Detection of small joint arthritis with corrected spectral images

Rogelj, Luka (author), Simončič, Urban (mentor)

The main aim of the dissertation was to detect rheumatoid arthritis using spectral images. The organization of the work follows the four specific aims: (I) development and validation of the hyperspectral and multispectral imaging systems with integrated 3D profilometry; (II) development of algorithms for spectral image correction and tissue physiological parameters extraction; (III) application of the developed methods for imaging arthritis patients; (IV) evaluation of the developed methods for arthritis diagnosis of the small joints. In the first chapter, the characteristics of arthritis disease are presented. The clinical picture and pathology are described along with treatment options. Possible options for arthritis detection are introduced, with emphasis on optical imaging techniques, following the brief theory of light and tissue interaction. The second chapter presents the characterization and development of multispectral and hyperspectral imaging systems. Spectral and spatial verification using common imaging resolution standards is described. In the third chapter, curvature and height correction methods are presented and applied to various objects. Tissue phantoms were used to evaluate the effect of the correction on the reflected light and, consequently, on the extracted optical parameters The fourth chapter describes a clinical study in which healthy and arthritic patients were imaged with diffuse reflectance and transmission imaging. Reflectance images are analyzed with statistical parametric mapping. For transmission images, the effective thickness as a measure of the disease severity is extracted from the images. In the last chapter, the results are discussed and compared with the existing research on similar topics. Finally, future plans for improving imaging systems and analysis methods are presented.

• Language: English
• Keywords: Arthritis, hyperspectral imaging, multispectral imaging, tissue phantoms, biomedical optics
• Work type: Doctoral dissertation
• Typology: 2.08 - Doctoral Dissertation
• Organization: FMF - Faculty of Mathematics and Physics
• Year: 2022
• PID: 20.500.12556/RUL-142775
• COBISS.SI-ID: 131068419
• Publication date in RUL: 25.11.2022

Secondary language

• Language: Slovenian
• Title: Zaznavanje artritisa malih sklepov s popravljenimi spektralnimi slikami

Abstract

Glavni cilj doktorskega dela je odkrivanje revmatoidnega artritisa z uporabo spektralnih slik. Delo je organizirano v skladu s štirimi specifičnimi cilji: (I) razvoj in karakterizacija hiperspektralnega in multispektralnega slikovnega sistema z integrirano 3D profilometrijo; (II) razvoj algoritmov za korekcijo spektralnih slik in ekstrakcijo fizioloških parametrov tkiva; (III) uporaba razvitih metod za slikanje artritičnih bolnikov; (IV) validacija razvitih metod za diagnostiko artritisa malih sklepov. V prvem poglavju predstavimo glavne značilnosti artritisa. Opišemo klinično sliko in patologijo, nato pa še možnosti zdravljenja bolezni. Zadnji del poglavja namenjamo slikovnim tehnikam za odkrivanje artritisa s poudarkom na optičnih slikovnih tehnikah. V drugem poglavju predstavimo razvoj in karakterizacijo multispektralnih in hiperspektralnih slikovnih sistemov. Sistema sta spektralno in prostorsko validirana z uporabo običajnih resolucijskih standardov. V tretjem poglavju predstavimo metode za korekcijo ukrivljenosti površine in višine. Metodo smo uporabili na različnih ukrivljenih vzorcih. Za oceno učinka korekcije na odbito svetlobo in posledično na pridobljene optične parametre smo uporabili tkivne fantome. V četrtem poglavju opišemo klinično študijo, pri kateri smo z uporabo odbojne in presevne tehnike slikali zdrave in bolne prostovoljce. Odbojne slike smo analizirali s statistično parametrično kartografijo. Pri presevnih slikah pa smo kot merilo resnosti bolezni izluščili efektivno debelino. V zadnjem poglavju diskutiramo o dobljenih rezultatih in jih primerjamo z že obstoječimi študijami. Na koncu predstavimo še načrte za prihodnost, ki bi potencialno lahko izboljšale slikovne sisteme in metode za analizo.

• Keywords: Artritis, hiperspektralno slikanje, multispektralno slikanje, tkivni fantomi, biomedicinska optika