With 50% of all interventional procedures in the US being minimally invasive, there is a need for objective tools to help guide surgeons in this challenging environment. Tissue oxygenation is a useful biomarker of tissue viability and suitable for surgical guidance. Here we present our efforts to perform real-time quantitative optical imaging through a rigid endoscope using Single Snapshot of Optical Properties (SSOP) imaging. In particular, in this work we introduce for the first time 3 dimensionally-corrected dual wavelength optical properties imaging using SSOP through an endoscope, allowing accurate oxygenation maps to be obtained on tissue simulating phantoms and in vivo samples. We compared the results with state-of-the-art wide-field spatial frequency domain imaging (SFDI). Overall, results from the novel endoscopic imaging system agreed within 10% in absorption, reduced scattering, and oxygenation. Moreover, we introduce here real-time, video-rate quantitative optical imaging with 3D profile correction through an endoscope. These results demonstrate the potential of endoscopic SSOP as an objective surgical guidance tool for the clinic.
Alzheimer's disease (AD) is one of the most frequent forms of dementia and an increasing challenging public health problem. In the last two decades, structural magnetic resonance imaging (MRI) has shown potential in distinguishing patients with Alzheimer's disease and elderly controls (CN). To obtain AD-specific biomarkers, previous research used either statistical testing to find statistically significant different regions between the two clinical groups, or ℓ1 sparse learning to select isolated features in the image domain. In this paper, we propose a new framework that uses structural MRI to simultaneously distinguish the two clinical groups and find the bio-markers of AD, using a group lasso support vector machine (SVM). The group lasso term (mixed ℓ1- ℓ2 norm) introduces anatomical information from the image domain into the feature domain, such that the resulting set of selected voxels are more meaningful than the ℓ1 sparse SVM. Because of large inter-structure size variation, we introduce a group specific normalization factor to deal with the structure size bias. Experiments have been performed on a well-designed AD vs. CN dataset1 to validate our method. Comparing to the ℓ1 sparse SVM approach, our method achieved better classification performance and a more meaningful biomarker selection. When we vary the training set, the selected regions by our method were more stable than the ℓ1 sparse SVM. Classification experiments showed that our group normalization lead to higher classification accuracy with fewer selected regions than the non-normalized method. Comparing to the state-of-art AD vs. CN classification methods, our approach not only obtains a high accuracy with the same dataset, but more importantly, we simultaneously find the brain anatomies that are closely related to the disease.
Preoperative imaging of the cerebral vessel tree is essential for planning therapy on intracranial stenoses and aneurysms. Usually, a magnetic resonance angiography (MRA) or computed tomography angiography (CTA) is acquired from which the cerebral vessel tree is segmented. Accurate analysis is helped by the labeling of the cerebral vessels, but labeling is non-trivial due to anatomical topological variability and missing branches due to acquisition issues. In recent literature, labeling the cerebral vasculature around the Circle of Willis has mainly been approached as a graph-based problem. The most successful method, however, requires the definition of all possible permutations of missing vessels, which limits application to subsets of the tree and ignores spatial information about the vessel locations.
This research aims to perform labeling using probabilistic atlases that model spatial vessel and label likelihoods. A cerebral vessel tree is aligned to a probabilistic atlas and subsequently each vessel is labeled by computing the maximum label likelihood per segment from label-specific atlases.
The proposed method was validated on 25 segmented cerebral vessel trees. Labeling accuracies were close to 100% for large vessels, but dropped to 50-60% for small vessels that were only present in less than 50% of the set.
With this work we showed that using solely spatial information of the vessel labels, vessel segments from stable vessels (>50% presence) were reliably classified. This spatial information will form the basis for a future labeling strategy with a very loose topological model.
The cohort size required in epidemiological imaging genetics studies often mandates the pooling of data from multiple hospitals. Patient data, however, is subject to strict privacy protection regimes, and physical data storage may be legally restricted to a hospital network. To enable biomarker discovery, fast data access and interactive data exploration must be combined with high-performance computing resources, while respecting privacy regulations. We present a system using fast and inherently secure light-paths to access distributed data, thereby obviating the need for a central data repository. A secure private cloud computing framework facilitates interactive, computationally intensive exploration of this geographically distributed, privacy sensitive data. As a proof of concept, MRI brain imaging data hosted at two remote sites were processed in response to a user command at a third site. The system was able to automatically start virtual machines, run a selected processing pipeline and write results to a user accessible database, while keeping data locally stored in the hospitals. Individual tasks took approximately 50% longer compared to a locally hosted blade server but the cloud infrastructure reduced the total elapsed time by a factor of 40 using 70 virtual machines in the cloud. We demonstrated that the combination light-path and private cloud is a viable means of building an analysis infrastructure for secure data analysis. The system requires further work in the areas of error handling, load balancing and secure support of multiple users.
Minimally invasive surgeries are approaching 50% of all interventional procedures in the US, yet there is a lack of
objective tools to assist surgeons in this limited sensing environment. In this preliminary work, we present a novel proof
of concept implementation of Single Snapshot of Optical Properties (SSOP) imaging through a rigid endoscope. In this
embodiment, a stereo rigid endoscope is used with one channel to project spatially modulated illumination and the
second channel to image the diffuse reflectance onto a CCD sensor. Optical property maps are then obtained for various
tissue simulating phantoms and validated against standard wide-field spatial frequency domain imaging (SFDI). The
implementation of endoscopic SSOP creates potential for practical use of endoscopic tissue constituent quantification.
The results show good agreement (within 5%) for endoscopic SSOP versus wide-field SFDI. However, endoscopic
SSOP acquisition allows for video-rate imaging, limited only by the exposure time of image capture. These results show
promise for an objective endoscopic tissue viability assessment tool being achievable in a clinical setting.
Wide-field optical tissue characterization has a large clinical potential that is currently not exploited due to the lack of realtime imaging methods. In this work we propose 3D single shot optical properties imaging (3D-SSOP) a new acquisition and processing method for obtaining surface profile corrected tissue absorption and reduced scattering coefficient maps from a single image. A profile is projected that is sensitive to both optical properties and surface profile. With image processing, the two responses are separated and surface profile corrected tissue optical properties as with profile corrected spatial frequency domain imaging (3D-SFDI). Overall, 3D-SSOP estimates showed a small bias of -1.2% in both μa and μ's in comparison with 3D-SFDI. Standard deviations on flat surfaces for 3D-SSOP were 7% (μa) and 17% (μ's) lower than for 3D-SFDI. However, 3D-SSOP showed significant artifacts near edges, where spatial averaging caused inaccuracies in diffuse reflectance estimates, as well as the surface profile. In an in-vivo experiment of a hand optical property estimates were equivalent, but processing artifacts suppressed smaller details with 3D-SSOP. To our knowledge, this method is the first method to estimate surface profile corrected tissue optical properties from a single image. Therefore we expect this method to be an important step in bringing real-time wide-field tissue characterization to the operating room.
KEYWORDS: Tumors, In vivo imaging, In vitro testing, Luminescence, Fluorescence lifetime imaging, Picosecond phenomena, Imaging systems, Tissues, Near infrared, Sensors
Excision of the whole tumor is crucial, but remains difficult for many tumor types. Fluorescence lifetime imaging could be helpful intraoperative to differentiate normal from tumor tissue. In this study we investigated the difference in fluorescence lifetime imaging of indocyanine green coupled to cyclic RGD free in solution/serum or bound to integrins e.g. in tumors. The U87-MG glioblastoma cell line, expressing high integrin levels, was cultured to use in vitro and to induce 4 subcutaneous tumors in a-thymic mice (n=4). Lifetimes of bound and unbound probe were measured with an experimental time-domain single-photon avalanche diode array (time resolution <100ps). In vivo measurements were taken 30-60 minutes after intravenous injection, and after 24 hours. The in vitro lifetime of the fluorophores was similar at different concentrations (20, 50 and 100μM) and showed a statistically significant higher lifetime (p<0.001) of bound probe compared to unbound probe. In vivo, lifetimes of the fluorophores in tumors were significantly higher (p<0.001) than at the control site (tail) at 30-60 minutes after probe injection. Lifetimes after 24 hours confirmed tumor-specific binding (also validated by fluorescence intensity images). Based on the difference in lifetime imaging, it can be concluded that it is feasible to separate between bound and unbound probes in vivo.
Wide-field oxygenation saturation (StO2) estimates can be clinically very advantageous. Particularly when implemented in a non-contact manner, applications such as intra-operative assessment of tissue perfusion are very promising. Nevertheless, wide-field optical oxygenation imaging did not yet successfully translate to the clinic.
In this work we compare four proposed methods for wide-field imaging that are based on different photon propagation models and that depend on different sets of assumed parameters such as absorption and reduced scattering coefficients. We investigated these for methods, with particular attention to sensitivities to errors in assumed parameters of calibration estimates. To this end we acquired an in vivo time series of a pig skin flap with a venous occlusion. StO2 estimates of all methods were compared to estimates from spatial frequency domain imaging of the same time series.
Correct assumptions on scatter power and accurate calibration were found to be the most important prerequisites for accurate StO2 estimates. Although all models were able to measure relative changes in StO2 when the occlusion was applied and released, only the models that incorporated assumed reduced scattering coefficients estimated StO2 values within 5% of the expected values (estimated using SFDI).
An important aspect of the compared methods is their ability to be used for real-time imaging. With the addition of real-time calibration and robust tissue scattering estimates, real-time wide-field imaging of oxygenation saturation can prove to provide important added value in the clinic.
Diffuse optical spectroscopy (DOS) may be advantageous for monitoring tumor response during chemotherapy treatment, particularly in the early treatment stages. In this paper we perform a second analysis on the data of a clinical trial with 25 breast cancer patients that received neoadjuvant chemotherapy. Patients were monitored using delayed contrast enhanced MRI and additionally with diffuse optical spectroscopy at baseline, after 1 cycle of chemotherapy, halfway therapy and before surgery.
In this analysis hemoglobin content between tumor tissue and healthy tissue of the same breast is compared on all four monitoring time points. Furthermore, the predictive power of the tumor-healthy tissue difference of HbO2 for non-responder prediction is assessed.
The difference in HbO2 content between tumor and healthy tissue was statistically significantly higher in responding tumors than in non-responding tumors at baseline (10.88 vs -0.57 μM, P=0.014) and after one cycle of chemotherapy (6.45 vs -1.31 μM, P=0.048). Before surgery this difference had diminished. In the data of this study, classification on the HbO2 difference between tumor and healthy tissue was able to predict tumor (non-)response at baseline and after 1 cycle with an area-under-curve of 0.95 and 0.88, respectively.
While this result suggests that tumor response can be predicted before chemotherapy onset, one should be very careful with interpreting these results. A larger patient population is needed to confirm this finding.
Diffuse optical spectroscopy imaging (DOSI) has shown great potential for the early detection of non-responding
tumors during neoadjuvant chemotherapy in breast cancer, already one day after therapy starts. Patients with rectal
cancer receive similar chemotherapy treatment. The rectum geometry and tissue properties of healthy and tumor
tissue in the rectum and the requirement of surface contact impose constraints on the probe design.
In this work we present the design of a DOSI probe with the aim of early chemotherapy/radiotherapy
effectiveness detection in rectal tumors. We show using Monte Carlo simulations and phantom measurements that
the colon tissue can be characterized reliably using a source-detector separation in the order of 10 mm. We present a
design and rapid prototype of a probe for DOSI measurements that can be mounted on a standard laparoscope and
that fits through a standard rectoscope. Using predominantly clinically approved components we aim at fast clinical
translation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.