Quantitative radioimmunoimaging for radioimmunotherapy treatment planning: effect of reduction in data sampling on dosimetric estimates.

Abstract

Quantitative radioimmunoimaging (serial anterior/posterior imaging and blood sampling) is useful for radioimmunotherapy treatment planning, but can be quite time consuming. To predict whether accurate radiation absorbed dose estimates can be maintained with a reduction in data sampling, 12 patients undergoing indium-111/yttrium-90 anti-CD20 monoclonal therapy for whom absorbed doses were estimated based on eight data samples (acquired at 0, 2, 4, 24, 48, 72, 96, and 144 h, respectively), were retrospectively reanalyzed using only five samples (0, 4, 24, 72, and 144 h, respectively). Calculated residence times (in h) and absorbed doses (in cGy), for the whole body, kidneys, liver, lungs, spleen, and red marrow were compared with the original values based on the eight samples using Student's paired t-test. Linear regression and Bland-Altman analysis of the two data sample groups was also performed. The mean residence times in the five- and eight-data samples groups were essentially the same (17.7 +/- 26.6 h [range, 0.3-79.0 h] versus 17.6 +/- 26.6 h [range, 0.3-79.5 h]; p = 0.72), as were the mean absorbed doses (336 +/- 411 cGy [range, 38-2434 cGy] versus 325 +/- 381 cGy [range, 39-2246 cGy]; p = 0.24). Also, the linear regressions were excellent (residence time y = 1.00x + 0.09 h [r = 0.99]; absorbed dose y = 1.06x - 7.74 cGy [r = 0.98]). Additionally, Bland-Altman analysis revealed no significant sample bias in residence time (0.03 +/- 0.68 h, 0.9% +/- 10.0) or absorbed dose (11 +/- 76 cGy, 1.0% +/- 9.3). These results demonstrate that reduced data sampling in quantitative radioimmunoimaging can be achieved without significantly altering radiation absorbed dose estimates, but with a significant savings in imaging, blood sampling, and processing time.