Abstract Background: Confocal microscopic strip mosaicing (CSM) provides noninvasive optical sectioning and high resolution, which allows for imaging of nuclear and morphological detail in freshly excised tissue. CSM can image large areas of tissue at micron-level resolution in minutes, which may offer an advantage over standard histology that requires days. We have conducted a preliminary investigation of the feasibility of this technology for the evaluation of breast tissue from surgical excision specimens. Design: In a prospective study, 80 fresh human breast tissue samples from surgical excision specimens of 24 patients were imaged using a prototype confocal strip scanner. Fresh tissue specimens were immersed in Acridine Orange (AO) for 45 seconds to stain the nuclei, then pressed against the glass imaging window and imaged with a 30X, 0.75 numerical aperture (measured) objective lens and a 488 nm laser. Images were acquired in two modes of contrast: in fluorescence (with AO), showing nuclear morphology, and in reflectance (endogenous), showing stroma. The use of fluorescence for nuclear staining mimics the use of hematoxylin in pathology, and the use of reflectance eosin. Use of two contrast modes allows the fluorescence image to be colorized purple and the reflectance image pink, producing confocal strip mosaics that mimic H&E histology in appearance. Specimens were subsequently fixed in formalin and routinely processed to obtain H&E stained sections. H&E and confocal images were compared by the study pathologist (M.M.) Results: Freshly excised breast tissue samples as large as 2 cm x 2 cm were imaged in less than five minutes, with 1-micron resolution and measured optical sectioning of 6 microns. We compared the CSM images against standard histopathology images. In our series we evaluated the following histologies: 12 invasive carcinoma (11 ductal, 1 lobular), 3 ductal carcinoma in-situ, 3 lobular carcinoma in situ, 1 atypical lobular hyperplasia, 1 atypical ductal hyperplasia and various benign lesions such as fat necrosis, fibrocystic changes, and ductal hyperplasia. In confocal images invasive and in situ carcinoma as well as benign ducts and lobules were distinguished from surrounding stromal tissue. Limitations that are typically encountered in standard histology, such as distinguishing low grade ductal carcinoma in situ (DCIS) from lobular carcinoma in situ (LCIS) or atypical proliferations were encountered in the grayscale confocal images as well. Conclusion: In this initial feasibility study, CSM produced images that could be diagnosed as benign or neoplastic by the study pathologist. Further study is needed to build an image library of breast histology and compare reproducibility of histologic diagnoses between CSM (grayscale and colorized images) and traditional optical microscopy, and assess interobserver reproducibility in diagnosis. CSM potentially provides rapid and noninvasive evaluation of breast parenchyma, and has a potential application for intraoperative margin assessment of resected breast specimens. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-03-03.
Read full abstract