FOCUS
Photonic Pathways
and two- and three-photon excited autofluorescence, to the clinic. In HHGM, two or three photons from the excitation laser combine to form a single photon with twice or thrice the energy. SHG arises in organized, non-centrosymmetric molecules such as collagen, while THG emerges at refractive index discontinuities, interfaces between cell membranes, nuclei, collagen, and elastin fibers. Combined with autofluorescence from NAD( P) H, flavins, retinol, and elastin, these contrasts yield richly detailed, label-free images that resemble histology in architecture, but contains information beyond that.
Generating efficient harmonic signals requires ultrashort femtosecond pulses. To achieve this with minimal tissue heating, Flash Pathology BV, uses a compact fiber-laser microscope operating at 1070 nm combined with pulse-picking technology: the system delivers bursts of five sub-80 fs pulses at 1 MHz, producing strong nonlinear excitation with only 5 mW of average power, well within safe limits for excised tissue. The resulting instrument, measuring 0.6 × 0.8 × 1.5 m ³ and weighing 40 kg, achieves 0.4 × 0.4 × 2.4 µ m ³ resolution and can generate diagnostic images within minutes. See figure 1 for the layout inside the microscope.
Figure 2. HHG( SHG-THG-2PEF) images compared to conventional histology images with hematoxylin and eosin stain( H & E), for healthy tissue: skin, lymphatic, adipose, cartilage, muscle and brain tissue.
FROM BENCH TO BEDSIDE: TESTING HHGM ACROSS CANCER TYPES We have validated HHGM across a range of clinical applications, see e. g.( 4, 5). In HHGM images of renal tumor resections, pathologists correctly identified normal versus tumor tissue in 97 % of cases. In lung biopsies from 47 patients( 109 samples), high-quality HHGM images were obtained within six minutes of excision, with 97 % deemed suitable for diagnostic interpretation and 87 % correctly classified by pathologists. In parathyroid tissue, 85 % sensitivity and 83 % specificity was achieved. Beyond structural imaging, 3D HHGM combined with artificial intelligence enables automated classification of immune cell types in bronchoalveolar lavage samples
32 www. photoniques. com I Photoniques 134