- Email: [email protected]
Diagnostic Utility of 3T Lung MRI in Children with Interstitial Lung Disease
ARTICLE IN PRESS
Preliminary Laboratory Investigation
Diagnostic Utility of 3T Lung MRI in Children with Interstitial Lung Disease: A Prospective Pilot Study Kushaljit Singh Sodhi, MD, PhD, Madhurima Sharma, MD, Edward Y. Lee, MD, MPH, Akshay Kumar Saxena, MD, Joseph L. Mathew, MD, Meenu Singh, MD, Niranjan Khandelwal, MD Rationale and Objectives: The objective of this study was to assess the diagnostic utility of 3-tesla (3T) magnetic resonance imaging (MRI) of lungs in the detection of interstitial lung disease (ILD) in pediatric patients. Materials and Methods: Twelve children (mean: 8.5 years, range: 4–12 years) with ILD were consecutively enrolled in this prospective study. HRCT and 3T lung MRI were performed in all patients within 2 days of each other. The sensitivity, the specificity, the positive predictive value, and the negative predictive value of detecting lung abnormalities related to ILD with 3T lung MRI were calculated, with high-resolution computed tomography (HRCT) as a standard of reference. Agreement between HRCT and 3T lung MRI, as well as between two reviewers, was calculated with the kappa coefficient. Results: 3T lung MRI had low sensitivity (66.67%) and high specificity (97.33%) in the detection of abnormalities related to ILD when compared to HRCT in children. Although 3T lung MRI performed well in the detection of consolidation, parenchymal bands and fissural thickening with a sensitivity of 100%, the sensitivity of 3T lung MRI in the detection of septal thickening, ground-glass opacity, nodules, and cysts was relatively low (50.0%, 50.0%, 66.67%, and 25.0%, respectively). Substantial agreement was seen between HRCT and 3T lung MRI (k = 0.7), whereas perfect agreement was seen between two reviewers in detecting abnormalities related to pediatric ILD (k = 0.9–1.0). Conclusions: In comparison to HRCT, 3T lung MRI with routinely available MRI protocols and sequences can also well detect abnormalities such as consolidation, parenchymal bands, and fissural thickening in children with ILD. However, evaluation of septal thickening, ground-glass opacity, nodules, and cysts is limited with 3T lung MRI. Key Words: Interstitial lung disease; imaging; MRI. © 2017 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
I
nterstitial lung diseases (ILDs) in pediatric populations are a heterogeneous group of disorders with diffuse lung involvement, frequently leading to significant mortality and morbidity (1,2). Unlike the adult population, the spectrum of ILD in children is entirely different, with certain diseases unique to the pediatric population (3). In addition, clinical manifestations of ILD in children are unfortunately nonspecific and variable, often leading to missed or delayed diagnosis. Acad Radiol 2017; ■:■■–■■ From the Department of Radiodiagnosis and Imaging, PGIMER, Sector-12, Chandigarh, 160012, India (K.S.S., M.S., A.K.S., N.K.); Departments of Radiology and Medicine, Pulmonary Division, Boston Children’s Hospital and Harvard Medical School, Boston, Massachusetts (E.Y.L.); Department of Pediatrics, PGIMER, Chandigarh, India (J.L.M., M.S.). Received July 7, 2017; revised September 1, 2017; accepted September 21, 2017. Source of Funding: This study has been sponsored by funds from Indian Council of Medical Research (2012-2081). Address correspondence to: K.S.S. e-mail: [email protected] © 2017 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.acra.2017.09.013
Main investigations employed for the diagnosis of ILD include imaging, pulmonary function tests, bronchoalveolar lavage, and lung biopsy. Pulmonary function tests are difficult to perform in children because of poor cooperation (4), and bronchoalveolar lavage is of diagnostic value in a substantially limited number of conditions (5–8). Tissue biopsy provides a definitive diagnosis but has disadvantage of being invasive. Considering the limitations of all these currently available investigations, imaging plays a crucial role in the diagnosis and evaluation of pediatric ILD. Currently, high-resolution computed tomography (HRCT) remains the imaging investigation of choice for the evaluation of ILD (9–11). HRCT is helpful not only in the diagnosis of ILD but also in monitoring disease activity and severity. However, the main concern with HRCT is the potentially harmful radiation, particularly considering chronicity of ILD and requirement for frequent follow-up examinations. Magnetic resonance imaging (MRI) is an attractive radiationfree alternative to computed tomography (CT). In recent years, lung MRI has evolved because of refinement in pulse sequences 1
SODHI ET AL
and technical advances. Few studies have highlighted the role of lung MRI in the evaluation of pediatric pulmonary infections (12–16). Only a few published reports on MRI in diffuse lung diseases in adult populations are available (17,18). However, there is no published data on the role of lung MRI, particularly with a 3-tesla (3T) MRI scanner, in pediatric ILD. Although the higher field strength of 3T MRI offers the advantage of improved contrast and increased signal from lung tissue, high-field strength susceptibility artifacts resulting in more pronounced image distortion and signal loss would be more prevalent on 3T MRI. Therefore, the purpose of our pilot study was to assess the diagnostic utility of 3T MRI of lungs in the detection of ILD in pediatric patients and to compare it with the current imaging gold standard, HRCT. Although this is a small prospective pilot study, it has the potential to pave the way for further research into the clinical application of lung MRI for ILD in children.
MATERIALS AND METHODS This was a prospective study, approved by our institutional ethics committee. Informed written consent was obtained from the parents or guardians of all pediatric patients enrolled in the present study.
Study Population
From October 2015 to January 2017, a total of 12 consecutive pediatric patients, referred to our department for HRCT of the chest for the evaluation of ILD, were enrolled in this prospective study. The inclusion criteria were pediatric patients in the age range of 4–14 years with clinical suspicion of ILD, supported by clinical symptoms and chest radiographic findings, who underwent HRCT for further evaluation. MRI was performed after HRCT in all patients within 2 days of each other. Children less than 4 years of age who required sedation were excluded from the present study.
Imaging Technique
HRCT Protocol HRCT of the chest was performed in all pediatric patients enrolled in this study using a 64-row multidetector CT scanner (Aquilion64; Toshiba America Medical Systems, Otawara, Japan). The scans were obtained from the lung apex to the base. Acquisition parameters for HRCT were 120 kVp, 100– 150 mA, 1-mm collimation, and 5- to 10-mm spacing, and a high-spatial resolution image reconstruction algorithm was used. Axial CT images were subsequently reconstructed using bone and soft tissue reconstruction algorithms and were evaluated on a Terrarecon workstation (Foster City, CA) with a standard lung window (level: 500 HU, width: 1500 HU) 2
Academic Radiology, Vol ■, No ■■, ■■ 2017
and mediastinal window settings (level: 350 HU, width: 50 HU). MRI Protocol MRI of the chest was performed in all pediatric patients enrolled in the present study using a 3T MRI scanner (Ingenia; Philips Health Systems, Amsterdam, The Netherlands). All MRI sequences were acquired with breath-holding technique and were nonrespiratory and nonelectrocardiogram gated. All MRI studies were obtained without sedation or contrast administration. All children were duly prepared for breath holding before MR acquisition. Children who were dyspneic or who could not hold their breadth or were uncooperative were not taken up for MRI study. The chest was scanned in the axial plane in the craniocaudal direction in all pediatric patients. The following MRI pulse sequences and parameters were used: (1) T2-weighted turbo spin echo (repetition time [TR] 1250 ms, echo time [TE] 80 ms, slice 4 mm, flip angle 90°, field of view [FOV] 300 mm, matrix 216 × 189); (2) balanced turbo field echo (TR 2.6 ms, TE 1.32 ms, slice 4 mm, flip angle 45°, FOV 270 mm, matrix 152 × 135); (3) Dixon T1 (TR 2.6 ms, TE 1.32 ms, slice 4 mm, flip angle 90°, FOV 250 mm; matrix 156 × 124); and (4) Multivane MVTX (Philips, Netherlands) (TR 2500 ms, TE 127 ms, slice 4 mm, flip angle 90°, FOV 275 mm, matrix 184 × 184). Multivane is a PROPELLER (Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction) sequence. The total MRI scanning time was approximately 7 minutes for all four MRI sequences combined. The total time on the magnetic resonance (MR) table varied from 20 to 25 minutes. HRCT and MR Image Evaluation HRCT and MR images were reviewed by two experienced pediatric radiologists, with 15 and 12 years of experience in interpreting thoracic HRCT and MRI. To minimize the potential bias, the reviewers were blinded to the clinical profile of the patients. Identification information from HRCT and MR images was removed before the review session. The order of HRCT and MR images was randomized. In addition, HRCT and MR images were reviewed with 3- to 4-week time intervals between review sessions. Both observers were blinded to reciprocal image assessment, and in case of doubting cases, a consensus agreement was pursued. First, the reviewers evaluated the quality of thoracic MRI studies. Thoracic MRI studies were evaluated and classified into three categories: 0 (not diagnostically acceptable), 1 (few artifacts, but images were diagnostically acceptable), and 2 (good diagnostic quality images without substantial artifacts). MR images with diagnostic quality category 1 or 2 were included and subsequently reviewed. Diagnostic Criteria
The lung findings were reviewed according to the glossary of terms from the Fleischner Society (19,20).
Academic Radiology, Vol ■, No ■■, ■■ 2017
Statistical Analysis
The sensitivity, the specificity, the positive predictive value (PPV), and the negative predictive value (NPV) of MRI for diagnosing lung abnormalities of 3T lung MRI studies were calculated using HRCT findings as gold standard with a chisquare test separately for each abnormality. The kappa test of agreement was used to determine the strength of agreement between the HRCT and MRI findings as well as between the two reviewers. The degree of agreement was classified as slight (k < 0.20), fair (k = 0.21–0.40), moderate (k = 0.41– 0.60), substantial (k = 0.61–0.80), and almost perfect (k = 0.81– 1.00). All statistical tests were two-sided and were performed at a significance level of <0.05. The statistical analysis was conducted using SPSS software for Windows (version 22.0; SPSS Inc., Chicago, IL).
3T LUNG MRI IN ILD
TABLE 1. HRCT and MRI Findings in Pediatric Patients with Interstitial Lung Disease
Imaging Findings Consolidation Ground-glass opacity Nodules Cyst Septal thickening Bronchiectasis Parenchymal bands Fissural thickening
Number and Percentage of Patients on HRCT
Number and Percentage of Patients on MRI
4/12 (33%) 2/12 (17%) 3/12 (25%) 4/12 (33%) 4/12 (33%) 0/12 (0%) 2/12 (17%) 2/12 (17%)
4/12 (33%) 2/12 (17%) 2/12 (17%) 1/12 (8%) 2/12 (17%) 1/12 (8%) 2/12 (17%) 2/12 (17%)
HRCT, high-resolution computed tomography; MRI, magnetic resonance imaging.
RESULTS Study Cohort
A total of 12 pediatric patients with suspected ILD were prospectively and consecutively enrolled in the present study. The mean age of the patients was 8.5 years (range: 4–12 years), with a standard deviation of 2.7. There were six males and six females. Presenting clinical symptoms include chronic cough, wheezing, hypoxia, and tachypnea. On physical examination, clubbing was present in one patient. Out of the 12 patients, 3 patients (25%) had associated connective tissue disorder including juvenile idiopathic arthritis (n = 1), mixed connective tissue disorder (n = 1), and undifferentiated connective tissue disorder (n = 1). The underlying diagnoses of the remaining 9 patients (75%) include eosinophilic lung disease (n = 2), sarcoidosis (n = 1), nonspecific interstitial pneumonia (n = 1), Langerhan cell histiocytosis (n = 1), hypersensitivity pneumonitis (n = 1), pulmonary vein atresia (n = 1), and lymphoid interstitial pneumonia (n = 1). HRCT Findings
The HRCT findings of patients are summarized in Table 1. HRCT did not show any abnormal findings, which could be suggestive of ILD in three patients (25%). In the remaining nine patients (75%) with lung abnormalities on HRCT (Figs 1–3), consolidation (n = 4, 33%), cyst (n = 4, 33%), septal thickening (n = 4, 33%), nodules (n = 3, 25%), ground-glass opacity (n = 2, 17%), parenchymal band (n = 2, 17%), and fissural thickening (n = 2, 17%) were detected. Bronchiectasis was not seen. MRI Findings
3T lung MRI was performed in all the patients without any complication. All MR images were diagnostically acceptable with diagnostic quality categories 1 (n = 8, 67%) and 2 (n = 4, 33%). The MRI findings of the patients are summarized
in Table 1. The 3T lung MRI did not show any abnormal findings, which could be suggestive of ILD in five patients (42%). In the remaining seven patients (58%) with lung abnormalities on 3T lung MRI (Figs 1–3), consolidation (n = 4, 33%), ground-glass opacity (n = 2, 17%), nodules (n = 2, 17%), septal thickening (n = 2, 17%), parenchymal band (n = 2, 17%), fissural thickening (n = 2, 17%), cyst (n = 1, 17%), and bronchiectasis (n = 1, 17%) were detected.
Diagnostic Performance of MRI
With HRCT as a standard of reference, the diagnostic performance of 3T lung MRI was the same as HRCT in the detection of consolidation, parenchymal bands, and fissural thickening with sensitivity, specificity, PPV, and NPV (100%, respectively) (Table 2). However, 3T lung MRI was less reliable for accurately detecting nodules, ground-glass opacity, and septal thickening. 3T lung MRI yielded one falsepositive finding and one false-negative finding for groundglass opacity. 3T lung MRI also yielded one false-positive result for bronchiectasis. The overall accuracies of 3T lung MRI for detecting lung abnormalities in pediatric patients were 66.7% (sensitivity), 97.3% (specificity), 87.5% (PPV), and 91.25% (NPV).
Agreement Between HRCT and MRI
There was perfect agreement between HRCT and 3T lung MRI in the detection of parenchymal bands, consolidation, and fissural thickening (k = 1). However, only fair agreement was present between HRCT and 3T lung MRI in the detection of cysts and ground-glass opacities (k = 0.2–0.4). Detailed results are tabulated in Table 3. The overall agreement between HRCT and 3T lung MRI was categorized as substantial (k = 0.7). 3
SODHI ET AL
Academic Radiology, Vol ■, No ■■, ■■ 2017
Figure 1. Eight-year-old boy with a mixed connective tissue disorder. High-resolution computed tomography (a) revealed diffuse extensive bilateral smooth and irregular septal thickening (arrow) and fissural thickening with few ill-defined nodules. 3.0-T lung magnetic resonance imaging performed with axial T2 (b), Multivane MVTX axial (c), and balanced turbo field echo sequence (d) reveal similar imaging findings.
Agreement Between Two Reviewers
There was perfect agreement (k = 0.9–1.0) between the two reviewers for abnormalities detected on HRCT and 3T lung MRI. The strength of agreement between the two reviewers was also high. The concordance of findings using the kappa statistics was 0.968 on CT (for all findings) and 0.906 on 3T lung MRI (for all findings).
DISCUSSION In our study, 3T lung MRI performed equivalent to HRCT in the detection of consolidation, fissural thickening, and parenchymal bands in pediatric patients with ILD. MRI had high specificity for the detection of these findings. However, sensitivity for the detection of cysts, septal thickening, and ground-glass opacity was substantially lower. Also MRI correctly detected nodules in only two-third of the cases. In one case, bronchiectasis was falsely detected on MRI but was not seen on HRCT, which was the reference standard in our study. Overall, MRI had moderate sensitivity, with high specificity and NPV for detecting abnormalities associated with ILD in 4
the pediatric population. Also, there was substantial agreement between HRCT and MRI in the detection of all the findings as well as between the two reviewers. The results of our study showed that 3T lung MRI has poor sensitivity for the detection of ground-glass opacity, cysts, and septal thickening. We believe that the low signal intensity of cysts and mild ground-glass opacities is most likely an underlying reason for poor detection by MRI (21,22). In regard to suboptimal detection of septal thickening with lung MRI, it could be ascribed to the decreased visibility related to less cellular edema and relatively lesser interstitial fluid component associated with septal thickening. Results of previously published studies evaluating the performance of lung MRI in ILD have been encouraging (17,18). It is noteworthy that both the previous studies obtained on adult populations did not compare HRCT and MRI for detection of each finding such as septal thickening and groundglass opacity. One study analyzed MR images for increased T2 signal intensity of lung parenchyma, which was considered a marker for disease activity (17). However, in another study, a diagnosis of diffuse lung disease was made on MRI, if any segment of lung showed abnormal high signal
Academic Radiology, Vol ■, No ■■, ■■ 2017
3T LUNG MRI IN ILD
Figure 3. Nine-year-old girl with sarcoidosis. High-resolution computed tomography scan (a) revealed extensive septal thickening and fissural thickening (arrow) in both lungs, which is not evident on axial T2-weighted magnetic resonance imaging (b) and axial Multivane MVTX (c) magnetic resonance images when seen in isolation.
Figure 2. Four-year-old boy with unilateral pulmonary vein atresia. High-resolution computed tomography (a) revealed diffuse smooth interlobular septal thickening (arrow) in the left lung, which is also well demonstrated at T2 axial magnetic resonance imaging (b) and Multivane MVTX T2 axial image (c).
intensity (18). Thus, both studies assessed only for lung parenchymal signal changes in ILD and did not classify the specific radiological abnormalities (as in our study). However, one recent study by Gorkem et al. found that contrast-enhanced MRI
is comparable to CT for the detection of parenchymal abnormalities in pediatric patients with sarcoidosis (21). In this study, the investigators also have reported the poor sensitivity of MRI for nodules less than 3 mm, mild bronchiectasis, mild ground-glass opacity, and subpleural cysts (21), similar to our study. The difference between our study and Gorkem et al.’s study is that their study used a 1.5-T MRI scanner whereas our investigation is based on the use of 3T MRI scanner, which is first for the evaluation of pediatric patients with ILD. Results of our preliminary study indicate that 5
Academic Radiology, Vol ■, No ■■, ■■ 2017
SODHI ET AL
TABLE 2. Sensitivity, Specificity, PPV, and NPV of Magnetic Resonance Imaging in Detecting Lung Abnormalities in Pediatric Patients with Interstitial Lung Disease
Imaging Finding Consolidation Parenchymal bands Fissural thickening Nodules Ground-glass opacity Septal thickening Cyst Bronchiectasis
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
100 100 100 66.7 50 50 25 0
100 100 100 100 90 100 100 92
100 100 100 100 50 100 100 0
100 100 100 90 90 80 73 100
NPV, negative predictive value; PPV, positive predictive value.
TABLE 3. Agreement Between CT and MRI
Finding Septal thickening Bronchiectasis Cyst Parenchymal bands Consolidation Fissural thickening Nodule Ground-glass opacity
Agreement Between CT and MRI (kappa) .571 0 .308 1.00 1.00 1.00 .75 .40
CT, computed tomography; MRI, magnetic resonance imaging.
T2-weighted turbo spin echo and Multivane sequences had the best yield in ILD and correlated best with the CT findings. Although there are a few promising reports in the adult population of the use of MRI in patients with systemic sclerosis and interstitial pulmonary findings (23–25), we have not come across any published data on diagnostic utility of 3T MRI in children with pediatric ILD. With improvement in knowledge of the underlying pathogenesis of diffuse lung diseases in recent years, the understanding of ILD in children has changed substantially. It is highly likely that imaging will continue to play an important role in the diagnosis and follow-up of these children. HRCT is currently the modality of choice for detection and categorization of ILD. With HRCT, radiation dose still remains a concern, considering these children may require repeated examinations for follow-up and monitoring of disease processes (26,27). MRI is an attractive, nonionizing radiological alternative study to CT. Recent technical advances and refined pulse sequences have increased the diagnostic quality of lung MRI. Various studies in pediatric population have highlighted the role of lung MRI in the diagnosis of pulmonary infections (12–16). Most of these studies have emphasized on detection of parenchymal abnormalities by lung MRI. In addition, the role of MRI in evaluation of airways has also been suggested (28–32). Only few studies have been performed to 6
evaluate the role of MRI in the evaluation of ILD (17,18,21–25), which has shown promising results. Our study has also shown substantial agreement between HRCT and MRI, with an overall sensitivity of 66.67% and specificity of 97.33% of 3T lung MRI for the detection of all abnormalities related to pediatric ILD. The main idea to use MRI as an alternative to CT is not only avoidance of radiation exposure but also the possibility to functional imaging, which can be significant in the context of ILD and can be a potential for further research. However, lung MRI has few limitations, which include availability, cost factor, the need for anesthesia in smaller children, and a relatively higher scan time as compared to CT scan. Patient cooperation is a must in lung MRI and any movement can significantly impair image quality. We acknowledge that there are several limitations in our study. First, the study population size is small. However, the incidence of ILD in children is low, and these are just initial results of our pilot study performed on a 3T MRI. Also, our study group included only children older than 4 years of age who can undergo lung MRI study without sedation or intubation. Infants and young children (<4 years old) were not included in the study because of the ethical issues of sedation requirement for both CT and MRI. Second, the underlying diagnoses of ILD included in our study were varied, which created a somewhat heterogeneous patient population. However, we would like to emphasize that the inclusion of pediatric patients with various underlying ILDs closely follow the current practice pattern in real clinical setting. Dixon sequence with the given parameters (TR and flip angle combination that is likely going to saturate water signal) in the present study was not of much use. Having a true T1weighted sequence might have yielded different results. For this reason, the present study is a limited representation of the current capabilities of commercially available protocols for pulmonary MRI. Also, the standard of reference in the present study was HRCT scan, which by itself can miss few lesions. However, HRCT scan is the current imaging investigation of choice at our institution. Lastly, the use of newer sequences such as ultrafast echo time sequence hold potential in the detection of ILD in children; however, our lung MRI techniques and sequences did not include newer sequences because the goal of our study was to use the MRI techniques currently widely used for the evaluation of lungs. In conclusion, lung MRI, obtained with 3T MRI scanner and routinely available MRI protocols and sequences, can well detect abnormalities such as consolidation, parenchymal bands, and fissural thickening in children with ILD. However, evaluation of septal thickening, ground-glass opacity, nodules, and cysts is limited with 3T lung MRI. Our pilot study shows that lung MRI, with current routinely available sequences, is not currently near to replacing HRCT with MRI for the evaluation of pediatric ILD in the immediate future. However, this is only a pilot study, and larger studies are required in this field. In view of its lack of potentially harmful ionizing
Academic Radiology, Vol ■, No ■■, ■■ 2017
radiation, further investigation, particularly with newer MRI sequences, is needed to validate the role of lung MRI in the evaluation of pediatric ILD. ACKNOWLEDGMENT The authors thank Dr. Amit Kumar Shrivastav, Senior Research fellow, Indian Council of Medical Research, for his assistance in conducting this study. REFERENCES 1. Dinwiddie R, Sharief N, Crawford O. Idiopathic interstitial pneumonitis in children: a national survey in the United Kingdom and Ireland. Pediatr Pulmonol 2002; 34:23–29. 2. Fan LL, Kozinetz CA. Factors influencing survival in children with chronic interstitial lung disease. Am J Respir Crit Care Med 1997; 156(3 Pt 1):939– 942. 3. Langston C, Fan LL. Diffuse interstitial lung disease in infants. Pediatr Pulmonol 2001; S23:74–76. 4. Clement A, Nathan N, Epaud R, et al. Interstitial lung diseases in children. Orphanet J Rare Dis 2010; 5:22. 5. deBlic J, Midulla F, Barbato A, et al. Bronchoalveolar lavage in children. ERS task force on bronchoalveolar lavage in children. European Respiratory Society. Eur Respir J 2000; 15:217–231. 6. Refabert L, Rambaud C, Mamou-Mani T, et al. Cd1a-positivecells in bronchoalveolar lavage samples from children with Langerhans cell histiocytosis. J Pediatr 1996; 129:913–915. 7. Midulla F, Strappini PM, Ascoli V, et al. Bronchoalveolar lavage cell analysis in a child with chronic lipid pneumonia. EurRespir J 1998; 11:239– 242. 8. Oermann CM, Panesar KS, Langston C, et al. Pulmonary infiltrates with eosinophilia syndromes in children. J Pediatr 2000; 136:351–358. 9. Copley SJ, Bush A. HRCT of paediatric lung disease. PaediatrRespir Rev 2000; 1:141–147. 10. Klusmann M, Owens C. HRCT in paediatric diffuse interstitial lung disease–a review for 2009. Pediatr Radiol 2009; 39(suppl 3):471–481. 11. Vrielynck S, Mamou-Mani T, Emond S, et al. Diagnostic value of highresolution CT in the evaluation of chronic infiltrative lung disease in children. AJR Am J Roentgenol 2008; 191:914–920. 12. Rupprecht T, Böwing B, Kuth R, et al. Steady-state free precession projection MRI as a potential alternative to the conventional chest X-ray in pediatric patients with suspected pneumonia. Eur Radiol 2002; 12:2752– 2756. 13. Yikilmaz A, Koc A, Coskun A, et al. Evaluation of pneumonia in children: comparison of MRI with fast imaging sequences at 1.5T with chest radiographs. Acta Radiol 2011; 52:914–919. 14. Abolmaali ND, Schmitt J, Krauss S, et al. MR imaging of lung parenchyma at 0.2 T: evaluation of imaging techniques, comparative study with chest radiography and interobserver analysis. Eur Radiol 2004; 14:703– 708.
3T LUNG MRI IN ILD
15. Sodhi KS, Khandelwal N, Saxena AK, et al. Rapid lung MRI in children with pulmonary infections: time to change our diagnostic algorithms. J Magn Reson Imaging 2016; 43:1196–1206. 16. Sodhi KS, Khandelwal N, Saxena AK, et al. Rapid lung MRI: paradigm shift in evaluation of febrile neutropenia in children with leukemia: a pilot study. Leuk Lymphoma 2016; 57:70–75. 17. Lutterbey G, Grohé C, Gieseke J, et al. Initial experience with lung-MRI at 3.0 T: comparison with CT and clinical data in the evaluation of interstitial lung disease activity. Eur J Radiol 2007; 61:256– 261. 18. Lutterbey G, Gieseke J, Von Falkenhausen M, et al. Lung MRI at 3.0 T: a comparison of helical CT and high-field MRI in the detection of diffuse lung disease. Eur Radiol 2005; 15:324–328. 19. Hansell DM, Bankier AA, MacMahon H, et al. Fleischner Society: glossary of terms for thoracic imaging 1. Radiology 2008; 246:697– 722. 20. Eichinger M, Optazaite DE, Kopp-Schneider A, et al. Morphologic and functional scoring of cystic fibrosis lung disease using MRI. Eur J Radiol 2012; 81:1321–1329. 21. Gorkem SB, Köse S, Lee EY, et al. Thoracic MRI evaluation of sarcoidosis in children. Pediatr Pulmonol 2017; 52:494–499. 22. Chung JH, Little BP, Forssen AV, et al. Proton MRI in the evaluation of pulmonary sarcoidosis: comparison to chest CT. Eur J Radiol 2013; 82:2378–2385. 23. Mirsadraee S, Tse M, Kershaw L, et al. T1 characteristics of interstitial pulmonary fibrosis on 3T MRI—a predictor of early interstitial change? Quant Imaging Med Surg 2016; 6:42–49. 24. Müller CS, Warszawiak D, Paiva ED, et al. Pulmonary magnetic resonance imaging is similar to chest tomography in detecting inflammation in patients with systemic sclerosis. Rev Bras Reumatol 2017; 57:419– 424. 25. Barreto MM, Rafful PP, Rodrigues RS, et al. Correlation between computed tomographic and magnetic resonance imaging findings of parenchymal lung diseases. Eur J Radiol. 2013; 82:e492–e501. 26. Sodhi KS, Krishna S, Saxena AK, et al. Clinical application of “Justification” and “Optimization” principle of ALARA in pediatric CT imaging: “how many children can be protected from unnecessary radiation?” Eur J Radiol 2015; 84:1752–1757. 27. Sodhi KS, Lee EY. What all physicians should know about the potential radiation risk that computed tomography poses for paediatric patients. Acta Paediatr 2014; 103:807–811. 28. Sodhi KS, Khandelwal N. Magnetic resonance imaging of lungs as a radiation-free technique for lung pathologies immunodeficient patients. J ClinImmunol 2016; 36:621–623. 29. Liszewski MC, Ciet P, Sodhi KS, et al. Updates on MRI evaluation of pediatric large airways. AJR Am J Roentgenol 2017; 1–11. doi:10.2214/AJR.16.17132; [Epub ahead of print]. 30. Garg MK, Gupta P, Agarwal R, et al. MRI: a new paradigm in imaging evaluation of allergic bronchopulmonary aspergillosis? Chest 2015; 147:e58–e59. 31. Sodhi KS, Sharma M, Saxena AK, et al. MRI in thoracic tuberculosis in children. Indian J Pediatr 2017; doi: 10.1007/s12098-017-2392-3. 32. Sodhi KS, Bhalla AS, Mahomed N, et al. Imaging of thoracic tuberculosis in children: current and future directions. Pediatr Radiol 2017; 47:1260–1268.
7
Preliminary Laboratory Investigation
Diagnostic Utility of 3T Lung MRI in Children with Interstitial Lung Disease: A Prospective Pilot Study Kushaljit Singh Sodhi, MD, PhD, Madhurima Sharma, MD, Edward Y. Lee, MD, MPH, Akshay Kumar Saxena, MD, Joseph L. Mathew, MD, Meenu Singh, MD, Niranjan Khandelwal, MD Rationale and Objectives: The objective of this study was to assess the diagnostic utility of 3-tesla (3T) magnetic resonance imaging (MRI) of lungs in the detection of interstitial lung disease (ILD) in pediatric patients. Materials and Methods: Twelve children (mean: 8.5 years, range: 4–12 years) with ILD were consecutively enrolled in this prospective study. HRCT and 3T lung MRI were performed in all patients within 2 days of each other. The sensitivity, the specificity, the positive predictive value, and the negative predictive value of detecting lung abnormalities related to ILD with 3T lung MRI were calculated, with high-resolution computed tomography (HRCT) as a standard of reference. Agreement between HRCT and 3T lung MRI, as well as between two reviewers, was calculated with the kappa coefficient. Results: 3T lung MRI had low sensitivity (66.67%) and high specificity (97.33%) in the detection of abnormalities related to ILD when compared to HRCT in children. Although 3T lung MRI performed well in the detection of consolidation, parenchymal bands and fissural thickening with a sensitivity of 100%, the sensitivity of 3T lung MRI in the detection of septal thickening, ground-glass opacity, nodules, and cysts was relatively low (50.0%, 50.0%, 66.67%, and 25.0%, respectively). Substantial agreement was seen between HRCT and 3T lung MRI (k = 0.7), whereas perfect agreement was seen between two reviewers in detecting abnormalities related to pediatric ILD (k = 0.9–1.0). Conclusions: In comparison to HRCT, 3T lung MRI with routinely available MRI protocols and sequences can also well detect abnormalities such as consolidation, parenchymal bands, and fissural thickening in children with ILD. However, evaluation of septal thickening, ground-glass opacity, nodules, and cysts is limited with 3T lung MRI. Key Words: Interstitial lung disease; imaging; MRI. © 2017 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
I
nterstitial lung diseases (ILDs) in pediatric populations are a heterogeneous group of disorders with diffuse lung involvement, frequently leading to significant mortality and morbidity (1,2). Unlike the adult population, the spectrum of ILD in children is entirely different, with certain diseases unique to the pediatric population (3). In addition, clinical manifestations of ILD in children are unfortunately nonspecific and variable, often leading to missed or delayed diagnosis. Acad Radiol 2017; ■:■■–■■ From the Department of Radiodiagnosis and Imaging, PGIMER, Sector-12, Chandigarh, 160012, India (K.S.S., M.S., A.K.S., N.K.); Departments of Radiology and Medicine, Pulmonary Division, Boston Children’s Hospital and Harvard Medical School, Boston, Massachusetts (E.Y.L.); Department of Pediatrics, PGIMER, Chandigarh, India (J.L.M., M.S.). Received July 7, 2017; revised September 1, 2017; accepted September 21, 2017. Source of Funding: This study has been sponsored by funds from Indian Council of Medical Research (2012-2081). Address correspondence to: K.S.S. e-mail: [email protected] © 2017 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.acra.2017.09.013
Main investigations employed for the diagnosis of ILD include imaging, pulmonary function tests, bronchoalveolar lavage, and lung biopsy. Pulmonary function tests are difficult to perform in children because of poor cooperation (4), and bronchoalveolar lavage is of diagnostic value in a substantially limited number of conditions (5–8). Tissue biopsy provides a definitive diagnosis but has disadvantage of being invasive. Considering the limitations of all these currently available investigations, imaging plays a crucial role in the diagnosis and evaluation of pediatric ILD. Currently, high-resolution computed tomography (HRCT) remains the imaging investigation of choice for the evaluation of ILD (9–11). HRCT is helpful not only in the diagnosis of ILD but also in monitoring disease activity and severity. However, the main concern with HRCT is the potentially harmful radiation, particularly considering chronicity of ILD and requirement for frequent follow-up examinations. Magnetic resonance imaging (MRI) is an attractive radiationfree alternative to computed tomography (CT). In recent years, lung MRI has evolved because of refinement in pulse sequences 1
SODHI ET AL
and technical advances. Few studies have highlighted the role of lung MRI in the evaluation of pediatric pulmonary infections (12–16). Only a few published reports on MRI in diffuse lung diseases in adult populations are available (17,18). However, there is no published data on the role of lung MRI, particularly with a 3-tesla (3T) MRI scanner, in pediatric ILD. Although the higher field strength of 3T MRI offers the advantage of improved contrast and increased signal from lung tissue, high-field strength susceptibility artifacts resulting in more pronounced image distortion and signal loss would be more prevalent on 3T MRI. Therefore, the purpose of our pilot study was to assess the diagnostic utility of 3T MRI of lungs in the detection of ILD in pediatric patients and to compare it with the current imaging gold standard, HRCT. Although this is a small prospective pilot study, it has the potential to pave the way for further research into the clinical application of lung MRI for ILD in children.
MATERIALS AND METHODS This was a prospective study, approved by our institutional ethics committee. Informed written consent was obtained from the parents or guardians of all pediatric patients enrolled in the present study.
Study Population
From October 2015 to January 2017, a total of 12 consecutive pediatric patients, referred to our department for HRCT of the chest for the evaluation of ILD, were enrolled in this prospective study. The inclusion criteria were pediatric patients in the age range of 4–14 years with clinical suspicion of ILD, supported by clinical symptoms and chest radiographic findings, who underwent HRCT for further evaluation. MRI was performed after HRCT in all patients within 2 days of each other. Children less than 4 years of age who required sedation were excluded from the present study.
Imaging Technique
HRCT Protocol HRCT of the chest was performed in all pediatric patients enrolled in this study using a 64-row multidetector CT scanner (Aquilion64; Toshiba America Medical Systems, Otawara, Japan). The scans were obtained from the lung apex to the base. Acquisition parameters for HRCT were 120 kVp, 100– 150 mA, 1-mm collimation, and 5- to 10-mm spacing, and a high-spatial resolution image reconstruction algorithm was used. Axial CT images were subsequently reconstructed using bone and soft tissue reconstruction algorithms and were evaluated on a Terrarecon workstation (Foster City, CA) with a standard lung window (level: 500 HU, width: 1500 HU) 2
Academic Radiology, Vol ■, No ■■, ■■ 2017
and mediastinal window settings (level: 350 HU, width: 50 HU). MRI Protocol MRI of the chest was performed in all pediatric patients enrolled in the present study using a 3T MRI scanner (Ingenia; Philips Health Systems, Amsterdam, The Netherlands). All MRI sequences were acquired with breath-holding technique and were nonrespiratory and nonelectrocardiogram gated. All MRI studies were obtained without sedation or contrast administration. All children were duly prepared for breath holding before MR acquisition. Children who were dyspneic or who could not hold their breadth or were uncooperative were not taken up for MRI study. The chest was scanned in the axial plane in the craniocaudal direction in all pediatric patients. The following MRI pulse sequences and parameters were used: (1) T2-weighted turbo spin echo (repetition time [TR] 1250 ms, echo time [TE] 80 ms, slice 4 mm, flip angle 90°, field of view [FOV] 300 mm, matrix 216 × 189); (2) balanced turbo field echo (TR 2.6 ms, TE 1.32 ms, slice 4 mm, flip angle 45°, FOV 270 mm, matrix 152 × 135); (3) Dixon T1 (TR 2.6 ms, TE 1.32 ms, slice 4 mm, flip angle 90°, FOV 250 mm; matrix 156 × 124); and (4) Multivane MVTX (Philips, Netherlands) (TR 2500 ms, TE 127 ms, slice 4 mm, flip angle 90°, FOV 275 mm, matrix 184 × 184). Multivane is a PROPELLER (Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction) sequence. The total MRI scanning time was approximately 7 minutes for all four MRI sequences combined. The total time on the magnetic resonance (MR) table varied from 20 to 25 minutes. HRCT and MR Image Evaluation HRCT and MR images were reviewed by two experienced pediatric radiologists, with 15 and 12 years of experience in interpreting thoracic HRCT and MRI. To minimize the potential bias, the reviewers were blinded to the clinical profile of the patients. Identification information from HRCT and MR images was removed before the review session. The order of HRCT and MR images was randomized. In addition, HRCT and MR images were reviewed with 3- to 4-week time intervals between review sessions. Both observers were blinded to reciprocal image assessment, and in case of doubting cases, a consensus agreement was pursued. First, the reviewers evaluated the quality of thoracic MRI studies. Thoracic MRI studies were evaluated and classified into three categories: 0 (not diagnostically acceptable), 1 (few artifacts, but images were diagnostically acceptable), and 2 (good diagnostic quality images without substantial artifacts). MR images with diagnostic quality category 1 or 2 were included and subsequently reviewed. Diagnostic Criteria
The lung findings were reviewed according to the glossary of terms from the Fleischner Society (19,20).
Academic Radiology, Vol ■, No ■■, ■■ 2017
Statistical Analysis
The sensitivity, the specificity, the positive predictive value (PPV), and the negative predictive value (NPV) of MRI for diagnosing lung abnormalities of 3T lung MRI studies were calculated using HRCT findings as gold standard with a chisquare test separately for each abnormality. The kappa test of agreement was used to determine the strength of agreement between the HRCT and MRI findings as well as between the two reviewers. The degree of agreement was classified as slight (k < 0.20), fair (k = 0.21–0.40), moderate (k = 0.41– 0.60), substantial (k = 0.61–0.80), and almost perfect (k = 0.81– 1.00). All statistical tests were two-sided and were performed at a significance level of <0.05. The statistical analysis was conducted using SPSS software for Windows (version 22.0; SPSS Inc., Chicago, IL).
3T LUNG MRI IN ILD
TABLE 1. HRCT and MRI Findings in Pediatric Patients with Interstitial Lung Disease
Imaging Findings Consolidation Ground-glass opacity Nodules Cyst Septal thickening Bronchiectasis Parenchymal bands Fissural thickening
Number and Percentage of Patients on HRCT
Number and Percentage of Patients on MRI
4/12 (33%) 2/12 (17%) 3/12 (25%) 4/12 (33%) 4/12 (33%) 0/12 (0%) 2/12 (17%) 2/12 (17%)
4/12 (33%) 2/12 (17%) 2/12 (17%) 1/12 (8%) 2/12 (17%) 1/12 (8%) 2/12 (17%) 2/12 (17%)
HRCT, high-resolution computed tomography; MRI, magnetic resonance imaging.
RESULTS Study Cohort
A total of 12 pediatric patients with suspected ILD were prospectively and consecutively enrolled in the present study. The mean age of the patients was 8.5 years (range: 4–12 years), with a standard deviation of 2.7. There were six males and six females. Presenting clinical symptoms include chronic cough, wheezing, hypoxia, and tachypnea. On physical examination, clubbing was present in one patient. Out of the 12 patients, 3 patients (25%) had associated connective tissue disorder including juvenile idiopathic arthritis (n = 1), mixed connective tissue disorder (n = 1), and undifferentiated connective tissue disorder (n = 1). The underlying diagnoses of the remaining 9 patients (75%) include eosinophilic lung disease (n = 2), sarcoidosis (n = 1), nonspecific interstitial pneumonia (n = 1), Langerhan cell histiocytosis (n = 1), hypersensitivity pneumonitis (n = 1), pulmonary vein atresia (n = 1), and lymphoid interstitial pneumonia (n = 1). HRCT Findings
The HRCT findings of patients are summarized in Table 1. HRCT did not show any abnormal findings, which could be suggestive of ILD in three patients (25%). In the remaining nine patients (75%) with lung abnormalities on HRCT (Figs 1–3), consolidation (n = 4, 33%), cyst (n = 4, 33%), septal thickening (n = 4, 33%), nodules (n = 3, 25%), ground-glass opacity (n = 2, 17%), parenchymal band (n = 2, 17%), and fissural thickening (n = 2, 17%) were detected. Bronchiectasis was not seen. MRI Findings
3T lung MRI was performed in all the patients without any complication. All MR images were diagnostically acceptable with diagnostic quality categories 1 (n = 8, 67%) and 2 (n = 4, 33%). The MRI findings of the patients are summarized
in Table 1. The 3T lung MRI did not show any abnormal findings, which could be suggestive of ILD in five patients (42%). In the remaining seven patients (58%) with lung abnormalities on 3T lung MRI (Figs 1–3), consolidation (n = 4, 33%), ground-glass opacity (n = 2, 17%), nodules (n = 2, 17%), septal thickening (n = 2, 17%), parenchymal band (n = 2, 17%), fissural thickening (n = 2, 17%), cyst (n = 1, 17%), and bronchiectasis (n = 1, 17%) were detected.
Diagnostic Performance of MRI
With HRCT as a standard of reference, the diagnostic performance of 3T lung MRI was the same as HRCT in the detection of consolidation, parenchymal bands, and fissural thickening with sensitivity, specificity, PPV, and NPV (100%, respectively) (Table 2). However, 3T lung MRI was less reliable for accurately detecting nodules, ground-glass opacity, and septal thickening. 3T lung MRI yielded one falsepositive finding and one false-negative finding for groundglass opacity. 3T lung MRI also yielded one false-positive result for bronchiectasis. The overall accuracies of 3T lung MRI for detecting lung abnormalities in pediatric patients were 66.7% (sensitivity), 97.3% (specificity), 87.5% (PPV), and 91.25% (NPV).
Agreement Between HRCT and MRI
There was perfect agreement between HRCT and 3T lung MRI in the detection of parenchymal bands, consolidation, and fissural thickening (k = 1). However, only fair agreement was present between HRCT and 3T lung MRI in the detection of cysts and ground-glass opacities (k = 0.2–0.4). Detailed results are tabulated in Table 3. The overall agreement between HRCT and 3T lung MRI was categorized as substantial (k = 0.7). 3
SODHI ET AL
Academic Radiology, Vol ■, No ■■, ■■ 2017
Figure 1. Eight-year-old boy with a mixed connective tissue disorder. High-resolution computed tomography (a) revealed diffuse extensive bilateral smooth and irregular septal thickening (arrow) and fissural thickening with few ill-defined nodules. 3.0-T lung magnetic resonance imaging performed with axial T2 (b), Multivane MVTX axial (c), and balanced turbo field echo sequence (d) reveal similar imaging findings.
Agreement Between Two Reviewers
There was perfect agreement (k = 0.9–1.0) between the two reviewers for abnormalities detected on HRCT and 3T lung MRI. The strength of agreement between the two reviewers was also high. The concordance of findings using the kappa statistics was 0.968 on CT (for all findings) and 0.906 on 3T lung MRI (for all findings).
DISCUSSION In our study, 3T lung MRI performed equivalent to HRCT in the detection of consolidation, fissural thickening, and parenchymal bands in pediatric patients with ILD. MRI had high specificity for the detection of these findings. However, sensitivity for the detection of cysts, septal thickening, and ground-glass opacity was substantially lower. Also MRI correctly detected nodules in only two-third of the cases. In one case, bronchiectasis was falsely detected on MRI but was not seen on HRCT, which was the reference standard in our study. Overall, MRI had moderate sensitivity, with high specificity and NPV for detecting abnormalities associated with ILD in 4
the pediatric population. Also, there was substantial agreement between HRCT and MRI in the detection of all the findings as well as between the two reviewers. The results of our study showed that 3T lung MRI has poor sensitivity for the detection of ground-glass opacity, cysts, and septal thickening. We believe that the low signal intensity of cysts and mild ground-glass opacities is most likely an underlying reason for poor detection by MRI (21,22). In regard to suboptimal detection of septal thickening with lung MRI, it could be ascribed to the decreased visibility related to less cellular edema and relatively lesser interstitial fluid component associated with septal thickening. Results of previously published studies evaluating the performance of lung MRI in ILD have been encouraging (17,18). It is noteworthy that both the previous studies obtained on adult populations did not compare HRCT and MRI for detection of each finding such as septal thickening and groundglass opacity. One study analyzed MR images for increased T2 signal intensity of lung parenchyma, which was considered a marker for disease activity (17). However, in another study, a diagnosis of diffuse lung disease was made on MRI, if any segment of lung showed abnormal high signal
Academic Radiology, Vol ■, No ■■, ■■ 2017
3T LUNG MRI IN ILD
Figure 3. Nine-year-old girl with sarcoidosis. High-resolution computed tomography scan (a) revealed extensive septal thickening and fissural thickening (arrow) in both lungs, which is not evident on axial T2-weighted magnetic resonance imaging (b) and axial Multivane MVTX (c) magnetic resonance images when seen in isolation.
Figure 2. Four-year-old boy with unilateral pulmonary vein atresia. High-resolution computed tomography (a) revealed diffuse smooth interlobular septal thickening (arrow) in the left lung, which is also well demonstrated at T2 axial magnetic resonance imaging (b) and Multivane MVTX T2 axial image (c).
intensity (18). Thus, both studies assessed only for lung parenchymal signal changes in ILD and did not classify the specific radiological abnormalities (as in our study). However, one recent study by Gorkem et al. found that contrast-enhanced MRI
is comparable to CT for the detection of parenchymal abnormalities in pediatric patients with sarcoidosis (21). In this study, the investigators also have reported the poor sensitivity of MRI for nodules less than 3 mm, mild bronchiectasis, mild ground-glass opacity, and subpleural cysts (21), similar to our study. The difference between our study and Gorkem et al.’s study is that their study used a 1.5-T MRI scanner whereas our investigation is based on the use of 3T MRI scanner, which is first for the evaluation of pediatric patients with ILD. Results of our preliminary study indicate that 5
Academic Radiology, Vol ■, No ■■, ■■ 2017
SODHI ET AL
TABLE 2. Sensitivity, Specificity, PPV, and NPV of Magnetic Resonance Imaging in Detecting Lung Abnormalities in Pediatric Patients with Interstitial Lung Disease
Imaging Finding Consolidation Parenchymal bands Fissural thickening Nodules Ground-glass opacity Septal thickening Cyst Bronchiectasis
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
100 100 100 66.7 50 50 25 0
100 100 100 100 90 100 100 92
100 100 100 100 50 100 100 0
100 100 100 90 90 80 73 100
NPV, negative predictive value; PPV, positive predictive value.
TABLE 3. Agreement Between CT and MRI
Finding Septal thickening Bronchiectasis Cyst Parenchymal bands Consolidation Fissural thickening Nodule Ground-glass opacity
Agreement Between CT and MRI (kappa) .571 0 .308 1.00 1.00 1.00 .75 .40
CT, computed tomography; MRI, magnetic resonance imaging.
T2-weighted turbo spin echo and Multivane sequences had the best yield in ILD and correlated best with the CT findings. Although there are a few promising reports in the adult population of the use of MRI in patients with systemic sclerosis and interstitial pulmonary findings (23–25), we have not come across any published data on diagnostic utility of 3T MRI in children with pediatric ILD. With improvement in knowledge of the underlying pathogenesis of diffuse lung diseases in recent years, the understanding of ILD in children has changed substantially. It is highly likely that imaging will continue to play an important role in the diagnosis and follow-up of these children. HRCT is currently the modality of choice for detection and categorization of ILD. With HRCT, radiation dose still remains a concern, considering these children may require repeated examinations for follow-up and monitoring of disease processes (26,27). MRI is an attractive, nonionizing radiological alternative study to CT. Recent technical advances and refined pulse sequences have increased the diagnostic quality of lung MRI. Various studies in pediatric population have highlighted the role of lung MRI in the diagnosis of pulmonary infections (12–16). Most of these studies have emphasized on detection of parenchymal abnormalities by lung MRI. In addition, the role of MRI in evaluation of airways has also been suggested (28–32). Only few studies have been performed to 6
evaluate the role of MRI in the evaluation of ILD (17,18,21–25), which has shown promising results. Our study has also shown substantial agreement between HRCT and MRI, with an overall sensitivity of 66.67% and specificity of 97.33% of 3T lung MRI for the detection of all abnormalities related to pediatric ILD. The main idea to use MRI as an alternative to CT is not only avoidance of radiation exposure but also the possibility to functional imaging, which can be significant in the context of ILD and can be a potential for further research. However, lung MRI has few limitations, which include availability, cost factor, the need for anesthesia in smaller children, and a relatively higher scan time as compared to CT scan. Patient cooperation is a must in lung MRI and any movement can significantly impair image quality. We acknowledge that there are several limitations in our study. First, the study population size is small. However, the incidence of ILD in children is low, and these are just initial results of our pilot study performed on a 3T MRI. Also, our study group included only children older than 4 years of age who can undergo lung MRI study without sedation or intubation. Infants and young children (<4 years old) were not included in the study because of the ethical issues of sedation requirement for both CT and MRI. Second, the underlying diagnoses of ILD included in our study were varied, which created a somewhat heterogeneous patient population. However, we would like to emphasize that the inclusion of pediatric patients with various underlying ILDs closely follow the current practice pattern in real clinical setting. Dixon sequence with the given parameters (TR and flip angle combination that is likely going to saturate water signal) in the present study was not of much use. Having a true T1weighted sequence might have yielded different results. For this reason, the present study is a limited representation of the current capabilities of commercially available protocols for pulmonary MRI. Also, the standard of reference in the present study was HRCT scan, which by itself can miss few lesions. However, HRCT scan is the current imaging investigation of choice at our institution. Lastly, the use of newer sequences such as ultrafast echo time sequence hold potential in the detection of ILD in children; however, our lung MRI techniques and sequences did not include newer sequences because the goal of our study was to use the MRI techniques currently widely used for the evaluation of lungs. In conclusion, lung MRI, obtained with 3T MRI scanner and routinely available MRI protocols and sequences, can well detect abnormalities such as consolidation, parenchymal bands, and fissural thickening in children with ILD. However, evaluation of septal thickening, ground-glass opacity, nodules, and cysts is limited with 3T lung MRI. Our pilot study shows that lung MRI, with current routinely available sequences, is not currently near to replacing HRCT with MRI for the evaluation of pediatric ILD in the immediate future. However, this is only a pilot study, and larger studies are required in this field. In view of its lack of potentially harmful ionizing
Academic Radiology, Vol ■, No ■■, ■■ 2017
radiation, further investigation, particularly with newer MRI sequences, is needed to validate the role of lung MRI in the evaluation of pediatric ILD. ACKNOWLEDGMENT The authors thank Dr. Amit Kumar Shrivastav, Senior Research fellow, Indian Council of Medical Research, for his assistance in conducting this study. REFERENCES 1. Dinwiddie R, Sharief N, Crawford O. Idiopathic interstitial pneumonitis in children: a national survey in the United Kingdom and Ireland. Pediatr Pulmonol 2002; 34:23–29. 2. Fan LL, Kozinetz CA. Factors influencing survival in children with chronic interstitial lung disease. Am J Respir Crit Care Med 1997; 156(3 Pt 1):939– 942. 3. Langston C, Fan LL. Diffuse interstitial lung disease in infants. Pediatr Pulmonol 2001; S23:74–76. 4. Clement A, Nathan N, Epaud R, et al. Interstitial lung diseases in children. Orphanet J Rare Dis 2010; 5:22. 5. deBlic J, Midulla F, Barbato A, et al. Bronchoalveolar lavage in children. ERS task force on bronchoalveolar lavage in children. European Respiratory Society. Eur Respir J 2000; 15:217–231. 6. Refabert L, Rambaud C, Mamou-Mani T, et al. Cd1a-positivecells in bronchoalveolar lavage samples from children with Langerhans cell histiocytosis. J Pediatr 1996; 129:913–915. 7. Midulla F, Strappini PM, Ascoli V, et al. Bronchoalveolar lavage cell analysis in a child with chronic lipid pneumonia. EurRespir J 1998; 11:239– 242. 8. Oermann CM, Panesar KS, Langston C, et al. Pulmonary infiltrates with eosinophilia syndromes in children. J Pediatr 2000; 136:351–358. 9. Copley SJ, Bush A. HRCT of paediatric lung disease. PaediatrRespir Rev 2000; 1:141–147. 10. Klusmann M, Owens C. HRCT in paediatric diffuse interstitial lung disease–a review for 2009. Pediatr Radiol 2009; 39(suppl 3):471–481. 11. Vrielynck S, Mamou-Mani T, Emond S, et al. Diagnostic value of highresolution CT in the evaluation of chronic infiltrative lung disease in children. AJR Am J Roentgenol 2008; 191:914–920. 12. Rupprecht T, Böwing B, Kuth R, et al. Steady-state free precession projection MRI as a potential alternative to the conventional chest X-ray in pediatric patients with suspected pneumonia. Eur Radiol 2002; 12:2752– 2756. 13. Yikilmaz A, Koc A, Coskun A, et al. Evaluation of pneumonia in children: comparison of MRI with fast imaging sequences at 1.5T with chest radiographs. Acta Radiol 2011; 52:914–919. 14. Abolmaali ND, Schmitt J, Krauss S, et al. MR imaging of lung parenchyma at 0.2 T: evaluation of imaging techniques, comparative study with chest radiography and interobserver analysis. Eur Radiol 2004; 14:703– 708.
3T LUNG MRI IN ILD
15. Sodhi KS, Khandelwal N, Saxena AK, et al. Rapid lung MRI in children with pulmonary infections: time to change our diagnostic algorithms. J Magn Reson Imaging 2016; 43:1196–1206. 16. Sodhi KS, Khandelwal N, Saxena AK, et al. Rapid lung MRI: paradigm shift in evaluation of febrile neutropenia in children with leukemia: a pilot study. Leuk Lymphoma 2016; 57:70–75. 17. Lutterbey G, Grohé C, Gieseke J, et al. Initial experience with lung-MRI at 3.0 T: comparison with CT and clinical data in the evaluation of interstitial lung disease activity. Eur J Radiol 2007; 61:256– 261. 18. Lutterbey G, Gieseke J, Von Falkenhausen M, et al. Lung MRI at 3.0 T: a comparison of helical CT and high-field MRI in the detection of diffuse lung disease. Eur Radiol 2005; 15:324–328. 19. Hansell DM, Bankier AA, MacMahon H, et al. Fleischner Society: glossary of terms for thoracic imaging 1. Radiology 2008; 246:697– 722. 20. Eichinger M, Optazaite DE, Kopp-Schneider A, et al. Morphologic and functional scoring of cystic fibrosis lung disease using MRI. Eur J Radiol 2012; 81:1321–1329. 21. Gorkem SB, Köse S, Lee EY, et al. Thoracic MRI evaluation of sarcoidosis in children. Pediatr Pulmonol 2017; 52:494–499. 22. Chung JH, Little BP, Forssen AV, et al. Proton MRI in the evaluation of pulmonary sarcoidosis: comparison to chest CT. Eur J Radiol 2013; 82:2378–2385. 23. Mirsadraee S, Tse M, Kershaw L, et al. T1 characteristics of interstitial pulmonary fibrosis on 3T MRI—a predictor of early interstitial change? Quant Imaging Med Surg 2016; 6:42–49. 24. Müller CS, Warszawiak D, Paiva ED, et al. Pulmonary magnetic resonance imaging is similar to chest tomography in detecting inflammation in patients with systemic sclerosis. Rev Bras Reumatol 2017; 57:419– 424. 25. Barreto MM, Rafful PP, Rodrigues RS, et al. Correlation between computed tomographic and magnetic resonance imaging findings of parenchymal lung diseases. Eur J Radiol. 2013; 82:e492–e501. 26. Sodhi KS, Krishna S, Saxena AK, et al. Clinical application of “Justification” and “Optimization” principle of ALARA in pediatric CT imaging: “how many children can be protected from unnecessary radiation?” Eur J Radiol 2015; 84:1752–1757. 27. Sodhi KS, Lee EY. What all physicians should know about the potential radiation risk that computed tomography poses for paediatric patients. Acta Paediatr 2014; 103:807–811. 28. Sodhi KS, Khandelwal N. Magnetic resonance imaging of lungs as a radiation-free technique for lung pathologies immunodeficient patients. J ClinImmunol 2016; 36:621–623. 29. Liszewski MC, Ciet P, Sodhi KS, et al. Updates on MRI evaluation of pediatric large airways. AJR Am J Roentgenol 2017; 1–11. doi:10.2214/AJR.16.17132; [Epub ahead of print]. 30. Garg MK, Gupta P, Agarwal R, et al. MRI: a new paradigm in imaging evaluation of allergic bronchopulmonary aspergillosis? Chest 2015; 147:e58–e59. 31. Sodhi KS, Sharma M, Saxena AK, et al. MRI in thoracic tuberculosis in children. Indian J Pediatr 2017; doi: 10.1007/s12098-017-2392-3. 32. Sodhi KS, Bhalla AS, Mahomed N, et al. Imaging of thoracic tuberculosis in children: current and future directions. Pediatr Radiol 2017; 47:1260–1268.
7