| Advances in targeted therapy hold the premise for the delivery of more effective treatments to lung cancer patients, given the ability to diagnose and identify patient specific cell types, small cell carcinoma, adenocarcinoma, or squamous cell carcinoma. Label free optical imaging techniques like the Coherent Anti-Stokes Raman Scattering microscopy (CARS) can provide physicians with minimally invasive access to tumors and allow diagnosis and sub-typing. Exploiting CARS requires developing data analysis methods that can rapidly and accurately analyze the new types of data they provide. In this study we designed an image processing framework that automatically and accurately, detects cancer cells in two and three dimensional CARS images. Moreover, we built upon this capability with new approaches to analyzing the segmented data, that provided significant information about the cancerous tissue that allowed for the automatic differential classification of non-small cell lung carcinoma cell types, overcoming the shortcomings of previous such approaches. |
