The researchers investigated the foetal dose for 18 pregnant patients who had previously undergone clinically-indicated CT scans. Such scans are sometimes necessary for pregnant women who have been in an accident, or a CT scan may have been performed before the patient knew that she was pregnant.

"We found a relationship between the radiation dose to the foetus and other values that are available on the CT console," UCLA medical physicist Michael McNitt-Gray reported at a press conference at last week's AAPM Annual Meeting. The values in question are the volume CT dose index (CTDIvol) and the patient size expressed as a water equivalent diameter (WED). CTDIvol is reported on the CT scanner console, while WED can be measured on the image containing the 3D geometric centroid of the foetus.

"This means that a reasonably accurate dose to the foetus can be obtained by knowing the scanner output and the size of the patient," said McNitt-Gray. "This work improves upon previous methods of estimating radiation dose by taking into account patient size and composition through the WED. This allows a better characterization of patient size and therefore a better estimate of the dose to the foetus."

"The information that we provide usually is not only comforting to the patient, it's comforting to the physician who is trying to comfort the patient," he added.

Dose modelling

McNitt-Gray and colleagues developed size-specific, scanner-independent CTDIvol-to-foetal dose conversion coefficients for abdominal and pelvic CT exams of pregnant patients. The patients had gestational ages ranging from 12 to 36 weeks.

They used the CT images to create models of maternal and foetal anatomy and estimated foetal dose using Monte Carlo simulations of fixed tube current helical scans for four manufacturer's 64-slice multidetector CT scanners. The foetal doses were normalized by scanner-specific 32 cm CTDIvol values and averaged to obtain a scanner-independent conversion coefficient for each patient.

The researchers calculated the coefficients of variation (CoV) across the four scanners for the simulated foetal dose before and after CTDIvol normalization, and found CoV values ranging from 25.10–30.16% and 5.60–9.48%, respectively. They also observed an exponential relationship between the CTDIvol-to-foetal dose conversion coefficients and patient size, with a coefficient of determination of 0.84.

The results indicate that foetal dose from CT exams of pregnant patients can be accurately estimated using foetal dose normalized by CTDIvol to account for scanner variation and WED to account for patient size variation.

The researchers have not yet constructed and published details of the full equations. First, they plan to account for additional factors such as automatic exposure control (AEC), rather than the fixed tube current used in this study. "We are happy to put the equations out there after we've published our work," added McNitt-Gray.

Related stories

• Enhanced models expand simulation options
• CT dose is not always what it seems
• Foetal phantoms offer organ-level definition
• Pregnant patients: image with caution