Geri Dön

Image analysis for contrast enhanced ultrasound carotid plaque imaging

Başlık çevirisi mevcut değil.

PDF İndir

  1. Tez No: 402224
  2. Yazar: ZEYNETTİN AKKUŞ
  3. Danışmanlar: PROF. J. G. BOSCH
  4. Tez Türü: Doktora
  5. Konular: Radyasyon Onkolojisi, Radyoloji ve Nükleer Tıp, Radiation Oncology, Radiology and Nuclear Medicine
  6. Anahtar Kelimeler: Belirtilmemiş.
  7. Yıl: 2014
  8. Dil: İngilizce
  9. Üniversite: Erasmus Universiteit Rotterdam
  10. Enstitü: Yurtdışı Enstitü
  11. Ana Bilim Dalı: Belirtilmemiş.
  12. Bilim Dalı: Belirtilmemiş.
  13. Sayfa Sayısı: 204

Özet

Özet yok.

Özet (Çeviri)

Intraplaque neovascularization (IPN) has been presented as an important biomarker for progressive atherosclerotic disease and plaque vulnerability in several pathological studies. Therefore, quantification of IPN may allow early prediction of plaque at risk of rupture and thus prevention of future cardiovascular events such as stroke. Contrast enhanced ultrasound (CEUS) enables us to detect and visualize IPN by use of ultrasound contrast agents. So, the degree of IPN can potentially be measured by quantitative imaging biomarkers derived from CEUS. Since quantification tools for IPN are scarce, so far mainly visual IPN scoring on CEUS clips has been used to assess IPN, which is subjective and tedious. Currently available commercial tools for contrast quantification, e.g. QLAB region of interest (ROI) quantification tool (Philips Medical Systems, Bothell, USA) and VueBox (Bracco Suisse SA, Geneva, Switzerland), are not suitable for quantitative analysis of IPN. These commercial quantification tools have been developed mainly for time intensity curve analysis (TIC) of large organs such as heart, liver and prostate, not for plaques. Plaques are very small and intermittently perfused. Therefore, the perfusion characteristics of plaques are quite different from those of large organs and TIC analysis as applied in large well-perfused organs is not applicable. Some IPN quantification approaches have been reported but they suffer from a number of limitations such as imaging artifacts and no or imperfect motion compensation. In this thesis work, we avoided the known limitations of IPN quantification methods reported in previous studies and developed and evaluated specialized IPN analysis tools for carotid CEUS image sequences. Chapter 1 gives an overview of diagnostic ultrasonography, contrast enhanced ultrasound, contrast detection techniques, and ultrasound image analysis. The scope and outline of the thesis are also described in more detail. Carotid images exhibit considerable motion such as breathing, pulsation, probe or patient motion. For accurate quantification of IPN, motion compensation is a prerequisite step. In Chapter 2, we present a motion compensation tool tailored for plaques. A rigid motion compensation of a local region based on block matching combined with multidimensional dynamic programming is introduced to follow the plaque ROI over time. It is shown that the method is sufficiently accurate and successful for in-vivo application. However, our rigid technique only compensates for translational movements but not for rotation and deformation, and also gives only a local motion estimate. In Chapter 3, we present an IPN quantification method which detects and tracks individual contrast spots to reconstruct microvascular network paths. This method allows distinguishing between stationary and moving objects based on their displacement over time. However, microvessels that cross the imaging plane almost perpendicularly might be considered as stationary objects or artifacts and this would lead to underestimation of IPN. In Chapter 4, we investigate plaque perfusion with motion compensated TIC analysis and time integrated parametric images such as maximum intensity projection (MIP). Also, several quantitative parameters, derived from TIC and MIP, are studied to replace qualitative visual IPN scores. However, if there are artifacts within the plaque ROI, the MIP image will include them as well. This will cause overestimation of IPN area. To prevent this, we analyze the structure of the microvasculature network in Chapter 3 and apply statistical contrast spot segmentation methods in Chapter 5. We introduce a statistical contrast spot segmentation method in Chapter 5. This method classifies intensities within the plaque ROI into 4 classes: background, intermediate, contrast spot and artifact class. Classifying intensities in each time frame and applying spatiotemporal analysis handles several issues raised by other methods and allows accurate quantification of IPN. It is shown that the method is more robust to artifacts in chapter 6. In Chapter 6, we present a dedicated software package that includes all the developed carotid IPN quantification tools, called Carotid Intraplaque Neovascularization Quantification Software (CINQS). CINQS is designed as a special-purpose platform for IPN quantification tools for carotid plaques. CINQS was developed in a modular and extensible way and provides a user-friendly tool for analyzing carotid IPN in CEUS. CINQS could also be used to check changes in neovascularization over time, and to check the outcome of novel therapies on neovascularization. To the best of our knowledge, this is the first software package dedicated to carotid IPN quantification. In Chapter 7, we present a nonrigid motion estimation method to estimate and compensate the complex motion in the complete B-mode ultrasound (BMUS) and CEUS images. This method exploits the combined information from simultaneously acquired BMUS&CEUS images to overcome the difficulties of the separate use of BMUS and CEUS. The method generates a detailed nonrigid deformation field for the complete BMUS and CEUS images. This not only allows us to track the plaques but also to improve image quality by time averaging of the registered image sequences. This provides“epitome”images of improved quality that we employ for an accurate segmentation of the lumen-intima (LI) interface in subjects with atherosclerotic arteries, using the joint intensity histogram of BMUS and CEUS epitomes and graph based segmentation. Carotid plaque segmentation in BMUS and CEUS images is crucial to assess plaque morphology and composition, which are linked to plaque vulnerability. In Chapter 8, we present a novel and fully automatic plaque segmentation technique in simultaneously acquired BMUS and CEUS images. The LI interface is segmented as described in Chapter 7. The media-adventitia (MA) interface from the BMUS epitome is segmented by using multidimensional dynamic programming for parallel curves. Our plaque segmentation method allows several applications such as fully automated IPN quantification, arterial distensibility, and plaque characterization. To the best of our knowledge, this is the first method segmenting carotid plaques in combined BMUS and CEUS images. In Chapter 9, we describe a clinical study, assessment of IPN in patients with familial hypercholesterolemia, using the software described in Chapter 6 and quantification tools described in Chapters 3-5. It was shown that irregular and ulcerated plaques exhibited significantly more IPN than plaques with a smooth surface. In Chapter 10, we discuss the merits of the developed IPN analysis tools and provide the future perspectives and conclusions of this thesis work.

Benzer Tezler

  1. Tez No

    438612

    Görüntü füzyonu yöntemleri ile karaciğer lezyon görüntülerinin değerlendirilmesi

    Evaluation of liver lesion images with image fusion methods

    SAİM ERVURAL

    Yüksek Lisans

    Türkçe

    Türkçe

    2016

    Elektrik ve Elektronik MühendisliğiSelçuk Üniversitesi

    Elektrik-Elektronik Mühendisliği Ana Bilim Dalı

    YRD. DOÇ. DR. MURAT CEYLAN

  2. Tez No

    223031

    Ses ötesi görüntülemenin benzetimi

    Simulation of ultrasound imaging

    LALEHAN CANDEMİR

    Yüksek Lisans

    Türkçe

    Türkçe

    2007

    Biyomühendislikİstanbul Teknik Üniversitesi

    Elektronik ve Haberleşme Mühendisliği Ana Bilim Dalı

    PROF.DR. İNCİ ÇİLESİZ

  3. Tez No

    115827

    Dinamik manyetik rezonans görüntüleme ve ultrason kontrast ajanlı power doppler ultrasonografi ile meme lezyonlarında benign ve malign ayrımı

    Başlık çevirisi yok

    FAİK SUNGURLU

    Tıpta Uzmanlık

    Türkçe

    Türkçe

    2002

    OnkolojiMarmara Üniversitesi

    Radyodiagnostik Ana Bilim Dalı

    DOÇ.DR. ERKİN ARIBAL

  4. Tez No

    365907

    Meme parankiminde kontrastlı manyetik rezonans inceleme ile tespit edilen kitlelerin yeni bir metod olan volume navigasyon tekniğiyle ultrasonografik görüntülemede yerinin tespit edilmesi ve bu teknikle lezyonların yerinin doğru bir şekilde tespit edilebilirliğinin ve yeterliliğinin araştırılması

    Evaluati̇on of feasi̇bi̇li̇ty and accuracy of a new method volume navi̇gati̇on techni̇que for a lesi̇on i̇n breast parenchyma detected by contrast-enhanced magneti̇c resonance i̇magi̇ng and fi̇ndi̇ng lesi̇on locali̇zati̇on i̇n ultrasound usi̇ng thi̇s techni̇que

    FİKRET KÜÇÜKKAYA

    Tıpta Uzmanlık

    Türkçe

    Türkçe

    2014

    Radyoloji ve Nükleer TıpMarmara Üniversitesi

    Dahili Tıp Bilimleri Bölümü

    PROF. DR. MUSTAFA ERKİN ARIBAL

  5. Tez No

    873723

    MR inceleme ile malignite şüphesi olan myoma uteri olgularının cerrahi sonrası doku tanılarının radyolojik bulgular ile karşılaştırılması

    Comparison of after surgical histopathological diagnoses of uterine myoma cases with suspi̇ci̇on of malignancy in MRI with radiological findings

    NÜKHET KILIÇ HİMMETOĞLU

    Tıpta Uzmanlık

    Türkçe

    Türkçe

    2024

    Kadın Hastalıkları ve DoğumSağlık Bilimleri Üniversitesi

    Kadın Hastalıkları ve Doğum Ana Bilim Dalı

    PROF. DR. MURAT EKİN

Geri Dön