Risa Chaisuparat

Objectives: To quantify the provincial heterogeneity incidence of oral potentially malignant disorders (OPMDs) and explore geographic disparities and ecological determinants in Thailand through Bayesian disease mapping. Methods: An observational ecological study analyzed clinically diagnosed OPMDs from Thailand’s national electronic health record database (2023 to 2024) except Bangkok. Spatial autocorrelation with spatial Bayesian hierarchical models were used to evaluate their associations with behavioral, socioeconomic, and oral-health determinants. Results: A total of 5,956 OPMDs diagnoses were identified across 76 provinces. The median age at diagnosis was 62 y, with females comprising 68.2%. Lichen planus (45.9%) and leukoplakia (39.5%) were the predominant subtypes. The age-standardized incidence rate was 1.9 per 100,000 (female: 2.5; male: 1.3), peaking at ages 60 to 69 y. Spatial analyses revealed high-risk clusters in northern Thailand and identified smokeless tobacco use, dentist density, and Gini coefficient as significant ecological predictors of the incidence of OPMDs. Conclusions: This population-based analysis demonstrates spatial disparities in the incidence of OPMDs. Contextual social determinants of health including socioeconomic inequality, smokeless tobacco behaviors, and workforce availability contribute to variations in disease distribution at the population level. Strengthening the integration of big data analytics within oral health surveillance systems can enhance timely, evidence-informed decision-making and support the strategic refinement and resource optimization of Thailand’s national oral cancer screening benefit package. Knowledge Transfer Statement: The results provide strategic insights for clinicians, public health professionals, and policymakers identifying high-risk geographic areas where preventive interventions and early detection efforts should be prioritized. By clarifying province-specific risk profiles, the study supports the development of targeted health promotion, prevention, and screening strategies aimed at reducing geographic inequalities and improving oral health outcomes.

Objectives: To define tumor immunoarchitectural patterns (IPs) and characterize the immune profile in salivary gland mucoepidermoid carcinoma (MEC) toward assessing MEC prognostic significance and implications for immunotherapy. Methods: This study analyzed 41 MEC cases, evaluating the tumor IPs and tumor‐infiltrating lymphocyte (TIL) parameters by using whole‐slide imaging and AI‐assisted assessment. Immunohistochemistry of CD3 and CD8 markers was performed to assess key lymphocyte subpopulations. Results: Immune‐rich (I‐rich) tumors were characterized by high TIL density and were associated with aggressive tumor behavior, particularly in histological high‐grade MEC, with a significant increase in TIL density observed in the inner invasive margin compartment ( p < 0.05). I‐rich tumor cases were linked to poorer disease‐free survival (DFS) (Breslow = 4.686, p = 0.03). Aggressive MEC cases displayed a highly heterogeneous TIL profile, however, no TIL patterns showed a significant impact on DFS ( p > 0.05). Conclusions: Despite the high immunogenicity suggested by the abundance of T lymphocytes, the protective role of CD8+ T lymphocytes was not prominent in MEC. I‐rich tumors appeared more aggressive and unfavorable MEC cases exhibited high heterogeneity in their TIL profiles. Interrogating the role of TILs in the inner invasive tumor margin may offer a new mechanistic understanding for future immunotherapy discovery.

Introduction: Oral squamous cell carcinoma (OSCC) is often preceded by oral epithelial dysplasia (OED). The role of ribosomal protein S6 (RPS6) and programmed cell death ligand-1 (PD-L1) in the progression of OED to OSCC remains unclear. This study aimed to investigate the expression of phosphorylated RPS6 (p-RPS6) and PD-L1 in OSCC and OED and to examine its relationship with clinicopathological features.Methods: Fifty-two OSCC and 48 OED cases were recruited for immunohistochemical analysis of p-RPS6 and PD-L1 expression. The expression of markers was correlated with clinicopathological features of OSCC and OED.Results: We found p-RPS6 expression in all cases of OSCC and OED, whereas PD-L1 was expressed in 42/48 (87%) OED and in 28/52 (53%) OSCC. The patients with mild OED presented higher expression level of PD-L1 and p-RPS6 significantly, when compared to moderate-differentiated OSCC patients (p < 0.05). Moreover, we found a significant positive correlation between PD-L1 and p-RPS6 expression in OED and OSCC patients (p < 0.01). The PD-L1 expression was significantly related to more than 2 cm tumor size in OSCC patients (p = 0.007).Discussion: Our findings suggest the upregulation of PD-L1 may be related with activation of the mTOR pathway in the early events of tumor progression and the pathogenesis of OSCC.

Oral cancer is one of the most commonly found cancers worldwide. Oral Epithelial Dysplasia (OED) is an Oral Potentially Malignant Disorder (OPMD) that can be characterized for preventive oral cancer screening. The standard for OED histological grading is conducted via the epithelial regions of tissue biopsies. However, this procedure is laborious, time-consuming, and subjective; consequently, it is prone to variability due to fatigue and limited expertise. Therefore, this study aims to explore the potential of using Convolutional Neural Network (CNN) and Transformer models for an automated epithelium segmentation algorithm directly from Whole Slide Images (WSIs). This approach can reduce the manual process and support pathologists in grading activities. Accordingly, candidate architectures based on CNN and Transformer are selected: UNet, ResNet50-UNet, VGG19-UNet, Swin-UNet, and MISSFormer. These models are trained using patch-based segmentation to mitigate the high computational cost caused by processing WSIs. The results indicate that UNet, optimized with the ADAM optimizer, demonstrates the best performance in patch-based segmentation with Intersection over Union (IoU) of 0.82 and Dice-Similarity Coefficient (DSC) of 0.87. Furthermore, this model achieves the highest IoU and DSC for tissue-level prediction, scoring 0.88 and 0.94, respectively. According to the experiment, overlapping and non-overlapping patching strategies perform similarly in most of the selected architectures. The latter approach, hence, is suggested for computational efficiency. These results can support enhancing automated epithelium segmentation to provide a reliable tool for assisting pathologists.

Background: Mucoepidermoid carcinoma is a rare salivary gland malignant tumour. This study aimed to investigate inflammatory and immune signatures of mucoepidermoid carcinoma by identifying potential proteo‐transcriptomic biomarkers towards the development of precision immuno‐oncology treatment strategies.Methods: A total of 30 biopsies obtained from patients diagnosed with mucoepidermoid carcinoma between 2013 and 2022 were analysed after H&E staining for scoring of histological inflammatory stroma subtypes and inflammatory hotspots with QuPath. Multiplex immunofluorescence staining and NanoString nCounter PanCancer IO 360™ panel were used to assess stroma and tumour inflammation signatures in high grade mucoepidermoid carcinoma cases in the tumour microenvironment via proteomics and transcriptomics, respectively.Results: Inflammatory cells within the histological inflammatory stroma inflammatory (HIS‐INF/hot) tumour neighbourhoods were greater compared to the histological inflammatory stroma‐immune desert (HIS‐ID/cold) (p = 0.001). A similar trend was observed between treatment non‐responders and responders in stroma neighbourhoods (p = 0.0625) and in stroma‐to‐interface inflammatory hotspots (p = 0.0081), indicating an augmented inflammatory response in hot tumours and non‐responders. Furthermore, there were striking differences in the expression of pan‐immune leukocyte marker CD45 between responders and non responders particularly in the tumour neighbourhoods (p = 0.0341), but such were not robust for PD‐1 and macrophage fractions. Additionally, transcriptomic analysis revealed key differences in leukocyte activation profiles between responders and non‐responders.Conclusion: This preliminary report unveils the importance of assessing immune leukocyte cellular fractions and pathways for future prognostic biomarker discoveries in mucoepidermoid carcinoma as per the involvement of CD45‐driven inflammatory and immune mediators in high grade mucoepidermoid carcinoma in non‐responders to treatment. These findings will potentially contribute to the development of novel personalised immunotherapies.

Salivary glands (SG) are exocrine organs with secretory units commonly injured by radiotherapy. Bio-engineered organoids and extracellular vesicles (EV) are currently under investigation as potential strategies for SG repair. Herein, three-dimensional (3D) cultures of SG functional organoids (SGo) and human dental pulp stem cells (hDPSC) were generated by magnetic 3D bioassembly (M3DB) platforms. Fibroblast growth factor 10 (FGF10) was used to enrich the SGo in secretory epithelial units. After 11 culture days via M3DB, SGo displayed SG-specific acinar epithelial units with functional properties upon neurostimulation. To consistently develop 3D hDPSC in vitro, 3 culture days were sufficient to maintain hDPSC undifferentiated genotype and phenotype for EV generation. EV isolation was performed via sequential centrifugation of the conditioned media of hDPSC and SGo cultures. EV were characterized by nanoparticle tracking analysis, electron microscopy and immunoblotting. EV were in the exosome range for hDPSC (diameter: 88.03 ± 15.60 nm) and for SGo (123.15 ± 63.06 nm). Upon ex vivo administration, exosomes derived from SGo significantly stimulated epithelial growth (up to 60%), mitosis, epithelial progenitors and neuronal growth in injured SG; however, such biological effects were less distinctive with the ones derived from hDPSC. Next, these exosome biological effects were investigated by proteomic arrays. Mass spectrometry profiling of SGo exosomes predicted that cellular growth, development and signaling was due to known and undocumented molecular targets downstream of FGF10. Semaphorins were identified as one of the novel targets requiring further investigations. Thus, M3DB platforms can generate exosomes with potential to ameliorate SG epithelial damage.

Background: Medication‐related osteonecrosis of the jaw (MRONJ) is a pathology condition of jaw bone caused by a side effect of medications prescribed for skeletal disease. The mechanism of MRONJ is still unknown now. Osteoclasts are cells directly influenced by the medication and the modification in cells metabolisms by the drugs lead to MRONJ. Thus, this study aimed to investigate the osteoclasts morphology, quantity, and comparing with other necrotic diseases.Methods: Thirty‐eight (38) subjects, including cases with MRONJ (n = 11), osteoradionecrosis of the jaw (n = 9), osteomyelitis of the jaw (n = 9), and normal jaw bone (n = 9), were studied. Hematoxylin‐and‐eosin‐stained slides of these diagnosed cases were used to evaluate osteoclasts’ characteristics. Immunohistochemistry of TRAP was performed to observed the function of osteoclasts. These characteristics of osteoclasts were also evaluated in the relationship with the histological features using regression analysis.Results: The results showed that osteoclasts in MRONJ enhance activity by increasing the size and the quantity (p < 0.05). The presence of osteoblasts, inflammatory cells, and bacterial colonies showed a strong correlation with the change in morphology and the number of osteoclasts (p < 0.05). However, the TRAP‐positive mean number and the TRAP intensity of osteoclasts in MRONJ did not show a significant difference with those in other groups (p > 0.05).Conclusion: In conclusion, osteoclasts in MRONJ increase the number and become bigger with multi‐nuclei which might relate to the presence of osteoblasts, inflammation, and microorganisms. This finding supports the idea osteoclasts might be the main key to investigate MRONJ.