ERROR ESTIMATION IN QUANTITATIVE MEDICAL IMAGE ANALYSIS

MADAN, HENNADII (author), Pernuš, Franjo (mentor)

An impressive improvement of the effectiveness of medical care evident in the recent decades is to a large extent driven by the progress made in the fields of medical imaging and analysis. A hallmark characteristic of this trend is the transition from a purely visual, qualitative assessment of the medical images to a computational and more quantitative assessment, which involves in vivo image-based measurements. In the domains of disease diagnosis and monitoring, treatment efficacy assessment, but also in surgery and radiotherapy planning and execution, the clinical workflow is becoming increasingly dependent on image-derived measurements (i.e. imaging biomarkers). Some of the quantitative imaging biomarkers have already become well established as surrogates of clinical outcomes. The values of these imaging biomarkes may directly impact the decision-making process | hence, the accuracy and precision of the methods that extract the measurements from the images need to be rigorously validated. Problems of objective validation and comparison of measurement methods feature prominently in the medical imaging discourse. In registration and segmentation | the two major fields of image analysis, the state of the art of method validation and comparison is based on reference measurements usually requiring some human involvement. In case of registration it is the detection and manual localization of fiducial markers. For segmentation it is the manual delineation of anatomical structures by expert radiologists. Certain problems are inherent in this approach: humans are subjective { measurements by different experts usually disagree, they are error prone | they get distracted and tired, and their time is costly. When human errors in validation standards propagate to medical practice they acquire a potential to cause costly damages. The patients, medical care establishments and the economy at large are all impacted by the consequences of these errors. Strategies to predicting and preventing the measurement errors and cutting the costs associated with validation and comparison of measurement methods are discussed in this Thesis. A direct strategy to alleviate the costs of the burdensome manual reference creation is through automation. Such strategy was applied in the first contribution of this Thesis using a novel automated computational approach to gold standard reference dataset creation for validating rigid-body registration of pre-operative 3D and intra-operative 2D images. Therein, the use of automatic image analysis pipeline eliminated the need for human interaction and manual input, previously required in a semi-automated approach. This has significantly improved the registration accuracy as validated on intra-operatively acquired 3D and 2D images of twenty patients with cerebral aneurysms and arteriovenous malformations. A different, more inventive, strategy is to validate the measurement methods without ever creating a reference, through advanced statistical inference. Two new reference-free Bayesian frameworks for estimating the systematic and random errors of an ensemble of (automated) measurement methods, are developed in this Thesis. They facilitate the validation and comparison of measurement methods without requiring costly reference measurements. A clear advantage of this strategy is that it eliminates the need for the reference measurements altogether and therefore annihilates the associated costs. For instance, in the image analysis domain, applying several automated methods to a certain dataset requires only computational resources, which is much cheaper than engaging an expert to manually create the reference. The two proposed frameworks were successfully validated on several synthetic and on relevant clinical datasets, involving imaging biomarkers of neurological diseases. Theoretical developments of one of the proposed frameworks allow to use it for advanced applications of estimation of latent true values of an unobserved quantity and selection of best predictors for it from a set of related biomarkers. In conclusion, the contributions of this Thesis do not only solve the practical problems of reference creation, but address the conceptual problems associated with reference based error estimation. The two proposed and validated Bayesian frameworks represent important theoretical advances in the emerging field of reference-free error estimation, making this methodology practical for measurement method validation, comparison and further beyond | for selection of best predictors of unobservable quantities and the their estimation.

• Language: English
• Keywords: -
• Work type: Doctoral dissertation
• Organization: FE - Faculty of Electrical Engineering
• Year: 2018
• PID: 20.500.12556/RUL-102002
• COBISS.SI-ID: 12105812
• Publication date in RUL: 19.07.2018

Secondary language

• Language: Slovenian
• Title: OCENJEVANJE NAPAK KVANTITATIVNE ANALIZE MEDICINSKIH SLIK

Abstract

Izjemen napredek v učinkovitosti medicinske oskrbe v zadnjih desetletjih v veliki meri poganjajo napredki v povezanih in vzporednih področjih medicinskih slikovnih tehnologij in računalniške analize slik. Ena izmed poglavitnih značilnosti tega trenda je prehod iz povsem vizualnega, kvalitativnega vrednotenja medicinskih slik v bolj računsko in kvantitativno vrednotenje. Slednje vključuje predvsem in vivo meritve računsko izluščene iz medicinskih slik. V kontekstu diagnoze in spremljanja razvoja bolezni ter vrednotenja učinkovitosti zdravljenja, pa tudi v kontekstu načrtovanja in izvedbe kirurških posegov ter radioterapije, klinični protokoli in smernice vedno bolj temeljijo na meritvah iz medicinskih slik (i.e. slikovni biomarkerji). Nekateri slikovnih biomarkerji so se že uveljavili kot nadomestki kliničnih ciljev. Vrednosti slikovnih biomarkerjev torej lahko neposredno vplivajo na odločanje v omenjenih kliničnih kontekstih | zato mora biti natančnost in točnost postopkov izločanja slikovnih biomarkerjev rigorozno validirana. Problematika objektivnega vrednotenja in primerjave postopkov merjenja je zelo izražena na področju medicinskih slikovnih tehnologij. Pri poravnavi in razgradnji slik | dveh glavnih metodoloških pristopov k analizi medicinskih slik, so uveljavljeni načini validacije in primerjave sposobnosti teh postopkov osnovani na uporabi referenčnih meritev. Referenčne meritve običajno pridobimo ročno s pomočjo eksperta. Pri poravnavi slik je to lahko očno zaznavanje in ročno označevanje oslonilnih markerjev na slikah, pri razgradnji pa je to na primer ročno obrisovanje anatomskih struktur, kar lahko naredi izkušen radiolog. V tem procesu je kritičen subjektiven doprinos posameznega eksperta | različni eksperti bodo izluščili različne vrednosti meritev, izključena ni niti možnost večjih napak in razhajanj, kot posledica utrujenosti in naključnih dejavnikov. čas, ki ga porabi ekspert je tudi zelo drag. Potencialno zelo drage so lahko tudi posledice prej omenjenih napak in razhajanj v meritvah, ker vplivajo na medicinsko prakso in lahko povzročijo resne posledice. Tako bolniki, kot bolnišnice in družba nasploh lahko čutijo posledice teh napak. Razprava in razvoj strategij za napovedovanje in preprečevanje merilnih napak in hkratno zmanjševanje stroškov pri validaciji in primerjavi postopkov merjenja predstavljajo jedro te doktorske disertacije. Direktna strategija manjšanja stroškov je preko manjšanja bremena bremena ustvarjanja reference, kar lahko dosežemo z avtomatizacijo. Zato je prvi prispevek te disertacije nov avtomatski računski pristop za ustvarjanje reference oziroma zlatega standarda za validacijo toge poravnave med pred-operativnimi 3D in med-operativnimi 2D slikami. Z uporabo verige avtomatskih postopkov analize slik smo odpravili potrebo po interakciji z operaterjem in morebitne ročne vnose, kar je bilo sicer potrebno pri predhodnem pol-avtomatskem pristopu. Na ta način smo signikantno izboljšali natančnost referenčne poravnave, kot kažejo rezultati validacije pristopa na med-operativno zajetih 3D in 2D slikah dvajsetih bolnikov z možganskimi anevrizmami in arteriovenoznimi malformacijami. Povsem drugačna in bolj inovativna strategija je validacija postopkov merjenja brez uporabe reference, in sicer z uporabo naprednega statističnega sklepanja. V disertaciji predlagamo dva nova Bayesianska pristopa za oceno sistematičnih in naključnih napak množice (avtomatskih) postopkov merjenja neke količine. Naprimer, v kontekstu slikovnih biomarkerjev je stroškovno precej bolj učinkovito na določeni zbirki slik uporabiti več različnih avtomatskih postopkov analize medicinskih slik kot pa pridobiti referenco s pomočjo eksperta. Pristopa smo uspešno validirali na več sintetičnih in kliničnih zbirkah podatkov, kjer so slednje vključevale meritve slikovnih biomarkerjev nevroloških bolezni. Teoretična dognanja v enem izmed predlaganih pristopov omogočajo tudi ocenjevanje vrednosti latentne količine in hkratno izbiro najboljših napovednih meritev te količine. Prispevki te disertacije ne le rešujejo praktične probleme pri ustvarjanju reference, pač pa naslavljajo tudi prikrite konceptualne probleme kot je napaka reference. Dva predlagana in validirana Bayesianska računska pristopa predstavljata pomembne teoretične preskoke v okviru novonastalega področja ocenjevanja napake brez reference in s katerima je ta postala praktično uporabna za namen validacije in primerjave sposobnosti postopkov merjenja nasplošno. še več, eden izmed pristopov omogoča tudi določanje napovedni vrednosti meritev glede na latentno količino in tudi oceno njene vrednosti.

• Keywords: -