The trouble with CTDI100

The computed tomography dose index (CTDI100) is typically measured using a 100 mm long pencil ion chamber with cylindrical polymethyl methacrylate (PMMA) dosimetry phantoms. While this metric was useful in the era of single slice CT scanners with collimated slice thicknesses of 10 mm or less, the efficiency of this metric in multi-slice CT scanners with wide (40 mm) collimated x-ray beams is unknown. Monte Carlo simulations were used to assess the efficiency of the CTDI100 parameter for wider beam collimations. The simulations utilized the geometry of a commercially available CT scanner, with modeled polyenergetic x-ray spectra. Dose spread functions (DSFs) were computed along the length of 12.4 mm diam rods placed at several radii in infinitely long 160 mm diam (head) and 320 mm diam (body) PMMA phantoms. The DSFs were used to compute radiation dose profiles for slice thicknesses from 1 to 400 mm. CTDI100 efficiency was defined as the fraction of the dose along a PMMA rod collected in a 100 mm length centered on the CT slice position, divided by the total dose deposited along an infinitely long PMMA rod. For a 10 mm slice thickness, a 120 kVp x-ray spectrum, and the PMMA head phantom, the efficiency of the CTDI100 was 82% and 90% for the center and peripheral holes, respectively. The corresponding efficiency values for the body phantom were 63% and 88%. These values are reduced by only 1% when a 40 mm slice thickness was studied, so the use of CTDI100 for 40 mm wide x-ray beams is no less valid than its use for 10 mm beam widths. However, these data illustrate that the efficiency of the CTDI100 measurement even with 10 mm beam widths is low and, consequently, dose computations which are derived from this metric may not be as accurate as desirable