Multivariate modeling using quantitative CT metrics may improve accuracy of diagnosis of bronchiolitis obliterans syndrome after lung transplantation
Introduction
Lung transplantation (LTx) is an established treatment for end-stage, irreversible pulmonary disease, most commonly due to chronic obstructive pulmonary disease (COPD) and interstitial lung disease (ILD). While continued improvements in surgical techniques and immunosuppressive medications have reduced the complication rates and increased short-term survival after the procedure, chronic allograft rejection (CLAD), mostly due to bronchiolitis obliterans (a fibrous obliterative disease of bronchioles), results in obstructive pulmonary physiology and remains the major cause of morbidity and mortality after six months following LTx. It is the greatest limitation to long-term survival after LTx [1], [2], [3], [4], [5], [6]. The patchy distribution of disease, with focal areas of abnormality surrounded by normal lung, may result in failure to establish the diagnosis, even with multiple transbronchial biopsies obtained via bronchoscopy. More invasive surgical lung biopsy may provide more tissue for analysis, but often is not feasible in a patient with abnormal or declining lung function. Consequently, a clinical diagnosis based on irreversible, sustained decline in spirometry after ruling out other causes has been advocated [3], [4], [5], termed bronchiolitis obliterans syndrome (BOS). BOS serves as a surrogate marker of this phenotype of chronic allograft rejection. In clinical practice, the diagnosis of BOS is raised when there is an unexplained and sustained decline in lung function measured by pulmonary function testing (PFT), of greater than 10% of baseline post-transplant forced expiratory volume in the first second (FEV1) and worsening respiratory symptoms (such as cough and dyspnea), in the absence of other causes such as pulmonary infection or congestive heart failure [5] – defined as BOS grade 0-p. BOS is the most common and classic form of CLAD. Other forms of CLAD include restrictive lung allograft syndrome (RAS), which was not evaluated in this study. Multidetector CT (MDCT) plays an important role by excluding alternative diagnoses and demonstrating low attenuation areas representing air trapping, particularly on expiratory images, which may correlate with bronchiolitis obliterans. Prior studies reported limited sensitivity for early diagnosis of bronchiolitis obliterans; however, these utilized semi-quantitative assessment of air trapping in non-volumetric data sets [7], [8], [9], [10], [11]. More recent studies have assessed the role of parametric response maps and density maps and correlated these quantitative metrics with progression of BOS and mortality [12], [13], [14].
There is emerging recognition of the need for quantitative CT (qCT) imaging to move beyond mere subjective interpretation, noting recent interest in applying this strategy to BOS [15], [16], [17] and COPD [18], [19], [20], [21], [22], [23], [24], [25]. Our primary hypothesis is that multivariate modeling of qCT metrics obtained from volumetric paired MDCT datasets acquired in full inspiration and end expiration may allow better prediction of BOS status in lung transplant recipients, when compared to PFT and to subjective semi-quantitative assessment of CT features associated with BOS. Prediction of BOS via qCT metrics may offer imaging biomarkers that may be more sensitive to detect onset and progression of BOS than PFT, potentially offering a superior method to test novel BOS therapies, with the aim of improving post LTx survival and quality of life.
Section snippets
Materials and methods
Approval from our Institutional Review Board and a Health Insurance Portability and Accountability Act waiver were obtained prior to study initiation.
Results
The median age of the 176 patients was 61 (range 20–72), and 115 (64.6%) were males. Among them, we found COPD in 63 (35.4%) patients based on Pre LTx diagnosis, UIP/IPF in 57 (32.0%), Cystic Fibrosis in 19 (10.7%), α1-AT deficiency in 11 (6.2%), sarcoidosis in 11 (6.2%), and other in 17 (9.6%) patients (other included unclassifiable ILD, NSIP, silicosis, berylliosis and pulmonary hypertension). 95 patients underwent bilateral LTx (53.4%) and 83 underwent unilateral LTx. 43 (24.1%) of the 83
Discussion
Prior studies have compared how well several SQS of pulmonary abnormalities on CT correlate with the presence of BOS and associated PFT abnormalities. The most common conclusion [7], [9], [11] was that air trapping detected on expiratory imaging is the principal MDCT finding of BOS and more strongly correlated with PFT metrics such as FEV1. Nonetheless, most studies demonstrated limited sensitivity and specificity of SQS derived from CT images to predict BOS status [8], [11]. Moreover, most
Contributions of each author
EB designed the study, collected all CT data, reviewed all CT images, performed statistical analysis and wrote the manuscript. JL collected clinical data and reviewed the manuscript. SS reviewed CT images. HS revised statistical analysis and reviewed the manuscript. NT and JG performed quantitative image analysis.
Acknowledgements
Rupal Shah, MD, Gang Song, PhD, John Wu, PhD.
References (36)
- et al.
Parametric response mapping of bronchiolitis obliterans syndrome progression after lung transplantation
Am. J. Transpl.
(2016 Nov) - et al.
Paired inspiratory/expiratory volumetric thin-slice CT scan for emphysema analysis: comparison of different quantitative evaluations and pulmonary function test
Chest
(2005) - et al.
Quantitative CT evaluation in patients with combined pulmonary fibrosis and emphysema: correlation with pulmonary function
Acad. Radiol.
(2015) - et al.
A comparative study of HRCT image metrics and PFT values for characterization of ILD and COPD
Acad. Radiol.
(2012 Jul) - et al.
Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration
Neuroimage
(2009) - et al.
Deaths: final data for 2007
Natl. Vital Stat. Rep.
(2010) - et al.
Bronchiolitis obliterans organizing pneumonia: pathogenesis, clinical features, imaging and therapy review
Ann. Thorac. Med.
(2008) - et al.
Obliterative bronchiolitis
N. Engl. J. Med.
(2014) - et al.
Chronic lung allograft dysfunction: evolving practice
Curr. Opin. Organ Transpl.
(2015) - et al.
An international ISHLT/ATS/ERS clinical practice guideline: summary for clinicians. Bronchiolitis obliterans syndrome complicating lung transplantation
Ann. Am. Thorac. Soc.
(2015)
Imaging of lung transplantation: review
AJR Am. J. Roentgenol.
Utility of high resolution computed tomography in predicting bronchiolitis obliterans syndrome following lung transplantation
J. Thorac. Imaging
Bronchiolitis obliterans syndrome in lung transplant recipients: correlation of computed tomography findings with bronchiolitis obliterans syndrome stage
J. Thorac. Imaging
Conventional high-resolution CT versus contiguous multidetector CT in the detection of bronchiolitis obliterans syndrome in lung transplant recipients
J. Thorac. Imaging
High-resolution CT findings of bronchiolitis obliterans syndrome after hematopoietic stem cell transplantation
J. Thorac. Imaging
Bronchiolitis obliterans syndrome in lung transplant recipients: can thin-section CT findings predict disease before its clinical appearance
Radiology
Parametric response mapping as an imaging biomarker in lung transplant recipients
Am. J. Respir. Crit. Care Med.
Quantification of pathologic air trapping in lung transplant patients using CT density mapping: comparison with other CT air trapping measures
PLoS One
Cited by (6)
Application of deep learning technique to manage COVID-19 in routine clinical practice using CT images: Results of 10 convolutional neural networks
2020, Computers in Biology and MedicineCitation Excerpt :Furthermore, CAD systems are a useful tool to monitor lung function after transplantation. In this regard, Barbosa et al. compared the quantitative CT (qCT), pulmonary function testing (PFT), and semi-quantitative image scores (SQS) metrics and found that pulmonary function testing and qCT metrics demonstrated the highest accuracy for monitoring bi- and unilateral lung transplantation with an AUC of 0.771 and 0.817, respectively [19]. The potential of CAD systems can go even further and help radiologists to detect lung abnormalities by computerized analysis of pulmonary sounds [20] and flow and gas concentration of breath [21].
Machine Learning Algorithms Utilizing Quantitative CT Features May Predict Eventual Onset of Bronchiolitis Obliterans Syndrome After Lung Transplantation
2018, Academic RadiologyCitation Excerpt :Multiple studies have suggested that HRCT could be useful in the diagnosis of BOS after LTx; however, its clinical value has not yet been proven because of inconsistent findings (7,16,17). Nonetheless, two more recent publications demonstrate promising results in using advanced quantitative analysis of HRCT in LTx patients to improve the accuracy of diagnosis of BOS (18) and in predicting functional parameters such as FEV1 from qCT metrics specifically obtained from the transplanted lung in unilateral LTx (19). In our study, CT scans obtained after LTx were quantitatively analyzed using FRI to investigate how multiple quantitative imaging parameters correlate with the onset of BOS and to differentiate the mechanism accounting for FEV1 decline, whether related to bronchiolitis obliterans or a reversible cause, such as a pulmonary infection or pulmonary edema.
Machine learning-based techniques to improve lung transplantation outcomes and complications: a systematic review
2022, BMC Medical Research MethodologyScene merging technology with high adaptability
2021, Advances in Intelligent Systems and Computing