Abstract:
The most encountered cancer type among women is the breast cancer. Advanced clinical treatment applications are currently present, and the effectiveness of treatment is closely related to early diagnosis. Detectability of sentinel lymph nodes (SLNs) in breast assumes noteworthy role in order to predict breast cancer before metastasis. This study intended to scan SLNs in breast by using SPECT/CT and to enhance image quality of SLNs and find out optimal values of different parameters by using the design of experiments methods. The SIMIND Monte Carlo Simulation package and Design of Experiments were used to evaluate SLNs in breast. A Zubal torso phantom was modelled. Typical imaging conditions were utilized from published guidelines and literature. The OSEM algorithms were used for reconstruction. Seven parameters were specified as attenuation correction(+ ; -), number of projections (120; 90), collimator type (low- energy-high-resolution (LEHR); low-energy-general-purpose (LEGP)), energy window (20%; 15%), iteration number (20; 10), subset ( 15 ; 5) and count activity. An orthogonal array with 8 experimental runs was used. Contrast-tonoise ratios were calculated from lesion and background statistics. From the response table, the highest CNR value was found 9.63 with a combination of LEHR collimator type, attenuation correction, 20% energy window, 20 iterations, 15 subsets, 120 projections and higher count activity. According to results, collimator type, number of projections and attenuation correction had the highest effect. More research is needed to include other correction algorithms.|Keywords : sentinel lymph nodes, SPECT/CT, Monte Carlo simulation, breast.
