Search results
(1 - 2 of 2)
- Title
- A three-dimensional tissue molecular imaging system based on angular domain optical projection tomography: Applications in lymph node biopsy
- Creator
- Torres, Veronica Calliste
- Date
- 2020
- Description
-
Sentinel lymph node biopsy is a good prognostic factor for several cancers as therapeutic decisions are often determined by the results....
Show moreSentinel lymph node biopsy is a good prognostic factor for several cancers as therapeutic decisions are often determined by the results. Despite this importance, false negatives remain common because of standard pathology procedures that aim only to detect macrometastases (> 2 mm diameter) and leave more than 99% of lymph node volumes unassessed. While it is possible to section tissue samples more thoroughly, a subsequent 10x increase in pathologist read time is undesirable. Therefore, a more sensitive and rapid approach for lymph node evaluation is warranted.Our proposed solution was the development of an angle-restricted optical projection tomography system to provide high resolution quantitative imaging of whole lymph nodes prior to conventional pathology. Two main strategies were employed: 1) early photon imaging achieved with angular restriction to minimize the number of detected multiply scattered photons that add to imaging blur; and 2) paired-agent molecular imaging, which can quantify targeted biomolecule concentrations through co-administration of targeted and control imaging agents.This thesis focused primarily on the first aspect; however, all work was performed with paired-agent imaging in mind, such that the technique can be implemented directly in future studies. The first chapter presents a proof-of-concept that verifies the utility of angle-domain imaging for evaluation of low scattering lymph nodes. Filtered backprojection and strict angle restriction for scatter rejection were sufficient to detect and localize clinically relevant metastases. In the second chapter, improvements were made to the system so that detection efficiency could be improved, and the system was more rigorously characterized in terms of reconstruction accuracy and limits of detection. Finally, the third chapter presents the investigation of alternate reconstruction techniques to push the limits of achievable resolution and image quality. The overall findings of this work demonstrate the potential for an angle-restricted tomography system to provide significant improvements of metastases detection sensitivity in excised lymph nodes compared to conventional pathology at a fraction of the time and cost.
Show less
- Title
- DEEP LEARNING IMAGE-DENOISING FOR IMPROVING DIAGNOSTIC ACCURACY IN CARDIAC SPECT
- Creator
- Liu, Junchi
- Date
- 2022
- Description
-
Myocardial perfusion imaging (MPI) using single-photon emission computed tomography (SPECT) is a noninvasive imaging modality widely utilized...
Show moreMyocardial perfusion imaging (MPI) using single-photon emission computed tomography (SPECT) is a noninvasive imaging modality widely utilized for diagnosis of coronary artery diseases (CAD) in nuclear medicine. Because of the concern of potential radiation risks, the imaging dose administered to patients is limited in SPECT-MPI. Due to the low count statistics in acquired data, SPECT images can suffer from high levels of noise. In this study, we investigate the potential benefit of applying deep learning (DL) techniques for denoising in SPECT-MPI studies. Owing to the lack of ground truth in clinical studies, we adopt a noise-to-noise (N2N) training approach for denoising in full-dose studies. Afterwards, we investigate the benefit of applying N2N DL on reduced-dose studies to improve the detection accuracy of perfusion defects. To address the great variability in noise level among different subjects, we propose a scheme to account for the inter-subject variabilities in training a DL denoising network to improve its generalizability. In addition, we propose a dose-blind training approach for denoising at multiple reduced-dose levels. Moreover, we investigate several training schemes to address the issue that defect and non-defect image regions are highly unbalanced in a data set, where the overwhelming majority by non-defect regions tends to have a more pronounced contribution to the conventional loss function. We investigate whether these training schemes can effectively improve preservation of perfusion defects and yield better defect detection accuracy. In the experiments we demonstrated the proposed approaches with a set of 895 clinical acquisitions. The results show promising performance in denoising and improving the detectability of perfusion-defects with the proposed approaches.
Show less