Ranish Deedar Ali Khawaja MD, Presenter: Nothing to Disclose

Michael Austin Blake MBBCh, Abstract Co-Author: Editor with royalties, Springer Science+Business Media Deutschland GmbH

Mukesh Gobind Harisinghani MD, Abstract Co-Author: Nothing to Disclose

Garry Choy MD, MS, Abstract Co-Author: Nothing to Disclose

Ali Devrim Karaosmanoglu MD, Abstract Co-Author: Nothing to Disclose

Mannudeep K. S. Kalra MD, Abstract Co-Author: Nothing to Disclose

Sarabjeet Singh MD, Abstract Co-Author: Research Grant, Siemens AG Research Grant, Toshiba Corporation Research Grant, General Electric Company Research Grant, Koninklijke Philips NV

Atul Padole MD, Abstract Co-Author: Nothing to Disclose

Sarvenaz Pourjabbar MD, Abstract Co-Author: Nothing to Disclose

Thomas Koehler PhD, Abstract Co-Author: Employee, Koninklijke Philips NV

To assess substantial dose reduction in abdominal MDCT using of a knowledge-based Iterative Model Reconstruction (IMR) technique compared with hybrid-based iDose and traditional filtered back projection (FBP) technique.

This IRB-approved prospective study included 41 patients (62 ± 12 years; BMI 28 ± 5 kg/m2) who underwent ultra-low dose (ULD) CT immediately after their standard-of-care (SD) CT on 256 MDCT (iCT,Philips Healthcare). Size-specific dose estimates for SD and SubmSv CT were 10 ±3 mGy (~6 mSv) and 1.5 ±0.4 mGy (~0.9 mSv), respectively. SD CT were reconstructed using filtered back projection (FBP), whereas ULD CT were with FBP, IMR and iDose. Four radiologists assessed subjective image quality independently, using 5-point scale (1=supraclinical;5=unacceptable). Lesions (true, pseudo or missed) were detected on ULD-FBP and compared to SD-FBP "reference-standard". Objective noise and CT numbers of soft tissue structures were measured. Noise spectral density (NSD) curves to assess noise in frequency domain were obtained. Friedman’s test, ANOVA and intraclass correlation coefficient were used for data analysis.

All true lesions (n=52) on SD-FBP were detected on ULD images. There were no missed or pseudo-lesions on ULD images. Mean intraclass correlation was 0.7. ULD-FBP was deemed unacceptable for subjective quality. Subjective ratings showed higher image quality for IMR for liver margins, soft-tissue structures, and retroperitoneal lymphadenopathy, compared to iDose in patients with a BMI ≤25kg/m2. For patients with BMI >26kg/m2, ULD IMR outperformed FBP and iDose for subjective ratings. Irrespective of patient BMI, subjective ratings for hepatic lesions, renal cysts, and colonic diverticula were significantly better with ULD IMR images. Objective noise for ULD FBP and iDose was 57-66% and 10-23% higher compared to SD-FBP, but 8-56% lower with ULD-IMR. NSD showed significantly lower noise in the frequency domain with IMR technique in all patients irrespective of BMI.  

Lesion detection is similar in standard-dose and ultra-low dose abdominal MDCT (~1.5 mGy). IMR considerably improved image quality compared to iDose and FBP with mean 85% dose reduction.

Knowledge-based Iterative Model Reconstruction technique enables substantial dose reduction in abdominal MDCT with uncompromised lesion detection compared to standard-of-care abdominal CT.

Khawaja, R, Blake, M, Harisinghani, M, Choy, G, Karaosmanoglu, A, Kalra, M, Singh, S, Padole, A, Pourjabbar, S, Koehler, T,  Knowledge-based Iterative Model Reconstruction Technique for Substantial Dose Reduction in Abdominal MDCT: Comparison with Hybrid and Traditional Filtered Back Projection in a Prospective Clinical Study.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14003248.html