Development of new gamma camera localization method for sentinel nodes by image fusion and navigation in lymphoscintigraphy

Full text: The objective of this study was the development of the localization technique for skin marking the lesions with low accumulating Tc-99m labeled radiopharmaceuticals. Fusion of high count static planar scan (base image) with real time acquired scan (fluent image) and with added moving point source (Tc-99m or Co-57) in the same patient position was performed for best overlapping the target lesions and point source spot. Special acquisition software in Windows (Oncology MedicView) was developed. Both images were pre-processed by online visual inspection and then fluently fused in the way that only point source spot is overlapped to the base image. Image normalization was done by linear, log or combined log/linear conversions followed by on-line contrasting of fused image by high sensitive color scaling and spatial contrast filtering. The localization was performed in several patient positions with fixed bed and patient. Navigation tools using audio and visual signals were continuously created from the 'lesion-point source spot' distance information. Localization accuracy for SLN(s) using this technique was considerably high. SLN detection improved from 76 % (95 patients, 72 detected SLNs, 23 undetectable SLNs) to 95 % (45 pts, 42 detected SLNs, 3 undetectable SLNs). Localization procedure was shortened for at least 3 times. The new technique substantially lowered the localization time and increased the lesion detection by on-line interactive optimization of fused images. It can be used for all radioisotope localizations in oncology diagnostics using simple analog or semi digital gamma cameras connected to low-cost IAEA acquisition module and specially developed acquisition/processing software. (author)