BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Using machine learning approaches to automate the diagnosis of sma
 ll intestinal biopsies   - Professor Elizabeth Soilleux\, MA\, MB BChir\, 
 PhD\, FRCPath\, PGDipMedEd\, SFHEA  Professor of Diagnostics and Biomarker
 s/ Honorary Consultant in Pathology\, University of Cambridge
DTSTART:20251124T120000Z
DTEND:20251124T123000Z
UID:TALK240184@talks.cam.ac.uk
CONTACT:Sam Nallaperuma-Herzberg
DESCRIPTION:Histopathology is a clinical discipline in which pathologists 
 look down microscopes at biopsies\, to make a diagnosis. Nowadays many dep
 artments scan the glass microscope slides\, so that pathologists can view 
 them on a screen. This opens up the possibility of automating histopatholo
 gical diagnosis\, particularly as there are international shortages of pat
 hologists\, leading to backlogs and delays. We chose the duodenum (small i
 ntestine) as a starting point due to its low medicolegal risk (<0.6% biops
 ies contain cancer)\, stereotyped nature of biopsies and the fact that jus
 t 2 diagnostic categories account for 85% of the total.  Furthermore\, the
  limited agreement among pathologists about these two diagnoses provides t
 he opportunity to improve accuracy. \n\nWe apply a series of processing st
 eps\, including artefact removal\, division into small tiles and colour no
 rmalisation\, before applying a multiple instance learning approach to bio
 psy classification\, leading to accuracy >97% against a carefully curated 
 ground truth. In order to facilitate adoption of this technology by pathol
 ogists and their clinical colleagues\, we are now working on making our ca
 tegorisation processes more explainable. We have spun out a company\, Lyze
 um Ltd\, to progress our software to market.
LOCATION:SS03 Seminar Room\, Willam Gates building (Department of Computer
  Science and Technology)
END:VEVENT
END:VCALENDAR