Automated Image Analysis System Research Articles

Exploring the Immune Microenvironment of Diffuse Large B Cell Lymphoma in a Tissue Microarray: Predicting Survival with a Score That Incorporates Macrophages, Cytotoxic and Regulatory T Cells

Abstract 951Gene expression profiling studies in Diffuse Large B cell Lymphoma (DLBCL) have described signatures enriched in genes representing the immune microenvironment. These signatures, together with the cell of origin classification have prognostic significance. However, a comprehensive functional understanding of the role of inflammation and the immune response in DLBCL is still lacking. The use of immunohistochemistry (IHC) to validate and elaborate upon molecular results is appealing. Our objective was to assess the cellular composition of the non-malignant population in DLBCL quantitatively and thoroughly in a single centre TMA of patients with DLBCL with corresponding clinical and survival data. Based on reported findings and known functional interactions a panel of antibodies were selected that included CD68, FOXP3 and TIA-1.IHC was performed on TMAs from high quality formalin-fixed paraffin-embedded diagnostic tissue biopsies of DLBCL from 1977 to 2009. Triplicate cores were made from areas of representative tumour tissue, arrayed onto glass slides and stained for a panel of antibodies, aiming to characterize immune cell subsets. The TMA included tissue from 218 patients: 128 males, 90 females, with a median age of 55 (18-94) years; 23.2% high-int/high risk IPI; with median follow up of 3,1 years. 33% of patients were treated in the rituximab era. Absolute numbers of positive cells were counted across all intact cores using an automated image analysis system (Ariol), confirmed by histopathologists, and means calculated and corrected to a 1 mm2 area. A training-validation set method was used to generate optimal cutoff values discriminating prognostic groups and comparisons performed using the chi-square test.Among nine antibodies examined, statistically significant prognostic information was provided using staining for FOXP3, TIA-1 and CD68. Using single cut-offs of positive cells/mm2 to define low and high marker density, two prognostic groups were defined for each marker. Patients with a more favourable prognosis had the highest cell density for FOXP3, TIA-1 and CD68.To clarify whether there was an additive prognostic relationship we developed a score that incorporated these three markers. This scoring system allowed us to discriminate two subgroups of patients with divergent survival parameters (fig 1).Marker (cutoff value high vs low)Overall Survival at 3yr (high vs low)Progression Free Survival at 3yr (high vs low)Disease Specific Survival at 3yr (high vs low)FOXP3 (450+ cells/mm2)64% vs 44% (p 0.005)55% vs 36% (p 0.001)69% vs 54% (0.01)CD68 (900+ cells/mm2)58% vs 28% (p 0.023)50% vs 20% (p 0.05)66% vs 36% (p 0.005)TIA-1 (2500+ cells/mm2)76% vs 45% (p 0.002)66% vs 41% (p 0.01)80% vs 56% (p 0.005)Combined Score68% vs 39% (p 0.0002)60% vs 32% (p 0.0003)73% vs 49% (p 0.0008) [Display omitted] A Cox regression model was developed, incorporating age, IPI, rituximab treatment and IHC score. The score results were validated as independent prognostic markers for OS, PFS and DSS, together with age, IPI (for OS and DSS) and rituximab (for PFS).We confirm that there is heterogeneity in immune cell infiltration at diagnosis in DLBCL, and we provide correlative prognostic data to suggest a corresponding biological relevance. Attributing single cut-off values for immune cell markers is feasible, attractive as a clinical tool, and appears to provide robust prognostic information. In our single centre cohort we found that higher numbers of macrophages, Tregs and cytotoxic T cells correlate independently with improved patient survival. CD68 has been associated with an adverse prognosis in lymphoid malignancies but phenotypic diversity of macrophages may account for this study's finding and will be explored using macrophage subset specific markers.Our multivariate analysis revealed that the IHC results have an independent impact on survival on this cohort of DLBCL patients. We are currently validating this approach in an independent TMA of R-CHOP treated patients. Combining these three different immune cell parameters allowed us to discriminate prognostic subgroups providing further evidence of the impact of the non-malignant immune cells in the biology of DLBCL and validating their role as potential therapeutic targets for intervention. Disclosures:Gribben:Gilead: Honoraria; Mundipharma: Honoraria; GSK: Honoraria; Celgene: Honoraria; Roche: Honoraria; Pharmacyclics: Honoraria.

Automated clarity assessment of retinal images using regionally based structural and statistical measures

An automated image analysis system for application in mass medical screening must assess the clarity of the images before analysing their content. This is the case in grading for diabetic retinopathy screening where the failure to assess clarity could result in retinal images of people with retinopathy being erroneously classed as normal. This paper compares methods of clarity assessment based on the degradation of visible structures and based on the deviation of image properties outside expected norms caused by clarity loss. Vessel visibility measures and statistical measures were determined at locations in the image which have high saliency and these were used to obtain an image clarity assessment using supervised classification. The usefulness of the measures as indicators of image clarity was assessed. Tests were performed on 987 disc-centred and macula-centred retinal photographs (347 with inadequate clarity) obtained from the English National Screening Programme. Images with inadequate clarity were detected with 92.6% sensitivity at 90% specificity. In a set of 2000 macula-centred images (200 with inadequate clarity) from the Scottish Screening Programme, inadequate clarity was detected with 96.7% sensitivity at 90% specificity. This study has shown that structural and statistical measures are equally useful for retinal image clarity assessment.

Comparison of parameter‐adapted segmentation methods for fluorescence micrographs

Interpreting images from fluorescence microscopy is often a time-consuming task with poor reproducibility. Various image processing routines that can help investigators evaluate the images are therefore useful. The critical aspect for a reliable automatic image analysis system is a robust segmentation algorithm that can perform accurate segmentation for different cell types. In this study, several image segmentation methods were therefore compared and evaluated in order to identify the most appropriate segmentation schemes that are usable with little new parameterization and robustly with different types of fluorescence-stained cells for various biological and biomedical tasks. The study investigated, compared, and enhanced four different methods for segmentation of cultured epithelial cells. The maximum-intensity linking (MIL) method, an improved MIL, a watershed method, and an improved watershed method based on morphological reconstruction were used. Three manually annotated datasets consisting of 261, 817, and 1,333 HeLa or L929 cells were used to compare the different algorithms. The comparisons and evaluations showed that the segmentation performance of methods based on the watershed transform was significantly superior to the performance of the MIL method. The results also indicate that using morphological opening by reconstruction can improve the segmentation of cells stained with a marker that exhibits the dotted surface of cells.

Canine differential leukocyte counting with the CellaVision DM96Vision, Sysmex XT‐2000iV, and Advia 2120 hematology analyzers and a manual method

For differential leukocyte counts, automated blood smear evaluation systems have been too slow or inaccurate to replace or supplement the manual differential count. The CellaVision DM96Vision (DM96V), a new instrument, is an automated image analysis system that is rapid and accurate enough to be used for enumerating human leukocytes and may be useful for analysis of canine blood. The aims of this study were to evaluate the performance of the DM96V in differential counting of canine leukocytes, to compare its performance with that of other methods, and to analyze interoperator variability. Four methods of determining the leukocyte differential count of 108 canine blood samples were compared based on agreement, precision, and errors as well as relative performance. Differential counts were obtained using the DM96V, the manual method, and automated methods performed by the Advia 2120 and Sysmex XT-2000iV. All leukocyte types were detected by the DM96V and the manual method, and all 4 methods had similar mean and median results in most cases. The automated methods were more precise than either the DM96V or manual method when comparing identification of a single type of leukocyte, especially neutrophils and lymphocytes. However, precision of the automated methods was only fair for monocytes, and the Advia and Sysmex failed to identify basophils. The Advia reported fewer monocytes and eosinophils than did the other methods. Significantly fewer lymphocytes were identified by the manual method than by the Sysmex, Advia, and DM96V. The DM96V occasionally presented duplicate images of the same neutrophils. The CellaVision DM96V is a satisfactory system for facilitating canine differential leukocyte counting. The DM96V differential count was more similar to the manual count than to automated counts, which were more precise but had errors and omissions in detecting some types of leukocytes.

Automatic identification of responses from porphyry intrusive systems within magnetic data using image analysis

Abstract Direct targeting of mineral deposits using magnetic data may be facilitated by hydrothermal alteration associated with the mineralising event if the alteration changes the magnetic properties of the host rock. Hydrothermal alteration associated with porphyry-style mineralisation typically comprises concentric near-circular alteration zones surrounding a roughly circular central intrusion. The intrusion itself and the proximal alteration zone are usually associated with positive magnetic anomalies whilst the outer alteration zones are much less magnetic. Because the country rocks are usually magnetic, this pattern of alteration produces a central magnetic ‘high’ surrounded by an annular magnetic ‘low’. This paper presents an automatic image analysis system for gridded data that provides an efficient, accurate and non-subjective way to seek the magnetic response of an idealised porphyry mineralising system within magnetic datasets. The method finds circular anomalies that are associated with the central intrusion and inner alteration zone of the porphyry system using a circular feature detection method called the radial symmetry transform. Next, their boundaries are traced using deformable splines that are drawn to the locations of maximum contrast between the amplitudes of the central ‘high’ and surrounding area of lower magnetisation. Experiments were conducted on magnetic data from Reko Diq, Pakistan; a region known to contain numerous occurrences of porphyry-style mineralisation. The predicted locations of porphyry systems closely match the locations of the known deposits in this region. This system is suitable as an initial screening tool for large geophysical datasets, therefore reducing the time and cost imposed by manual data inspection in the exploration process. The same principles can be applied to the search for circular magnetic responses with different amplitude characteristics.

Effect of baseline magnetic resonance imaging (MRI) apparent diffusion coefficient lesion volume on functional outcome in ischemic stroke

<title/>Objective: We explored the relationship between predicted infarct core, predicted ischemic penumbras and predicted final infarct volumes obtained though apparent diffusion coefficient (ADC)-based method, as well as other clinical variables, and functional outcome.Methods: Patients with acute cerebral ischemic stroke were retrospectively recruited. The National Institutes of Health Stroke Scale score was evaluated at baseline and the modified Rankin Scale (mRS) at day 90. Favorable outcome was defined as an mRS score of 0 to 2, and unfavorable outcome as 3 to 6. Multimodal stroke magnetic resonance imaging was carried out at presentation. The volumes of diffusion-weighted imaging (DWI) and perfusion-weighted imaging (PWI) were measured using the regions of interest (ROI) method. The volumes of predicted infarct core, predicted ischemic penumbra and predicted final infarct were obtained by an automated image analysis system based on baseline ADC maps. The association between baseline magnetic resonance imaging volumes, baseline clinical variables, and functional outcome was statistically analyzed.Results: The study included 30 males and 20 females (mean±SD age, 56±10 years). Baseline DWI, PWI and PWI-DWI mismatch volumes were not correlated with day-90 mRS (P>0·05). Predicted infarct core, predicted ischemic penumbra and predicted final infarct through ADC-based method were all correlated with day-90 mRS (P<0·05). A better outcome was associated with a smaller predicted volume. Low baseline National Institutes of Health Stroke Scale and recanalization also demonstrated a trend toward a favorable outcome. Receiver operating characteristic analysis showed that the area under the curve of predicted final infarct volume and recanalization were higher with statistical significance (P<0·001).Discussion: Predicted volumes obtained from ADC-based methods, especially predicted final infarct volume, as well as baseline National Institutes of Health Stroke Scale and recanalization may have effect on functional outcome in acute ischemic stroke.

Quantitative Automated Image Analysis System with Automated Debris Filtering for the Detection of Breast Carcinoma Cells

Objective: To develop an intraoperative method for margin status evaluation during breast conservation therapy (BCT) using an automated analysis of imprint cytology specimens. Study Design: Imprint cytology samples were prospectively taken from 47 patients undergoing either BCT or breast reduction surgery. Touch preparations from BCT patients were taken on cut sections through the tumor to generate positive margin controls. For breast reduction patients, slide imprints were taken at cuts through the center of excised tissue. Analysis results from the presented technique were compared against standard pathologic diagnosis. Slides were stained with cytokeratin and Hoechst, imaged with an automated fluorescent microscope, and analyzed with a fast algorithm to automate discrimination between epithelial cells and noncellular debris. Results: The accuracy of the automated analysis was 95% for identifying invasive cancers compared against final pathologic diagnosis. The overall sensitivity was 87% while specificity was 100% (no false positives). This is comparable to the best reported results from manual examination of intraoperative imprint cytology slides while reducing the need for direct input from a cytopathologist. Conclusion: This work demonstrates a proof of concept for developing a highly accurate and automated system for the intraoperative evaluation of margin status to guide surgical decisions and lower positive margin rates.

Expression of organic anion-transporting polypeptides 1B3, 1B1, and 1A2 in human pancreatic cancer reveals a new class of potential therapeutic targets

Background:Organic anion-transporting polypeptides (OATPs) are influx transporters that mediate intracellular uptake of selective endogenous and xenobiotic compounds. Identification of new molecular targets and discovery of novel targeted therapies is top priority for pancreatic cancer, which lacks any effective therapy.Materials and methods:We studied expression of OATP 1A2, 1B1, and 1B3 in pancreatic cancer tissue and in cell lines. Formalin-fixed paraffin-embedded biopsy material of 12 human pancreatic cancers was immunohistochemically assessed for protein expression of the three studied influx transporters. Immunohistochemistry was evaluated by experienced pathologists and quantified by use of an automated image analysis system. BxPC-3 and MIA PaCa-2 pancreatic cancer cell lines were used to quantify transcripts of OATP 1B1 and 1B3.Results:OATP 1A2, 1B1, and 1B3 proteins were found ubiquitously expressed in all studied cases. Quantification performed by HistoQuest system revealed that mean intensity was 53 for 1A2, 45 for 1B1, and 167 for OATP 1B1/1B3 on a range scale 0–250 units. At mRNA level, 1B1 and 1B3 were overexpressed in both studied cancer cell lines but not in normal pancreatic tissue.Conclusion:OATPs 1A2, 1B1, and 1B3 are highly expressed in pancreatic adenocarcinoma. We suggest that expression of these transporters in pancreatic cancer justify research efforts towards discovery of novel therapeutics targeting OATPs.

The expression of p53 in preventing restenosis after rabbit carotid artery balloon injury

：Objective To observe theeffect of p53 protein on smooth muscle cell(VSMC)in rabbit artery balloon injury.MethodsRestenosis model of carotid artery after balloon injury was established in rabbits.30rabbits were divided into 2 groups,the sham group(n = 6)and the vascular injury group(n =24).With H.E.staining and automatic image analysis system,we investigated arterymorphology alteration and measured the area of arterial intima and media.The expressionsof p53 protein were detected by immunohistochemical analysis.Results With H.E.staining andautomatic image analysis and immunohistochemistry,the results showed that the expressionof p53 was significantly reduced and the intima area was increased in model group comparedwith the sham group(P ＜0.01).But the expression of p53 inmedia was remarkably reduced compared with intima(P ＜ 0.01).Conclusions The possible mechanism of preventingarterial restenosis in the balloon injury might be related with p53,which may be throughinhibiting neointimal proliferation in arterial restenosis . Key words: Tumorsuppressor protein p53/ME; Carotid artery injuries/CO/ME; Constriction,pathologic/CO/ME; Angioplasty,transluminal,percutaneous coronary/AE

Microscopic image analysis for quantitative characterization of muscle fiber type composition

Skeletal muscles consist of muscle fibers that are responsible for contracting and generating force. Skeletal muscle fibers are categorized into distinct subtypes based on several characteristics such as contraction time, force production and resistance to fatigue. The composition of distinct muscle fibers in terms of their number and cross-sectional areas is characterized by a histological examination. However, manual delineation of individual muscle fibers from digitized muscle histology tissue sections is extremely time-consuming. In this study, we propose an automated image analysis system for quantitative characterization of muscle fiber type composition. The proposed system operates on digitized histological muscle tissue slides and consists of the following steps: segmentation of muscle fibers, registration of successive slides with distinct stains, and classification of muscle fibers into distinct subtypes. The performance of the proposed approach was tested on a dataset consisting of 25 image pairs of successive muscle histological cross-sections with different ATPase stain. Experimental results demonstrate a promising overall segmentation and classification accuracy of 89.1% in identifying muscle fibers of distinct subtypes.

Emotion categorization using affective-pLSA model

Emotion categorization of natural scene images represents a very useful task for automatic image analysis systems. Psychological experiments have shown that visual information at the emotion level is aggregated according to a set of rules. Hence, we attempt to discover the emotion descriptors based on the composition of visual word representation. First, the composition of visual word representation models each image as a matrix, where elements record the correlations of pairwise visual words. In this way, an image collection is modeled as a third-order tensor. Then we discover the emotion descriptors using a novel affective-probabilistic latent semantic analysis (affective-pLSA) model, which is an extension of the pLSA model, on this tensor representation. Considering that the natural scene image may evoke multiple emotional feelings, emotion categorization is carried out using the multilabel k-nearest-neighbor approach based on emotion descriptors. The proposed approach has been tested on the International Affective Picture System and a collection of social images from the Flickr website. The experimental results have demonstrated the effectiveness of the proposed method for eliciting image emotions.

Pattern Recognition Software and Techniques for Biological Image Analysis

The increasing prevalence of automated image acquisition systems is enabling new types of microscopy experiments that generate large image datasets. However, there is a perceived lack of robust image analysis systems required to process these diverse datasets. Most automated image analysis systems are tailored for specific types of microscopy, contrast methods, probes, and even cell types. This imposes significant constraints on experimental design, limiting their application to the narrow set of imaging methods for which they were designed. One of the approaches to address these limitations is pattern recognition, which was originally developed for remote sensing, and is increasingly being applied to the biology domain. This approach relies on training a computer to recognize patterns in images rather than developing algorithms or tuning parameters for specific image processing tasks. The generality of this approach promises to enable data mining in extensive image repositories, and provide objective and quantitative imaging assays for routine use. Here, we provide a brief overview of the technologies behind pattern recognition and its use in computer vision for biological and biomedical imaging. We list available software tools that can be used by biologists and suggest practical experimental considerations to make the best use of pattern recognition techniques for imaging assays.

An Immunohistochemical Score Based On CD68 and FOXP3 Expression In the Tumour Microenvironment at Diagnosis Defines Prognostic Groups In Both Early and Advanced Stage Classical Hodgkin Lymphoma

AbstractAbstract 750The non-malignant immune infiltrate comprises the bulk of pathologic tissue in classical Hodgkin lymphoma (CHL). This microenvironment has the potential to induce both malignant cell suppression and support. Increased macrophage infiltration assessed by CD68 expression has been shown to confer adverse prognostic significance (Steidl et al. N Engl J Med, 2010; 362:875-85) while increased FOXP3 expressing T cells are beneficial in this disease (Tzankov et al. Haematologica, 2008; 93: 193–200). However no histological score routinely leads to modification of treatment. Assessing outcomes by the parameters of overall survival (OS), disease specific survival (DSS) and freedom from first line treatment failure (FFTF) at 5 years in a cohort of patients treated at St Bartholomew's Hospital (Barts), London we developed a prognostic score based upon expression of both FOXP3 and CD68 in the CHL microenvironment which defined poor and good risk groups in both early (Stages I and IIA) and advanced stage disease, including a ‘poor risk’ early stage group with 25% FFTF and a ‘good risk’ advanced stage group with 90% FFTF. Immunohistochemical analysis was performed on tissue microarrays (TMAs) from previously untreated patients' diagnostic lymph node biopsies in whom clinical outcome was available. From all 1056 adult patients with HL diagnosed at Barts between 1972 and 2005, high quality formalin-fixed paraffin-embedded tissue was available for 122 (12%), with characteristics representative of the whole group. Median age of the 122 patients was 30 (range 18–80) years, 35% female, 71% advanced stage with median follow up 16.5 (range 2–40) years. Triplicate cores were made from areas of high cellularity, containing malignant cells and avoiding fibrotic, acellular portions, arrayed onto glass slides and stained immunohistochemically for FOXP3 or CD68. Absolute and proportional numbers of FOXP3+ nuclei and CD68+ cell bodies were counted across all intact cores using an automated image analysis system (Ariol), confirmed by expert histopathologists, and means calculated and corrected to a single high powered field (hpf). Recursive partitioning was used to generate optimal cutoff values discriminating prognostic groups and comparisons performed by the chi-square test. Using cutoffs of <5%, 5–15% and >15% to define low, intermediate and high CD68 density, 3 prognostic groups were defined, the favourable group having the lowest CD68+ density. OS for low, intermediate and high groups were 89%, 80% and 65% respectively (p=0.02), with FFTF 82%, 64% and 29% (p=0.001). Prognostic significance was maintained in subgroups presenting with advanced stage (FFTF 73%, 63% and 33%, p=0.03), as well as early stage disease (FFTF 92%, 70% and 20%, p=0.01). Using cutoffs of <12.5, 12.5–50 and >50 nuclei/hpf to define low, intermediate and high FOXP3+ cell density, 3 prognostic groups were defined, the favourable group having the highest FOXP3+ density. OS for low, intermediate and high groups were 68%, 80% and 94% respectively (p=0.006), with FFTF 50%, 62%, and 84% (p=0.002). Prognostic significance was maintained for both advanced (FFTF: 48%, 60% and 72%, p=0.04) and early stage disease (FFTF: 57%, 67% and 100%, p=0.04). A combined ‘FOXP3/CD68 score’ derived from the patient's prognostic group for both markers, for which suitable cores were available on 98 patients, further improved the predictive value of each individually (See Figure). In this model, favourable, intermediate and unfavourable groups had 5 year FFTF of 93%, 62% and 47% (p=0.0002), DSS 93%, 82% and 63% (p=0.03) and OS 93%, 82% and 59% (p=0.002). The score retained prognostic significance for 5 year FFTF and OS in the subgroup of patients presenting with advanced (FFTF: 90%, 59% and 46%, p=0.008; OS: 90%, 80% and 54%, p=0.004) as well as early stage disease (FFTF: 100%, 71% and 25%, p=0.005; OS: 100%, 82% and 75%, trend only, p=ns). We conclude that FOXP3 and CD68 are important independent factors, which in combination have considerable predictive power and will now be validated prospectively. [Display omitted] Disclosures:No relevant conflicts of interest to declare.

Comparative Study of Peripheral Blood Morphology by Conventional Microscopy and Cellavision DM96 In Hematological and non Hematological Diseases

AbstractAbstract 4737 Objective:Microscopic examination of peripheral blood cells is an important diagnostic tool. We evaluated the CellaVision DM96 (CellaVision AB, Lund, Sweden), an automated image analysis system for digital peripheral blood cell analysis, comparing the results with direct manual microscopy. The system obtains digital images of the blood cells at high magnification and these images are analyzed using a neural network based on a large database of cells. Material and Methods:We analyzed 234 PB films stained with May-Grünwald-Giemsa from patients of the Hospital Clínic of Barcelona. Leukocyte values were from 1.12 to 282 × 109/L (Advia 2120, Siemens Healthcare Diagnostics SL). 177 of the PB films were from patients with hematological diseases: lymphoid neoplasias: 83, acute leukemias: 52, chronic myeloproliferative diseases: 20, myelodisplastic syndromes: 18, paroxysmal nocturnal hemoglobinuria: 2, hemoglobin S: 1 and thrombotic thrombocytopenic purpura: 1. Manual differentials were performed using standard microscopy. After the microscopic analysis the slides were loaded into the CellaVision DM96 obtaining digital images of preclassified cells, which were verified or corrected when necessary by the hematologist (DM96POST). WBC differentials were abnormal in 120 cases. Statistical analysis was performed using correlation (Pearson) and concordance (Lin) tests. Results:Correlation coefficients between results obtained from the CellaVision DM96 preclassification and by conventional direct microscopy were excellent for segmented neutrophils, lymphocytes, monocytes and blasts (r>0.87<0.94 and p<0.0001) and good for band neutrophils, eosinophils, basophils and plasma cells (r>0.74<0.81 and p<0.0001). Concordance coefficients were higher than 0.7 for all of the white blood cell subtypes preclassified by the DM96. Pearson and Lin coefficients were higher when we compared DM96POST values. After the reclassification of the cells very good concordance coefficients were observed for promyelocytes and myelocytes (> 0.7), intermediates for reactive lymphocytes and erythroblasts (>0.5 and <0,7) and low (<0.5) for metamyelocytes. Whatever the pathology and the number of blasts on the manual review films all 97 patients were positive for blast detection on DM96. Pathological cells such as prolymphocytes, large granular lymphocytes, hairy cells, Sézary cells and other atypical lymphocytes were reclassified by the user. Digital images showed dysplastic features or inclusions in blood cells and morphologic alterations in red cells or platelets were easily identified. Conclusion:Comparison of morphological classification of blood cells by the automated system DM96 shows good correlation and concordance values with respect to manual differentials. Advantages of the Cellavision DM96 over direct microscopy includes that requires less time than manual differentiation, allows the verification of the results by an expert from another location (telehematology), digital images can be stored and therefore available for re-evaluation, is a good tool for educational purposes and can improve the efficiency of a modern Hematology Laboratory. Disclosures:No relevant conflicts of interest to declare.

Prediction of Infarct Core and Salvageable Ischemic Tissue Volumes by Analyzing Apparent Diffusion Coefficient Without Intravenous Contrast Material

To investigate whether baseline apparent diffusion coefficient (ADC) maps can be employed to predict both infarct core and salvageable ischemic tissue volumes in acute ischemic stroke. An automated image analysis system based on baseline ADC maps was tested against 30 patients with acute ischemic stroke of anterior circulation to predict both infarct core and salvageable ischemic tissue volumes. The predicted infarct core and predicted salvageable ischemic tissue were quantitatively and qualitatively compared with follow-up imaging data in recanalization and no recanalization groups, respectively. Direct comparisons with perfusion- and diffusion- weighted magnetic resonance imaging measures were also made. Wilcoxon signed-rank test, Spearman rank correlation, and Bland-Altman plots were performed. In the recanalization group, the predicted infarct core volume was significantly correlated with the final infarct volume (r = 0. 868, P < .001). In the no recanalization group, the predicted final infarct volume (sum of the predicted infarct core and salvageable ischemic tissue volumes), as well as the predicted salvageable ischemic tissue volume, was also significantly correlated with the true final infarct volume (r = 0.955, P < .001) and infarct growth (r = 0.918, P < .001), respectively. The volumes of perfusion-diffusion mismatch were significantly larger than those of infarct growth and predicted salvageable ischemic tissue. Good agreement between predicted and true final infarct lesions was visualized by Bland-Altman plots in two groups. Direct visual comparative analysis revealed good qualitative agreement between the true final infarct and predicted lesions in 21 patients. The proposed ADC based approach may be a feasible and practical tool to predict the volumes of infarct core and salvageable ischemic tissue without intravenous contrast media-enhanced perfusion-weighted imaging at baseline.

In vivo reflectance confocal microscopy: automated diagnostic image analysis of melanocytic skin tumours

In vivo reflectance confocal microscopy (RCM) has been shown to be a valuable imaging tool in the diagnosis of melanocytic skin tumours. However, diagnostic image analysis performed by automated systems is to date quite rare. In this study, we investigated the applicability of an automated image analysis system using a machine learning algorithm on diagnostic discrimination of benign and malignant melanocytic skin tumours in RCM. Overall, 16,269 RCM tumour images were evaluated. Image analysis was based on features of the wavelet transform. A learning set of 6147 images was used to establish a classification tree algorithm and an independent test set of 10, 122 images was applied to validate the tree model (grouping method 1). Additionally, randomly generated 'new' learning and test sets, tumour images only and different skin layers were evaluated (grouping method 2, 3 and 4). The classification tree analysis correctly classified 93.60% of the melanoma and 90.40% of the nevi images of the learning set. When the classification tree was applied to the independent test set 46.71 ± 19.97% (range 7.81-83.87%) of the tumour images in benign melanocytic skin lesions were classified as 'malignant', in contrast to 55.68 ± 14.58% (range 30.65-83.59%; t-test: P < 0.036) in malignant melanocytic skin lesions (grouping method 1). Further investigations could not improve the results significantly (grouping method 2, 3 and 4). The automated RCM image analysis procedure holds promise for further investigations. However, to date our system cannot be applied to routine skin tumour screening.

SPEX2: automated concise extraction of spatial gene expression patterns from Fly embryo ISH images

Motivation: Microarray profiling of mRNA abundance is often ill suited for temporal–spatial analysis of gene expressions in multicellular organisms such as Drosophila. Recent progress in image-based genome-scale profiling of whole-body mRNA patterns via in situ hybridization (ISH) calls for development of accurate and automatic image analysis systems to facilitate efficient mining of complex temporal–spatial mRNA patterns, which will be essential for functional genomics and network inference in higher organisms.Results: We present SPEX2, an automatic system for embryonic ISH image processing, which can extract, transform, compare, classify and cluster spatial gene expression patterns in Drosophila embryos. Our pipeline for gene expression pattern extraction outputs the precise spatial locations and strengths of the gene expression. We performed experiments on the largest publicly available collection of Drosophila ISH images, and show that our method achieves excellent performance in automatic image annotation, and also finds clusters that are significantly enriched, both for gene ontology functional annotations, and for annotation terms from a controlled vocabulary used by human curators to describe these images.Availability: Software will be available at http://www.sailing.cs.cmu.edu/Contact: epxing@cs.cmu.eduSupplementary information: Supplementary data are avilable at Bioinformatics online.

Fast-neutron spectroscopy studies using induced-proton tracks in PADC track detectors

In this work, a simple and adequate method for fast-neutron spectroscopy is proposed. This method was performed by free-in-air fast-neutron irradiation of CR-39 Nuclear Track Detectors (NTD) using an Am-Be source. Detectors were then chemically etched to remove few layers up to a thickness of 6.25 μm. By using an automatic image analyzer system for studying the registration of the induced-proton tracks in the NTD, the obtained data were analyzed via two tracks shapes. In the first one, the elliptical tracks were eliminated from the calculation and only the circular ones were considered in developing the response function. In the second method all registered tracks were considered and the corresponding response function was obtained. The rate of energy loss of the protons as a function of V[(dE/dX) - V] was calculated using the Monte Carlo simulation. The induced-proton energy was extracted from the corresponding dE/dX in NTD using a computer program based on the Bethe-Bloch function. The energy of the incident particles was up to few hundred MeV/nucleon. The energy of the interacting neutrons was then estimated by means of the extracted induced-proton energies and the scattering angle. It was found that the present resulting energy distribution of the fast-neutron spectrum from the Am-Be source was similar to that given in the literature where an average neutron energy of 4.6MeV was obtained.

“FISHing” to detect urinary and other cancers

The UroVysion Bladder Cancer Kit detects amplifications of chromosomes 3, 7, and 17, and the deletion of 9p21, by fluorescence in situ hybridization (FISH). Because manual interpretation of UroVysion FISH is time consuming and can be challenged by variable probe signal strengths and background labeling, the authors investigated an automated image analysis system to improve throughput, productivity, quality control, and accuracy. The authors evaluated the interactive BioView Duet imaging system as an aid to UroVysion FISH interpretation in a 2-armed, blinded comparison with manual screens of the same 135 consecutive cases. Manual and Duet-assisted interpretations were compared with respect to concordance, reproducibility, and timing. Eighty-one cases were interpreted as positive or negative with 94% concordance and a kappa value of 0.84 between manual and Duet-aided interpretations. Three cases that ultimately were judged positive were detected with the aid of Duet but were missed with a manual screen. A final interpretation could not be given for approximately 25% of Duet-scanned cases. Duet-aided interpretation was highly reproducible for patient and control slides. Pathologist evaluation time per case was 4 minutes compared with 30 minutes for manual interpretation. Cytotechnologist involvement added 18 minutes for a total of 22 minutes, a savings of 8 minutes per case. Duet-aided interpretations were at least equivalent to manual interpretations. The system permitted interactive review of abnormal cells and had the ability to evaluate the same cells for brightfield cytology followed by FISH. The image processing and analysis tools of the Duet system enhanced the morphology skills of cytology professionals in providing accurate interpretations.

Learning natural scene categories by selective multi-scale feature extraction

Natural scene categorization from images represents a very useful task for automatic image analysis systems. In the literature, several methods have been proposed facing this issue with excellent results. Typically, features of several types are clustered so as to generate a vocabulary able to describe in a multi-faceted way the considered image collection. This vocabulary is formed by a discrete set of visual codewords whose co-occurrence and/or composition allows to classify the scene category. A common drawback of these methods is that features are usually extracted from the whole image, actually disregarding whether they derive properly from the natural scene to be classified or from foreground objects, possibly present in it, which are not peculiar for the scene. As quoted by perceptual studies, objects present in an image are not useful to natural scene categorization, indeed bringing an important source of clutter, in dependence of their size. In this paper, a novel, multi-scale, statistical approach for image representation aimed at scene categorization is presented. The method is able to select, at different levels, sets of features that represent exclusively the scene disregarding other non-characteristic, clutter, elements. The proposed procedure, based on a generative model, is then able to produce a robust representation scheme, useful for image classification. The obtained results are very convincing and prove the goodness of the approach even by just considering simple features like local color image histograms.