DESIGN AND ASSESSMENT OF A SCINTIGRAPHY-GUIDED BIPLANE LOCALIZATION TECHNIQUE FOR BREAST-TUMORS - A PHANTOM STUDY

The aim of this study was to design and assess a scintigraphy-guided s tereotaxic localization technique for breast phantoms. We wished to de velop and validate scintigraphically aided biopsy of non-palpable brea st masses identified only by sestamibi scintimammography. A biopsy tab le was built for patients to lie prone with the breast in a dependent position. The breast can be compressed in the cranial and caudal direc tions by fenestrated paddles. Three freely adjustable radioactive refe rence lines, placed along the x, y and z axes, each containing about 3 0 MBq (800 mu Ci) Tc-99(m), were mounted on sliding rules on the exter nal frame which surrounded the phantom. The breast phantom was a semi- square-shaped sponge. Background activity was provided by a sponge clo th containing 37 MBq (1 mCi) Tc-99(m) solution. Non-palpable lesions w ere carved out of the same sponge and wrapped in thin plastic film, an d labelled with about 11 MBq (300 mu Ci) Tl-201. The lesions, 3-15 mm in size, were placed at random at phantom depths of 2-6 cm by an indiv idual different from the person performing the localization. Scintigra phy-guided three-dimensional localization of the lesion was performed by acquiring two orthogonal images and superimposing the reference bar s over the lesion image and thus identifying the exact x, y and z coor dinates of the lesion. Using these coordinates, a 22 gauge needle, con taining about 37 MBq (1 mCi) Tc-99(m) within its lumen, was stereotact ically placed into the phantom, and the tracer contained in the needle injected into the lesion. Needle placement was primarily guided by th e exact coordinates, but also by real-time visualization of the radioa ctive needle. Pre- and post-localization images were acquired and regi ons of interest (ROIs) defined. Also, after the tracer was injected in to the lesion, images of the phantom with and without the lesion were obtained to calculate the percentage of tracer injected outside the le sion. The results of 30 consecutive localization attempts included 25 exact localizations with less than 20% tracer injected outside the les ion, 2 'near misses' with 37-44% injected outside the lesion, and 3 'm isses' with 60, 85 and 100% of the tracer being injected outside the l esion. The missed localizations were all in lesions at least 4 cm deep , and all had partially superimposed ROIs, which indicates the needle came very close to the lesion. To conclude, our scintigraphy-guided bi plane localization technique for breast phantom lesions successfully l ocalized 90% of all lesions.