Effect of different reconstruction algorithms on computer-aided diagnosis (CAD) performance in ultra-low dose CT colonography

Coronary computed tomography angiography using model-based iterative reconstruction algorithms in the detection of significant coronary stenosis: how the plaque type influences the diagnostic performance

Purpose

To evaluate the ability of coronary computed tomography angiography (CCTA) with model-based iterative reconstruction (MBIR) algorithm in detecting significant coronary artery stenosis compared with invasive coronary angiography (ICA).

Material and methods

We retrospectively identified 55 patients who underwent CCTA using the MBIR algorithm with evidence of at least one significant stenosis (≥ 50%) and an ICA within three months. Patients were stratified based on calcium score; stenoses were classified by type and by coronary segment involved. Dose-length-product was compared with the literature data obtained with previous reconstruction algorithms. Coronary artery stenosis was estimated on ICAs based on a qualitative method.

Results

CCTA data were confirmed by ICA in 89% of subjects, and in 73% and 94% of patients with CS < 400 and ≥ 400, respectively. ICA confirmed 81% of calcific stenoses, 91% of mixed, and 67% of soft plaques. Both the dose exposure of patients with prospective acquisition (34) and the exposure of the whole population were significantly lower than the standard of reference (p < 0.001 and p = 0.007).

Conclusions

CCTA with MBIR is valuable in detecting significant coronary artery stenosis with a solid reduction of radiation dose. Diagnostic performance was influenced by plaque composition, being lower compared with ICA for patients with lower CAC score and soft plaques; the visualisation of an intraluminal hypodensity could cause false positives, particularly in D1 and MO segments.

Low-Dose Computed Tomography Colonography Technique

Significant anxiety has been expressed by some over the radiation risks associated with computed tomography (CT), particularly when it applies to a screening examination such as CT colonography. These theoretic risks are far outweighed by the significant benefits colorectal cancer screening offers. Regardless of how significant the theoretic risk of CT radiation is in the older population, the ALARA principle maintains that radiation dose should be reduced to As Low As Reasonably Achievable. This article will discuss various strategies that may be utilized to reduce radiation dose and mitigate any increase in image noise that may occur.

Effects of Iterative Reconstruction Algorithms on Computer-assisted Detection (CAD) Software for Lung Nodules in Ultra-low-dose CT for Lung Cancer Screening

RATIONALE AND OBJECTIVES

This study aimed to evaluate the effects of iterative reconstruction (IR) algorithms on computer-assisted detection (CAD) software for lung nodules in ultra-low-dose computed tomography (ULD-CT) for lung cancer screening.

MATERIALS AND METHODS

We selected 85 subjects who underwent both a low-dose CT (LD-CT) scan and an additional ULD-CT scan in our lung cancer screening program for high-risk populations. The LD-CT scans were reconstructed with filtered back projection (FBP; LD-FBP). The ULD-CT scans were reconstructed with FBP (ULD-FBP), adaptive iterative dose reduction 3D (AIDR 3D; ULD-AIDR 3D), and forward projected model-based IR solution (FIRST; ULD-FIRST). CAD software for lung nodules was applied to each image dataset, and the performance of the CAD software was compared among the different IR algorithms.

RESULTS

The mean volume CT dose indexes were 3.02 mGy (LD-CT) and 0.30 mGy (ULD-CT). For overall nodules, the sensitivities of CAD software at 3.0 false positives per case were 78.7% (LD-FBP), 9.3% (ULD-FBP), 69.4% (ULD-AIDR 3D), and 77.8% (ULD-FIRST). Statistical analysis showed that the sensitivities of ULD-AIDR 3D and ULD-FIRST were significantly higher than that of ULD-FBP (P < .001).

CONCLUSIONS

The performance of CAD software in ULD-CT was improved by using IR algorithms. In particular, the performance of CAD in ULD-FIRST was almost equivalent to that in LD-FBP.

One-mSv CT colonography: Effect of different iterative reconstruction algorithms on radiologists' performance

PURPOSE

To analyze the effect of different reconstruction algorithms on image noise and radiologists' performance at ultra-low dose CT colonography (CTC) in human subjects.

MATERIALS AND METHODS

This retrospective study had institutional review board approval, with waiver of the need to obtain informed consent. CTC and subsequent colonoscopy were performed at the same day in 28 patients. CTC was scanned at the supine/prone positions using 120/100kVp and fixed 10mAs, and reconstructed using filtered back projection (FBP), adaptive statistical iterative reconstruction (ASIR), and model-based IR (Veo) algorithms. Size-specific dose estimates (SSDE) and effective radiation doses were recorded. Image noise was compared among the three datasets using repeated measures analysis of variance (ANOVA). Per-polyp sensitivity and figure-of-merits were compared among the datasets using the McNemar test and jackknife alternative free-response receiver operating characteristic (JAFROC) analysis, respectively, by one novice and one expert reviewer in CTC.

RESULTS

Mean SSDE and effective radiation dose of CTC were 1.732mGy and 1.002mSv, respectively. Mean image noise at supine/prone position datasets was significantly lowest with Veo (17.2/13.3), followed by ASIR (52.4/38.9) and FBP (69.9/50.8) (P<0.0001). Forty-two polyps in 25 patients were reference polyps. For both readers, per-polyp sensitivity of all 42 polyps was highest with Veo reconstruction (81.0%, 64.3%), followed by ASIR (73.8%, 54.8%) and FBP (57.1%, 50.0%) with statistical significance between Veo and FBP for reader 1 (P=0.002). JAFROC analysis revealed that the figure-of-merit for the detection of polyps was highest with Veo (0.917, 0.786), followed by ASIR (0.881, 0.750) and FBP (0.750, 0.746) with statistical significances between Veo or ASIR and FBP for reader 1 (P<0.05).

CONCLUSION

One-mSv CTC was not feasible using the standard FBP algorithm. However, diagnostic performance expressed as per-polyp sensitivity and figures-of-merit can be improved with the application of IR algorithms, particularly Veo.