Vibrational imaging for label-free cancer diagnosis and classification

The first key step in the detection and classification of most cancers is the microscopic assessment of thin tissue slices, the so-called "histopathology". This procedure is still nowadays, similarly to 150 years ago, performed by staining the tissue with two or more dyes able to bind to specific biological structures, followed by visual inspection by the histopathologist under the bright-field optical microscope. This approach involves long manual procedures which can be accompanied by human errors, subjectivity, and lack of reproducibility. Vibrational microscopies are capable of directly providing chemical and biomolecular information on tissues, identifying them through their fingerprint vibrational spectra without the need of staining and, thus, constitute powerful tools for label-free and objective tumour identification. The two most established techniques, spontaneous Raman microscopy and infrared absorption microscopy, suffer, respectively, from long acquisition times and low spatial resolution. These limitations can be overcome by novel and more technically demanding approaches such coherent Raman scattering and photothermal infrared microscopy. Here we present an extended overview of the major advances in the field of vibrational imaging for cancer diagnosis. We start from a detailed description of the different technologies and then present examples of their applications to tissue imaging for cancer assessment. We critically compare the presented approaches, discussing the steps required to bring these powerful technologies from bench to bedside.