Of the major issues that dermatopathology will face in the immediate future, two powerful challenges loom large. The first is the application of novel nondestructive imaging technologies to in vivo diagnosis in humans. The second is the application of molecular technologies to a diagnostic arena which formerly belonged exclusively to the light microscopist. The first to be considered in this context is the application of near infrared spectroscopy to the noninvasive in vivo diagnosis of neoplastic skin disease. The second will be a discussion of application, methodology and the current state of the art in microarray technologies as they apply to neoplastic dermatopathology and, in particular, the diagnosis and prognostication of melanoma. Modern Pathology (2006) 19, S155-S163. doi:10.1038/modpathol.3800513Keywords: in vivo microscope; tissue microarray; basal cell carcinoma; infrared spectroscopy
Noninvasive assessment of skin lesions by near infrared (IR) spectroscopyIn the late 1990s, working with Dr Laura McIntosh and colleagues at the National Research Council of Canada, the University of Manitoba, Central Medical Laboratories and the Misericordia General Hospital in Winnipeg, Canada, we designed and patented a noninvasive tool for the diagnosis of skin tumors using visible and near IR spectroscopy in the 400-2500 nm size wavelength range. 1-5 Initially, we excised neoplasms, and processed them in the fresh state with sections for formaldehyde fixation and paraffin embedding matched to tissue elements snap frozen in liquid nitrogen and stored at À801F. Thick sections from the frozen tissue elements were interrogated by mid-IR wavelength light (Figure 1). The transmitted light generated significant spectral differences; water, hemoglobin, cytochromes, lipids and proteins all absorb light at specific frequencies ( Figure 2). In particular, the mid-IR range is rich with information about proteins including in the context of collagen and RNA. When analyzed by a sophisticated leave-one-out hierarchical classification algorithm, distinction between microanatomic compartments of the skin could be made (Figure 3). Using this methodology, we were able with an accuracy of 90-95% to distinguish basal cell carcinoma (BCC) from melanocytic nevi, seborrheic keratoses and squamous cell carcinomata in vitro. Melanocytic nevi could be subdivided into banal vs dysplastic nevi based upon their spectral differences and melanomas could be separately recognized as well. Furthermore, the different types of lesion were shown to have distinct mid-IR signatures when compared to adjacent normal epidermal and dermal compartments.Unfortunately, the diagnostic potential of mid-IR spectroscopy for in vivo applications is limited, as complete absorption of mid-IR light occurs with samples greater than 10-15 mm in thickness. In contrast, near-IR light scatters to a much greater extent than it absorbs and in consequence, tissues are relatively transparent in the near-IR region which permits the examination of larger tissue volumes and led to...