With personalized chest CT protocols, radiation dose for certain clinical indications can be reduced without reducing interpretability, according to research conducted at University Hospital Erlangen in Germany.
"We found that personalized chest CT protocols lead to significant dose reduction," said presenter Markus Kopp, MD, consultant, University Hospital Erlangen. "This study provides a suggestion for optimized CT protocol for detecting pneumonia, especially with the COVID-19 pandemic, and for nodules."
The researchers first evaluated 10 tin filter-based low-dose CT protocols in a pre-clinical setting, on a human cadaver without known lung disease. Four protocols were optimized for high-resolution structures, pulmonary nodules, infectious disease and one minimum dose protocol with less but still sufficient quality to detect lesions, were selected for clinical application.
Radiation dose and image quality for the four protocols — high-resolution CT (HR-CT), pulmonary nodule CT (PN-CT), infectious disease CT (ID-CT) and minimum dose CT (MIN-CT) — were prospectively evaluated using a volume scanner (32 x 0.7 mm z-coverage, 0.8 seconds rotation time, tin pre-filtration).
The volume CT dose index (CTDI) in the clinical setting were 5.4 milligray (mGy) for HR-CT, 1.2 mGy for PN-CT, and 0.6 mGy for ID-CT (P < 0.001). The 0.2 mGy difference between ID-CT and MIN-CT wasn't significant (P = 0.262). For the upper reference protocol, the dose settings of a preexisting high-resolution CT were used. The radiation dose for MIN-CT was comparable to a chest radiograph in two planes.
In the clinical setting, the researchers applied the corresponding CT protocol to 88 patients based on the primary clinical indication, including 38 HR-CT, 21 PN-CT, 20 ID-CT, and nine MIN-CT. Radiation dose, and subjective and objective image quality were compared.
"Which CT protocol was selected for which patient was based on the primary clinical indication and made by the radiologist in charge," Dr. Kopp said. "For example, a patient with suspected infectious disease would receive the infectious disease protocol. However, because a patient may have multiple conditions — a patient with suspected pulmonary fibrosis, for example, may also have infectious disease, or vice versa — most of the CT images were evaluated for several lung lesions and patterns."
The reduction in radiation dose in the clinical setting ranged from 50% between the PN-CT and ID-CT protocols, to a maximum reduction of 96% between the HR-CT and MIN-CT protocols. The reduction was 78% between HR-CT and PN-CT, and 67% between ID-CT and MIN-CT.
"This is a really substantial and also significant dose reduction between high-resolution CT and pulmonary nodule CT, and pulmonary nodule CT and infectious disease CT," Dr. Kopp said.
Image Quality Informs Dose Recommendations
Optimal image quality was defined as a Likert scale rating of 4 (good delineation) or 5 (perfect delineation). A Likert score of 3 was still diagnostic but limited.
"In all 20 examinations with ID-CT, image quality was sufficient to accurately characterize the infectious disease patterns," Dr. Kopp said. "In our opinion, there is no reason we should provide more radiation dose than necessary."
Appropriate, personalized radiation dose while maintaining image clarity is important for patient safety and to prevent repeat imaging.
Some pulmonary nodules detected via ID-CT scored an unacceptable 3 or 3.5 on the Likert scale for ID-CT.
"We don't want to miss any relevant findings," Dr. Kopp said. "Our study shows that radiologists should use slightly higher dose levels for dedicated examination of pulmonary nodules."
With MIN-CT, detection of pulmonary nodules and infectious diseases was still possible, but the characterization of the lung lesions became more difficult.
Access the presentation, "Personalized Chest CT: Optimal Examination Protocols for the Detection of Fibrosis, Nodules and Pneumonia," (SSCH06) on demand at Meeting.RSNA.org.