Search

Currently, diagnosis of colon cancer is based on manual examination of histopathological images by a pathologist. This can be time consuming and interpretation of the images is subject to inter- and intra-observer variability. This may be improved by introducing a computer-aided diagnosis (CAD) system for automatic detection of cancer tissue within whole slide hematoxylin and eosin (H&E) stains. Cancer disrupts the normal control mechanisms of cell proliferation and differentiation, affecting the structure and appearance of the cells. Therefore, extracting features from segmented cell nuclei structures may provide useful information to detect cancer tissue. A framework for automatic classification of regions of interest (ROI) containing either benign or cancerous colon tissue extracted from whole slide H&E stained images using cell nuclei features was proposed. A total of 1,596 ROI's were extracted from 87 whole slide H&E stains (44 benign and 43 cancer). A cell nuclei segmentation algorithm consisting of color deconvolution, k-means clustering, local adaptive thresholding, and cell separation was performed within the ROI's to extract cell nuclei features. From the segmented cell nuclei structures a total of 750 texture and intensity-based features were extracted for classification of the ROI's. The nine most discriminative cell nuclei features were used in a random forest classifier to determine if the ROI's contained benign or cancer tissue. The ROI classification obtained an area under the curve (AUC) of 0.96, sensitivity of 0.88, specificity of 0.92, and accuracy of 0.91 using an optimized threshold. The developed framework showed promising results in using cell nuclei features to classify ROIs into containing benign or cancer tissue in H&E stained tissue samples. © 2017 International Society for Advancement of Cytometry.

Currently, colon cancer diagnosis is based on manual assessment of tissue samples stained with hematoxylin and eosin (H&E). This is a high volume, time consuming, and subjective task which could be aided by automatic cancer detection. We propose an algorithm for automatic cancer detection within WSI H&E stains using a multi class colon tissue classifier based on features extracted from 5 different color representations. Approx. 32000 tissue patches were extracted for the classifier from manual annotations of 9 representative colon tissue types from 74 WSI H&E stains. Colon tissue classifiers based on gray level or color features were trained using leave-one-out forward selection. The best colon tissue classifier was based on color texture features obtaining an average tissue precision-recall (PR) area under the curve (AUC) of 0.886 and a cancer PR-AUC of 0.950 on 20 validation WSI H&E stains.

Currently, diagnosis of colon cancer is based on manual examination of histopathological images by a pathologist. This can be time consuming and interpretation of the images is subject to inter- and intra-observer variability. This may be improved by introducing a computer-aided diagnosis (CAD) system for automatic detection of cancer tissue within whole slide hematoxylin and eosin (HE) stains. Cancer disrupts the normal control mechanisms of cell proliferation and differentiation, affecting the structure and appearance of the cells. Therefore, extracting features from segmented cell nuclei structures may provide useful information to detect cancer tissue. A framework for automatic classification of regions of interest (ROI) containing either benign or cancerous colon tissue extracted from whole slide HE stained images using cell nuclei features was proposed. A total of 1,596 ROI's were extracted from 87 whole slide HE stains (44 benign and 43 cancer). A cell nuclei segmentation algorithm consisting of color deconvolution, k-means clustering, local adaptive thresholding, and cell separation was performed within the ROI's to extract cell nuclei features. From the segmented cell nuclei structures a total of 750 texture and intensity-based features were extracted for classification of the ROI's. The nine most discriminative cell nuclei features were used in a random forest classifier to determine if the ROI's contained benign or cancer tissue. The ROI classification obtained an area under the curve (AUC) of 0.96, sensitivity of 0.88, specificity of 0.92, and accuracy of 0.91 using an optimized threshold. The developed framework showed promising results in using cell nuclei features to classify ROIs into containing benign or cancer tissue in HE stained tissue samples. © 2017 International Society for Advancement of Cytometry.

A novel closed-loop system for improving gait in hemiparetic patients by supporting the production of the swing phase using electrical stimulations evoking the nociceptive withdrawal reflex was designed. The system exploits the modular organization of the nociceptive withdrawal reflex and its stimulation site- and gait-phase modulation in order to evoke movements of the hip, knee, and ankle joints during the swing phase. A modified model reference adaptive controller (MRAC) was designed to select the best stimulation parameters from a set of 12 combinations of four electrode locations on the sole of the foot and three different stimulation onset times between heel-off and toe-off. It was hypothesized that the MRAC system would result in a better walking pattern compared with an open-loop preprogrammed fixed pattern of stimulation (FPS) controller. Thirteen chronic or subacute hemiparetic subjects participated in a study to compare the performance of the two control schemes. Both control schemes resulted in a more functional gait compared to no stimulation (P

A novel closed loop method for improving gait in hemiplegic patients by supporting the production of the swing phase using electrical stimulations to evoke the nociceptive withdrawal reflex was developed and evaluated in one chronic hemiplegic subject. Electrical stimulations were delivered to 4 locations on the sole of the foot at 3 different times between heel-off and toe-off. The system exploits the modular organization of the nociceptive withdrawal reflex and its stimulation site- and gait phase-modulation in order to evoke optimal flexion of the hip, knee and ankle joints in the early swing phase. A Rule-based Model Reference Adaptive Controller (MRAC) was designed to select the optimal stimulation parameters. It was hypothesized that the MRAC-system would result in a better walking pattern compared with a preprogrammed fixed stimulation pattern controller. Based on the unperturbed gait and withdrawal strategies of the patient, an individual controller target for hip, knee and ankle flexion was set. The patient walked 10 min with the MRAC-system, 10 min with the fixed pre-programmed stimulation pattern, and 10 min with no-stimulation. The results indicate that both stimulation paradigms resulted in a more functional gait compared with no-stimulation and that the control strategy in the MRAC system is superior suggesting that it will be able to adapt better to the varying needs during rehabilitation therapy.

The objective of this study was to investigate the modulation of the nociceptive withdrawal reflex during gait measured using Force Sensitive Resistors (FSR). Electrical stimulation was delivered to four locations on the sole of the foot at three different time points between heel-off and toe-off. Peak force changes were measured by FSRs attached to the big toe, distal to the first and fourth metatarsophalangeal joints, and the medial process of the calcaneus on both feet. Force changes were assessed in five gait sub-phases. The painful stimulation led to increased ipsilateral unloading (10 +/- 1 N) and contralateral loading (12 +/- 1 N), which were dependent on stimulation site and phase. In contrast, the hallux of the ipsilateral foot plantar flexed, thus facilitating the push-off. The highest degree of plantar flexion (23 +/- 10 N; range, 8-44 N) was seen in the second double support phase following the stimulation. Site and phase modulation of the reflex were detected in the force signals from all selected anatomical landmarks. In the kinematic responses, both site and phase modulation were observed. For stimulations near toe-off, withdrawal was primarily accomplished by ankle dorsiflexion, while the strategy for stimulations at heel-off was flexion of the knee and hip joints.

The objective of this study was to investigate the capability of Force Sensitive Resistors (FSR) to characterize perturbations of gait produced by electrically evoked withdrawal reflexes. Reflexes were evoked by cutaneous electrical stimulation at 4 locations on the sole of the left foot and delivered at 3 different phases in early swing. Force changes were measured by 4 FSR's attached to each foot. The vertical ground reaction force was estimated by summing the force signals for each foot separately. Response measures were mean peak force and peak slope evaluated in 5 gait sub-phases immediately after the stimulus: 1st double support, 1st right single support, 2nd double support, 1st left single support and 3rd double support. The results showed that perturbation of gait could be detected by the selected features and that stimulation led to a weight shift from the ipsilateral to contralateral leg starting with a more rapid unloading of the ipsilatteral leg, supported by a rapid loading of the contralateral leg in the 1st double support.