Semi-supervised Self-training Approaches in Small and Unbalanced Datasets: Application to Xerostomia Radiation Side-Effect

Machine learning and modeling: Data, validation, communication challenges

“…Additionally, in semi-supervised learning unlabeled data could be used to infer high-order representations of data to aid the supervised component of the learning task 14 with application examples such as interactive prostate segmentation 15 and xerostomia (dry mouth) prediction in head and neck cancer. 16 Although there are several ongoing efforts to provide guidelines for developing and reporting ML results, 17,18 with the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statement receiving wide range endorsements for predictive modeling, 17 there are yet no universal consensus recommendations for ML in general or in the setting of medical physics and radiation oncology specifically. This white paper aims to (a) further promote progress in the new ML field in radiation oncology by highlighting its untapped advantages and potential for clinical advancement to newcomers; (b) present current challenges and open questions for further research by newcomers and practitioners; and (c) provide general recommendations to active researchers to avoid common pitfalls and suggest guidelines for transparent and informative reporting of ML results for medical physics and radiation oncology applications.…”

“…The part of the data, which is labeled, could be used to infer the unlabeled portion (e.g., text/image retrieval systems) through transductive learning, or to induce the general mapping from input to output by inductive learning. Additionally, in semi‐supervised learning unlabeled data could be used to infer high‐order representations of data to aid the supervised component of the learning task with application examples such as interactive prostate segmentation and xerostomia (dry mouth) prediction in head and neck cancer …”