BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260605T154540Z
LOCATION:Bldg. 6 - Room 104
DTSTART;TZID=Europe/Stockholm:20260701T093000
DTEND;TZID=Europe/Stockholm:20260701T100000
UID:submissions.pasc-conference.org_PASC26_sess154_msa252@linklings.com
SUMMARY:Trust and Transparency From Pipeline to Practice: Foundations for 
 Robust Clinical AI
DESCRIPTION:Caroline Chung (UT MD Anderson Cancer Center)\n\nDeployment of
  AI in clinical environments demands more than computational capabilities,
  it requires systematic commitment to data integrity and a deliberate stra
 tegy for building clinician trust. This talk examines two interdependent p
 illars of responsible clinical AI: the critical role of data in context ac
 ross the full model lifecycle, and the collaborative practices that transl
 ate algorithmic performance into clinical confidence.\n\nTransparency abou
 t  data is as important as transparency about models. Robust clinical AI r
 equires explicit documentation of the characteristics, provenance, and lim
 itations of training, validation, and test datasets. Performance at implem
 entation is directly impacted by how closely real-world data resembles the
  development data, including patient populations, case-mix, and technical 
 heterogeneity. Rich contextual metadata, capturing how, when, and where da
 ta were collected, deepens understanding of the performance variation acro
 ss contexts to anticipate where models will generalize or fail. Making the
 se dependencies visible to developers and clinicians is a prerequisite for
  trustworthy deployment. Clinicians who understand a model's intended scop
 e, known failure modes, and uncertainty are better positioned to use AI ou
 tputs appropriately. Rigorous data stewardship and structured trust-buildi
 ng together offer a pathway to clinical AI that is not only performant but
  genuinely fit for practice.\n\nDomain: Applied Social Sciences and Humani
 ties, Life Sciences, Computational Methods and Applied Mathematics\n\nSess
 ion Chairs: Justin M. Wozniak (Argonne National Laboratory, University of 
 Chicago) and Thomas Brettin (Argonne National Laboratory, University of Ch
 icago)\n\n
END:VEVENT
END:VCALENDAR