Rapid and Label-Free Histopathology of Oral Lesions Using Deep Learning Applied to Optical and Infrared Spectroscopic Imaging Data

Abstract

Oral potentially malignant disorders (OPMDs) are precursors to over 80% of oral cancers. Hematoxylin and eosin (H&E) staining, followed by pathologist interpretation of tissue and cellular morphology, is the current gold standard for diagnosis. However, this method is qualitative, can result in errors during the multi-step diagnostic process, and results may have significant inter-observer variability. Chemical imaging (CI) offers a promising alternative, wherein label-free imaging is used to record both the morphology and the composition of tissue and artificial intelligence (AI) is used to objectively assign histologic information. Here, we employ quantum cascade laser (QCL)-based discrete frequency infrared (DFIR) chemical imaging to record data from oral tissues. In this proof-of-concept study, we focused on achieving tissue segmentation into three classes (connective tissue, dysplastic epithelium, and normal epithelium) using a convolutional neural network (CNN) applied to three bands of label-free DFIR data with paired darkfield visible imaging. Using pathologist-annotated H&E images as the ground truth, we demonstrate results that are 94.5% accurate with the ground truth using combined information from IR and darkfield microscopy in a deep learning framework. This chemical-imaging-based workflow for OPMD classification has the potential to enhance the efficiency and accuracy of clinical oral precancer diagnosis.

Challenges in Optimizing Nanoplatforms Used for Local and Systemic Delivery in the Oral Cavity

“…The oral cavity has several functional structures, including the tongue, salivary glands, cheeks, lips, and palate including teeth, all of which play an important role in physiological processes [35,36]. Histologically, the oral mucosa is a multilayered structure consisting of the epithelium (basement membrane and chorion), and the epithelium presents the stratum basale (basal layer), stratum spinosum (spinous layer), stratum granulosum (granular layer), and stratum corneum [35][36][37][38]:…”

“…These cells have intensely colored voluminous nuclei located in the basal third of the cells. Basal cells exhibit intense metabolism and frequent mitoses that provide flaking turnovers [35][36][37][38]; -The stratum spinosum is arranged above the basal layer and presents 7-8 rows of well-defined polyhedral cells with visible intercellular spaces; between the cells of the stratum spinosum there are fine cytoplasmic filaments arranged in the form of thorns. The nuclei of the cells of the spinosum stratum stain less intensely than the nuclei of the cells of the stratum basale, so that the stratum spinosum is well delineated from the stratum basale.…”

“…The nuclei of the cells of the spinosum stratum stain less intensely than the nuclei of the cells of the stratum basale, so that the stratum spinosum is well delineated from the stratum basale. The cells in the deep areas of the stratum spinosum show mitosis and together with the basal cells form the germinative zone of the epithelium, and towards the surface, the cells of the stratum spinosum become flatter [35][36][37][38]; -…”

“…The stratum granulosum occurs when there is also the stratum corneum, and in the absence of the stratum corneum, may be absent. It is located above the stratum spinosum and consists of 2-3 rows of flattened cells that present kerato-hyaline granules in the cytoplasm [35][36][37][38]; -…”

“…The cells of the stratum corneum are flattened and degenerated with nuclear pycnosis, or without nuclei, they are weakly joined to each other. The ep-ithelium is separated from the dermis by the basement membrane, which presents itself as a condensation of the fundamental substance of the underlying connective tissue [35][36][37][38].…”