Dataset Information

Machine learning application for prediction of locoregional recurrences in early oral tongue cancer: a Web-based prognostic tool.

ABSTRACT: Estimation of risk of recurrence in early-stage oral tongue squamous cell carcinoma (OTSCC) remains a challenge in the field of head and neck oncology. We examined the use of artificial neural networks (ANNs) to predict recurrences in early-stage OTSCC. A Web-based tool available for public use was also developed. A feedforward neural network was trained for prediction of locoregional recurrences in early OTSCC. The trained network was used to evaluate several prognostic parameters (age, gender, T stage, WHO histologic grade, depth of invasion, tumor budding, worst pattern of invasion, perineural invasion, and lymphocytic host response). Our neural network model identified tumor budding and depth of invasion as the most important prognosticators to predict locoregional recurrence. The accuracy of the neural network was 92.7%, which was higher than that of the logistic regression model (86.5%). Our online tool provided 88.2% accuracy, 71.2% sensitivity, and 98.9% specificity. In conclusion, ANN seems to offer a unique decision-making support predicting recurrences and thus adding value for the management of early OTSCC. To the best of our knowledge, this is the first study that applied ANN for prediction of recurrence in early OTSCC and provided a Web-based tool.

SUBMITTER: Alabi RO

PROVIDER: S-EPMC6828835 | biostudies-literature | 2019 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning application for prediction of locoregional recurrences in early oral tongue cancer: a Web-based prognostic tool.

Alabi Rasheed Omobolaji RO Elmusrati Mohammed M Sawazaki-Calone Iris I Kowalski Luiz Paulo LP Haglund Caj C Coletta Ricardo D RD Mäkitie Antti A AA Salo Tuula T Leivo Ilmo I Almangush Alhadi A

Virchows Archiv : an international journal of pathology 20190817 4

Estimation of risk of recurrence in early-stage oral tongue squamous cell carcinoma (OTSCC) remains a challenge in the field of head and neck oncology. We examined the use of artificial neural networks (ANNs) to predict recurrences in early-stage OTSCC. A Web-based tool available for public use was also developed. A feedforward neural network was trained for prediction of locoregional recurrences in early OTSCC. The trained network was used to evaluate several prognostic parameters (age, gender, ...[more]

PMID: 31422502

Similar Datasets

Project description: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.

Project description:ImportanceThe indications for adjuvant therapy in resected oral tongue cancers are based on both clinical and pathological factors, with clear evidence for adjuvant radiation in patients with pathologically positive neck lymph nodes, positive margins, and extracapsular extension, but the data for patients with no nodal disease are sparse.ObjectiveTo investigate determinants of failure and survival in patients with node-negative oral tongue cancer.Design, setting, and participantsMedical records for patients with oral tongue cancer treated with definitive surgery from 2003 to 2013 were reviewed. All patients were cN0 negative and classified as pathologically node-negative (pN0) if a neck dissection was performed. Patients received adjuvant radiotherapy (RT) or chemoradiotherapy (CRT) based on standard clinical and pathological determinants.Main outcomes and measuresKaplan-Meier and multivariable (MVA) logistic regression and Cox proportional hazard regression analyses were performed to identify patient, tumor, and treatment characteristics predictive of locoregional control (LRC) and overall survival (OS).ResultsA total of 180 patients met entry criteria, with a median follow-up time of 4.9 years (range, 0.9-12.5 years); 102 patients (56.7%) were female and 42 patients (23.3%) were younger than 45 years at diagnosis. One hundred fifty-three patients (85%) had T1/T2 tumors, and 112 patients (62%) had elective neck dissections with confirmed pN0. Lymphovascular space invasion (LVSI) was present in 36 patients (20%). On MVA, LVSI (OR, 0.06; 95% CI, 0.02-0.19; P < .01) was associated with worse LRC. Elective neck dissection (odds ratio [OR], 2.99; 95% CI, 1.16-7.73; P = .02) and receipt of RT (OR, 7.74; 95% CI, 2.27-26.42; P < .01) were associated with improved LRC. Three-year LRC rates were significantly lower for patients with LVSI (38.8%; 95% CI, 22.8%, 54.6%) than those without LVSI (81.9%; 95% CI, 74.4%, 87.4%). On MVA, only LVSI (hazard ratio, 2.20; 95% CI, 1.19-4.06; P = .01) and age greater than 44 years (hazard ratio, 4.38; 95% CI, 1.34-14.27; P = .01) were associated with worse OS. Three-year OS rates were significantly lower in patients with LVSI (71.3%; 95% CI, 53.2%-83.4%) than those without LVSI (90.3%; 95% CI, 83.8%-94.3%).Conclusions and relevanceLymphovascular space invasion in patients with node-negative oral tongue cancer treated with upfront definitive surgery is associated with worse LRC and OS. Node-negative oral cavity cancers with LVSI warrant consideration of further adjuvant therapy, which should be further evaluated in a prospective setting.

Project description:DNA methylation profiling of heterogeneous head and neck squamous cell carcinoma (HNSCC) cohorts has been reported to predict patient outcome. We investigated if a prognostic DNA methylation profile could be found in tumour tissue from a single uniform subsite, the oral tongue. The methylation status of 109 comprehensively annotated oral tongue squamous cell carcinoma (OTSCC) formalin-fixed paraffin-embedded (FFPE) samples from a single institution were examined with the Illumina HumanMethylation450K (HM450K) array. Data pre-processing, quality control and analysis were performed using R packages. Probes mapping to SNPs, sex chromosomes and unreliable probes were accounted for prior to downstream analyses. The relationship between methylation and patient survival was examined using both agnostic approaches and feature selection. The cohort was enlarged by incorporation of 331 The Cancer Genome Atlas (TCGA) HNSCC samples, which included 91 TCGA OTSCC samples with HM450K and survival data available.Given the use of FFPE-derived DNA, we defined different cohorts for separate analyses. Overall, similar results were found between cohorts. With an unsupervised approach, no distinct hypermethylated group of samples was identified and nor was a prognostic methylation profile identified. The use of multiple downstream feature selection approaches, including a linear models for microarray data (LIMMA), centroid feature selection (CFS), and recursive feature elimination (RFE) support vector machines, similarly failed to identify a significant methylation signature informative for patient prognosis or any clinicopathological data available. Furthermore, we were unable to confirm the prognostic methylation profiles or specific prognostic loci reported within the literature for HNSCC.With genome-scale assessment of DNA methylation using HM450K in one of the largest OTSCC cohorts to date, we were unable to identify a hypermethylated group of tumours or a prognostic methylation signature. This suggests that either DNA methylation in isolation is not likely to be of prognostic value or larger cohorts are required to identify such a biomarker for OTSCC.

Dataset Information

Machine learning application for prediction of locoregional recurrences in early oral tongue cancer: a Web-based prognostic tool.

Publications

Machine learning application for prediction of locoregional recurrences in early oral tongue cancer: a Web-based prognostic tool.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets