An Accurate Ensemble Classifier for Medical Volume Analysis: Phantom and Clinical PET Study

Abstract

The predominant application of positron emission tomography (PET) in the field of oncology and radiotherapy and the significant amount of medical imaging research have led to an imminent need for effective approaches for PET volume analysis and the development of accurate and robust volume analysis techniques to support oncologists in their clinical practice, diagnosis procedure, arrangement of the right radiotherapy treatment and evaluation of patients’ response to the provided therapy. This paper proposes an efficient optimised ensemble classifier to tackle the analysis problem of squamous cell carcinoma in patients’ PET volumes. This optimised classifier is based on an artificial neural network (ANN), fuzzy C-means (FCM), an adaptive neuro-fuzzy inference system (ANFIS), K-means and a selforganising map (SOM). Four ensemble classifier machines are proposed in this study. The first three are built using a voting approach, an averaging technique and weighted averaging, respectively. The fourth novel ensemble classifier machine is based on the combination of a modified particle swarm optimisation (PSO) approach and a weighted averaging approach. Experimental national electrical manufacturers association and international electrotechnical commission (NEMA IEC) body phantom and clinical PET studies of participants with laryngeal squamous cell carcinoma are utilised for the evaluation of the proposed approach. Superior results are achieved using the new optimised ensemble classifier when compared with the results from the investigated classifiers and the non-optimised ensemble classifiers. The proposed approach can identify the region of interest class (tumour), with an average accuracy of 98.11% achieved in the studies of participants with laryngeal tumour. This system underpins the expertise of clinicians for PET tumour analysis.