A six-year Australian and New Zealand experience of artificial intelligence techniques applied to MRI liver data – tricks and traps

Congress:

2018 ASM

Poster Number:

R-0041

Type:

Educational Exhibit

Keywords:

Computer Applications-Detection, diagnosis, Neural networks, Computer applications, Image verification

Authors:

M. Blake¹, T. Cooper², S. Khorshid¹, T. St Pierre¹; ¹WA/AU, ²NZ

DOI:

10.1594/ranzcr2018/R-0041

DOI-Link:

https://dx.doi.org/10.1594/ranzcr2018/R-0041

Fig. 1: Assigning degree of fibrosis to MR images of the liver - contrast enhanced 3D...

Fig. 2: Classification of 28 subjects by the SVM model on 3 min post contrast 3D VIBE...

Fig. 3: Schematic diagram illustrating an overfitted model (green line) for...

Fig. 4: Learning curve showing the root mean squared error (RMSE) vs numbers of...

Fig. 5: Predictions of liver iron concentration from the transfer learning neural...

Fig. 6: Predictions of liver iron concentration from the convolutional neural networks...

Fig. 7: Series of single spin echo images of an iron loaded liver with increasing echo...

Fig. 8: Spin echo image together with saliency map. Note how not only pixels in the...

Fig. 9: Spin echo image together with saliency map. Note how the CNN uses pixels in the...

Learning objectives

This paper presents a six-year experience in an Australian and New Zealand setting of applying computer algorithms to automate the analysis of MRI images of the liver for the quantification of iron and fibrosis, and the insights and learnings achieved over this period to transform these algorithms into a regulated medical device.

Background

The following terms are used either throughout the paper or are essential background knowledge to an understanding of the methodologies relating to image processing and computer based artificial intelligence: Machine Learning is the use of algorithms to allow computers to learn models from data without being explicitly programmed. Support Vector Machines – use supervised learning methods (see below) for classification. The SVM is presented with examples of images that have been classified into one of two categories. The SVM builds a model that predicts whether...

Imaging findings OR Procedure details

Part One: Support Vector Machine Learning - Liver Fibrosis In 2012, we applied support vector machine learning (SVM) to a cohort of 28 patients with hepatitis C to assess textural differences in MRI images [2] to build a predictive model for categorising the patients’ Ishak fibrosis scores [3]. Figure 1 shows an example liver image used in the study. The assumptions and limitations of the modelling of small patient numbers, that include cross-validation and overfitting, are presented. The small subject population was divided into both...

Conclusion

Other things to consider when building a neural network: This paper reports a six-year experience applying machine learning models to small and large image datasets. The key elements that have resulted in an acceptable model are the growth in computer processing power, access to advanced machine learning algorithm libraries, and access to large volumes of consistently acquired and well labelled data. Although the majority of the analysis in this paper has focussed on the application of computer algorithms to automate the analysis of MR images...

Personal information

Martin Blake: MBBS, FRANZCR, FAANMS, MBA, FAICD Martin is a Radiologist and Nuclear Medicine Specialist, with 30 years medical imaging experience and has been a partner of Perth Radiological Clinic since 1997 and Chairman since 2009. He has also been a non-executive director of Resonance Health (ASX:RHT) since 2007 and Chairman since 2012. Tony Cooper: BSc (Hons), MS (Statistics) Tony is a Data Scientist and is a founder of Precision AI which provides artificial intelligence solutions to European and Australasian industry. He has a distinguished...

References

Brownlee J. Supervised and Unsupervised Machine Learning Algorithms,https://machinelearningmastery.com/supervised-and-unsupervised-machine-learning-algorithms/ House M.J., Bangma S.J., Thomas M., Gan E.K., Ayonrinde O.T., Adams L.A., Olynyk J.K., St. Pierre T.G. (2015) Texture-based classification of liver fibrosis using MRI. Journal of Magnetic Resonance Imaging 41: 322-328. St Pierre T.G., House M.J., Mian A., Bangma S., Burgess G., Standish R.A., Casey S., Hornsey E., Angus P.W. (2015) Machine-learned Image Analysis Models for Classifying Liver Fibrosis Stage from Magnetic Resonance Images. Hepatology 62: 607A. Pan S.J. and Yanq Q., "A Survey on Transfer...