Multimodal characterization of radiologically detectable lung lesions.

Abstract

Diagnosing radiologically detectable lung lesions on the basis of cyto/histopathological staining often suffers from ambiguity, leading to faulty detection of lung diseases, especially cancer. Present study attempted to perform a multimodal characterzation of clinical samples from patients with radiologically detected lung lesions, in order to classify diseases with higher precision. The study included analysing nuclear morphometric and intensity based differences between benign and malignant lung lesions in a quantitative way. Simultaneous analysis of serum samples taken from similar sets of patients was performed by Fourier transform infra red (FTIR) spectroscopy which provided differential global biochemical insights of normal and diseased states. Both nuclear and serum biochemical features were helpful in accomplishing high classification accuracy in distinguishing benign and different malignant cases.