Lung disease assessment in primary ciliary dyskinesia: a comparison between chest high-field magnetic resonance imaging and high-resolution computed tomography findings

Primary ciliary dyskinesia (PCD; MIM #244400) is a rare (1:15–30 000 live births) and usually autosomal recessive disease associated with situs viscerum inversus (Kartagener syndrome) in nearly half the cases [1]. Impaired mucociliary clearance due to defective motility of cilia is the hallmark of the condition [2]. Early clinical events are represented by continuous rhinorrhoea from the first days of life, respiratory distress or neonatal pneumonia with no obvious predisposing cause, and chronic or recurrent lower and upper airway infections, e.g. otitis media and purulent rhinosinusitis [3]. Three lower airway diseases are associated with PCD: pneumonia, bronchiectasis, and asthma [3].

Computed tomography (CT) of the chest, particularly high-resolution CT (HRCT), has become the method of choice to evaluate chronic lung disease at any age [4, 5]. Nevertheless, CT entails exposure to ionizing radiation, thereby increasing the risk of cancer in exposed individuals [6, 7]. Magnetic resonance imaging (MRI) of the chest has been proposed as a potential radiation-free technique in several chest disorders [8–13]. However, its application in lung disease has long been limited by technical problems, namely a low signal-to-noise ratio because of the low proton density of the lung, and artifacts due to cardiac and breathing motion or to air/soft tissue transition [14]. Nevertheless, the results of chest MRI in assessment of lung disease were found to be comparable with those of conventional chest X-ray and CT, mainly in patients with cystic fibrosis (CF) [15–17]. Lung disease in PCD is similar to lung disease in CF, although changes can be milder in PCD [18]. Recent studies showed that CT scoring systems adequately describe the extent and severity of PCD lung changes [18–20]. Patients with PCD may develop chronic lower airways symptoms and/or signs at any time [21]. Therefore, a sensitive radiation-free imaging tool is highly desirable for the longitudinal assessment of structural lung damage.

To our knowledge, no study has compared chest CT to MRI in PCD. The primary aim of this pilot study was to assess whether MRI is as effective as CT in identifying pulmonary abnormalities in patients with PCD. Our secondary aim was to investigate the relationships between the severity and extension of lung disease, identified with HRCT and MRI, and pulmonary function tests (PFTs).

Median HRCT and MRI scores are summarized in Table 4. Since only one or two patients had sacculations/abscesses, bullae or emphysema, agreement between HRCT and MRI scores was not calculated for these categories. Moreover, agreement for generation of bronchial divisions involved by bronchiectasis or plugging was not computable because of the constant value assigned to this category at HRCT and at MRI in all the subjects. However, the same score for generation of bronchial divisions involved by the above mentioned changes was attributed in each patient to both HRCT and MRI scans. For all the other categories, agreement between the two techniques was good or excellent. Figure 1 shows the excellent correlation (r = 0.95) between HRCT and MRI total scores.

	HRCT	MRI	r
Severity of bronchiectasis	2 (0–3)	2 (0–3)	0.87
Severity of peribronchial wall thickening	2 (1–2)	2 (0–2)	1
Extent of bronchiectasis	2 (0–3)	2 (0–3)	0.97
Extent of mucous plugging	2 (0–3)	2 (1–3)	0.94
Extent of sacculations or abscesses	0 (0–1)	0 (0–1)	NA
Generation of bronchial divisions involved (bronchiectasis or plugging)	3*	3*	NA
Severity of bullae	0 (0–3)	0*	NA
Severity of emphysema	0 (0–1)	0 (0–1)	NA
Severity of collapse or consolidation	2 (0–2)	2 (0–2)	1
Total score	12 (6–20)	12 (5–17)	0.95

NA: not applicable (see text).

Values in parentheses are ranges.

* No range values are given because all patients had the same score. For generation of bronchial divisions involved by bronchiectasis or plugging, the same score was attributed in each patient to both HRCT and MRI scans.

Median FVC and FEV₁ were 96% predicted (range, 57–121) and 85% predicted (range, 47–108), respectively. Children had higher values than adults for both FVC (96% predicted [range, 82–120] versus 80% predicted [range, 57–121], respectively) and FEV₁ (89% predicted [range, 72–108] versus 60% predicted [range, 47–103], respectively). Total HRCT and MRI scores were significantly related to FVC (r = -0.5, p = 0.05; and r = -0.7, p = 0.009, respectively) and FEV₁ (r = -0.6, p = 0.03; and r = -0.7, p = 0.009, respectively). Fifty-four percent of patients had a normal FEV₁.

The most common pathogen at deep throat or sputum culture was Haemophilus influenzae in both adults and children (60% and 50%, respectively). Pseudomonas aeruginosa was isolated only in one children and in one adult, whereas Staphylococcus aureus was found in one boy. In three cases no pathogens were cultured.

Figure 2 shows an area of consolidation at HRCT (panel A) and MRI (panel B) scans of a boy with PCD and situs viscerum inversus (Kartagener syndrome).

In this study, we compared the efficacy of chest HRCT and high-field MRI in the assessment of the severity and extent of lung abnormalities in children and young adults with PCD. The agreement between the two techniques was good or excellent for most of the abnormalities considered. Bronchiectasis, peribronchial wall thickening, mucous plugging, and collapse/consolidation were the most frequent abnormalities. HRCT and MRI total scores were significantly related to PFTs.

At present, the management of patients with PCD is based on clinical assessment and regular sputum cultures combined with pulmonary function evaluation [3]. Pulmonary function measurement is generally used to assess lung disease in PCD. It has been previously reported that lung function assessment, and particularly static lung volumes evaluation, may provide suggestive information about peripheral airway disease and its evolution over time [3, 21, 24–26]. However, PFTs are relatively insensitive markers of early disease and fail to detect regional structural changes because these tools reflect the function of the lung as a whole and gives no information about localized abnormalities. Chest CT provides a more accurate picture of the type, distribution and nature of lung changes than either PFTs or chest radiography [4, 27]. In fact, we recently found that 40% of PCD patients had a normal FEV₁ despite CT evidence of significant structural lung damage [18]. Our MRI results reported herein confirm this finding in that pulmonary abnormalities were easily recognized also in patients with normal lung function.

Chest CT has the substantial advantage of revealing early and/or localized structural changes [14, 28]. Nevertheless, despite its potential benefits, CT is not regularly used in PCD, and there is no consensus as to when it is best obtained [3]. However, although the outcome of PCD is not universally unfavourable, complications of bronchiectasis and severe pulmonary impairment become more severe with age [21]. Early assessment of structural lung changes is then crucial in PCD, especially in candidates for lobectomy and/or lung transplantation [29]. CT has been criticized for its ionizing radiation burden and the possible consequences of cumulative doses, particularly when used for frequent follow-up examinations in patients with chronic lung diseases, in pregnancy, or in children [6, 7]. Several surveys on the effects of repeated CT examinations have demonstrated that lifetime cancer risks are cumulative and not negligible, even though the consequences of low levels of exposure have not been clearly elucidated [30, 31]. It has been hypothesized that, at low doses, there is a linear dose-response relationship between the exposure to ionizing radiation and the development of solid cancers in humans [32]. It is unlikely that there is a threshold below which cancers are not induced, but at low doses the number of radiation-induced cancers might be small. In addition to this, other health end-points have been linked to radiation exposure. In particular, statistically significant associations have been found with heart diseases, stroke, and diseases of the digestive, respiratory, and hematopoietic systems [32]. In this scenario, a sensitive technique that could assess PCD lung disease without ionizing radiation is highly desirable.

To determine the clinical value of chest MRI scanning and its potential as a surrogate outcome measure in PCD, we assessed the comparability of MR and HRCT images using a quantitative scoring system, i.e. a modified Helbich score [16]. Although the Helbich score was devised for the CF setting [22], we used it because there is no scoring system specific for PCD and because we found it to be easily adaptable to MRI. Here we show that in PCD patients the results of chest MRI coincide with chest CT findings. Indeed, we found an excellent agreement between HRCT and MRI for most of the categories of the scoring system, thereby supporting the concept that MRI might be an alternative, radiation-free method for pulmonary assessment in PCD as well as in CF [15–17]. This issue is particularly relevant in PCD because of the presumed long life expectancy of these patients.

Chest MRI has several advantages over CT. First, being radiation-free, it can be repeated (e.g., in case of artifacts), and is thus suitable for long-term disease monitoring. Second, MRI identifies various characteristics of lung tissue, allows a precise characterization of the lesion and assesses function, e.g. lung perfusion and/or ventilation, and respiratory mechanics [14]. Thus, structural information and functional data are obtained in a single examination, thereby reducing imaging costs and increasing the patient's compliance [33, 34]. Finally, unlike CT, chest MRI easily distinguishes between mucous plugging and bronchial wall thickening even in the peripheral airways [15]. This finding is relevant because peripheral mucous plugging is a sensitive marker of early small airway disease in PCD [18]. None of these features is found (or is only partly found) in spirometry and other surrogate outcome measures, including CT.

On the other hand, the spatial resolution of MRI is lower than that of CT and slight morphological changes such as peripheral bronchiectasis without bronchial wall thickening are not consistently visualized by MRI [15]. However, this did not seem to affect our findings as shown by the comparable HRCT and MRI scores. Other potential drawbacks of MRI are limited access to the technology, long acquisition times and high costs. Long acquisition times represented a concern in our study, because our population included children who were unable to hold their breath for a long time, despite training. Therefore, we used respiratory-gating to overcome the lack of cooperation in children, and applied it also to adults for the sake of uniformity. We believe that the benefits of MRI outweigh the greater discomfort for the patient and the higher costs. Furthermore, technical progress will hopefully lead to lower costs, shorter examination times and higher image resolution.

To our knowledge, this is the first time that 3T MRI has been used for the quantitative analysis of pulmonary abnormalities in children and adults with PCD. Moreover, this study is the first to compare typical PCD findings at chest CT with MRI. It is also the first evaluation of chest high-field 3T morphological MRI in patients with lung disorders due to abnormal mucociliary clearance. Although a field strength of 1.5T is generally used in clinical practice, MR whole-body units operating at field strengths of 3T and beyond are increasingly being installed in research institutions and in clinical facilities [35–37]. The scanner we used has several advantages over 1.5T: 1) it has a high-speed and a high-strength gradient system; 2) it is equipped with multiple phased-array coils and receiver channels; and 3) it has acquisition acceleration techniques, such as parallel imaging. Parallel acquisition techniques improve image quality by shortening the echo times of single-shot sequences [38, 39]. Despite the high-field used, shorter echo times result in decreased blurring artifacts and less signal decay caused by T2* effects. Another advantage of the system we used is that cardiac- and respiratory-gating reduce artifacts due to heart/great vessels/chest wall motion thereby overcoming the need for sedation even in poorly cooperating subjects. Thus, patient's discomfort was less with this system. However, our study is limited by the fact that CT scans were performed for clinical purposes, images were read in consensus, functional MRI data were not provided, infants were not included, and longitudinal data were not obtained. In addition to this, MRI failed to detect bullae in two of our patients. In both of them these lesions were small in size, and it has been previously reported that small pulmonary bullae might pose a problem to chest MRI [40]. However, the prevalence of bullae in our study population was very low, and other authors failed to detect them at HRCT in larger PCD populations [19, 20]. Given the very low prevalence of bullae in PCD and the small size of the lesions found in our patients, the failure in the detection of bullae unlikely entails a significant clinical impact. Nevertheless, further studies on chest MRI in larger PCD populations are needed.

Chest high-field 3T MRI appears to be as effective as HRCT in the assessment of extent and severity of PCD lung disease. Therefore, it might be considered a reliable radiation-free option to HRCT in PCD and be proposed for follow-up examinations. This non-ionizing radiation technique could prove to be a useful tool in monitoring lung disease, guiding clinical management and assessing the consequences of therapy in PCD as well as in other pulmonary disorders in children, adults and the elderly.