BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260728T174355EDT-1725BN4GWM@132.216.98.100
DTSTAMP:20260728T214355Z
DESCRIPTION:Abstract\n\nThis thesis presents\, in sum\, the first image-bas
 ed longitudinal deep learning precision medicine model. The model uses mod
 ern deep learning methods and classical treatment effect estimation theory
  to estimate individual treatment effects (ITE) using high-dimensional pat
 ient MRI sequences as input. Before\, treatment effect analysis focused on
  differences across entire populations of treated patients or subgroups of
  patients with simple distinguishing factors\, e.g.\, sex. With unique pat
 ient data\, such as MRI sequences\, treatment effects can be estimated for
  the individual\, which offers a fine-grained estimate of patient response
 . Currently\, the models built using high-dimensional imaging data have on
 ly focused on predicting disease outcomes for only a single treatment. The
 se models help doctors manage patient risk\, but predicting a single outco
 me only is insufficient for treatment effect estimation. Treatment effect 
 analysis takes a step beyond single outcome prediction by predicting outco
 mes for all potential therapies.\n\nTo achieve the goal of personalized tr
 eatment predictions\, we developed the first imaging-based precision medic
 ine model. The model is trained on data spanning different stages of multi
 ple sclerosis (MS) and has been designed to predict several clinically rel
 evant outcomes. Our first advancement estimates the number of future brain
  lesions for the Relapsing Remitting form of MS (RRMS) on five treatments 
 and placebo. We demonstrate accurate prediction when the model makes a pre
 diction for the patient's assigned therapy (factual outcome) and on the id
 entification of groups that have heterogeneous responses to treatment. We 
 deem such patients responders. In some cases\, we found subgroups of respo
 nders whose outcomes were statistically significantly improved\, even for 
 drugs that did not have significant effects at the group level.\n\nOur nex
 t advancement better aids clinical decision-making by quantifying and vali
 dating the uncertainty in predictions of treatment effect. Uncertainty qua
 ntification is critical in high-risk applications\, such as treatment reco
 mmendation\, as time spent on non-working treatments can have long-lasting
  irreparable effects. We show that treatment recommendations made with add
 itional uncertainty information improves decision-making and leads to bett
 er outcomes when treating a population of patients.\n\nThe final contribut
 ion models the disease trajectory over time. Modeling the trajectory of pa
 tients becomes far more important when considering longitudinal diseases\,
  where the intermediate outcomes can be just as important as the outcomes 
 at the end. As MS is a chronic disease\, predicting the entire trajectory 
 better estimates the patient's experience throughout the disease. We valid
 ate these predictions on new lesion development and an outcome that better
  reflects the patient's quality of life\, the Expanded Disability Status S
 cale (EDSS). The EDSS is more difficult to predict as it is not an image-b
 ased measure of MS\; however\, it is the primary outcome of interest in th
 e Second Progressive (SPMS) and Primary Progressive (PPMS) subtypes of MS.
  Finally\, it should be noted that while this work focuses on multiple scl
 erosis outcomes\, the framework and design considerations can be applied b
 roadly for any personalized medicine tools built using complex data.\n
DTSTART:20250314T140000Z
DTEND:20250314T160000Z
LOCATION:Room 603\, McConnell Engineering Building\, CA\, QC\, Montreal\, H
 3A 0E9\, 3480 rue University
SUMMARY:PhD defence of Joshua Durso-Finley – Uncertainty Aware Causal Model
 s for Prediction of Individual Treatment Effects for Image-based Precision
  Medicine
URL:/ece/channels/event/phd-defence-joshua-durso-finle
 y-uncertainty-aware-causal-models-prediction-individual-treatment-363950
END:VEVENT
END:VCALENDAR