Image quality comparison between model-based iterative reconstruction and adaptive statistical iterative reconstruction chest computed tomography in cystic fibrosis patients.

INTRODUCTION

Cystic fibrosis (CF) predominantly affects young adults. Accurate radiological assessment of pulmonary disease is vital for predicting exacerbations, one of the leading causes of morbidity and mortality. We evaluated the image quality of model-based iterative reconstruction (MBIR) ultra-low-dose CT chest (ULD-CT) in CF evaluation.

METHODS

We compared ULD-CT with standard adaptive statistical iterative reconstruction (ASIR) low-dose CT (LD-CT). Subjective assessment of contrast and noise were performed for each study. Background noise, signal to noise ratio (SNR) and contrast to noise ratio (CNR) were calculated and compared between the CT studies. Conspicuity of major structures was assessed. These aspects of image quality were compared to determine whether ULD-CT was superior to LD-CT in assessment of CF.

RESULTS

The ULD-CT achieved median effective dose of 0.073 mSv, comparable to one standard chest radiograph and significantly lower than the median LD-CT dose of 1.22 mSv. ULD-CT had lower subjective contrast and higher subjective noise when compared to LD-CT. Objectively measured background noise was lower in ULD-CT (16.33 HU vs 38.53 HU, P < 0.0001) compared to LD-CT. ULD-CT had higher median CNR (52.65 vs 22.09, P < 0.0001) and SNR in lung (9.08 vs 7.29, P = 0.002) compared to LD-CT. ULD-CT was equal to LD-CT in identification of trachea, bronchi, pleural and pericardium. Interobserver reliability showed agreement of 80-92%.

CONCLUSIONS

The image quality of ULD-CT is similar to LD-CT, at 1/16th the dose. MBIR constructed ULD-CT is an effective imaging modality for CF surveillance, with potential applications in other disease settings.