Data Science in Radiology: A Path Forward

Artificial intelligence (AI), especially deep learning, has the potential to fundamentally alter clinical radiology. AI algorithms, which excel in quantifying complex patterns in data, have shown remarkable progress in applications ranging from self-driving cars to speech recognition. The AI application within radiology, known as radiomics, can provide detailed quantifications of the radiographic characteristics of underlying tissues. This information can be used throughout the clinical care path to improve diagnosis and treatment planning, as well as assess treatment response. This tremendous potential for clinical translation has led to a vast increase in the number of research studies being conducted in the field, a number that is expected to rise sharply in the future. Many studies have reported robust and meaningful findings; however, a growing number also suffer from flawed experimental or analytic designs. Such errors could not only result in invalid discoveries, but also may lead others to perpetuate similar flaws in their own work. This perspective article aims to increase awareness of the issue, identify potential reasons why this is happening, and provide a path forward. Clin Cancer Res; 24(3); 532–4. ©2017 AACR.

[1]  N. Lee,et al.  A concise guide to cDNA microarray analysis. , 2000, BioTechniques.

[2]  E. Topol,et al.  Adapting to Artificial Intelligence: Radiologists and Pathologists as Information Specialists. , 2016, JAMA.

[3]  Mia K Markey,et al.  Breast cancer CADx based on BI-RAds descriptors from two mammographic views. , 2006, Medical physics.

[4]  H. Aerts.  Semantics Features : Phenotype Quantification by a Radiologist ’ s Expert Eye , 2016 .

[5]  Stuart A. Taylor,et al.  Imaging biomarker roadmap for cancer studies , 2016, Nature Reviews Clinical Oncology.

[6]  P. Marsden,et al.  False Discovery Rates in PET and CT Studies with Texture Features: A Systematic Review , 2015, PloS one.

[7]  Dov Stekel.  Microarray Bioinformatics: Data Standards, Storage and Sharing , 2003 .

[8]  John Quackenbush.  Microarray data normalization and transformation , 2002, Nature Genetics.

[9]  David M. Simcha,et al.  Tackling the widespread and critical impact of batch effects in high-throughput data , 2010, Nature Reviews Genetics.

[10]  P. Lambin,et al.  Defining the biological basis of radiomic phenotypes in lung cancer , 2017, eLife.

[11]  L. Schwartz,et al.  Defining a Radiomic Response Phenotype: A Pilot Study using targeted therapy in NSCLC , 2016, Scientific Reports.

[12]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[13]  R. Gillies,et al.  Radiological Image Traits Predictive of Cancer Status in Pulmonary Nodules , 2016, Clinical Cancer Research.

[14]  Kenji Suzuki.  Machine Learning in Computer-Aided Diagnosis: Medical Imaging Intelligence and Analysis , 2012 .

[15]  Ashish Sharma,et al.  The public cancer radiology imaging collections of The Cancer Imaging Archive , 2017, Scientific Data.

[16]  P. Lambin,et al.  Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach , 2014, Nature Communications.

[17]  Ronald M. Summers,et al.  Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning , 2016, IEEE Transactions on Medical Imaging.

[18]  Ricardo A. M. Valentim,et al.  Computer-aided detection system for lung cancer in computed tomography scans: Review and future prospects , 2014, BioMedical Engineering OnLine.

[19]  Keith L. Ligon,et al.  Somatic mutations associated with MRI-derived volumetric features in glioblastoma , 2015, Neuroradiology.

[20]  Rita Strack,et al.  Highly multiplexed imaging , 2015, Nature Methods.

[21]  Paul Kinahan,et al.  Radiomics: Images Are More than Pictures, They Are Data , 2015, Radiology.

[22]  H. Aerts,et al.  Applications and limitations of radiomics , 2016, Physics in medicine and biology.

[23]  Patrick Granton,et al.  Radiomics: extracting more information from medical images using advanced feature analysis. , 2012, European journal of cancer.

[24]  John Quackenbush,et al.  Somatic Mutations Drive Distinct Imaging Phenotypes in Lung Cancer. , 2017, Cancer research.

[25]  Matthew T. Freedman,et al.  Mammographic CADx system using an image library with an intelligent agent: a pattern matching approach , 2006, SPIE Medical Imaging.

[26]  Harold P Lehmann,et al.  After Detection: The Improved Accuracy of Lung Cancer Assessment Using Radiologic Computer-aided Diagnosis. , 2016, Academic radiology.

[27]  D. Allison,et al.  Microarray data analysis: from disarray to consolidation and consensus , 2006, Nature Reviews Genetics.

[28]  Ronald M. Summers.  Deep Learning and Computer-Aided Diagnosis for Medical Image Processing: A Personal Perspective , 2017, Deep Learning and Convolutional Neural Networks for Medical Image Computing.

[29]  M. Giger,et al.  Robustness Studies of Ultrasound CADx in Breast Cancer Diagnosis , 2012 .

[30]  Brian J. Kolowitz,et al.  Technical Challenges in the Clinical Application of Radiomics. , 2017, JCO clinical cancer informatics.

[31]  Werner Dubitzky,et al.  A Practical Approach to Microarray Data Analysis , 2003, Springer US.

[32]  I. Kohane,et al.  Translating Artificial Intelligence Into Clinical Care. , 2016, JAMA.

[33]  G. A. Whitmore,et al.  Importance of replication in microarray gene expression studies: statistical methods and evidence from repetitive cDNA hybridizations. , 2000, Proceedings of the National Academy of Sciences of the United States of America.

[34]  John Quackenbush,et al.  Computational genetics: Computational analysis of microarray data , 2001, Nature Reviews Genetics.

[35]  Jason E. Stewart,et al.  Minimum information about a microarray experiment (MIAME)—toward standards for microarray data , 2001, Nature Genetics.