Abstract

RATIONALE
As a result of respiratory muscle weakness, patients with Spinal Muscular Atrophy (SMA) face difficulties performing a good spirometry test based on the ATS/ERS 2019 standards. We aimed to evaluate if SMA patients can meet these criteria, and further explore if an AI-based software can correctly support the assessment of spirometry quality for these patients.


METHODS
Spirometry data of 51 SMA patients (138 spirometry curves) from the UK Biobank are analyzed by an AI-based software (ArtiQ.QC v1.4.0, ArtiQ, BE). The results are reviewed by a skilled spirometry operator. 


RESULTS
Patients can reach an acceptable FEV1 for 62% (N=86/138) of the curves and an acceptable FVC for 49% (N=67/138). The most common artefacts in the curves detected by ArtiQ.QC are a submaximal blast (N=10/138, 7%) and coughing (N=27/138, 20%). On session level, 55% of them have an acceptable FEV1 (grade A, B or C) and 47% have an acceptable grade for FVC. The review of results confirms that ArtiQ.QC can robustly assess the quality of curves of SMA patients, with an agreement of 98,6% (N=136/138) for FEV1 labels and 97,8% (N=135/138) for FVC labels. 


CONCLUSIONS
The observed prevalence of acceptable curves and sessions for these SMA patients is low. However, this is still similar to other studied populations, indicating that SMA does not prevent patients to meet the standards. This study also indicates that extra attention to coughing is useful when taking a spirometry measurement from SMA patients. Finally, it shows that ArtiQ.QC can accurately assess spirometry curves from SMA patients.