Computer-assisted analysis of cell proliferation markers in oral lesions

A CNN-based approach for joint segmentation and quantification of nuclei and NORs in AgNOR-stained images

BACKGROUND AND OBJECTIVE

Oral cancer is the sixth most common kind of human cancer. Brush cytology for counting Argyrophilic Nucleolar Organizer Regions (AgNORs) can help early mouth cancer detection, lowering patient mortality. However, the manual counting of AgNORs still in use today is time-consuming, labor-intensive, and error-prone. The goal of our work is to address these shortcomings by proposing a convolutional neural network (CNN) based method to automatically segment individual nuclei and AgNORs in microscope slide images and count the number of AgNORs within each nucleus.

METHODS

We systematically defined, trained and tested 102 CNNs in the search for a high-performing solution. This included the evaluation of 51 network architectures combining 17 encoders with 3 decoders and 2 loss functions. These CNNs were trained and evaluated on a new AgNOR-stained image dataset of epithelial cells from oral mucosa containing 1,171 images from 48 patients, with ground truth annotated by specialists. The annotations were greatly facilitated by a semi-automatic procedure developed in our project. Overlapping nuclei, which tend to hide AgNORs, thus affecting their true count, were discarded using an automatic solution also developed in our project. Besides the evaluation on the test dataset, the robustness of the best performing model was evaluated against the results produced by a group of human experts on a second dataset.

RESULTS

The best performing CNN model on the test dataset consisted of a DenseNet-169 + LinkNet with Focal Loss (DenseNet-169 as encoder and LinkNet as decoder). It obtained a Dice score of 0.90 and intersection over union (IoU) of 0.84. The counting of nuclei and AgNORs achieved precision and recall of 0.94 and 0.90 for nuclei, and 0.82 and 0.74 for AgNORs, respectively. Our solution achieved a performance similar to human experts on a set of 291 images from 6 new patients, obtaining Intraclass Correlation Coefficient (ICC) of 0.91 for nuclei and 0.81 for AgNORs with 95% confidence intervals of [0.89, 0.93] and [0.77, 0.84], respectively, and p-values < 0.001, confirming its statistical significance. Our AgNOR-stained image dataset is the most diverse publicly available AgNOR-stained image dataset in terms of number of patients and the first for oral cells.

CONCLUSIONS

CNN-based joint segmentation and quantification of nuclei and NORs in AgNOR-stained images achieves expert-like performance levels, while being orders of magnitude faster than the later. Our solution demonstrated this by showing strong agreement with the results produced by a group of specialists, highlighting its potential to accelerate diagnostic workflows. Our trained model, code, and dataset are available and can stimulate new research in early oral cancer detection.

BMI-1 expression increases in oral leukoplakias and correlates with cell proliferation

Oral leukoplakia (OL) is a white lesion of an indeterminate risk not related to any excluded (other) known diseases or disorders that carry no increased risk for cancer. Many biological markers have been used in an attempt to predict malignant transformation; however, no reliable markers have been established so far.

Plasma-Derived Inflammatory Proteins Predict Oral Squamous Cell Carcinoma

Oral squamous cell carcinoma (OSCC) is a major concern with high morbidity and mortality worldwide, even with the current knowledge and the advancement in treatment. OSCCs diagnosed at late-stage often require wide-excision with or without neck dissection, radiotherapy, or chemotherapy. When deemed successful, treatment often results in diminished quality of life, impaired function, and disfigurement. Strategies for early detection are urgently needed for patients afflicted with this disease. Inflammatory protein plasma biomarkers have shown to be potential tests for early detection and disease monitoring in several cancers. There has been no study on inflammation-related plasma biomarkers in OSCC. The objectives of the study were to use a multiplex approach to screen plasma-derived biomarkers and to examine the association of measurable proteins with OSCC. A total of 260 plasma samples (210 OSCC and 50 normal controls) were collected to measure for concentration of inflammatory related biomarkers using electrochemiluminescence multiplex assay. After screening of 82 potential biomarkers of the first 160 OSCC, 16 cytokines, chemokines, and growth factors were identified and verified in the second set of samples containing 50 OSCC and 50 normal. After adjustment of age and batch effects, the adjusted differential expression analysis showed that the OSCCs were markedly lower in 14 biomarkers and significantly higher level of interleukin 1 receptor antagonist (IL1Ra). By performing unsupervised clustering analysis, we observed distinctive groups of normal and two subgroups of OSCC. Linear regression of IL2, IL1Ra, and macrophage inhibitory factor (MIF) showed high accuracy in classifying OSCC with sensitivity of 0.96 and specificity of 0.92. In conclusion, this is the first paper to identify potential inflammatory plasma protein biomarkers of patients with OSCC. With further validation, the set of biomarkers can potentially be used to assist in early detection of OSCC when the disease is localized and in more treatable stage.

Evaluation of AgNORs in Oral Potentially Malignant Lesions

Oral squamous cell carcinoma (OSCC) is usually preceded by detectable mucosal changes, as leukoplakias and erythroplakia. Histologically, these lesions can range from hyperkeratosis and acanthosis to epithelial dysplasia and even OSCC. The aim of this study was to investigate the proliferative activity, using AgNORs quantification proteins, in low- and high-risk oral epithelial dysplasia, OSCC, and nondysplastic epithelium (inflammatory fibrous hyperplasia). The sample was divided into 4 groups: G1: 10 cases of inflammatory fibrous hyperplasia (IFH), G2: 11 cases of low-risk epithelial dysplasia (LD), G3: 10 cases of high-risk epithelial dysplasia (HD), and G4: 11 cases of OSCC. The quantitative analysis was performed using an image processing software in photomicrographs at 1000x magnification. The one-way ANOVA was used for comparison of the mean AgNORs counts between the study groups. The mean AgNORs count was significantly higher (P ≤ 0.01) in OSCC when compared to IFH and the LD; however, it was not statistically different from HD. The mean number of LD was significantly lower than the HD and OSCC, with no difference related to IFH. AgNORs quantification can be an important and cheap method to help in the determination of the degree of epithelial dysplasia and, consequently, in the analysis of their potential for malignant transformation.

Computer based correlation of the texture of P63 expressed nuclei with histological tumour grade, in laryngeal carcinomas

Background. P63 immunostaining has been considered as potential prognostic factor in laryngeal cancer. Considering that P63 is mainly nuclear stain, a possible correlation between the texture of P63-stained nuclei and the tumor's grade could be of value to diagnosis, since this may be related to biologic information imprinted as texture on P63 expressed nuclei. Objective. To investigate the association between P63 stained nuclei and histologic grade in laryngeal tumor lesions. Methods. Biopsy specimens from laryngeal tumour lesions of 55 patients diagnosed with laryngeal squamous cell carcinomas were immunohistochemically (IHC) stained for P63 expression. Four images were digitized from each patient's IHC specimens. P63 positively expressed nuclei were identified, the percentage of P63 expressed nuclei was computed, and 118 textural, morphological, shape, and architectural features were calculated from each one of the 55 laryngeal lesions. Data were split into the low grade (21 grade I lesions) and high grade (34 grade II and grade III lesions) classes for statistical analysis. Results. With advancing grade, P63 expression decreased, P63 stained nuclei appeared of lower image intensity, more inhomogeneous, of higher local contrast, contained smaller randomly distributed dissimilar structures and had irregular shape. Conclusion. P63 expressed nuclei contain important information related to histologic grade.

Immunohistochemical evaluation of oral epithelial dysplasia using cyclin-D1, p27 and p63 expression as predictors of malignant transformation

OBJECTIVE

To evaluate the degree of expression of cyclin-D1, p27 and p63 in mild, moderate and severe dysplasia using immunohistochemical evaluation in order to illustrate their prognostic value and attempt to propose a molecular grading system for oral epithelial dysplasia.

MATERIALS AND METHODS

The analysis included thirty cases of mild, moderate and severe dysplasia from Department of Oral and Maxillofacial Pathology, Saveetha Dental College, Chennai after a critical review of the Hematoxylin and Eosin (H and E) stained sections. They were subjected to immunohistochemical evaluation using the markers cyclin-D1, p27 and p63. The assessment of the expression based on staining intensity and distribution of immunohistochemical staining of the various markers was analyzed followed by statistical analysis.

RESULTS

A highly significant increase in the expression of cyclin-D1 (P < 0.000) and p63 (P < 0.001) and a moderately significant decrease in the expression of p27 (P < 0.012) with the increasing severity of dysplasia was observed in our study.

CONCLUSIONS

The result of our research affirms the fact that the increase in the expression of markers of cell cycle regulators such as cyclin D1, decrease in the expression of cell cycle inhibitors like p27 and increased expression of p63 in parallel with the increasing severity of dysplasia, emphasizes the use of immunohistochemical markers cyclin D1, p27 and p63 as prognostic markers for better understanding the behaviour of these potentially malignant disorders aiming towards proposing a molecular grading system for oral epithelial dysplasia to enable timely management prior to their possible malignant transformation.

Cellular proliferation markers in peripheral and central fibromas: a comparative study

OBJECTIVE

To perform a comparative study of the cellular proliferation in the peripheral and central fibromas.

MATERIAL AND METHODS

Immunohistochemistry for PCNA and the AgNOR technique were performed in 9 cases of peripheral odontogenic fibroma (POF), in 4 cases of odontogenic fibroma (OdF), in 8 cases of peripheral ossifying fibroma (PEOF) and 7 cases of ossifying fibroma (OsF). The Kruskal-Wallis and Mann-Whitney tests were used for the statistical analyses.

RESULTS

Mesenchymal component of the central lesions presented a higher mean number of AgNOR per nucleus and PCNA index than did the peripheral lesions (P≤0.05). The mean number of AgNOR per nucleus in the epithelial component proved to be higher in the OdF than in the POF (P≤0.05). The mesenchymal and epithelial components presented similar mean numbers of AgNOR per nucleus and PCNA index in the OdF, as well as a similar mean number of AgNOR per nucleus in the POF.

CONCLUSIONS

The mesenchymal component may well play a role in the differences between the biological behaviour of the central lesions as compared to the peripheral lesions. Moreover, considering that the epithelial and mesenchymal components in odontogenic fibromas presented a similar proliferation index, more research is warranted to understand the true role of the epithelial components, which are believed to be inactive in nature, as well as in the development and biological behaviour of these lesions.

Oncoapoptotic signaling and deregulated target genes in cancers: special reference to oral cancer

Cancer is a class of diseases characterized by uncontrolled cell growth. The development of cancer takes place in a multi-step process during which cells acquire a series of mutations that eventually lead to unrestrained cell growth and division, inhibition of cell differentiation, and evasion of cell death. Dysregulation of oncoapoptotic genes, growth factors, receptors and their downstream signaling pathway components represent a central driving force in tumor development. The detailed studies of signal transduction pathways for mechanisms of cell growth and apoptosis have significantly advanced our understanding of human cancers, subsequently leading to more effective treatments. Oral squamous cell carcinoma represents a classic example of multi-stage carcinogenesis. It gradually evolves through transitional precursor lesions from normal epithelium to a full-blown metastatic phenotype. Genetic alterations in many genes encoding crucial proteins, which regulate cell proliferation, differentiation, survival and apoptosis, have been implicated in oral cancer. As like other solid tumors, in oral cancer these genes include the ones coding for cell cycle regulators or oncoproteins (e.g. Ras, Myc, cyclins, CDKs, and CKIs), tumor suppressors (e.g. p53 and pRb), pro-survival proteins (e.g. telomerase, growth factors or their receptors), anti-apoptotic proteins (e.g. Bcl2 family, IAPs, and NF-kB), pro-apoptotic proteins (e.g. Bax and BH-3 family, Fas, TNF-R, and caspases), and the genes encoding key transcription factors or elements for signal transduction leading to cell growth and apoptosis. Here we discuss the current knowledge of oncoapoptotic regulation in human cancers with special reference to oral cancers.

Abnormal cell-cycle expression of the proteins p27, mdm2 and cathepsin B in oral squamous-cell carcinoma infected with human papillomavirus

Since the role of human papillomavirus (HPV) infection in oral carcinogenesis is still unclear, the purpose of this study was to verify the association between the expression of p27, mdm2 and cathepsin B and by HPV-related oral lesions. Fifty-five oral biopsies were studied and HPV detection and typing (6/11, 16, 18, 31 and 33) were performed using polymerase chain reaction techniques. The distribution p27, mdm2 and cathepsin B was determined by immunohistochemistry. Twenty-one (38%) out of the 55 oral lesions tested positive for HPV, of which 6 (33%) were HPV 6/11, 1 (5%) was HPV 16, 14 (72%) were HPV 18 and none was HPV 33/31. Among the 55 biopsies, immunopositivity for p27, mdm2 and cathepsin B was observed in 17 (30.9%), 37 (67.2%) and 37 (67.2%), respectively. Among 21 HPV-positive oral lesions, immunopositivity of mdm2, p27 and cathepsin B was found, respectively, in 6 (33%) out of 18 benign lesions (BL), 4 (22%) out of 18 potential malignant epithelial lesions (PMEL) and 11 (57.9%) out of 19 malignant lesions (ML). High-risk HPV types may be associated with oral carcinoma, by cell-cycle control dysregulation, contributing to oral carcinogenesis and the overexpression of mdm2, p27 and cathepsin B.

Galectin-3 overexpression in invasive laryngeal carcinoma, assessed by computer-assisted analysis

The larynx is the most common site of malignancy in the upper aerodigestive tract. In Brazil, malignant laryngeal lesions represent 2% of all cancers, with approximately 3000 annual deaths. The association between human papillomavirus (HPV) and laryngeal cancer is still controversial. The aim of the present retrospective study was to determine the expression of galectin-3 immunoperoxidase in laryngeal carcinoma by examining paraffin-embedded larynx biopsies from 65 patients, 10 in situ laryngeal carcinomas, 27 laryngeal carcinomas without metastases, and 28 with metastases. Twenty-eight cervical lymph nodes from patients with metastatic lesions were also evaluated. Nested PCR was performed to detect and type HPV DNA. Galectin-3 expression was assessed by immunohistochemistry using a computer-assisted system. Among 65 patients, 55 (84.6%) were positive to beta-globin (internal control); 10 (15.4%) patients were beta-globin negative and were excluded from the HPV evaluation. Thus, 7 (12.7%) out of 55 patients were HPV positive and 48 (87.3%) out of 55 patients were HPV negative. High expression of galectin-3 was observed in invasive laryngeal tumors, suggesting that galectin-3 could be associated with the invasiveness and aggressiveness of laryngeal carcinomas.