NRDS can occur in term neonates owing to delayed surfactant synthesis, aberrant pulmonary vascular remodeling, or maternal diabetes [24].

Multivariable logistic regression further showed that lower VI^MV is independently associated with higher NRDS risk. Each 1% increase in VI^MV corresponded to a 65%-73% reduction in the odds of NRDS (whole lung OR, 0.27; P=0.003; peripheral lung OR, 0.35; P=0.002), indicating that VI^MV provides predictive information beyond traditional perinatal risk factors such as gestational age and birth weight.

Notably, peripheral-lung VI^MV outperformed the whole-lung measure. Whereas whole-lung VI^MV showed very low sensitivity (12%) and perfect specificity (100%), peripheral-lung VI^MV achieved balanced performance with sensitivity of approximately 79% and specificity of approximately 83% for NRDS prediction. This pattern suggests that the peripheral microvasculature is more sensitive to delayed maturation; focusing on this region reveals clinically meaningful differences in lung maturity. Although the PLR did not reach the ideal threshold, the data support the potential utility of peripheral VI^MV as a screening tool. Larger samples and gestational age–stratified models are needed to refine diagnostic thresholds and optimize clinical pathways.

In comparison with other methods for assessing fetal lung maturity, peripheral-lung VI^MV showed a balance between sensitivity and specificity: in the present study, its sensitivity (~79%) was slightly lower than that of texture analysis (90%) [12], but its specificity (83%) outperformed gray level histogram width (72%) [25], and its PPV (78%) exceeded that of QuantusFLM (50%) [12]. Although the NPV was marginally lower than that of QuantusFLM, it remained above 86%, which is acceptable for clinical screening. Crucially, VI^MV can be measured on widely available ultrasound platforms without specialized sequences or proprietary post-processing, making it a simple, non-invasive, real-time technique suitable for routine practice.

Several limitations merit consideration. First, the sample size for the NRDS-prediction analysis was limited, which may constrain generalizability of the diagnostic estimates. Moreover, because the cohort mainly included high-risk deliveries, findings may not directly extrapolate to low-risk or routine obstetric populations. Larger, multicenter studies are needed to provide more precise estimates of sensitivity, specificity, and optimal VI^MV cutoffs. Second, this single-center study used specific ultrasound settings; variations in operator technique and machine tuning can influence Doppler-based measurements. Although the protocol was standardized and inter-observer agreement was high, further work is needed to enhance cross-site comparability. The development of standardized MV-Flow imaging guidelines would facilitate consistent implementation. Finally, the current manual ROI delineation is operator-dependent. Future studies should evaluate AI-based approaches for automated ROI definition and VI^MV computation to improve consistency and scalability.

MV-Flow imaging enables real-time visualization of the fetal pulmonary microcirculation, providing a novel in utero approach to assessing fetal lung maturity. It was demonstrated that VI^MV increases with gestational age—reflecting normal pulmonary development—and is significantly lower in fetuses who later develop NRDS. Low prenatal VI^MV emerged as an independent predictor of NRDS, suggesting that subtle deficits in fetal pulmonary microvascular development contribute to postnatal respiratory morbidity. Taken together, these findings highlight the clinical value of fetal lung MV-Flow imaging: it offers a feasible and reproducible, non-invasive biomarker of lung maturity that may improve NRDS risk stratification. Incorporating this imaging-based assessment into practice could guide timely interventions to prevent neonatal respiratory complications and improve outcomes. Despite study limitations, ongoing technical optimization and broader validation support the potential of fetal lung microvascular indicators for assessing maturity and predicting neonatal respiratory outcomes. These results provide a foundation for larger-scale studies and point to new directions for clinical prenatal prediction and intervention in neonatal respiratory distress syndrome.

Supplementary Material

Article information Continued

Notes