Added value of a chest multidetector computed tomography scoring system in the assessment of chronic lung disease among pediatric patients

Pediatric chronic lung disease (CLD) represents a heterogeneous group of several distinct clinical entities, with its prevalence increasing over the last decade. The current study aimed to identify the role of chest multidetector computed tomography (MDCT) using modified Bhalla scoring for the early diagnosis of CLD in pediatric patients and determine the most common chest MDCT findings. This prospective study involved 45 pediatric patients with chronic respiratory symptoms, all of whom underwent MDCT. Thereafter, data were analyzed using the modified Bhalla score. Chronic lung diseases were classified according to their radiological and clinical criteria. The total CT score, which was the most significant factor for chronic inhalation and chronic recurrent inflammatory lung diseases, varied between 2 and 21 points, with those having autoimmune diseases exhibiting the largest value for the mean CT score. The clinical severity of symptoms was not correlated with CT score. Our findings showed that MDCT was a useful tool for diagnosing pediatric CLD and assessing disease extent, severity, and superimposed complications. The modified Bhalla CT scoring system allowed for systematic primary and follow-up assessments of various lung lesions in cases with varying CLD etiologies.


Background
Pediatric chronic lung disease (CLD) represents a heterogeneous group of several distinct clinical entities. The prevalence of CLD has increased over the last decade due to more advanced and intensive respiratory support provided to at-risk infants, as well as the overall improvement in the survival of premature infants. The most common causes of CLD in children include cystic fibrosis (CF), with other causes including bronchiectasis (e.g., immunodeficiency in third-world countries and postinfectious bronchiectasis, for instance after measles), bronchopulmonary dysplasia (BPD), chronic gastroesophageal reflux disease, aspiration pneumonitis, obstructive bronchiolitis disease, and childhood interstitial lung disease (chILD) [1,2].
With the advances in multidetector computed tomography (MDCT), it is now possible to image the lung for ≤ 10 s and visualize the bronchial tree, lobes, and airways from the fifth -to the seventh-generation bronchi, as well as lung regions. This allows the accurate characterization of attributes, such as density, texture, aeration, and Hafez et al. Egypt J Radiol Nucl Med (2022) 53:263 distribution of the lungs [3]. MDCT can cover large volumes during simple breath-holding, with good longitudinal, intrinsic spatial resolution and temporal resolution. The results obtained from the dataset allow the generation of enhanced multi-surface and three-dimensional images of the airway, including those obtained from techniques developed specifically for airway imaging, such as virtual bronchography and virtual bronchoscopy [4]. CT can reveal lung damage, such as bronchial wall thickening, bronchiectasis, mucus obstruction, and areas of consolidation or atelectasis of the lung, which are difficult to detect using a pulmonary function test (PFT). Recent studies in children with CF have shown that structural damage visible on MDCT precedes changes revealed via PFT [5].
Since 1991, several chest MDCT scoring systems have been developed to observe the extent and severity of CF. This system had originally been introduced by Bhalla et al. and has since been modified by several authors [6,7]. The images are scored using a modified Bhalla scoring system, which includes an evaluation of acinar nodules, interlobular septa thickening, ground glass and mosaic perfusion on inspiratory images, and air trapping on expiratory images. These changes are commonly observed and included in present scoring systems for CF lung disease. The total score is then derived by adding the scores for each abnormality and can range from 0 to 37 [8].
The current study aimed to identify the role of chest MDCT using a modified Bhalla scoring for the early diagnosis of CLD in pediatric patients and determine the most common chest MDCT findings and their correlation with clinical severity.

Methods
This prospective cohort study involved 45 patients referred from the Department of Pediatrics who underwent CT of the chest at the radiology department. The planned study duration was 6 months. This study was conducted according to the guidelines of the Ethical Committee and was approved by its institutional review board.
The inclusion criteria were pediatric patients who had a history of chronic respiratory symptoms throughout a period of 6 months, from January 2021 to June 2021; all cases with an acceptable quality of current chest CT; and those diagnosed with chronic lung injury after following up on previous diagnostic test. The exclusion criteria were patients with the chronic respiratory disorder who were unable to hold their breath during MDCT scanning.
Patients were clinically assessed at the pediatric chest clinic where they underwent comprehensive history taking, clinical examination, and relevant laboratory examinations (e.g., complete blood count, C-reactive protein, sputum culture). MDCT of the chest was performed using a GE Brightspeed Elite 16 MDCT scanner, with the following parameters: helical scan; 100 kV; 199 mA; field of view, 320 mm from the root of the neck to level of the renal arteries; slice thickness, 1.25 mm; slice interval, 5 mm. The total exposure time was 16 s, with a matrix size of 512 × 512.
Patients were positioned supine, with fixed inspiration and arms extended overhead. The area from the top of lungs through the bottom of lungs was scanned while instructing cooperative patients to hold their breath at inspiration. However, for uncooperative children, quiet respiration was determined after administering a light universally approved sedative, such as chloral hydrate.
All volumetric chest CT images were assessed at a lung window of WW 1500, WL 500 and a mediastinal window of WW 400, WL 40.

Data analysis
Two radiologists, both having 16 years of experience in chest and pediatric imaging, reviewed all the chest MDCT images separately along with the clinical data, laboratory data, and previous radiology reports. The location of the lesions was assessed for each chest CT. Chest CT assessment was based on the Bhalla scoring system described by Judge et al. [8] (Table 1), which evaluates the severity of bronchiectasis, peribronchial thickening, extent of bronchiectasis (number of pulmonary segments), the extent of mucus plugging, sacculations or abscesses, generation of bronchial divisions involved, number of bullae, emphysema, and collapse/consolidation. The total score for each item ranged from 0 to 3 except for emphysema, consolidation/collapse, mosaic attenuation, air trapping, and acinar nodules, which have a total score ranging from 0 to 2. All scores for each patient were then added in order to obtain the overall Bhalla score, which ranged up to 40.

Statistical analysis
Data were coded and entered using the Statistical Package for the Social Sciences version 26 (IBM Corp., Armonk, NY, USA). Categorical data were expressed as numbers and percentages. Spearman's correlations were determined, with P values < 0.05 indicating statistical significance.

Demographics
A total of 45 patients were included in our final analysis, among whom 34 (75.6%) and 11 were males and females (24.4%), respectively. The patients had a mean age of 7.17 ± 4.56 years (range, 1-18 years).

Bhalla CT score
The total CT score varied between 2 and 21 points among the included patients, with a mean score of 6.7 ± 4.66. The largest proportion of patients (20%) had a score of 3, whereas 17.8%, 11.1%, and 11.1% had a score of 5, 2, and 4, respectively ( Fig. 1).

Chronic lung disease distribution in the studied cases: correlation between disease extent and chest CT components of the Bhalla score
Bilateral disease was present in 75.6% of the cases. The most affected lung lobes were the lower zone (84.4%), followed by the upper zone (73.3%) (Fig. 2).
After determining the correlation between various chest CT findings and different etiological diagnosis, we found that the severity, extent, and generalities of bronchial division involvement of bronchiectasis were correlated with disease bilaterality (P = 0.013, 0.006, and 0.007, respectively). Moreover, the bilateral disease was significantly correlated with collapse/consolidation, which has been mainly noted in chronic recurrent infection (P = 0.008). Furthermore, the increase in the total CT score was positively correlated with disease bilaterality and negatively correlated with unilateral disease (P = 0.004 and 0.025, respectively).
Acinar nodules were significantly correlated with upper and middle lung lobe localization (P = 0.048 and 0.003, respectively).

Etiology of chronic lung disease in the studied cases
The most commonly encountered chronic lung disease (CLD) was chronic recurrent inflammatory diseases (66.7%), which had the highest total CT score, followed by interstitial lung disease and chronic airway disease (20% for each). The sub-etiology for each etiological diagnosis throughout the study duration is illustrated in Table 2. Associated etiologies encountered here included CF with chronic airway disease, infection with chronic airway disease, autoimmune disease with chronic inflammatory disease in a case, and ILD in another, as well as aspiration pneumonitis with sequestrated lung. The calculated mean and total modified Bhalla CT score for various etiologies are displayed in Table 3, and the correlation for their significance is displayed in Table 4.

Correlation between etiological diagnosis and chest CT components of the Bhalla score
Inhalation lung disease was significantly correlated with mosaic attenuation, acinar nodules, and mean CT score (P = 0.023, 0.004, and 0.006, respectively). Meanwhile, chronic recurrent inflammation was significantly correlated with collapse/consolidation, mosaic attenuation, and fibrosis/interlobular septal thickening (P = 0.01, 0.008, and 0.004) but was negatively correlated with mean CT score (P = 0.006). Moreover, autoimmune disease was significantly correlated with the presence of air bubbles and emphysema (P = 0.001 and 0.001, respectively), while hereditary lung disease was significantly correlated with the severity of bronchiectasis and bronchial wall thickening (P = 0.007 and 0.016, respectively). Interstitial lung disease was significantly correlated with emphysema, fibrosis, and honeycombing (P = 0.044, 0.00, and 0.044, respectively). Chronic airway disease was significantly correlated with air trapping (P = 0.026), while chronic vascular disease was significantly correlated with abscess/sacculation (P = 0.001).

Correlation between CT score and clinical severity
Among the included cases, seven had type I, II, and III respiratory distress, the CT scores of whom varied from 2 to 21, with a mean score of 7.5. One case had known cardiac disease.

Discussion
Noticeable confusion has existed regarding the nomenclature, classification, and management of pediatric diffuse lung diseases [9]. In this study, CLD was classified as inhalation lung disease, chronic recurrent inflammatory, The Bhalla score, first proposed by Bhalla et al., was created to adjust for pulmonary structural damage observed on MDCT in pediatric cases diagnosed with CF in the year 1991 [6]. Thereafter, Judge et al. proposed modifications to the score by including more CT features described in the disease, namely mosaic attenuation/perfusion, air trapping, acinar nodule, intralobular septal thickening, and ground-glass infiltrate [8]. In the current study, the significance of CT scores in inflammatory and inhalation lung diseases suggests the utility of this scoring system for such entities as well.
The American Thoracic Society classified Childhood interstitial lung disease (chILD) by the age-group from infancy to later childhood presentation [10]. In the current study, 34 patients (75.6%) were diagnosed as chILD with mainly later childhood presentation, including inhalation, post-inflammatory, aspiration pneumonitis, autoimmune diseases, and ILD (including infants with NEHI and children with UIP, NSIP, and hemosiderosis). Clinical evaluation through further tests, such as high-resolution CT, spirometer tests, bronchoalveolar lavage, and biopsy, is needed [10].
Aspiration syndrome (Fig. 4) is a common finding in children with recurrent lower respiratory tract infections and is predicted to be present in 26-49% of children with chILD. One study showed consolidation or airway changes in the dependent and basal portions of both lungs on CT [11]. In the current study, aspiration   [12]. Neuroendocrine cell hyperplasia (NEHI) has been found to present more frequently among males in the first 2 years of life. MDCT, which has a sensitivity of 78% and specificity of 100% for NEHI, typically shows a mosaic attenuation pattern affecting at least four lobes with geographic ground-glass opacities that are most conspicuous on the right middle lobe and lingula and areas of hyperlucency associated with air trapping. Confirmatory biopsy may not be necessary for patients with a classic clinical presentation and characteristic radiologic findings [10]. The study by Deutsch et al. showed that NEHI was diagnosed in 10% of chILD cases [9]. These findings were consistent with our clinical and radiological data on confirmed NEHI cases, although they accounted for only 2.9% of our chILD cases (Fig. 5).
There are several causes of lung hemorrhage, including vasculitis, good pasture syndrome, and systemic lupus erythematosus. In the acute phase, CT findings show an ill-defined fluffy appearing centrilobular ground glass that represents diffuse alveolar hemorrhage. With repeated hemorrhage, mild interstitial fibrosis has been found to cause interlobular septal thickening [10]. In the current study, similar changes were noted in a single case with repeated hemorrhage, who developed pulmonary alveolar hemosiderosis.
Langerhans cell histiocytosis is one of the chILD diseases that is not specific to infancy, with portions of the systemic disease with pulmonary involvement presenting in 20-50% of the cases [10]. CT shows nodules of varying size and bizarre pulmonary cyst shapes, with upper lobe predominance [13]. This was consistent with our case diagnosed with Langerhans cell histiocytosis (Fig. 6).
Our findings showed that ILD was significantly correlated with the presence of lung honeycombing (P = 0.044), similar to the case with usual interstitial pneumonia (Fig. 7). This has been considered a common finding in fibrotic lung disease that leads to traction of bronchioles and formation of a honeycomb appearance. Studies have shown a significant correlation between ILD and honeycomb appearance on MDCT [14].
In the present study, 8.9% of the included cases presented with autoimmune disease, which was significantly correlated with bilateral lung affection and a higher number of emphysematous bullae. These findings were consistent with those presented in several studies including children diagnosed with systemic autoimmune disease, which found that autoimmune disease was significantly correlated with bilateral lung pathologies [15][16][17]. The presence of autoimmune lung disease had been commonly associated with pulmonary emphysema based on the available literature [18].
In the present cohort, the severity of lung disease and peribranchial thickening was significantly correlated with overall hereditary lung anomalies. These findings were consistent with those presented in studies conducted on children diagnosed with cystic fibrosis, which revealed that children with hereditary malformations of the bronchial tree and ciliary dyskinesia were predisposed to severe manifestations of bronchiectasis [19,20].
In the current study, the three cases (6.7%) who had CF all presented with variable severity of bronchiectasis and peribronchial thickening, with two showing air trapping, one showing mucus plug formation, and another case   showing collapse and ground-glass attenuation (Fig. 8). Sasihuseyinoglu et al. showed that out of 36 patients, 96.4% exhibited peribronchial thickening, 50 showed mucus plug formation, and 32.1% exhibited collapse/consolidation. Emphysema was identified in 18 patients [21].
The current study showed that the total CT score was significantly correlated with the severity and extent of bronchiectasis, along with peribronchial thickening. These findings were consistent with those presented in a multicenter prospective observational study by Martínez-García et al. in 99 consecutive patients with moderateto-severe bronchiectasis, who showed that the severity of CLD was significantly correlated with the Bhalla CT score [22]. Similarly, another study on the Bhalla score showed an almost ninefold increase in the risk of having malacia/obliterative-like combination lesions in the bronchial tree [23].
Mosaic perfusion/attenuation and intralobular septal thickening was significantly correlated with the presence of inflammatory processes in lungs (P = 0.008 and 0.004, respectively). These findings were confirmed by Teufel and his team, who stated that mosaic attenuation was significantly correlated with the presence of lung inflammation [24]. However, this was inconsistent with another study conducted on patients with wet cough, which showed that inflammation was not significantly associated with the total Bhalla score [25].
The present study showed that chronic airway disease was associated with bronchial disease in 77.8%, mucous plugging in 22.2%, mosaic attenuation in 33.3%, and centrilobular nodules in 55.6% and was significantly correlated with air trapping (P = 0.026) (Fig. 9). These findings were consistent with those presented in several studies conducted to evaluate the extent of expiratory air trapping on MDCT and spirometry, which concluded that air trapping was a prominent feature of chronic airway diseases, such as asthma and chronic bronchitis [26][27][28].
Thickening of the interlobular septa was found more frequently in cases with CLD mainly secondary to inflammatory and interstitial lung disease. One study showed that the prominent inflammatory infiltration into the submucosa of patients with idiopathic bronchiectasis might lead to lymphatic congestion and thus thickening of the interlobular septa [29].
The current study showed that 17.8% of cases exhibited respiratory distress, with a severity score ranging from 2 to 21, and that the clinical severity of symptoms was not correlated with the CT score. Douros et al. [23] stated that MDCT scanning detected airway wall thickening and bronchiectasis, with the severity of the findings correlating positively with the length of clinical symptoms. Moreover, they found that the clinical picture was incompatible with MDCT findings.