Spectrochemical analysis highlights similarities in adenocarcinomas irrespective of primary tissue origins

Abstract

e13551

Background: Identification of primary tissue origin of brain metastases can be histopathologically challenging. Initial diagnosis of malignancy can be provided on intraoperative tissue smears; however, it requires formal histopathological examination along with immunohistochemistry to reach a final diagnosis. On occasion this will still fail to determine primary origin. This study aims to demonstrate if Raman spectroscopy is able to determine primary tissue origin of brain metastasis of three common primary tumours. Methods: Formalin fixed paraffin embedded tissue from twenty brain metastasis comprising colorectal adenocarcinomas (n= 7), lung adenocarcinomas (n= 7) and melanomas (n= 6) were obtained from the Brain Tumour NorthWest tissue bank. Sections (10-μm thick) were placed onto glass slides covered in aluminium foil and de-waxed prior to spectral acquisition. 25 spectra per section were collected at random, using a 785 nm laser at 1200 g mm-1grating with an acquisition time of 30 sec. Computational analysis within a MatLab environment was then conducted. Results: Following PCA-LDC analysis, classification accuracy of the three groups was; colorectal adenocarcinoma 71.3%, lung adenocarcinoma 71% and melanoma 70%. A one-way ANOVA showed statistically significant differences between the three groups, p = 0.0014. On combining adenocarcinoma groups accuracy increased to 87.8%, whilst melanoma fell to 68.9%. A student’s t-test confirmed statistical significance between the two groups, p < 0.0001. Conclusions: Raman spectroscopy can classify different tumours by type, though sensitivity and specificity is diminished if used to classify primary tissue origin of adenocarcinomas. Therefore for clinical use, such new tools may aid the clinician to determine tumour type intra-operatively but classic histopathology is still required for tissue origin confirmation. Such technologies may provide a useful adjunct to more conventional approaches; for instance, decreasing the number of immunohistochemical stains required to determine tissue origin. These results also suggest strong similarities between adenocarcinomas of different primary origins.