Abstract
Breast lesion detection employing state of the art microwave systems provide a safe, non-ionizing technique that can differentiate healthy and non-healthy tissues by exploiting their dielectric properties. In this paper, a microwave apparatus for breast lesion detection is used to accumulate clinical data from subjects undergoing breast examinations at the Department of Diagnostic Imaging, Perugia Hospital, Perugia, Italy. This paper presents the first ever clinical demonstration and comparison of a microwave ultra-wideband (UWB) device augmented by machine learning with subjects who are simultaneously undergoing conventional breast examinations. Non-ionizing microwave signals are transmitted through the breast tissue and the scattering parameters (S-parameter) are received via a dedicated moving transmitting and receiving antenna set-up. The output of a parallel radiologist study for the same subjects, performed using conventional techniques, is taken to pre-process microwave data and create suitable data for the machine intelligence system. These data are used to train and investigate several suitable supervised machine learning algorithms nearest neighbour (NN), multi-layer perceptron (MLP) neural network, and support vector machine (SVM) to create an intelligent classification system towards supporting clinicians to recognise breasts with lesions. The results are rigorously analysed, validated through statistical measurements, and found the quadratic kernel of SVM can classify the breast data with 98% accuracy.
Highlights
IntroductionWww.nature.com/scientificreports breast tissues, with a lower contrast (lower than 10% in dielectric properties) is found between healthy fibro glandular and malignant tissues[11,12,13]
The clinical trial UWB data have been collected at Perugia Hospital, Italy, using the microwave apparatus named “MammoWave”, a non-ionizing and X-ray free mammogram invented by UBT Srl, Italy
We have investigated the prospect of employing Machine learning (ML) algorithms for computer-aided breast lesion detection to support clinicians, by reducing overhead and increasing the speed in decision making between healthy and non-healthy lesion patterns from the clinically collected data through the current microwave apparatus
Summary
Www.nature.com/scientificreports breast tissues, with a lower contrast (lower than 10% in dielectric properties) is found between healthy fibro glandular and malignant tissues[11,12,13]. Machine learning (ML) can be explicitly used to make decisions based on learned patterns (available datasets) and can automatically create an analytical model for future predictions without direct human intervention Various methods such as nearest neighbor, neural networks, naive bayes, decision trees, conventional ML algorithms, and some hybrid approaches have been used for classification purpose. We have investigated the prospect of employing ML algorithms for computer-aided breast lesion detection to support clinicians, by reducing overhead and increasing the speed in decision making between healthy and non-healthy lesion patterns from the clinically collected data through the current microwave apparatus. Three popular methods, k-nearest neighbor (kNN), multi-layer perceptron (MLP) neural network, and support vector machine (SVM) are explored here to analyze the acquired labeled MammoWave data thoroughly These experiments have been performed to fit the labeled training data with the optimal model parameters for predicting the presence of a lesion. Preliminary results show the proposed method produces minimal false-positives and false-negatives compared to other state-of-art methods and develop a viable anonymize method for mass screening breast lesion detection in future
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
