Application of novel iterative reconstruction algorithms (IRA) and noise reduction filters (NRF) for reducing CT radiation dose

Background: Worldwide there have been concerns about increased risk of cancer with radiation dose from CT scanning. Reduction of radiation dose from CT will also decrease the risk of radiation induced cancer. Therefore, several techniques such as automatic exposure control techniques, and bow tie filters have been developed and assessed to reduce radiation dose with CT. Despite these developments, the radiation dose with CT scanning continues to increase each year as number of CT examinations performed each year keeps on increasing. Purpose and scientific questions: The aim of our study is to acquire low radiation dose CT data and assess if iterative reconstruction algorithms (IRA) and advanced noise reduction filters (ANRF) techniques can help improve image quality and acceptability of low dose CT images. Most important variables: CMIV has recently obtained access to novel IRA and NRF techniques for improving image quality of low dose CT images. Dose reduced CT images typically have higher noise and lower signal to noise ratio. We believe that IRA and ANRF, which work in different data domains to improve image quality and enable acquisition of low radiation dose CT. After acquisition of CT image data at different lower dose levels, we will independently process the data with IRA and ANRF and see if there is an improvement in the image quality with these techniques. The intent will be to see if variables such as image noise, artifacts, image contrast, sharpness as well as lesion conspicuity on low dose post processed images are similar to unprocessed higher dose images. In addition, quantitative measures of image quality such as quantitative image noise, and contrast to noise ratio will be performed. Advances in Knowledge and significance: This study will give information on use of IRA and ANRF for reducing radiation dose to patients undergoing CT scanning and quantify need and advantage of IRA and ANRF over unprocessed CT images reconstructed using conventional filtered back projection technique. If found useful, these techniques will help cut the radiation dose without sacrificing image quality, a result that may help save radiation dose from CT scanning.

• Staff:

Mannudeep Kalra , MD. PhD				Mass.general Hosp Boston
Michael Sandborg , phD				IMH/LIU

• Former Staff:

• Project Description:

1. Can novel ANRF technique make image quality of low radiation dose CT images similar to higher dose unprocessed filtered images?
2. Can IRA technique make image quality of low radiation dose CT images similar to higher dose filtered back projection (FBP) images?

In the present study, we will obtain additional CT images on subjects undergoing CT scans for a clinical reason as recommended by their physicians. Therefore, in our study, only CT images will be used. We will personally acquire the data in a manner described in the appendix of this application so as to ensure the reliability of acquired data. These images will be coded to remove all patient identifying information. The images will be then used for two independently used for two projects, one to assess use of IRA and other for assessing role of ANRF in dose reducing. The former technique is available presently at the CMIV so we will post process the CT source data to reconstruct images with IRA. Following image reconstruction, we will compare the standard and low dose FBP images with low dose IRA images. The latter technique of ANRF is not available presently at the CMIV, so low dose DICOM images will be given over to SharpView CT, Inc, Linkoping, Sweden, for improving image quality using ANRF technique. Following post processing, unprocessed FBP and ANRF processed FBP images will be compared for image quality and lesion characteristics.