Limits...
Comparison of image quality from filtered back projection, statistical iterative reconstruction, and model-based iterative reconstruction algorithms in abdominal computed tomography.

Kuo Y, Lin YY, Lee RC, Lin CJ, Chiou YY, Guo WY - Medicine (Baltimore) (2016)

Bottom Line: Subjective image quality was evaluated and analyzed using Wilcoxon signed-rank test with Bonferroni correction.Objective analysis revealed a reduction in image noise for statistical IR compared with that for FBP, with no significant differences in SNR.In the urinary bladder group, IMR achieved up to 53.7% noise reduction, demonstrating a superior performance to that of statistical IR.IMR also yielded a significantly superior SNR to that of statistical IR.

View Article: PubMed Central - PubMed

Affiliation: aDepartment of radiology, Taipei Veterans General Hospital bSchool of Medicine, National Yang-Ming University, Taipei, Taiwan, R.O.C.

ABSTRACT
The purpose of this study was to compare the image noise-reducing abilities of iterative model reconstruction (IMR) with those of traditional filtered back projection (FBP) and statistical iterative reconstruction (IR) in abdominal computed tomography (CT) imagesThis institutional review board-approved retrospective study enrolled 103 patients; informed consent was waived. Urinary bladder (n = 83) and renal cysts (n = 44) were used as targets for evaluating imaging quality. Raw data were retrospectively reconstructed using FBP, statistical IR, and IMR. Objective image noise and signal-to-noise ratio (SNR) were calculated and analyzed using one-way analysis of variance. Subjective image quality was evaluated and analyzed using Wilcoxon signed-rank test with Bonferroni correction.Objective analysis revealed a reduction in image noise for statistical IR compared with that for FBP, with no significant differences in SNR. In the urinary bladder group, IMR achieved up to 53.7% noise reduction, demonstrating a superior performance to that of statistical IR. IMR also yielded a significantly superior SNR to that of statistical IR. Similar results were obtained in the cyst group. Subjective analysis revealed reduced image noise for IMR, without inferior margin delineation or diagnostic confidence.IMR reduced noise and increased SNR to greater degrees than did FBP and statistical IR. Applying the IMR technique to abdominal CT imaging has potential for reducing the radiation dose without sacrificing imaging quality.

No MeSH data available.


Related in: MedlinePlus

Comparison of abdominal computed tomography image quality focusing on the urinary bladder of a 63-year-old male patient. (A) FBP; (B) IMR-R-L1; (C) IMR-R-L2; (D) IMR-R-L3; (E) iDose4; (F) IMR-ST-L1; (G) IMR-ST-L2; and (H) IMR-ST-L3. Noise granules were evident in the FBP (A) and iDose4 (E) reconstruction algorithms. Noise was reduced after applying IMR (B–H), yielding images that were more homogenous. FBP = filtered back projection, IMR = iterative model reconstruction, L = level, R = routine, ST = soft tissue.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4979832&req=5

Figure 2: Comparison of abdominal computed tomography image quality focusing on the urinary bladder of a 63-year-old male patient. (A) FBP; (B) IMR-R-L1; (C) IMR-R-L2; (D) IMR-R-L3; (E) iDose4; (F) IMR-ST-L1; (G) IMR-ST-L2; and (H) IMR-ST-L3. Noise granules were evident in the FBP (A) and iDose4 (E) reconstruction algorithms. Noise was reduced after applying IMR (B–H), yielding images that were more homogenous. FBP = filtered back projection, IMR = iterative model reconstruction, L = level, R = routine, ST = soft tissue.

Mentions: For the bladder group (Fig. 2), FBP demonstrated the highest mean image noise (16.5 ± 2.7), whereas IMR-ST-L3 demonstrated the lowest mean image noise (5.3 ± 1.3). FBP image noise was significantly (P < 0.001) higher than that of iDose4 (11.5 ± 1.9) and all IMR algorithms (P < 0.001). Furthermore, the image noise of iDose4 was significantly higher than that of all IMR algorithms (P < 0.001). iDose4, IMR-R-L1 (9.2 ± 1.5), and IMR-ST-L3 (5.3 ± 1.3) demonstrated 30.5%, 44.5%, and 67.8% noise reduction, respectively, evidencing higher performance than that of the FBP algorithm. IMR-R-L1 and IMR-ST-L3 respectively achieved 20.1% and 53.7% noise reduction, evidencing higher performance than that of iDose4. All IMR-R algorithms demonstrated a significantly (P ≤ 0.001) higher image noise than did the IMR-ST group. However, there was no significant difference in image noise between IMR-ST-L1 and IMR-ST-L2 or between IMR-ST-L2 and IMR-ST-L3.


Comparison of image quality from filtered back projection, statistical iterative reconstruction, and model-based iterative reconstruction algorithms in abdominal computed tomography.

Kuo Y, Lin YY, Lee RC, Lin CJ, Chiou YY, Guo WY - Medicine (Baltimore) (2016)

Comparison of abdominal computed tomography image quality focusing on the urinary bladder of a 63-year-old male patient. (A) FBP; (B) IMR-R-L1; (C) IMR-R-L2; (D) IMR-R-L3; (E) iDose4; (F) IMR-ST-L1; (G) IMR-ST-L2; and (H) IMR-ST-L3. Noise granules were evident in the FBP (A) and iDose4 (E) reconstruction algorithms. Noise was reduced after applying IMR (B–H), yielding images that were more homogenous. FBP = filtered back projection, IMR = iterative model reconstruction, L = level, R = routine, ST = soft tissue.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4979832&req=5

Figure 2: Comparison of abdominal computed tomography image quality focusing on the urinary bladder of a 63-year-old male patient. (A) FBP; (B) IMR-R-L1; (C) IMR-R-L2; (D) IMR-R-L3; (E) iDose4; (F) IMR-ST-L1; (G) IMR-ST-L2; and (H) IMR-ST-L3. Noise granules were evident in the FBP (A) and iDose4 (E) reconstruction algorithms. Noise was reduced after applying IMR (B–H), yielding images that were more homogenous. FBP = filtered back projection, IMR = iterative model reconstruction, L = level, R = routine, ST = soft tissue.
Mentions: For the bladder group (Fig. 2), FBP demonstrated the highest mean image noise (16.5 ± 2.7), whereas IMR-ST-L3 demonstrated the lowest mean image noise (5.3 ± 1.3). FBP image noise was significantly (P < 0.001) higher than that of iDose4 (11.5 ± 1.9) and all IMR algorithms (P < 0.001). Furthermore, the image noise of iDose4 was significantly higher than that of all IMR algorithms (P < 0.001). iDose4, IMR-R-L1 (9.2 ± 1.5), and IMR-ST-L3 (5.3 ± 1.3) demonstrated 30.5%, 44.5%, and 67.8% noise reduction, respectively, evidencing higher performance than that of the FBP algorithm. IMR-R-L1 and IMR-ST-L3 respectively achieved 20.1% and 53.7% noise reduction, evidencing higher performance than that of iDose4. All IMR-R algorithms demonstrated a significantly (P ≤ 0.001) higher image noise than did the IMR-ST group. However, there was no significant difference in image noise between IMR-ST-L1 and IMR-ST-L2 or between IMR-ST-L2 and IMR-ST-L3.

Bottom Line: Subjective image quality was evaluated and analyzed using Wilcoxon signed-rank test with Bonferroni correction.Objective analysis revealed a reduction in image noise for statistical IR compared with that for FBP, with no significant differences in SNR.In the urinary bladder group, IMR achieved up to 53.7% noise reduction, demonstrating a superior performance to that of statistical IR.IMR also yielded a significantly superior SNR to that of statistical IR.

View Article: PubMed Central - PubMed

Affiliation: aDepartment of radiology, Taipei Veterans General Hospital bSchool of Medicine, National Yang-Ming University, Taipei, Taiwan, R.O.C.

ABSTRACT
The purpose of this study was to compare the image noise-reducing abilities of iterative model reconstruction (IMR) with those of traditional filtered back projection (FBP) and statistical iterative reconstruction (IR) in abdominal computed tomography (CT) imagesThis institutional review board-approved retrospective study enrolled 103 patients; informed consent was waived. Urinary bladder (n = 83) and renal cysts (n = 44) were used as targets for evaluating imaging quality. Raw data were retrospectively reconstructed using FBP, statistical IR, and IMR. Objective image noise and signal-to-noise ratio (SNR) were calculated and analyzed using one-way analysis of variance. Subjective image quality was evaluated and analyzed using Wilcoxon signed-rank test with Bonferroni correction.Objective analysis revealed a reduction in image noise for statistical IR compared with that for FBP, with no significant differences in SNR. In the urinary bladder group, IMR achieved up to 53.7% noise reduction, demonstrating a superior performance to that of statistical IR. IMR also yielded a significantly superior SNR to that of statistical IR. Similar results were obtained in the cyst group. Subjective analysis revealed reduced image noise for IMR, without inferior margin delineation or diagnostic confidence.IMR reduced noise and increased SNR to greater degrees than did FBP and statistical IR. Applying the IMR technique to abdominal CT imaging has potential for reducing the radiation dose without sacrificing imaging quality.

No MeSH data available.


Related in: MedlinePlus