CAD systems CAD systems and and e-learning e-learning using using in in radiologist radiologist training training HealthCom 2013 César Suárez Ortega cesar.suarez@externos.ciemat.es CETA-CIEMAT Student designed by Piotrek Chuchla from The Noun Project

Breast cancer Breast cancer

Breast cancer Breast cancer high mortality high mortality

Breast cancer Breast cancer high mortality high mortality 1 1 of of 6 6 deaths from cancer in women * deaths from cancer in women * * I.C Moreira et al. “Inbreast: Toward a Full-field Digital Mammographic Database”, Academic Radiology. 19 (2012) 236-248 * I.C Moreira et al. “Inbreast: Toward a Full-field Digital Mammographic Database”, Academic Radiology. 19 (2012) 236-248

early detection early detection Best practice to reduce mortality * C. H. Lee, “Screening mammography: proven benefit, continued controversy”, Radiologic clinics of North America, vol. 40 * C. H. Lee, “Screening mammography: proven benefit, continued controversy”, Radiologic clinics of North America, vol. 40 * *

Double reading Double reading of mammograms of mammograms

Double reading Double reading of mammograms of mammograms Costly! Costly! Money designed by Atelier Iceberg from The Noun Project

Double reading Double reading of mammograms of mammograms Not enough Not enough radiologists! radiologists! Doctor designed by Piotrek Chuchla from The Noun Project * * * Kolb TM et al.; “Comparison of the performance screening mammography, physical examination, and * Kolb TM et al.; “Comparison of the performance screening mammography, physical examination, and breast US and evaluation of factors that influenve them: an analysis of 27, 825 patient evaluations”. Radiology. breast US and evaluation of factors that influenve them: an analysis of 27, 825 patient evaluations”. Radiology. 2002; 225(1): 165-175 2002; 225(1): 165-175

Any alternatives?

C Computer omputer D Diagnosis iagnosis A Aided ided

C Computer omputer D Diagnosis iagnosis A Aided ided Cancer Cancer detection detection rate rate Thumbs Up designed by Alv Jørgen Bovolden from The Noun Project

C Computer omputer D Diagnosis iagnosis A Aided ided RECALL RECALL rate rate Thumbs Up designed by Alv Jørgen Bovolden from The Noun Project

“ “Medical student attitudes toward radiology are Medical student attitudes toward radiology are significantly affected by early exposure to significantly affected by early exposure to didactic teaching in radiology didactic teaching in radiology” ” I.C Moreira et al. “Inbreast: Toward a Full-field Digital Mammographic Database”, Academic Radiology. 19 (2012) 236-248 I.C Moreira et al. “Inbreast: Toward a Full-field Digital Mammographic Database”, Academic Radiology. 19 (2012) 236-248 Students designed by Piotrek Chuchla from The Noun Project

E-learning & Medical education » Proven benefits Proven benefits » Increases Increases the the retention rate retention rate » Saves Saves costs costs » Improves the Improves the efficiency efficiency using using ICT ICT » Medical Informatics Medical Informatics » Buzzwords: Buzzwords: » Immediate Evaluation Immediate Evaluation » Simulated Cases Simulated Cases » Videoconferencing Videoconferencing » Virtual Patients Virtual Patients » Custom Custom applications needed applications needed Education designed by Patrick Trouvé from The Noun Project

E-learning & radiology » Lecture Lecture of radiographies of radiographies » Repository with Repository with real cases real cases » Auto-evaluation Auto-evaluation tools tools » Simulation Simulation cases cases » There is There is a need a need of applications of applications

E-learning & Radiology & CAD SYSTEMS Idea from The Noun Project

E-learning & CAD SYSTEMS » Very Very powerful powerful » Auto-evaluation Auto-evaluation » Familiriaty Familiriaty with CAD Systems with CAD Systems » Recall rate not so critical Recall rate not so critical » Don't replace Don't replace teachers teachers

aloe

aloe A platFORm for radiologist training In breast cancer diagnosis

Aloe objectives Aloe objectives » E-learning E-learning platform platform » Auto-evaluation Auto-evaluation using Machine Learning Classifiers using Machine Learning Classifiers » Modulable Modulable Multiple → Multiple → frontends frontends + One + One backend backend » Lots of Lots of anonymized real cases anonymized real cases! ! » Workflow Workflow designed specifically for radiologists designed specifically for radiologists Responsive Design designed by Diego Naive from The Noun Project Target designed by Julia Stoffer from The Noun Project

Aloe pillars Aloe pillars B Breast reast C Cancer ancer D Database atabase R Repository epository M Mammography ammography I Image mage W Workstation for orkstation for A Analysis and nalysis and D Diagnosis iagnosis M Machine achine L Learning earning C Classifiers lassifiers Pillar designed by Cengiz SARI from The Noun Project * * * http://bcdr.eu * http://bcdr.eu

Currently Currently » REST backend → REST backend → 50% 50% » MLCs → MLCs → 100% 100% » Web frontend → Web frontend → 10% 10% » Workflow → Workflow → 5% 5% » Release a first version Release a first version » UX tests UX tests » Virtual classroom Virtual classroom » Test it with students! Test it with students! To Be Done To Be Done Worker designed by James Fenton from The Noun Project

conclusions conclusions » CAD systems are a CAD systems are a solid technology solid technology. . » Highly-skilled Highly-skilled radiologists are needed radiologists are needed. . » Better tools for training Better tools for training are needed. are needed. » CAD systems for auto-evaluation CAD systems for auto-evaluation. . » Further research Further research is needed. is needed. Education designed by Pete Fecteau from The Noun Project

PARTNERS PARTNERS acknowledgments to acknowledgments to

Any questions? Thanks! Question designed by Martin Delin from The Noun Project