Neonatal Pneumonia Imaging

Pneumonia is the single largest infectious cause of death in children worldwide, accounting for 740,180 deaths in children younger than 5 years in 2019, 14% of all deaths in children younger than 5 years, and 22% of all deaths in children aged 1-5 years. ^{[1, 2, 3]}

Imaging modalities

Other than hematologic testing, blood biochemistry, and searches for the offending microorganism, chest radiographic imaging is considered to be an essential component in making the diagnosis of neonatal pneumonia, despite the potentially limited predictive value of radiographic and laboratory findings. ^[4] Attempts to identify and culture the causative microorganisms are often unsuccessful.

Diagnostic imaging is involved not only to initially assess the neonate's condition and to establish a diagnosis but also to monitor the progress of the disease and the effects of interventional therapeutic measures. ^{[5, 6, 7, 8, 9, 10]} Bedside studies obtained using portable equipment are often limited but can provide much useful information when a careful, detailed approach is used to produce the radiograph and interpret the results. Conventional chest radiography remains the mainstay of diagnosis in neonates with respiratory symptoms (see the images below). ^{[11, 12, 13, 14, 15]}

Ultrasound is a useful technique in the detection of pleural fluid and guided aspiration of pleural effusions. CT scanning or MRI may be helpful in excluding tumors, aberrant vessels, sequestered lobes, or other primary pulmonary anomalies (see the images below). These studies may also help establish the presence of an infiltrate. Routine follow-up chest radiographs are not needed in childhood community-acquired pneumonia if the child has a clinically uneventful recovery. ^[5]

The findings on chest radiography are nonspecific and have a wide differential diagnosis. Several lung dysplasias may mimic viral pneumopathy, especially in newborns. ^[16] Ultrasound detects pleural effusions with a wide differential diagnosis. CT scanning and MRI are expensive tests, and neonates may require heavy sedation or general anesthetic to undergo them. The radiation dose from CT scanning is high. In addition, although intravenous contrast may not be required to detect pulmonary abnormalities, when used, it can have nephrotoxic effects.

Wahlgren et al looked at 346 children to evaluate whether radiologic findings and healing time in children with pneumonia correlated to etiologic agent and suggested that conclusions about etiology could not be drawn from the chest radiograph findings and that laboratory examinations must be used to confirm the radiologic diagnosis of viral pneumonias. ^[6]  Radiologic findings did not differ between etiologic groups but correlated significantly with patient age. The radiologic healing frequency on follow-up radiographs was significantly lower in children with mixed bacterial and viral etiology than in children from each of the other groups and to the material as a whole.

MRI may be helpful to exclude tumors, aberrant vessels, sequestered lobes, or congenital hernias. Magnetic resonance angiography (MRA) and computed tomographic angiography (CTA) are relatively noninvasive techniques that may be substituted for angiography if required. More experience is required in using MRI to characterize neonatal intrathoracic abnormalities. MRI is not indicated in lung parenchymal abnormalities. Sequestrated lung segment is considered one of the differential diagnoses of neonatal pneumonia. Angiography represents the criterion standard in the detection and characterization of sequestrated lung segment but is being replaced with the relatively less invasive computed tomographic angiography (CTA) and the noninvasive technique of color Doppler ultrasonography. Angiography is an invasive technique, and a false-negative examination is possible with atypical arterial supply and venous drainage.

Lung pneumatoceles can be a life-threatening complication of pneumonia in infants, especially if the pneumatoceles do not spontaneously reabsorb. Imaging is important in the diagnosis, characterization, and follow-up of pneumatoceles, and image-guided percutaneous drainage is considered to be safe and effective in children, although it is not widely used in newborns and infants. On plain chest radiography and CT scan, a pneumatocele manifests as an almost round, thin-walled airspace with adjacent consolidations or ground-glass opacities. ^[17]

In a study of high-frequency bedside lung ultrasonography for neonatal pneumonia, the sensitivity of ultrasound was 96.6%, specificity 93.3%, positive predictive value 93.5%, and negative predictive value 96.5%. The sensitivity of chest radiography was 93.3%. ^[18]  In another study, lung ultrasound in 115 newborns with ventilator-associated pneumonia (VAP) had a sensitivity and specificity of 94.7 and 89.6%, respectively, in evaluating the outcome of weaning from the ventilator. ^[19]  VAP is the second most common cause of antibiotic use in the neonatal intensive care unit and is associated with a higher incidence of bronchopulmonary dysplasia, prolonged mechanical ventilation, and increased hospital stay. ^{[20, 21]}

Procedures

Thoracentesis is indicated when significant pleural fluid is detected with chest radiography or ultrasound. The procedure may be therapeutic or diagnostic. Diagnostic aspiration is usually performed for Gram stain, culture, and biochemical tests, while a therapeutic aspiration is indicated in neonates with respiratory distress. ^[22]  US guidance can aid in the procedure. Stretching the skin down over the entry site may minimize the risk of pneumothorax or injury to the intercostal vessels: release after the procedure permits the return of tissues to their usual location to occlude the path of the needle. Needle puncture over the superior rib margin reduces the chance of laceration of the intercostal vessels. Care is taken to avoid further needle penetration once a fluid sample has been obtained. Aspiration should be continuous once the skin is penetrated.

Bronchoscopy may be used to obtain biopsy material or brush specimens or to perform guided aspiration. A direct rigid bronchoscope may be used in larger infants; however, the fiberoptic bronchoscope technique is preferred in smaller infants or in infants in whom distal sites cannot be reached with rigid bronchoscopes. Noncontaminated specimens may be difficult to obtain because of oral or airway commensals. It may be difficult to obtain samples from distal bronchial sites.

Direct lung aspiration is a useful method to sample lung infiltrate for culture or biopsy analysis. This procedure is associated with a higher risk of air leaks, as it usually requires a larger-bore needle than is used to obtain pleural fluid. Because of this risk, the procedure is reserved for (1) infants in whom empiric therapy has failed and (2) infants who continue to deteriorate and in whom less invasive methods have failed to obtain microbial cultures. With advances in surgical techniques, clinicians may perform open surgical biopsy, as the success of obtaining adequate samples is much greater and complications may be dealt with directly.

In neonates, most chest radiographs are obtained supine and in the anteroposterior projection (see the images below). Radiographs should be well centered and at the right penetration.

Other views may be required to explain anatomic relationships and detect air-fluid levels. Although neonatal pneumonia has no characteristic appearance, many chest radiographic findings are consistent with neonatal pneumonia. Some of the patterns described include the following:

• Diffuse opacification of the lung parenchyma resembling the ground-glass appearances of respiratory distress syndrome. ^[22] This pattern is encountered with a hematogenous process; however, the appearances are nonspecific, and aspiration of infected fluid with secondary bloodstream infection can give rise to a similar appearance.

• Patchy opacification or consolidation is generally regarded as a complication of antepartum or intrapartum aspiration, particularly when the peripheral parts of the lungs are involved. Patchy densities at the lung bases, more pronounced on the right, suggest postnatal aspiration.

• The presence of pneumatoceles associated with pleural effusions indicates an infective pneumonic process.

• Hyperinflation associated with patchy consolidation suggests partial airway obstruction from a mucous plug or inflammatory debris.

• An air bronchogram usually suggests extensive consolidation, but the appearances are nonspecific and may be encountered in pulmonary hemorrhage or edema.

• A tumefactive-type lobar consolidation in neonatal pneumonia is rare.

Images reflect conditions only at the instant in which the study is performed. Since neonatal lung diseases, including pneumonia, are dynamic, initially suggestive images may require reassessment based on the subsequent clinical course and findings in later studies.

Neonates in intensive care units frequently undergo radiographic examinations. It is of great importance that, because of greater radiosensitivity and longer life expectancy of the neonates and premature babies, the radiation dose is optimized and kept to a minimum.

Faghihi et al studied 50 neonates, mostly premature, with different weights and gestational ages in 5 hospitals, mostly suffering from respiratory distress syndrome and pneumonia, and calculated the radiation dose involved using the values of entrance skin dose (ESD), dose area products (DAPs), energy imparted (EI), whole-body dose, and effective dose. ^[23] The risk of childhood cancer was estimated using 3 methods, including direct method (using thermoluminescence dosimetry [TLD] chips), indirect method (using tube output), and Monte Carlo (MC) method (using MCNP4C code).

In the first step, the ESD of the neonates was directly measured using TLD-100 chips. The values of ESD to neonates were indirectly obtained from the tube output in different imaging techniques. The results indicated that the mean ESD per radiograph estimated by the direct, indirect, and MC methods were 56.6±4.1, 50.1±3.1, and 54.5±3.3 μGy, respectively. The mean risk of childhood cancer estimated in this study varied between 4.21×10^-7 and 2.72×10^-6. ^[23]

Tracheoesophageal fistula is a rare malformation but should be suspected in neonates who present with coughing bouts and cyanosis after feeding and when radiography shows aspiration pneumonia, atelectasis, and gas within the colon. ^[24]

In a study of 1724 young (≤60 days) febrile infants (≥38°C) who had chest radiographs performed in pediatric emergency departments, 2.7% (n=46) had definite pneumonia, and 3.9% (n=67) had possible pneumonia. Median white blood cell count (WBC), absolute neutrophil count (ANC), and procalcitonin (PCT) were all higher in the definite-pneumonia group. ^[25]

Radiographic findings in neonatal pneumonia

In a study of the chest radiographs of 30 infants with autopsy-proven pulmonary infections, the most common abnormality identified was bilateral alveolar densities (77%). Of these patients, one third had characteristically extensive, dense alveolar changes with numerous air bronchograms. A pattern of radiographic abnormalities consistent with transient tachypnea of the newborn was found in 17% of patients, and a second pattern resembling hyaline membrane disease was found in 13%. The authors suggested that the presence of a pleural effusion (in cases of the hyaline membrane disease pattern) and persistence beyond 1-2 days (in the transient tachypnea pattern) are helpful features in the diagnosis of neonatal pneumonia. Recognition of the spectrum of expected radiographic changes can aid in the diagnosis of neonatal pneumonia, particularly if this information is correlated with the clinical features. ^[26]

Radiologic pulmonary changes during gram-negative bacillary nosocomial bloodstream infection in premature infants

Cordero et al studied the changes that occur in chest radiographs at the time of gram-negative bacilli (GNB) nosocomial bloodstream infection (BSI) and determined the contribution of bronchopulmonary dysplasia (BPD). Radiographic signs of air space disease accompanied by the recovery of GNB respiratory pathogens from the blood and from a previously uncolonized airway strongly supported the clinical diagnosis of GNB nosocomial pneumonia. Radiologic signs of BPD were found to be stable in relation to nosocomial bloodstream infection (BSI) caused by GNB, but BPD radiologic scores were higher in infants who also had a newly acquired respiratory GNB. In newborns on ventilation, BSI, new respiratory tract GNB, and BPD were found to be critical associations for the clinical interpretation of radiographic changes. ^[27]

Miliary pattern in neonatal pneumonia

Flores described 10 newborn babies who developed respiratory distress and whose chest radiographs demonstrated a miliary nodular pattern. All patients received a diagnosis of bacterial pneumonia and appeared to respond favorably to antibiotic therapy. The pulmonary abnormalities resolved. The children were clinically well in less than 3 weeks. Flores suggested that hematogenous bacterial dissemination may produce the miliary pattern, one of the radiologic patterns of neonatal pneumonia. ^[28]

Radiographic findings of neonatal herpes simplex pneumonia

Dominguez et al described radiologic findings in the lungs of 16 hospitalized neonates with disseminated herpes simplex infection. ^[29] The authors devised a sequential picture of 4 stages in the evolution of the pneumonitis of this hematogenous infection. The radiologic stages were as follows:

• Stage I, normal chest

• Stage II, prominent perihilar interstitial markings

• Stage III, coalescent areas of pulmonary infiltrates

• Stage IV, diffuse alveolar and interstitial disease ("white-out" lungs)

In general, the pulmonary abnormalities were widespread and without air trapping. Pleural effusions were noted in one case. All affected neonates died, and antemortem clinical and radiologic findings were correlated with multiple-organ postmortem histopathologic evidence of viral infection, especially with the associated pneumonia.

Korppi et al examined 61 children with microbiologically verified viral pneumonia or radiologically verified bacterial pneumonia. ^[30] Their results suggested that the presence of an alveolar infiltrate on a chest radiograph is a specific but insensitive indicator of bacterial pneumonia. The authors concluded that patients with alveolar pneumonia should be treated with antibiotics. However, in patients with interstitial pneumonia, a viral or bacterial etiology is possible.

CT scanning (see the images below) may be helpful to exclude tumors, aberrant vessels, sequestered lobes, or other primary pulmonary anomalies and to establish the presence of an infiltrate.

Chronic pneumonitis of infancy (CPI) is a rare entity. Olsen et al described chest radiography and high-resolution CT (HRCT) findings in a child with histologically confirmed chronic pneumonitis of infancy. The child was admitted for intensive care 18 hours after birth and died at age 39 days. Diffuse ground-glass change, interlobular septal thickening, and discrete centrilobular nodules were observed on HRCT. ^[7] An accurate diagnosis is crucial for correct management; however, several entities with the same HRCT findings are recognized.

CT scanning is also useful in the diagnosis of pulmonary sequestration. Wassia et al described a 2-day-old newborn with extrapulmonary sequestration that manifested as persistent Staphylococcus epidermidis pneumonia and high-output cardiac failure. ^[8] CT scanning of the chest without angiography established the diagnosis. Radiographic evaluation of pulmonary sequestration includes delineation of the aberrant systemic feeding artery and venous drainage. Color Doppler ultrasound, spiral CT scanning, and MR angiography are often sufficient to make the diagnosis.

HRCT in children remains a technically challenging procedure, both to perform optimally and to interpret correctly. Much remains to be learned about the optimal application of HRCT. However, HRCT can clarify often confusing or nonspecific chest radiographs, as well as clarify the diagnosis and evolution of both common and unusual pediatric lung diseases. As new therapies become available for these disorders and CT becomes faster and easier to perform, HRCT will be used more often for more accurate diagnosis and for better evaluation of therapeutic effects. ^[9]

Ultrasonography (US) is a useful adjunct to radiography in selected circumstances. ^[31] Sonograms are particularly useful for identifying and localizing fluid in the pleural and pericardial spaces. US is a noninvasive technique ideally suited to neonates, as sedation is seldom required. ^[32] US has high sensitivity in the detection of pleural effusions and detects consolidation at the lung bases. No radiation is involved, and the procedure may be repeated many times without untoward effects.

US is not ideal to detect air space disease. It is an operator-dependent procedure, and the modality produces many artifacts, which may result in false-positive and false-negative results. For example, soft tissue surgical emphysema following pleural intubation may degrade US examination quality.

Bober and Swietlinski assessed the diagnostic use of ultrasonography in respiratory distress syndrome (RDS) in 131 consecutive newborns and concluded that in RDS, 100% sensitivity and 92% specificity characterize US examination. Nevertheless, US cannot replace chest radiography in the evaluation for respiratory failure because it overestimates the diagnosis; however, US may be useful in excluding RDS as a cause of respiratory insufficiency in newborns. ^{[10, 33]}

In a study of lung ultrasound (LUS) versus chest radiography (CXR) in 115 newborns with ventilator-associated pneumonia (VAP), the sensitivity and specificity of LUS (94.7 and 89.6%, respectively) in evaluating the outcome of weaning from the ventilator were higher than those of CXR (73.7 and 84.4%, respectively). The finding of lung consolidation with air bronchogram on LUS had a higher sensitivity, specificity, and accuracy for the diagnosis of neonatal VAP than those of other LUS and CXR findings and showed better consistency with the clinical diagnosis. LUS findings in the 115 VAP neonatal patients included lung consolidation with air bronchogram in 111, alveolar-interstitial syndrome in 113, pleural effusion in 12, pleural line abnormalities in 114, and lung pulse in 15. ^[19]

In an Egyptian study of 50 infants younger than 1 year who presented with clinical symptoms and signs suggestive of community-acquired pneumonia (CAP), lung ultrasonography detected pneumonia in 49 patients (98%) and chest radiography diagnosed pneumonia in 36 (72%). ^[34]

In a study of 366 children in Nepal younger than 5 years (median age, 16.5 mo) with cough, fever, or difficulty breathing who received a chest radiograph and lung ultrasound, 84 patients (23%) were diagnosed with pneumonia by chest radiography. Sensitivity, specificity, and positive and negative likelihood ratios for lung ultrasonography were 89.3%, 86.1%, 6.4, and 0.12 respectively. ^[35]