KEYWORDS: Tissues, Breast cancer, Imaging spectroscopy, Image segmentation, Breast, Cancer, Tumors, Image classification, In vivo imaging, Library classification systems
Spectral imaging has recently been introduced in the biomedical field as a noninvasive, quantitative means of studying biological tissues. Many of its potential applications have been demonstrated (in vitro and, to a lesser degree, in vivo) with the use of stains or dyes. Successful translation to the clinical environment has been largely lagging, due to safety considerations and regulatory limitations preventing use of contrast agents in humans. We report experiments showing the feasibility of high-resolution spectral imaging of breast cancer without the use of contrast agents, thus completing the continuum of translational research, to in vivo imaging that will be directly applicable in the clinical environment. Our initial work focused on image acquisition using Fourier transform microinterferometry and subsequent segmentation of both stained and unstained breast cancer slides-derived image sets. We then applied our techniques to imaging fresh unstained ex vivo specimens of rat breast cancer and sentinel lymph nodes. We also investigated multiple methods of classification to optimize our image analyses, and preliminary results for the best algorithm tested yielded an overall sensitivity of 96%, and a specificity of 92% for cancer detection. Using spectral imaging and classification techniques, we were able to demonstrate that reliable detection of breast cancer in fixed and fresh unstained specimens of breast tissue is possible.
We introduce a novel node-structural map method for digital image-based analysis of angiogenesis, and apply it to quantitate the effect of pro-angiogenic (e.g. VEGF) and anti-angiogenic (e.g. anti-VEGF) treatments on blood vessel patterns in the quail chorioallantoic membrane (CAM) in vivo model. After vessel segmentation and skeletonization, a node-structural map is derived to represent the structural property of each pixel in the skeletonized image, such as end nodes (root and leaf), branch and furcation/junction nodes. By combining the node-structural map with vascular thickness map, our proposed method provides more detailed morphometric and structural measurement of vascular angiogenesis, such as average diameter, average branch length, branch thickness distribution, density of branch and furcation nodes, etc. Furthermore, the concept of vascular generations will be proposed. Accordingly, more detailed measurements related to the generations of the vascular tree can be easily evaluated. From the quantitative analysis results, it correctly shows the fact that VEGF/anti-VEGF can modify the blood vessel pattern and increase/decrease the vessel density in CAM significantly, compared to the phosphate-buffered saline treated controls.
Biological specimens are three-dimensional structures. However, when capturing their images through a microscope, there is only one plane in the field of view that is in focus, and out-of-focus portions of the specimen affect image quality in the in-focus plane. It is well-established that the microscope’s point spread function (PSF) can be used for blur quantitation, for the restoration of real images. However, this is an ill-posed problem, with no unique solution and with high computational complexity. In this work, instead of estimating and using the PSF, we studied focus quantitation in
multi-spectral image sets. A gradient map we designed was used to evaluate the sharpness degree of each pixel, in order to identify blurred areas not to be considered. Experiments with realistic multi-spectral Pap smear images showed that measurement of their sharp gradients can provide depth information roughly comparable to human perception (through a microscope), while avoiding PSF estimation. Spectrum and morphometrics-based statistical analysis for abnormal cell detection can then be implemented in an image database where the axial structure has been refined.
Finding malignant cells in Pap smear images is a "needle in a haystack"-type problem, tedious, labor-intensive and error-prone. It is therefore desirable to have an automatic screening tool in order that human experts can concentrate on the evaluation of the more difficult cases. Most research on automatic cervical screening tries to extract morphometric and texture features at the cell level, in accordance with the NIH "The Bethesda System" rules. Due to variances in image quality and features, such as brightness, magnification and focus, morphometric and texture analysis is insufficient to provide robust cervical cancer detection.
Using a microscopic spectral imaging system, we have produced a set of multispectral Pap smear images with wavelengths from 400 nm to 690 nm, containing both spectral signatures and spatial attributes. We describe a novel scheme that combines spatial information (including texture and morphometric features) with spectral information to significantly improve abnormal cell detection. Three kinds of wavelet features, orthogonal, bi-orthogonal and non-orthogonal, are carefully chosen to optimize recognition performance. Multispectral feature sets are then extracted in the wavelet domain. Using a Back-Propagation Neural Network classifier that greatly decreases the influence of spurious events, we obtain a classification error rate of 5%. Cell morphometric features, such as area and shape, are then used to eliminate most remaining small artifacts. We report initial results from 149 cells from 40 separate image sets, in which only one abnormal cell was missed (TPR = 97.6%) and one normal cell was falsely classified as cancerous (FPR = 1%).
Efficient computer-aided cervical cancer detection can improve both the accuracy and the productivity of cytotechnologists and pathologists. Nuclear segmentation is essential to automated screening, and is still a challenge. We propose and demonstrate a novel approach to improving segmentation performance by multispectral imaging followed by unsupervised nuclear segmentation relying on selecting a useful subset of spectral or derived image features. In the absence of prior knowledge, feature selection can be negatively affected by the bias, present in most unsupervised segmentation, to erroneously segment out small objects, yielding ill-balanced class samples. To address this issue, we first introduce a new measurement, Criterion Vector (CV), measuring the distances between the segmentation result and the original data. This efficiently reduces the bias generated by feature selection. Second, we apply a novel recursive feature selection scheme, to generate a new feature subset based on the corresponding CV, ensuring that the correct part of the initial segmentation results is used to obtain better feature subsets. We studied the speed and accuracy of our two-step algorithm in analyzing a number of multispectral Pap smear image sets. The results show high accuracy of segmentation, as well as great reduction of spectral redundancy. The nuclear segmentation accuracy can reach over 90%, by selecting as few as 4 distinct spectra out of 30.
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.