Oral squamous cell carcinoma is the most common oral cancer. In this paper, we present a performance analysis of four different deep learning-based pixel-wise methods for lesion segmentation on oral carcinoma images. Two diverse image datasets, one for training and another one for testing, are used to generate and evaluate the models used for segmenting the images, thus allowing to assess the generalization capability of the considered deep network architectures. An important contribution of this work is the creation of the Oral Cancer Annotated (ORCA) dataset, containing ground-truth data derived from the well-known Cancer Genome Atlas (TCGA) dataset.

Uveal melanoma (UM) is the most common intraocular tumor in adults. Despite sharing the name and similar morphological features with cutaneous melanoma (CM), it is an entirely different neoplasia with a particular genetic background and clinical behavior. CDKN2A is a gene located at chromosome 9p21, encoding for P16INK4a and P14(ARF) proteins, whose role as a tumor suppressor has been clearly defined in many malignant tumors. CDKN2A frequently presents germline mutations in familial CM and epigenetic downregulation in a considerable percentage of sporadic CM. It has been hypothesized that CDKN2A alterations are early events in CM development, playing a central role in the malignant transformation of melanocytes. Alterations of the CDKN2A gene reduce the expression of P16INK4a in most CM subtypes. Immunohistochemical evaluation of P16INK4a is currently used, in association with Ki67 and HMB45, in pathology practice to discriminate between dysplastic nevi and melanoma. On the other hand, CKDN2A is rarely mutated in UM, and the immunohistochemical expression of P16INK4a has only been reported in small case series. We tested P16INK4a expression on paraffin-embedded tissue sections from 9 tissue microarrays (TMAs), built with 2 mm cores derived from 133 uveal melanoma FFPE blocks, collected from 1990 to 2018, and from selected paraffin-blocks of 3 UM liver metastases. The immunohistochemical expression of P16INK4a was assessed with a visual evaluation by light microscopy and then with a digital approach. Both approaches, with an acceptable concordance rate, revealed P16INK4a expression in a large proportion of UM cases and all liver metastases, opening new possibilities of using it in the differential diagnosis between cutaneous and uveal melanoma metastases in cases of unknown primary tumor or patients with two different primary melanomas.

We introduce a machine learning-based analysis to predict the immunohistochemical (IHC) labeling index for the cell proliferation marker Ki67/MIB1 on cancer tissues based on morphometrical features extracted from hematoxylin and eosin (H&E)-stained formalin-fixed, paraffin-embedded (FFPE) tumor tissue samples. We provided a proof-of-concept prediction of the Ki67/MIB1 IHC positivity of cancer cells through the definition and quantitation of single nuclear features. In the first instance, we set our digital framework on Ki67/MIB1-stained OSCC (oral squamous cell carcinoma) tissue sample whole slide images, using QuPath as a working platform and its integrated algorithms, and we built a classifier in order to distinguish tumor and stroma classes and, within them, Ki67-positive and Ki67-negative cells; then, we sorted the morphometric features of tumor cells related to their Ki67 IHC status. Among the evaluated features, nuclear hematoxylin mean optical density (NHMOD) presented as the best one to distinguish Ki67/MIB1 positive from negative cells. We confirmed our findings in a single-cell level analysis of H&E staining on Ki67-immunostained/H&E-decolored tissue samples. Finally, we tested our digital framework on a case series of oral squamous cell carcinomas (OSCC), arranged in tissue microarrays; we selected two consecutive sections of each OSCC FFPE TMA (tissue microarray) block, respectively stained with H&E and immuno-stained for Ki67/MIB1. We automatically detected tumor cells in H&E slides and generated a “false color map” (FCM) based on NHMOD through the QuPath measurements map tool. FCM nearly coincided with the actual immunohistochemical result, allowing the prediction of Ki67/MIB1 positive cells in a direct visual fashion. Our proposed approach provides the pathologist with a fast method of identifying the proliferating compartment of the tumor through a quantitative assessment of the nuclear features on H&E slides, readily appreciable by visual inspection. Although this technique needs to be fine-tuned and tested on larger series of tumors, the digital analysis approach appears to be a promising tool to quickly forecast the tumor’s proliferation fraction directly on routinely H&E-stained digital sections.

Head and neck squamous cell carcinoma (HNSCC) includes a group of aggressive malignancies characterized by the overexpression of the epidermal growth factor receptor (EGFR) in 90% of cases. Neuropilin-1 (NRP-1) acts as an EGFR co-receptor, enhancing, upon ligand stimulation, EGFR signaling in several cellular models. However, NRP-1 remains poorly characterized in HNSCC. By utilizing in vitro cellular models of HNSCC, we report that NRP-1 is involved in the regulation of EGFR signaling. In fact, NRP-1 can lead to cisplatin-induced EGFR phosphorylation, an escape mechanism activated by cancer cells upon cytotoxic stress. Furthermore, we evaluated Neuropilin-1 staining in tissue samples of an HNSCC case series (n = 218), unraveling a prognostic value for the Neuropilin-1 tissue expression. These data suggest a potential role for NRP-1 in HNSCC cancer progression, expanding the repertoire of signaling in which NRP-1 is involved and eliciting the need for further investigations on NRP-1 as a suitable target for HNSCC novel therapeutic approaches.