Résumé | Tissue biopsies and fine-needle aspirates (FNA) of patients with suspected benign or malignant thyroid neoplasms were investigated by infrared microspectroscopy and multivariate statistical methods. Unsupervised cluster analysis revealed four different spectral patterns for the aspirates analyzed, corresponding to colloid goiter, adenoma, carcinomas and negative diagnoses. Infrared microspectroscopic measurements of neoplastic cells on infrared transparent slides provide a potentially new tool for diagnostic screening of these FNA. Biopsy material obtained during surgical removal of gland tissue, was successfully used to generate statistically significant criteria for the distinction of neoplastic from normal tissue. Bivariate histogram plots demonstrate that two selected parameters, DNA and protein, are sufficient to separate control tissue from adenoma and carcinomas. |
---|