A Transfer Learning Approach for Malignant Prostate Lesion Detection on Multiparametric MRI

Abstract

Purpose: In prostate focal therapy, it is important to accurately localize malignant lesions in order to increase biological effect of the tumor region while achieving a reduction in dose to noncancerous tissue. In this work, we proposed a transfer learning–based deep learning approach, for classification of prostate lesions in multiparametric magnetic resonance imaging images. Methods: Magnetic resonance imaging images were preprocessed to remove bias artifact and normalize the data. Two state-of-the-art deep convolutional neural network models, InceptionV3 and VGG-16, were pretrained on ImageNet data set and retuned on the multiparametric magnetic resonance imaging data set. As lesion appearances differ by the prostate zone that it resides in, separate models were trained. Ensembling was performed on each prostate zone to improve area under the curve. In addition, the predictions from lesions on each prostate zone were scaled separately to increase the area under the curve for all lesions combined. Results: The models were tuned to produce the highest area under the curve on validation data set. When it was applied to the unseen test data set, the transferred InceptionV3 model achieved an area under the curve of 0.81 and the transferred VGG-16 model achieved an area under the curve of 0.83. This was the third best score among the 72 methods from 33 participating groups in ProstateX competition. Conclusion: The transfer learning approach is a promising method for prostate cancer detection on multiparametric magnetic resonance imaging images. Features learned from ImageNet data set can be useful for medical images.

Automated Classification of Significant Prostate Cancer on MRI: A Systematic Review on the Performance of Machine Learning Applications

“…The flow diagram is depicted in Figure 1 . In total, 27 articles were eligible for inclusion in this review [ 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. From these, 13 studies reported enough information to perform a meta-analysis.…”

“…Nevertheless, the details regarding the selected populations were heterogenous between studies. Some papers limited their description to the number of positive and negative samples [27,31], while others mentioned the PCa lesion distribution per GS [11,23] and/or the volume distribution [16,21]. Both characteristics describe the type of population used and whether the PCa lesions are clinically relevant [2].…”

Applications of Artificial Intelligence to Prostate Multiparametric MRI (mpMRI): Current and Emerging Trends

“…The SVM method used discriminative features in training that resulted in an AUC score of 0.89 [95]. All of the studies listed in Table 4 used radiologists to determine their ground truth [77,[92][93][94][95][97][98][99][100]. These studies highlight the ability of DL algorithms to predict the likelihood of a lesion's malignancy based upon Gleason scores.…”

Progress towards Patient-Specific, Spatially-Continuous Radiobiological Dose Prescription and Planning in Prostate Cancer IMRT: An Overview

“…These features can then be used to develop predictive models using machine learning methods when "ground-truth" voxel-level data is available. Many investigators have explored and yielded promising results from using radiomics and machine learning of mpMRI for PCa detection [64][65][66][67][68], tumour aggressiveness (Gleason score) classification, and staging [62,[69][70][71][72][73][74]. Voxel-level tumour biology distributions generated by predictive models using mpMRI can then be used to drive biological optimisation of prostate radiotherapy as demonstrated by the bottom row of Figure 1.…”