Search

Background: The optimal management of oropharyngeal squamous cell carcinoma (OPSCC) includes both sur-gical and non-surgical, that is, (chemo)radiotherapy treatment options and their combinations. These approaches carry a risk of specific treatment-related side effects. HPV-positive OPSCC has been reported to be more sensitive to (chemo)radiotherapy-based treatment modalities. Objectives: This study aims to demonstrate how machine learning can aid in classifying OPSCC patients into risk groups (low-chance or high-chance) for overall survival. We examined the input variables using permutation feature importance. Furthermore, we provided explanations and interpretations using the Local Interpretable Model Agnostic Explanations (LIME) and SHapley Additive Explanation (SHAP) frameworks. Methods: The machine learning model for 3164 OPSCC patients was built using data obtained from the Surveillance, Epidemiology, and End Results (SEER) program database. A total of five variants of tree-based machine learning algorithms (voting ensemble, light GBM, XGBoost, Random Forest, and Extreme Random Trees) were used to divide the patients into risk groups. The developed model with the best predictive performance was temporally validated with a different cohort. Results: The voting ensemble machine learning algorithm showed an accuracy of 88.3%, Mathews' correlation coefficient of 0.72, and weighted area under curve of 0.93, when temporally validated. Human papillomavirus (HPV) status, age of the patients, T stage, marital status, N stage, and the treatment modality (surgery with postoperative radiotherapy) were found to have the most significant effects on the ability of the machine learning model to predict overall survival. Similarly, for the individual patients with SHAP framework, HPV status, gender, and treatment modality (surgery with postoperative radiotherapy) were the input features that improved the model's prediction. Conclusion: The proposed stratification of OPSCC patients into risk groups by machine learning techniques can provide accurate predictions and thus aid clinicians in administering early and personalized interventions. Clinicians could utilize the predicted risk with the explanations offered by the SHAP and LIME frameworks to understand previously undetected relationships between prognostic variables to make informed clinical decisions and effective interventions.

Simple Summary Patients with head and neck squamous cell carcinoma (HNSCC) might present with different clinical behaviors, even when classified at the same stage. This perspective highlights the recent findings on prognostic biomarkers that can aid in improving the staging of HNSCC. This was conducted with an aim of subsequently improving prognostic stratification and, hence, treatment planning. Abstract Tumor-node-metastasis (TNM) staging system is the cornerstone for treatment planning of head and neck squamous cell carcinoma (HNSCC). Many prognostic biomarkers have been introduced as modifiers to further improve the TNM classification of HNSCC. Here, we provide an overview on the use of the recent prognostic biomarkers, with a focus on histopathologic parameters, in improving the risk stratification of HNSCC and their application in the next generation of HNSCC staging systems.

While the relevance of the World Health Organization histopathological grading system as a prognostic tool for oral squamous cell carcinoma has received many critics, other histopathological features such as tumor-stroma ratio, tumor-infiltrating lymphocytes, and tumor budding are displaying promising results. Here, we evaluated the prognostic impact of the incorporation of tumor-stroma ratio, tumor-infiltrating lymphocytes, and tumor budding into World Health Organization histopathological grading for patients with oral squamous cell carcinoma.A total of 95 patients with early-stage oral squamous cell carcinoma were enrolled in the study, and World Health Organization tumor grading, tumor-stroma ratio, tumor-infiltrating lymphocytes, and tumor budding were evaluated in surgical slides stained with hematoxylin and eosin. Survival analyses for cancer-specific survival and disease-free survival were performed using Cox regression models, and receiver operating characteristic curves were applied for assessment of the performance of the combinations.Tumor-stroma ratio (stroma-rich) was significantly and independently associated with both shortened cancer-specific survival and poor disease-free survival, individually and in combination with World Health Organization histopathological grading. The combination of tumor-stroma ratio with World Health Organization grading did not improve the discriminatory ability compared to tumor-stroma ratio alone. Although low tumor-infiltrating lymphocytes were associated with shortened cancer-specific survival, the association did not withstand multivariate analysis. However, in combination with World Health Organization grading, low tumor-infiltrating lymphocytes were independently associated with poor cancer-specific survival. The combination of tumor-infiltrating lymphocytes and World Health Organization histopathological grading displayed a better discrimination of poor cancer-specific survival than tumor-infiltrating lymphocytes alone, but not at a significant level.Our findings support tumor-stroma ratio as a potential prognostic marker for patients with oral squamous cell carcinoma, and the incorporation of tumor-infiltrating lymphocytes into the World Health Organization grading system improves the prognostic ability of the tumor grading alone.

Background: Machine learning models have been reported to assist in the proper management of cancer through accurate prognostication. Integrating such models as a web-based prognostic tool or calculator may help to improve cancer care and assist clinicians in making oral cancer management-related decisions. However, none of these models have been recommended in daily practices of oral cancer due to concerns related to machine learning methodologies and clinical implementation challenges. An instance of the concerns inherent to the science of machine learning is explainability. Objectives: This study measures the usability and explainability of a machine learning-based web prognostic tool that was designed for prediction of oral tongue cancer. We used the System Usability Scale (SUS) and System Causability Scale (SCS) to evaluate the explainability of the prognostic tool. In addition, we propose a framework for the evaluation of post hoc explainability of web-based prognostic tools. Methods: A SUS- and SCS-based questionnaire was administered amongst pathologists, radiologists, cancer and machine learning researchers and surgeons (n = 11) to evaluate the quality of explanations offered by the machine learning-based web prognostic tool to address the concern of explainability and usability of these models for cancer management. The examined web-based tool was developed by our group and is freely available online. Results: In terms of the usability of the web-based tool using the SUS, 81.9% (45.5% strongly agreed; 36.4% agreed) agreed that neither the support of a technical assistant nor a need to learn many things were required to use the web-based tool. Furthermore, 81.8% agreed that the evaluated web-based tool was not cumbersome to use (usability). The average score for the SCS (explainability) was 0.74. A total of 91.0% of the participants strongly agreed that the web-based tool can assist in clinical decision-making. These scores indicated that the examined web-based tool offers a significant level of usability and explanations about the outcome of interest. Conclusions: Integrating the trained and internally and externally validated model as a web-based tool or calculator is poised to offer an effective and easy approach towards the usage and acceptance of these models in the future daily practice. This approach has received significant attention in recent years. Thus, it is important that the usability and explainability of these models are measured to achieve such touted benefits. A usable and well-explained web-based tool further brings the use of these web-based tools closer to everyday clinical practices. Thus, the concept of more personalized and precision oncology can be achieved.

Background The evaluation of immune response can aid in prediction of cancer behaviour. Here, we assessed the prognostic significance of tumour-infiltrating lymphocytes (TILs) in oropharyngeal squamous cell carcinoma (OPSCC). Methods A total of 182 patients treated for OPSCC were included in this study. Assessment of TILs was conducted on tumour sections stained with standard haematoxylin and eosin (HE) staining. We used the scoring criteria proposed by the International Immuno-Oncology Biomarker Working Group. Results The multivariable analysis showed that TILs associated with disease-specific survival with a hazard ratio (HR) of 2.13 (95% CI 1.14–3.96; P = 0.017). Similarly, TILs associated significantly with overall survival with HR of 1.87 (95% CI 1.11–3.13; P = 0.018). In a sub-analysis of HPV-positive and HPV-negative cases separately, TILs showed a significant prognostic value in both groups (P Conclusion The evaluation of TILs as proposed by the International Immuno-Oncology Biomarker Working Group is a simple and promising method in prediction of survival of OPSCC. It is easily applicable and after further validation can be implemented in the routine pathological report as a basic immune parameter.

Although there are many histopathologic prognosticators, grading of early oral tongue squamous cell carcinoma (OTSCC) is still based on morphological cell differentiation which has low prognostic value. Here we summarize the emerging histopathological markers showing powerful prognostic value, but are not included in pathology reports. Using PubMed, Scopus, Ovid Medline, and Web of Science databases, a systematic literature search was preformed to identify early OTSCC studies that investigated the prognostic significance of hematoxylin-eosin-based histopathologic markers. Our meta-analysis showed that tumor budding was associated with overall survival (hazard ratio [HR] 2.32; 95% CI 1.40-3.84; p 0.01) and disease-specific survival (DSS) (1.89; 95% CI 1.13-3.15; p = 0.02). Worst pattern of invasion was associated with disease-free survival (DFS) (1.95; 95% CI 1.04-3.64; p = 0.04). Tumor-stroma ratio was also associated with DFS (1.75, 95% CI 1.24-2.48; p 0.01) and DSS (1.69; 95% CI 1.19-2.42; p 0.01). Tumor budding, worst pattern of invasion, and tumor-stroma ratio have a promising prognostic value in early OTSCC. The evaluation and reporting of these markers is cost-effective and can be incorporated in daily practice.