Preliminary Results
Qualitative assessment
Effect of tube load
For all the image criteria assessed,
when comparing 30% and 70% tube loads,
improved image quality was observed as a significant strong effect for the higher tube load regardless of MBIR strength (p<0.001) (Fig.1–5,
Table 2).
Similar results are observed with significant strong effect of higher tube load when comparing the two strengths of the algorithm and same slice thickness individually (Table 4).
Effect of iterative reconstruction algorithm
Using MBIR strength 5 instead of 3 resulted in significantly better image quality for criteria C3 kidney and proximal ureter (Fig. 3) and C4 lymph node<15mm diameter (Fig. 4) (p<0.01),
but significantly lower for liver parenchyma (Fig. 1) and overall image quality (Fig. 5).
For criterium C2 pancreas contour (Fig. 2),
there were marginal differences in image quality between strength 5 compared to strength 3 but these were not statistically significant.
Effect of slice thickness
When comparing slice thickness,
independent of the algorithm strength,
better image quality was observed for all criteria assessed in favor of slice thicknesses of 2mm and 3mm compared to 1mm (Table 2).
No significant differences in image quality were found for comparison between 3mm and 2mm slice thicknesses (Fig.1–5,
Table 2).
The potential dose reductions for increase in slice thickness from 1mm to 2mm ranged from 24% to 35% and 25% to 41% for increase in slice thickness from 1mm to 3mm (criteria C2 pancreas contour and C5 overall image quality respectively) (Table 3).
An individual comparison of the effect of MBIR strength 3 and 5 and slice thickness,
one at a time,
significant better image quality is observed for criteria C2–C4 (Fig.
2–4) in favor of MBIR strength 5 at slice thickness 1mm (p=0.001).
Criteria C1 and C5 (Fig. 1,
Fig. 5),
however,
were significantly inferior (p=0.001,
p=0.05) for all slice thickness comparisons.
For slice thicknesses 2mm and 3mm, image quality for MBIR 5 was significantly better for criterium C4 (Fig. 4) at 2mm slice thickness (p=0.01) and for criterium C2 (Fig. 2) at 3mm slice thickness,
where it was found to be significantly inferior.
It can also be noted that with thicker slice thicknesses,
the effects of tube load,
and in most cases also reconstruction algorithm,
were weaker (regression coefficients closer to 0),
although still strongly significant (Table 4).
Quantitative assessment
Effect of tube load,
reconstruction algorithm and slice thickness
Noise (SD) decreases with increase in tube load and with increase in MBIR strength from 3 to 5 compared to Filtered Back Projection (FBP) (Fig. 6).
Consequently,
the CNR increases with increase in tube load and increase in MBIR strength 3 to 5 compared to FBP (Fig. 7).
Compared to FBP,
no change in spatial resolution is observed with increase in dose for all slice thicknesses of MBIR strengths 3 and 5.
Noise-Power Spectrum (NPS)
The noise power decreases with increasing strength of the MBIR compared to Filtered Back Projection (FBP) (Fig. 8),
when all other parameters remain constant.
The decrease in noise power is,
however,
more pronounced at the higher spatial frequencies when using MBIR strength 3 and particularly with strength 5 compared to FBP.
As expected,
the noise power decreases with increasing tube load and slice thickness. An example is depicted in Fig. 9 for MBIR strength 3 and slice thickness variation from 1mm to 2mm to 3mm.