Efficacy of Lung Ultrasound vs Chest X-Ray in Detecting Lung Consolidation and Edema in Premature Infants in the NICU

Introduction

Traditionally, the presence of air in the lungs has been considered a limitation in ultrasound scanning due to various artifacts [1,2]. However, in recent years, these artifacts generated by air have been discovered to contain valuable information and are linked to lung pathology [3,4]. Lung ultrasound (LUS) is gaining increasing recognition among clinicians, particularly in the diagnosis of neonatal lung diseases. Pediatricians prefer the non-radiation and bedside convenience of LUS [5–7], which may alter perceptions of characteristic pulmonary manifestations in certain neonatal conditions [8–10]. Differences in imaging characteristics between bedside LUS and CXR may contribute significantly to these changes [8–11]. This study evaluated the diagnostic efficacy of bedside LUS and CXR for the 2 most common indicators of pulmonary consolidation and edema in premature infants.

Material and Methods

ETHICS STATEMENT:

This retrospective observational study was conducted in the neonatal intensive care unit (NICU). The utilization of patients’ electronic records in this study was approved by the Hospital’s Ethics Committee (Approval Date: September 6, 2021, Approval Number: KY 202109006). LUS and CXR images were retrieved from the Picture Archiving and Communication Systems (PACS) and anonymized before analysis.

SCREENING CRITERIA:

The LUS, CXR, and clinical records of premature newborns (gestational age less than 37 weeks) admitted to the NICU were screened from November 1, 2019 to December 31, 2021. Both CXR and LUS scans for each newborn had to be conducted within a single day. During the interval between the 2 examinations, there should be no significant changes in the clinical presentation and treatment measures for the newborns. Their stable condition was evidenced by parameters such as blood oxygen saturation, pulmonary respiratory sound, and ongoing treatment strategy (ventilator parameters, dosage of pulmonary surfactant and other medications), all of which were retrieved from the electronic medical record system. Exclusion criteria encompassed congenital pulmonary and cardiac abnormalities, such as patent ductus arteriosus (PDA) affecting cardiac function, pulmonary vascular malformation, and complex congenital heart disease. Additionally, patients were excluded if they experienced significant clinical changes between the 2 examinations, including but not limited to acute respiratory distress, marked fluctuations in blood oxygen saturation, or the need for emergency medical interventions. The enrolled neonates were classified into normal, mild, moderate, and severe categories based on the severity of respiratory distress (RD) on the day of examinations.

INSTRUMENTS AND SETTING:

All LUS images were captured using Philips CX50 United States instruments (Philips Healthcare, Best, the Netherlands) with a high-frequency shallow probe (12 MHz) set to a depth of 4–5 cm. LUS scans were conducted in both horizontal and vertical orientations. A 6-area division method was employed during LUS scanning, dividing each lung into anterior, posterior, and lateral areas, with the anterior and posterior axillary lines as boundaries [12,13]. The scanning procedures were carried out by 2 trained sonographers.

CXR images were obtained using a Siemens Mobilett Mira Max machine (X chest ap baby) set at 52 kV and 1.0 mAs. All examinations were performed while the newborns were in a calm state.

IMAGE INTERPRETATION:

The LUS images were extracted from the PACS system and analyzed by a sonographer specialized in lung ultrasound interpretation. The extent of lung consolidation across the 6 areas was assessed, measuring the maximum range and identifying whether the consolidation was singular or multiple. The findings were categorized into 2 groups: the small consolidation group, indicating that the maximum diameter of consolidation was less than 2.0 cm; and the larger or multiple consolidations group, where the maximum diameter of consolidation exceeded 2.0 cm or multiple consolidations were present.

Simultaneously, through analysis of ultrasound images from the 6 lung areas, the region with the highest concentration of B-lines was identified as indicative of interstitial edema. Based on the number and density of B-lines, results were categorized into 4 types [12,13]:

The CXR images were analyzed by a neonatologist., who examined the CXR scans and documented any findings related to consolidation, reduction in lung transparency, or alterations in lung markings on the images. Both the sonographer and the neonatologist responsible for interpreting the images were unaware of the patient’s condition and any other imaging diagnostic results. Two observers conducted separate evaluations of all recordings.

STATISTICAL ANALYSIS:

Numerical variables were reported as median (interquartile range). Categorical variables were presented as counts and percentages. A chi-square test or Fisher’s exact test was used to compare categorical variables between the 2 imaging diagnostic methods, especially when data were limited. Results with a P value of less than 0.05 were deemed statistically significant. Statistical analyses were performed using SPSS version 23.0 (IBM Corp).

Results

We screened 259 premature newborns admitted to the NICU between November 2019 and December 2021. After screening, a total of 75 LUS images and their corresponding 75 CXR images from 75 premature newborns were included in the study. The screening process is shown in the flowchart (Figure 1). The 75 newborns had a mean birth weight of 1404.67±473.45 grams and a gestational age of 29.75±2.34 weeks. Among these newborns, 47 (62.7%) were male, and 28 (37.3%) were female. Of these 75 infants, the most common diagnoses were respiratory distress syndrome (RDS; 23 patients; 30.7%), pneumonia (including intrauterine infectious pneumonia and community-acquired pneumonia; 19 patients; 25.3%), transient tachypnea of the newborn (TTN; 13 patients; 17.3%), meconium aspiration syndrome (MAS; 9 patients; 12.0%), hyperbilirubinemia (6 patients; 8.0%), pulmonary hemorrhage (PH; 3 patients; 4.0%), and pneumothorax (2 patients; 2.7%). Among the 75 preterm infants, 14 were classified as normal (18.7%), 37 as mild (49.3%), 16 as moderate (21.3%), and 8 as severe (10.7%) based on the severity of RD during examination (Table 1).

Out of the 75 cases, 34 cases exhibited lung consolidation on LUS (45.3%), comprising 23 instances of subpleural consolidation and 11 cases of consolidations greater than 2.0 cm in maximum diameter or multiple consolidations. Among the 75 CXR images, 14 images displayed consolidation (18.7%). The detection rate of LUS for identifying consolidations was significantly higher compared to CXR (34/75 vs 14/75, P<0.001). In the 23 cases demonstrating consolidation on LUS but not on CXR, there were 16 instances of subpleural consolidation and 7 cases of consolidations greater than 2.0 cm or multiple consolidations on LUS.

Among the 75 cases, 22 cases showed B-lines on LUS, which exhibited normal lung markings or only slight changes on CXR. The numbers of cases with confluent B-lines, alveolar-interstitial syndrome, compact B-lines, or white lung detected on LUS were 29, 15, and 9, respectively (29/75, 15/75, 9/75). Among the 53 cases displaying confluent B-lines, alveolar-interstitial syndrome, compact B-lines, or white lung on LUS, the CXR images did not clearly discriminate the severity of lung edema. Only 21 out of these 53 cases showed reduced transparency on CXR. There was a significant disparity between LUS and CXR in detecting interstitial edema (53/75 vs 21/75, P<0.001). Regarding the 9 cases with compact B-lines or white lung on LUS, 3 of them solely demonstrated increased lung markings and a slight decrease in lung transparency for both lungs (Table 2).

Discussion

Lung disease is a major cause of respiratory distress in neonates, especially in premature infants who exhibit lung immaturity, both structurally and functionally. Severe lung conditions can lead to respiratory failure and potentially fatal outcomes. Xirouchaki et al [14] compared the diagnostic performance of LUS and CXR in detecting pathologic abnormalities in critically ill patients, using thoracic CT as the criterion standard. Forty-two mechanically ventilated patients were evaluated for consolidation, interstitial syndrome, pneumothorax, and pleural effusion. LUS showed higher sensitivity, specificity, and diagnostic accuracy than CXR for most conditions. Nevertheless, even a low dose of CXR (<0.2 mGy) can be harmful to newborns, particularly premature infants, as their rapidly dividing cells lack the ability to efficiently repair mutated DNA, rendering them more vulnerable to the effects of X-ray radiation [15]. Given their often-extended hospital stays, subjecting premature newborns to repeated X-ray examinations is not advisable.

The appearance of LUS findings hinges on the relative aeration of alveoli. Premature newborns, with their underdeveloped gestational age and thin chest walls devoid of significant fat interference, present clearer image quality during ultrasound scanning. Because the lung maturation process in premature newborns differs from that in full-term infants, children, or adults, the imaging characteristics observed in LUS may vary. It is essential to consider premature newborns as a distinct group when assessing lung conditions. In this context, the utilization of LUS has gained popularity in premature newborns due to its user-friendly imaging, real-time capabilities, and radiation-free nature [6,10,11,16,17].

In the current study, the detection rate of LUS for consolidation was notably higher than that of CXR (34/75 vs 14/75, P<0.001), possibly due to 2 main factors. Firstly, the superior image resolution of ultrasound equipment, coupled with the favorable acoustic window in preterm infants, enables clear visualization of lung consolidations, particularly those situated beneath the pleura, on ultrasound images. In contrast, CXR exhibits lower sensitivity for identifying subpleural consolidations [17,18]. Additionally, prior research has indicated that CXR can only detect lung consolidations greater than or equal to 1.5 cm due to tissue image blending [17]. Interestingly, in this study, 7 cases displaying consolidations exceeding 2.0 cm or multiple consolidations on LUS were not identified on CXR. Of these 7 cases, 3 had a single consolidation with a diameter greater than 2.0 cm, and 4 had multiple consolidations. All 7 neonates clinically presented with mild and moderate respiratory distress. Consolidations on LUS were accompanied by varying degrees of air bronchograms. The reason for CXR failing to detect consolidations could have included improper patient positioning, obscuration by the thymus, or consolidations located near the edge of the diaphragm (Figures 2–4). Additionally, when the degree of lung consolidation is mild, the density difference is not significant enough, which may make air bronchograms difficult to identify. Notably, out of the 14 cases showing consolidation on CXR, 3 cases were not detected on LUS (all 3 cases exhibited diffuse alveolar-interstitial syndrome in B-line patterns). Two of them presented with severe respiratory distress. This discrepancy might be due to the presence of air interposition, rendering ultrasound less sensitive to consolidations present in deep or perihilar areas. This is also one of the disadvantages of lung ultrasound.

In pulmonary ultrasound, the extent of interstitial edema is typically assessed based on the number and distribution of B-lines, which are artifacts formed by repeated reflections due to an increased concentration of physiological or pathological fluid [19]. Among the 75 cases, a small number of B-lines were identified in 22 cases (29.3%), showing either normal lung markings or only slight changes on CXR. In this context, both imaging modalities were found to be highly consistent in detecting normal or mildly abnormal cases. For the 53 cases demonstrating interstitial edema on LUS, LUS proved effective in diagnosing the number of B-lines indicative of interstitial edema, the degree of fusion, and the extent of involvement. LUS can provide insights into the water content across different lung regions, aiding in assessment of pulmonary conditions in premature newborns. However, only 21 out of the 53 cases exhibited abnormalities on CXR, and CXR was unable to clearly differentiate the severity of lung edema (Figure 5A, 5B). This disparity may be due to bedside CXR chest films being limited to an anteroposterior position, resulting in tissue image blending, whereas lung ultrasound permits dynamic analysis of each intercostal space during breathing [17].

Several limitations of this study need to be mentioned. As it was a retrospective study without a prospective design, certain confounding factors could not be entirely eliminated. Firstly, 2 sonographers performed the LUS and collected images. Even if sonographers received the same training and standardize operational procedures, there may be slight technical differences in practice among different sonographers. This study did not evaluate the consistency between the 2 sonographers, which may introduce some bias into the results. Secondly, LUS and CXR were conducted within 24 hours, rather than simultaneously. Although strict screening can reduce some confounding factors, minor, undetected fluctuations in the newborn’s condition may still occur within 24 hours, potentially affecting the comparability of the examination results. Future research should apply both examinations simultaneously to all newborns to improve the generalizability of the findings. The last limitation was the absence of a criterion standard for directly comparing LUS with CXR images. Ideally, comparing the data with computed tomography (CT) images could potentially elucidate the differing appearances seen on LUS and CXR. Nevertheless, subjecting newborns to CT scans is typically unnecessary in most cases, although this might be applicable in animal experiments.

Conclusions

LUS is superior to CXR in distinguishing the extent of consolidation and edema. Given its accessibility, radiation-free nature, and bedside applicability, LUS could emerge as a superior diagnostic tool in premature newborns compared to bedside CXR.