Confocal microscopy in oral cancer and oral potentially malignant disorders: A systematic review

Abstract

ObjectiveTo systematically identify and summarise current research on the utility of confocal microscopy in oral squamous cell carcinoma and oral epithelial dysplasia in oral potentially malignant disorders.MethodsDatabases Medline, Embase, Evidence‐Based Medicine, and Web of Science were searched with articles screened and included if their primary objective was the use of a confocal microscope in diagnosis of oral cancer or epithelial dysplasia, in vivo or ex vivo.Results and DiscussionTwenty‐eight relevant studies were identified of which 21 studies included oral squamous cell carcinoma specimens. Fifteen studies included in vivo use. The studies included both qualitative and fluorescence confocal microscope and reflectance confocal microscope analysis along with quantitative analysis of carcinoma and dysplasia. Thirteen studies reported the predictive value of their confocal device in the diagnosis of dysplasia and carcinoma. The quantitative software‐based studies show promise in objectifying the diagnostic process for identifying abnormalities within the microstructure of the oral mucosa.ConclusionsThere was heterogeneity in the criteria for diagnosis of dysplasia and oral squamous cell carcinoma with experience levels of assessors impacting method efficacy. Both qualitative and quantitative confocal assessment methodologies have been explored, the latter highlighting the potential of future machine‐augmented diagnostic precision.

Dynamic real‐time optical microscopy of oral mucosal lesions using confocal laser endomicroscopy

“…In this study, we demonstrate the performance of fsdCLE as a chairside technology for diagnostic assessment of oral mucosal lesions and determine its accuracy in the diagnosis of dysplasia/carcinoma compared with conventional histopathology. Several previous studies using ppCLE technology have demonstrated the possibility and limitations of applying CLE in vivo to malignancy and premalignancy in the head and neck region 10,13 but none have determined the ability of CLE to be used as a point-of-care diagnostic device offering on-thefly microscopic imaging data. Additionally, none have determined the use of acriflavine hydrochloride as contrast imaging agent for determination of cellular and subcellular features of oral tissues.…”

“…Reflectance and fluorescence ppCLE devices have been used in pilot studies of oral lesions including oral leukoplakia and OSCC (reviewed in 13 ). These have either performed imaging ex vivo using freshly excised tissue, 14 or produced inadequate images during capture requiring development and validation of a classification system for diagnosis for OSCC 15 .…”

Recent progress of second near-infrared (NIR-II) fluorescence microscopy in bioimaging

“…So far, confocal microscopy and its non-linear counterpart multiphoton microscopy have been considered the gold standard techniques for fluorescence bioimaging ( Ramani et al, 2022 ; Teo et al, 2022 ). Confocal microscopes can eliminate out-of-focus signals by placing a pinhole aperture at the conjugate point of the focus of the detecting objective lens and the focus of the concentrator.…”

Revolutionizing oral pathology and medicine: The artificial intelligence advantage

“…For instance, in the diagnosis of oral diseases, neural networks can be trained on large datasets of macrographs, micrographs, histopathology slide images, and other sources of patient data. 7 These networks can learn complex patterns and features from this data, enabling them to make accurate diagnoses with high sensitivity and specificity. 8 Upon analyzing patient data, such as medical and dental histories, genetic information, and imaging results, these deep learning models could recommend personalized treatment strategies.…”

“…In oral medicine, deep learning and neural networks can be utilized for various tasks such as disease diagnosis, treatment planning, and outcome prediction. For instance, in the diagnosis of oral diseases, neural networks can be trained on large datasets of macrographs, micrographs, histopathology slide images, and other sources of patient data 7 . These networks can learn complex patterns and features from this data, enabling them to make accurate diagnoses with high sensitivity and specificity 8 .…”