Feasibility of contrast-enhanced CT with knowledge-based iterative model reconstruction algorithm for the evaluation of parotid gland masses

Abstract

Purpose: The purpose of this study was to assess the diagnostic utility of knowledge-based iterative model reconstruction (IMR) algorithm for CT evaluation of parotid gland masses and compare the finding with those of filtered back projection (FBP) and hybrid iterative reconstruction (iDose4) using the same CT datasets.

Materials and methods: Forty-two consecutive patients with palpable parotid masses who underwent contrast-enhanced CT were enrolled. The same data were reconstructed using FBP, iDose4, and knowledge-based IMR algorithms. Quantitative imaging parameters, including background noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR), were measured. Subjective assessment for noise and artifact, parotid mass delineation, and blotchy, pixelated appearance was also performed by two independent radiologists using a five-point grading system.

Results: The background noise was significantly lower with the IMR algorithm than with the other two algorithms (p<0.001), while the SNR and CNR were significantly higher with the former than with the latter (p<0.001). The IMR algorithms resulted in significantly lesser noise and artifacts (p<0.001) and better image quality for parotid mass delineation (p<0.001) compared with the other two algorithms, although it was assoctiated more frequently with a blotchy, pixelated appearance that would not or slightly affect the diagnosis (p < 0.001). Conclusion: CT with knowledge-based IMR provides excellent image quality and decreases image noise and artifacts. Therefore, it can be considered clinically feasible for the assessment of parotid gland mass.