New Directions in Dermatopathology: In Vivo Confocal Microscopy in Clinical Practice

In-vivo optical microscopy is advancing into routine clinical practice for non-invasively guiding diagnosis and treatment of cancer and other diseases, and thus beginning to reduce the need for traditional biopsy. However, reading and analysis of the optical microscopic images are generally still qualitative, relying mainly on visual examination. Here we present an automated semantic segmentation method called “Multiscale Encoder-Decoder Network (MED-Net)” that provides pixel-wise labeling into classes of patterns in a quantitative manner. The novelty in our approach is the modeling of textural patterns at multiple scales (magnifications, resolutions). This mimics the traditional procedure for examining pathology images, which routinely starts with low magnification (low resolution, large field of view) followed by closer inspection of suspicious areas with higher magnification (higher resolution, smaller fields of view). We trained and tested our model on non-overlapping partitions of 117 reflectance confocal microscopy (RCM) mosaics of melanocytic lesions, an extensive dataset for this application, collected at four clinics in the US, and two in Italy. With patient-wise cross-validation, we achieved pixel-wise mean sensitivity and specificity of 74% and 92%, respectively, with 0.74 Dice coefficient over six classes. In the scenario, we partitioned the data clinic-wise and tested the generalizability of the model over multiple clinics. In this setting, we achieved pixel-wise mean sensitivity and specificity of 77% and 94%, respectively, with 0.77 Dice coefficient. We compared MED-Net against the state-of-the-art semantic segmentation models and achieved better quantitative segmentation performance. Our results also suggest that, due to its nested multiscale architecture, the MED-Net model annotated RCM mosaics more coherently, avoiding unrealistic-fragmented annotations.

Tanto el diagnóstico como el seguimiento en la dermatología dependerán en gran medida de una evaluación visual. Cuando no se puede llegar al diagnóstico, se procede a realizar una biopsia cutánea, un procedimiento invasivo y que con frecuencia los resultados retrasan debido a la demora del transporte, así como al procesamiento de la muestra. Por otro lado, las biopsias pueden destruir las lesiones, lo que imposibilitará realizar un seguimiento clínico de la evolución de la lesión a estudio. La microscopia confocal de reflectancia in vivo (MCR) ofrece una alternativa diagnóstica a la biopsia cutánea. Esta técnica captura en tiempo real imágenes de alta resolución y se utilizará en la evaluación de diversas condiciones dermatológicas. La identificación de características específicas en la MCR podría ayudar a diferenciar dermatosis con una morfología clínica similar. La dermatitis alérgica de contacto (DAC) y la dermatitis irritativa de contacto (DIC) se diagnostican generalmente gracias a pruebas epicutáneas; estas, a su vez, se basarán en un sistema de valoración subjetiva. El uso de la MCR tanto en la detección temprana, como en la clasificación de la intensidad de la DC, ha sido estudiado en múltiples estudios. Existirán algunas características comunes en la MCR de la DAC como de la DIC. Entre estas podemos señalar la disrupción del estrato córneo, la formación de vesículas, la exocitosis, la espongiosis y la paraqueratosis. Las características específicas para la DAC serán la presencia de vasodilatación, el aumento del grosor epidérmico, edema intercelular y acantosis. Las características específicas de la DIC son la presencia de los corneocitos disgregados y queratinocitos en diana. Esta revisión resume la utilidad de la MCR en las condiciones eccematosas de contacto cutáneas y busca incentivar futuras investigaciones, así como incrementar el interés clínico en esta prometedora técnica.

Dermatologic diagnosis and monitoring have been dependent largely on visual grading. A skin biopsy is performed in case of diagnostic uncertainty, but can be traumatic, and results are delayed due to time for specimen transport and processing. Biopsies also destroy specimens, prohibiting lesion evolution monitoring. In vivo reflectance confocal microscopy (RCM) offers a diagnostic alternative to skin biopsy. RCM captures real-time, high-resolution images, and has been piloted for the evaluation of various dermatologic conditions. Identification of unique RCM features may distinguish dermatoses with similar clinical morphologies. Allergic contact dermatitis (ACD) and irritant contact dermatitis (ICD) are diagnosed by patch testing that currently uses a subjective scoring system. RCM has increasingly been studied for early detection and severity grading of CD. Common RCM features shared by ACD and ICD are stratum corneum disruption, vesicle formation, exocytosis, spongiosis, and parakeratosis. Features unique to ACD are vasodilation, increased epidermal thickness, intercellular edema, and acanthosis. Features unique to ICD are detached corneocytes and targetoid keratinocytes. This review summarizes the use of RCM in evaluating contact eccematous conditions and aims to spark future research and interest in this promising tool.