The long-term objectives of this research are to establish optimum methods for the accurate calculation of tumor radiation dose from 131-I through imaging and to obtain a better correspondence between tumor radiation dose and response to therapy in 131-I anti-B1 radioimmunotherapy of patients. Trials with this therapy show that for salvage-therapy patients there is a 71 percent response rate and a 34 percent complete response rate and that for previously-untreated patients there is a 100 percent response rate and 71 percent complete response (CR) rate. For previously-untreated patients, the radiation dose for individual tumors was calculated based on CT-SPECT fusion, volumes of interest from CT, and activity quantification that employed empirical corrections for extra-tumor background and tumor size. To date, the radiation doses for 30 abdominal and pelvic tumors in 5 patients who achieved a partial response (PR) have been compared with the radiation doses for 56 tumors in 15 patients who achieved a CR. There is a statistically significant difference in log10 of the radiation dose for tumors of PR patients versus that for tumors of CR patients (p=0.04). However, values for dose predicting response are, at best, only 80 percent (positive) and 80 percent (negative). In this application, the specific aims are to: 1) Test the hypothesis that one can improve SPECT iterative reconstruction methods, and/or parallel-hole collimation, from present practice to produce activity estimates that have less bias and variance. 2) Compare SPECT versus conjugate views for the task of estimating tumor activity, augment the Computer Assisted Matrix Inversion (CAMI) method of conjugate views by a fusion of 3-D CT to 2-D projection, and evaluate the augmented CAMI method for tumor activity estimation even when the lesions overlap in the anterior-posterior projection. 3) Test CT to SPECT fusion by assessing its reproducibility and its ability to maximize counts in tumors. Also develop a fusion algorithms based on edge detection in CT and SPECT plus a joint maximum-likelihood estimation of both the edges and the Euler angles of the 3-D rotation that registers the objects. 4) Evaluate the effectiveness of the new reconstruction methods with patient image sets. We expect to obtain the means to provide a basis for the superiority of the labeled anti-B1 antibody with predose compared to the unlabeled antibody, to produce an early and reliable indicator of degree of response in anti-B1 therapy, and to ultimately predict which patients to treat with anti-B1. Applications of the methods to other therapies and to imaging with other multi-gamma-emission radionuclides such as 111-In are also possible.
