Fluorescence In Situ Hybridization Images Research Articles

Deep learning to automatically evaluate HER2 gene amplification status from fluorescence in situ hybridization images

Human epidermal growth factor receptor 2 (HER2) gene amplification helps identify breast cancer patients who may respond to targeted anti-HER2 therapy. This study aims to develop an automated method for quantifying HER2 fluorescence in situ hybridization (FISH) signals and improve the working efficiency of pathologists. An Aitrox artificial intelligence (AI) model based on deep learning was constructed, and a comparison between the AI model and traditional manual counting was performed. In total, 918 FISH images from 320 consecutive invasive breast cancers were analysed and automatically classified into 5 groups according to the 2018 ASCO/CAP guidelines. The overall classification accuracy was 85.33% (157/184) with a mean average precision of 0.735. In Group 5, the most common group, the consistency was as high as 95.90% (117/122), while the consistency was low in the other groups due to the limited number of cases. The causes of this inconsistency, including clustered HER2 signals, coarse CEP17 signals and some section quality problems, were analysed. The developed AI model is a reliable tool for evaluating HER2 amplification statuses, especially for breast cancer in Group 5; additional cases from multiple centres could further improve the accuracy achieved for other groups.

Abstract A010: ANO7 expression changes in prostate cancer progression — Association with an aggressive phenotype?

Abstract Introduction The expression level and single nucleotide polymorphisms of the prostate-specific gene ANO7 have been shown to correlate with prostate cancer (PrCa) outcome. Very little is known about ANO7, apart from its expression profile in tissues and that the protein is either a calcium-activated ion channel or a lipid scramblase. Our results suggest that ANO7 likely has a role in PrCa progression and that loss of ANO7 is associated with malignant PrCa. Materials and methods Sections of a prostate cancer formalin-fixed paraffin-embedded tissue microarray were used for ANO7 mRNA and protein detection. mRNA was detected with fluorescent in situ hybridization (FISH). Protein was detected with immunohistochemistry (IHC) using a polyclonal anti-ANO7 antibody (HPA035730). mRNA from fresh radical prostatectomy primary tumor samples were sequenced and subsequent gene expression levels were analyzed for co-expression with ANO7. The sample set consisted of a matched tumor and benign tissue sample (15+15) from each patient. Results The FISH images indicated explicitly that ANO7 mRNA is exclusively expressed in the luminal cells of the prostatic glandular epithelium. ANO7 mRNA is expressed in all cancerous lumina-forming glandular structures, but not in poorly-formed glands or sheets of cancer, in which ANO7 mRNA expression is very low or non-existent. Morphologically benign glands amidst cancerous tissue showed no expression, in contrast to tissue in which benign glands are predominating. The IHC showed a similar staining pattern to FISH.ANO7 co-expression gene set enrichment analysis resulted in 19 positively correlating and 77 negatively correlating pathways in KEGG (Kyoto Encyclopedia of Genes and Genomes) terms. Most of the positively correlating terms relate to metabolism, whereas among the negatively correlating terms, there were many signal transduction and cell growth and death-related pathways, such as Hippo, Ras, and p53 signaling pathways. Conclusions ANO7 mRNA and protein expression correlate spatially in the tissue, showing that the expression is specific to differentiated luminal cells and highly reduced in high-grade cancer. This notion is supported by the co-expressed gene-set enrichment analysis, which is showing a negative correlation of ANO7 with many pathways that are associated with aggressive PrCa. Mainly metabolism-related processes were positively correlated with ANO7, which could indicate that ANO7 is linked to the reprogramming of metabolism in the cells during cancer progression. Further research aims to identify the pathways in which ANO7 plays a significant role and to elucidate how the ANO7 protein functionally alters signaling related to PrCa progression. Citation Format: Olli Metsälä, Gudrun Wahlström, Pekka Taimen, Johanna Schleutker. ANO7 expression changes in prostate cancer progression — Association with an aggressive phenotype? [abstract]. In: Proceedings of the AACR Special Conference: Advances in Prostate Cancer Research; 2023 Mar 15-18; Denver, Colorado. Philadelphia (PA): AACR; Cancer Res 2023;83(11 Suppl):Abstract nr A010.

A Lightweight and Robust Framework for Circulating Genetically Abnormal Cells (CACs) Identification Using 4-Color Fluorescence In Situ Hybridization (FISH) Image and Deep Refined Learning.

Circulating genetically abnormal cells (CACs) constitute an important biomarker for cancer diagnosis and prognosis. This biomarker offers high safety, low cost, and high repeatability, which can serve as a key reference in clinical diagnosis. These cells are identified by counting fluorescence signals using 4-color fluorescence in situ hybridization (FISH) technology, which has a high level of stability, sensitivity, and specificity. However, there are some challenges in CACs identification, due to the difference in the morphology and intensity of staining signals. In this concern, we developed a deep learning network (FISH-Net) based on 4-color FISH image for CACs identification. Firstly, a lightweight object detection network based on the statistical information of signal size was designed to improve the clinical detection rate. Secondly, the rotated Gaussian heatmap with a covariance matrix was defined to standardize the staining signals with different morphologies. Then, the heatmap refinement model was proposed to solve the fluorescent noise interference of 4-color FISH image. Finally, an online repetitive training strategy was used to improve the model's feature extraction ability for hard samples (i.e., fracture signal, weak signal, and adjacent signals). The results showed that the precision was superior to 96%, and the sensitivity was higher than 98%, for fluorescent signal detection. Additionally, validation was performed using the clinical samples of 853 patients from 10 centers. The sensitivity was 97.18% (CI 96.72-97.64%) for CACs identification. The number of parameters of FISH-Net was 2.24M, compared to 36.9M for the popularly used lightweight network (YOLO-V7s). The detection speed was about 800 times greater than that of a pathologist. In summary, the proposed network was lightweight and robust for CACs identification. It could greatly increase the review accuracy, enhance the efficiency of reviewers, and reduce the review turnaround time during CACs identification.

Detection of ERBB2 and CEN17 signals in fluorescent in situ hybridization and dual in situ hybridization for guiding breast cancer HER2 target therapy

The overexpression of the human epidermal growth factor receptor 2 (HER2) is a predictive biomarker in therapeutic effects for metastatic breast cancer. Accurate HER2 testing is critical for determining the most suitable treatment for patients. Fluorescent in situ hybridization (FISH) and dual in situ hybridization (DISH) have been recognized as FDA-approved methods to determine HER2 overexpression. However, analysis of HER2 overexpression is challenging. Firstly, the boundaries of cells are often unclear and blurry, with large variations in cell shapes and signals, making it challenging to identify the precise areas of HER2-related cells. Secondly, the use of sparsely labeled data, where some unlabeled HER2-related cells are classified as background, can significantly confuse fully supervised AI learning and result in unsatisfactory model outcomes. In this study, we present a weakly supervised Cascade R-CNN (W-CRCNN) model to automatically detect HER2 overexpression in HER2 DISH and FISH images acquired from clinical breast cancer samples. The experimental results demonstrate that the proposed W-CRCNN achieves excellent results in identification of HER2 amplification in three datasets, including two DISH datasets and a FISH dataset. For the FISH dataset, the proposed W-CRCNN achieves an accuracy of 0.970±0.022, precision of 0.974±0.028, recall of 0.917±0.065, F1-score of 0.943±0.042 and Jaccard Index of 0.899±0.073. For DISH datasets, the proposed W-CRCNN achieves an accuracy of 0.971±0.024, precision of 0.969±0.015, recall of 0.925±0.020, F1-score of 0.947±0.036 and Jaccard Index of 0.884±0.103 for dataset 1, and an accuracy of 0.978±0.011, precision of 0.975±0.011, recall of 0.918±0.038, F1-score of 0.946±0.030 and Jaccard Index of 0.884±0.052 for dataset 2, respectively. In comparison with the benchmark methods, the proposed W-CRCNN significantly outperforms all the benchmark approaches in identification of HER2 overexpression in FISH and DISH datasets (p<0.05). With the high degree of accuracy, precision and recall , the results show that the proposed method in DISH analysis for assessment of HER2 overexpression in breast cancer patients has significant potential to assist precision medicine.

A Soft Label Deep Learning to Assist Breast Cancer Target Therapy and Thyroid Cancer Diagnosis.

According to the World Health Organization Report 2022, cancer is the most common cause of death contributing to nearly one out of six deaths worldwide. Early cancer diagnosis and prognosis have become essential in reducing the mortality rate. On the other hand, cancer detection is a challenging task in cancer pathology. Trained pathologists can detect cancer, but their decisions are subjective to high intra- and inter-observer variability, which can lead to poor patient care owing to false-positive and false-negative results. In this study, we present a soft label fully convolutional network (SL-FCN) to assist in breast cancer target therapy and thyroid cancer diagnosis, using four datasets. To aid in breast cancer target therapy, the proposed method automatically segments human epidermal growth factor receptor 2 (HER2) amplification in fluorescence in situ hybridization (FISH) and dual in situ hybridization (DISH) images. To help in thyroid cancer diagnosis, the proposed method automatically segments papillary thyroid carcinoma (PTC) on Papanicolaou-stained fine needle aspiration and thin prep whole slide images (WSIs). In the evaluation of segmentation of HER2 amplification in FISH and DISH images, we compare the proposed method with thirteen deep learning approaches, including U-Net, U-Net with InceptionV5, Ensemble of U-Net with Inception-v4, Inception-Resnet-v2 encoder, and ResNet-34 encoder, SegNet, FCN, modified FCN, YOLOv5, CPN, SOLOv2, BCNet, and DeepLabv3+ with three different backbones, including MobileNet, ResNet, and Xception, on three clinical datasets, including two DISH datasets on two different magnification levels and a FISH dataset. The result on DISH breast dataset 1 shows that the proposed method achieves high accuracy of 87.77 ± 14.97%, recall of 91.20 ± 7.72%, and F1-score of 81.67 ± 17.76%, while, on DISH breast dataset 2, the proposed method achieves high accuracy of 94.64 ± 2.23%, recall of 83.78 ± 6.42%, and F1-score of 85.14 ± 6.61% and, on the FISH breast dataset, the proposed method achieves high accuracy of 93.54 ± 5.24%, recall of 83.52 ± 13.15%, and F1-score of 86.98 ± 9.85%, respectively. Furthermore, the proposed method outperforms most of the benchmark approaches by a significant margin (p&amp;nbsp;&lt;0.001). In evaluation of segmentation of PTC on Papanicolaou-stained WSIs, the proposed method is compared with three deep learning methods, including Modified FCN, U-Net, and SegNet. The experimental result demonstrates that the proposed method achieves high accuracy of 99.99 ± 0.01%, precision of 92.02 ± 16.6%, recall of 90.90 ± 14.25%, and F1-score of 89.82 ± 14.92% and significantly outperforms the baseline methods, including U-Net and FCN (p&amp;nbsp;&lt;0.001). With the high degree of accuracy, precision, and recall, the results show that the proposed method could be used in assisting breast cancer target therapy and thyroid cancer diagnosis with faster evaluation and minimizing human judgment errors.

Lighting Up Nucleic Acid Modifications in Single Cells with DNA-Encoded Amplification.

Nucleic acids are naturally decorated with various chemical modifications at nucleobases. Most nucleic acid modifications (NAMs) do not alter Watson-Crick base pairing but can regulate gene expression known as "epigenetics". Their abundances present a very wide range, approximately 10-2 to 10-6 of total bases. Different NAMs may coexist in spatial proximity (e.g., <20 nm) in the crowded intracellular environment. Considering the highly dynamic chromatin accessibility (physical access to DNA), the NAMs in inaccessible DNA probably plays different roles. These multilayered features of NAMs vary from cell to cell. Our understanding of the function and mechanism of NAMs in biological processes and disease states has largely been driven by the expanding array of sequencing-based methodologies. However, an underexplored aspect is the measurement of the subcellular distribution, spatial proximity, and inaccessibility of NAMs in single cells. In recent years, we have developed new approaches that light up single-cell NAMs with single-site sensitivity. These methods are mainly based on the integration of chemical or chemoenzymatic tools, DNA amplification and nanotechnology, and/or microfluidics. An overview of these methods together with conventional methods such as immunofluorescence (IF) and fluorescence in situ hybridization (FISH) is provided in this Account.Our laboratory has proposed DNA-encoded amplification (DEA) as the main strategy for developing a set of single-cell NAM imaging methods. In brief, DEA transforms the different features of NAMs into unique DNA primers for rolling circle amplification (RCA) followed by FISH imaging. The first method is base-encoded amplifying FISH (BEA-FISH), in which we convert individual NAMs into RCA primers via chemoselective labeling and click bioconjugation. It enables the in situ visualization of low-abundance NAMs (e.g., 5hmU), which is impracticable by conventional methods. We subsequently developed pairwise proximity-differentiated amplifying FISH (PPDA-FISH), which integrates BEA-FISH with DNA nanotechnology. PPDA-FISH utilizes proximity ligation and toehold strand displacement to label the adjacent site of two different NAMs (one-to-one proximity) and their respective residual sites with three unique RCA probes. It achieves simultaneous counting of the above-mentioned three types of modified sites in the same cells. The third method is cellular macromolecule-tethered DNA walking indexing (Cell-TALKING) to probe more than two NAMs within the same nanoenvironments. Cell-TALKING uses dynamic DNA proximity cleavage to continuously activate different preblocked RCA primers (for each NAM) near one walking probe (for one target molecule). We have explored three NAMs around one histone (one-to-many proximity) in different cancer cell lines and clinical specimens. Then, we describe a single-cell hydrogel encoding amplification (scHEA) method by integrating droplet microfluidics with BEA-FISH. This method generates hydrogel beads that encapsulate single cells and their genomic DNA after cell lysis. It realizes the analysis of global (accessible and inaccessible) DNA from the same cells. We find that the global levels of both 5hmC and 5hmU in single cells can distinguish different breast cancer cells. Finally, the current limitations of these strategies and the future development directions are also discussed. We hope that this Account can spark new ideas and invite new efforts from different disciplines for single-cell NAM analysis.

Multimodal Registration of Fish and Nanosims Images Using Convolutional Neural Network Models

Nanoscale secondary ion mass spectrometry (nanoSIMS) and fluorescence in situ hybridization (FISH) microscopy provide high-resolution, multimodal image representations of the identity and cell activity respectively of targeted microbial communities in microbiological research. Despite its importance to microbiologists, multimodal registration of FISH and nanoSIMS images is challenging given the morphological distortion and background noise in both images. In this study, we use convolutional neural networks (CNNs) for multiscale feature extraction, shape context for computation of the minimum transformation cost feature matching and the thin-plate spline (TPS) model for multimodal registration of the FISH and nanoSIMS images. Registration accuracy was quantitatively assessed against manually registered images, at both, the pixel and structural levels using standard metrics. Although all six tested CNN models performed well, ResNet18 was observed to outperform VGG16, VGG19, GoogLeNet and ShuffleNet and ResNet101 based on most metrics. This study demonstrates the utility of CNNs in the registration of multimodal images with significant background noise and morphology distortion. We also show aggregate shape, preserved by binarization, to be a robust feature for registering multimodal microbiology-related images.

Preparation of long single-strand DNA concatemers for high-level fluorescence in situ hybridization

Fluorescence in situ hybridization (FISH) is a powerful tool to visualize transcripts in fixed cells and tissues. Despite the recent advances in FISH detection methods, it remains challenging to achieve high-level FISH imaging with a simple workflow. Here, we introduce a method to prepare long single-strand DNA concatemers (lssDNAc) through a controllable rolling-circle amplification (CRCA). Prepared lssDNAcs are used to develop AmpFISH workflows. In addition, we present its applications in different scenarios. AmpFISH shows the following advantages: 1) enhanced FISH signal-to-noise ratio (SNR) up to 160-fold compared with single-molecule FISH; 2) simultaneous detection of FISH signals and fluorescent proteins or immunofluorescence (IF) in tissues; 3) simple workflows; and 4) cost-efficiency. In brief, AmpFISH provides convenient and versatile tools for sensitive RNA/DNA detection and to gain useful information on cellular molecules using simple workflows.

Towards computer aided diagnosis of infective endocarditis in whole-slide images of heart valve tissue using FISH

Abstract Infective endocarditis (IE) is an infection of the endocardium, and the heart valves associated with high morbidity and mortality. Fluorescence in situ Hybridization (FISH) is a molecular imaging technique used for diagnosis of IE based on histological heart valve tissue sections. FISH allows detection and identification of microorganisms and gives information about their quantity and spatial distribution. This information is important to guide appropriate antibiotic treatment. However, as manual FISH image analysis is time- and costexpensive, an automated image analysis pipeline (consisting of tissue segmentation, bacteria detection, and spot detection modules) is proposed to assist locating potential regions with microorganisms. The proposed approach was evaluated in a study, where five observers manually assessed a set of 171 fields-of-view (FoVs) captured in 400-fold magnification from 10 randomly chosen WSI for the presence of microorganisms, morphologically detected by the nucleic acid stain DAPI. The task of the observers was to mark the presented image using a 2-class score (‘positive/questionable’ or ‘negative’). The human assessment was compared to the results suggested by the algorithm. The proposed algorithm locates and ranks potential regions with microorganisms in heart valve sections so that experts can validate them in higher power FoVs for the presence of bacteria and identify their species. The automated system for preselecting and recommending adequate FoVs is thus a starting point to support experts and save human resources. It is now ready to be further developed for the detection of bacteria by FISH.

Applying fluorescence in situ hybridization to aquatic systems with cyanobacteria blooms: Autofluorescence suppression and high‐throughput image analysis

AbstractCyanobacterial Harmful Algal Blooms (CyanoHABs) are expanding geographically in both fresh and marine water bodies due to coastal eutrophication and global climate change and are restructuring the microbial ecology of these systems. Cyanobacterial autofluorescence can pose a significant impediment to accurately identifying prokaryotic taxonomic groups in environmental samples using fluorescence in situ hybridization (FISH). This can hinder our ability to accurately quantify, and therefore fully understand ecological changes. As abundances of FISH target cells and autofluorescent cells can often be of the same the order of magnitude, simply subtracting average autofluorescent cell concentrations—determined from enumerating unhybridized samples—yields apparent concentrations of target cells with unacceptably large analytical uncertainty. Here we present a CuSO4/EtOH chemical pretreatment protocol that significantly reduces undesirable autofluorescence in hybridized environmental samples. We apply a novel data filtration routine to FISH images that efficiently removes residual autofluorescent cells from final cell counts. We then subject images to an automated image analysis routine that accurately enumerates probe‐positive cells. This method is inexpensive and easy to implement as part of a routine FISH workflow. By applying this method to cyanobacteria rich samples, we can better understand how microbial community changes are contributing to globally changing biogeochemical cycles.

Optimizing the hybridization chain reaction-fluorescence in situ hybridization (HCR-FISH) protocol for detection of microbes in sediments

Fluorescence in situ hybridization (FISH) is a canonical tool commonly used in environmental microbiology research to visualize targeted cells. However, the problems of low signal intensity and false-positive signals impede its widespread application. Alternatively, the signal intensity can be amplified by incorporating Hybridization Chain Reaction (HCR) with FISH, while the specificity can be improved through protocol modification and proper counterstaining. Here we optimized the HCR-FISH protocol for studying microbes in environmental samples, particularly marine sediments. Firstly, five sets of HCR initiator/amplifier pairs were tested on the laboratory-cultured bacterium Escherichia coli and the archaeon Methanococcoides methylutens, and two sets displayed high hybridization efficiency and specificity. Secondly, we tried to find the best combination of sample pretreatment methods and HCR-FISH protocol for environmental sample analysis with the aim of producing less false positive signals. Various detachment methods, extraction methods and formulas of hybridization buffer were tested using sediment samples. Thirdly, an image processing method was developed to enhance the DAPI signal of microbial cells against that of abiotic particles, providing a reliable reference for FISH imaging. In summary, our optimized HCR-FISH protocol showed promise to serve as an addendum to traditional FISH for research on environmental microbes.

The Myoepithelial Cells of Salivary Intercalated Duct-type Intraductal Carcinoma Are Neoplastic: A Study Using Combined Whole-slide Imaging, Immunofluorescence, and RET Fluorescence In Situ Hybridization.

Intraductal carcinoma (IDC) is a salivary gland tumor currently believed to be analogous to breast ductal carcinoma in situ, consisting of a complex neoplastic epithelial proliferation surrounded by a continuous layer of myoepithelial cells presumed to be native and non-neoplastic. Recent molecular insights have shown that there are at least 3 different types of IDC: (1) intercalated duct-like, with frequent NCOA4-RET fusions; (2) apocrine, with multiple mutations similar to salivary duct carcinoma; and (3) mixed intercalated duct-like and apocrine with frequent RET fusions, especially TRIM27-RET. Recent observations (eg, IDC occurring in lymph nodes) have challenged the notion that the myoepithelial cells of IDC are non-neoplastic. Five IDCs with known RET fusions by RNA sequencing were retrieved from the authors' archives, including 4 intercalated duct-like IDCs with NCOA4-RET, and 1 mixed intercalated duct-like/apocrine IDC with TRIM27-RET. A panel of immunohistochemistry antibodies (S100 protein, p63 or p40, mammaglobin, smooth muscle actin, calponin, androgen receptor) was tested. To precisely localize RET split-positive cells, each case was subjected to sequential retrieval of whole-slide imaging data of hematoxylin and eosin (HE) staining, immunofluorescence staining for calponin, and fluorescence in situ hybridization (FISH) for RET. Because NCOA4-RET is an inversion difficult to visualize on conventional RET FISH, a novel 3-color FISH technique was utilized to demonstrate it clearly. In all 5 cases, the proliferative ducts were completely surrounded by a layer of myoepithelial cells that were positive for p63 or p40, smooth muscle actin, and calponin. Using combined HE, calponin immunofluorescence, and RET FISH imaging, the positive signals were unmistakably identified in both calponin-negative ductal cells and peripheral, calponin-positive myoepithelial cells in all 5 cases. Utilizing combined HE, calponin immunofluorescence, and RET FISH imaging, we demonstrated that IDCs with RET fusions harbored this alteration in both the ductal and myoepithelial cells. This is compelling evidence that the myoepithelial cells of IDC are not mere bystanders, but are rather a component of the neoplasm itself, similar to other biphasic salivary gland neoplasms like pleomorphic adenoma and epithelial-myoepithelial carcinoma. This finding raises questions about the appropriate terminology, classification, and staging of IDC.

Chromosomal mapping of 5S and 18S-5.8S-25S rRNA genes in Saccharina japonica (Phaeophyceae) as visualized by dual-color fluorescence in situ hybridization

It has been reported that there was a linkage of 5S rRNA gene to 18S-5.8S-25S rRNA gene in a few of species in Ochrophyta. In regard to the usual two positions of linked 5S rDNA to the 3' end of 25S rDNA, two pairs of primers were designed for amplification to verify this linkage of two genes in a kelp cultivar of Saccharina japonica, one of species in Ochrophyta. This result supplemented the previous report that 5S rDNA was unlinked to 25S rDNA in this kelp. In order to simultaneously visualize this unlinkage of two genes, dual-color fluorescence in situ hybridization (FISH) technique was applied to the cytogenetics of S. japonica. Dual-color FISH images showed that two and four hybridization signals were present in the kelp gametophyte and sporophyte, respectively, metaphase nuclei hybridized simultaneously with the labeled probes of 18S rDNA and 5S rDNA. Both haploid and diploid karyotypes in decreasing length of chromosomes showed that 18S-5.8S-25S rDNA was localized at the interstitial region of Chromosome 23, whereas 5S rDNA resided at the sub-telomeric region of Chromosome 27. These karyotypes suggested that the kelp nuclear genome had only one locus of each rRNA gene, and their loci on different chromosomes indicated the physical unlinkage of 5S rDNA to 18S-5.8S-25S rDNA in this kelp. Therefore, dual-color FISH seems to be a powerful technique for the discrimination and pairing of chromosomes featured in both small size and nearly identical shape in S. japonica.

Artificial intelligence and multivariate statistics for comprehensive assessment of filamentous bacteria in wastewater treatment plants experiencing sludge bulking

Abstract Sludge bulking is an operational hurdle that affects the solid–liquid separation in wastewater treatment plants (WWTPs) worldwide. In this study, filamentous bulking issues were investigated in seven WWTPs located in South Africa using artificial neural network (ANN) and multivariate statistics. The microbial community belonging to sludge bulking was determined using staining and microscopic methods, with a further confirmed identification of selected species via fluorescent in situ hybridization (FISH). Based on a filament index scale from 1 (None filament) to 7 (Excessive filament), the developed ANN could predict the sludge volume index (SVI) using the abundances of ten inputs of filamentous species. Eikelboom Type 0041 attained the highest impact on SVI, followed by Gordonia spp., Nostocoida limicola, and Thiothrix spp. Principal component analysis (PCA) combined with FISH images showed that most WWTPs experienced inadequate sludge settling properties; however, the application of an efficient aeration system (i.e., diffusion) in the three-stage Phoredox process improved the settling characteristics of bio-flocs. Operational conditions that caused filament overgrowth in each WWTP were also determined. The study outputs would provide a scientific basis to control the proliferation of filamentous bacteria in other WWTPs located in similar environmental conditions to South Africa.

Preparing Maize Synaptonemal Complex Spreads and Sequential Immunofluorescence and Fluorescence In Situ Hybridization.

Immunofluorescence and fluorescence in situ hybridization (FISH) can be used to locate specific proteins and DNA sequences, respectively, in chromosomes by light microscopy. Here we describe sequential use of these techniques on spreads of maize synaptonemal complexes (SCs) to determine whether crossing over can occur in knob heterochromatin. We used antibodies to AFD1, an SC protein, and MLH1, a class I (interference-sensitive) crossover protein found in most recombination nodules (RNs) to identify crossovers (COs) along SCs. Next, we used FISH to localize a 180bp knob-specific tandem repeat. Combining immunofluorescence and FISH images of the same SC spreads showed that heterochromatic knobs do not prohibit class I COs. This technique is broadly applicable to investigations of plant prophase I chromosomes where meiotic recombination takes place.

Automated detection of the HER2 gene amplification status in Fluorescence in situ hybridization images for the diagnostics of cancer tissues

The human epidermal growth factor receptor 2 (HER2) gene amplification status is a crucial marker for evaluating clinical therapies of breast or gastric cancer. We propose a deep learning-based pipeline for the detection, localization and classification of interphase nuclei depending on their HER2 gene amplification state in Fluorescence in situ hybridization (FISH) images. Our pipeline combines two RetinaNet-based object localization networks which are trained (1) to detect and classify interphase nuclei into distinct classes normal, low-grade and high-grade and (2) to detect and classify FISH signals into distinct classes HER2 or centromere of chromosome 17 (CEN17). By independently classifying each nucleus twice, the two-step pipeline provides both robustness and interpretability for the automated detection of the HER2 amplification status. The accuracy of our deep learning-based pipeline is on par with that of three pathologists and a set of 57 validation images containing several hundreds of nuclei are accurately classified. The automatic pipeline is a first step towards assisting pathologists in evaluating the HER2 status of tumors using FISH images, for analyzing FISH images in retrospective studies, and for optimizing the documentation of each tumor sample by automatically annotating and reporting of the HER2 gene amplification specificities.

PTEN loss in Gleason grade 7 prostate tumors exhibits intratumoral heterogeneity and is associated with unfavorable pathological features

BackgroundPTEN loss is observed in 20–30% of prostate cancers and is associated with a poor outcome, but clinical details of the impact of this biomarker are unclear for intermediate grade tumors.MethodsWe investigated 43 radical prostatectomy-derived grade 7 prostate tumors from the Clinics Hospital of Ribeirão Preto. Tissue microarray (TMA) blocks were constructed and PTEN copy number status was determined for all patients through fluorescence in situ hybridization (FISH). To determine the presence of PTEN protein loss in our study cohort, we performed immunohistochemistry (IHC) in TMA sections. We then developed an automated algorithm in HALO™ to identify regions of PTEN protein loss in whole prostate scanned sections from ten patients with known PTEN deletion status by FISH. Clinical analyses were conducted to determine the associations between PTEN loss and patient outcome. All statistical analyses were conducted in R v3.4.3 with P-values below 0.05 being considered statistically significant.ResultsIn this study of 43 grade 7 tumors, we found PTEN deletions by FISH in 18.9% of tumors, and PTEN protein loss by IHC in 16.3% of tumors. Both techniques were highly concordant and complementary. Clinical analysis demonstrated that PTEN deletion by FISH was significantly associated with positive margin invasion (P = 0.04) and Gleason score upgrade (P = 0.001). Digital image analysis of ten representative tumors demonstrated distinct intratumoral heterogeneity for PTEN protein loss in four tumors.ConclusionsThis study shows that PTEN loss in Gleason grade 7 tumors can be heterogeneous and that a systematic analysis of this biomarker using a combination of FISH, IHC, and digital imaging may identify patients with a greater risk of poor outcome.

691P - Inter-patient and intra-tumoral heterogeneity of oncogenic copy number alterations (CNAs) in gastric and gastroesophageal junction (GEJ) adenocarcinomas

Background: Intra-tumoral heterogeneity is well recognized to be inherent in biomarker discovery for gastric cancer since the initial reporting of HER2 overexpression. The Cancer Genome Atlas (TCGA) has laid the groundwork for CNAs comprising the mutational landscape of gastric cancer. We aimed to investigate through a genomic single nucleotide polymorphism (SNP) array panel and fluorescence in-situ hybridization (FISH) inter-patient and intra-tumoral spatial heterogeneity of CNAs. Methods: 41 gastric adenocarcinoma patient samples treated with upfront surgical resection were retrospectively identified from the City of Hope Biospecimen Repository. CNAs of 891 cancer-related genes at 50-100 Kb resolution and detection of 74 frequent somatic mutations in 9 genes of interest (BRAF, KRAS, EGFR, IDH1, IDH2, PTEN, PIK3CA, NRAS, TP53) were assayed using the Affymetrix OncoScan™ platform. Genome wide coverage outside of the 891 cancer genes were provided at 300 Kb resolution along with genome wide LOH provided at 3-10 Mb resolution. For samples with multiple CNAs detected, FISH was pursued to define down to a single-cell level the spatial distribution of CNAs. Results: Detectable percentage (%) genomic changes ranged from 0.03 to 73.90%. Lauren intestinal subtype histology correlated strongly with higher % genomic changes compared to diffuse subtype histology (p = 0.0012). Tumors located in the GEJ/cardia/proximal stomach also correlated with higher % genomic changes compared to gastric body/antrum tumors (p = 0.0028). A variety of oncogenic CNAs were observed across patients including high copy gains in EGFR, JAK2, FGFR2, MET, VEGFA, KRAS, NRAS, and PIK3CA. One sample exhibited co-amplification of CD274 and PDCD1LG2 (encoding PD-L1 and PD-L2), in addition to concurrent amplification of ERBB2, JAK2, FGFR2, MET, KRAS, and PIK3CA. FISH images of the spatial distribution of CNAs will be presented. Conclusions: The inherent spatial intra-tumoral heterogeneity of oncogenic CNAs with de novo disease presentation illustrates the challenges in gastric cancer therapy. Further study will offer insight into strategies on combinatorial and/or sequential targeted and immunotherapeutic approaches. Legal entity responsible for the study: Joseph Chao Funding: United States National Institutes of Health - National Cancer Institute Disclosure: All authors have declared no conflicts of interest.

Evaluation of a confocal WSI scanner for FISH slide imaging and image analysis

Background: Technological advances contribute to the maturation of digital pathology in clinical and research applications. However, there are only few reports on fluorescence scanning especially on confocal fluorescence imaging technology in digital pathology, which has superior depth resolution compared to wide-field fluorescence imaging. Here, we explored the features of a confocal WSI scanner for typical diagnostic and research imaging applications of fluorescence in situ hybridization (FISH) assay.Methods: Multi-layer stacking (Z-stack) which stores all image information from each layer, and extended focus which stores the optimal image information from all scanned layers, featured in the Pannoramic Confocal scanner (3DHISTECH Ltd., Budapest, Hungary) were employed in digitizing 14 FISH slides (ALK, EGFR, and multi-gene). The slides were scanned with a 40× water immersion objective producing a final image with pixel resolution of 0.1625 µm/pixel. Z-stack and extended focus were used for N=6, 13, and 26 multiple layers scanning at 1, 0.4, and 0.2 µm depth intervals respectively. Single-layer scanning was also done for comparison. Scanning time and resultant file size were recorded. The CaseViewer from 3DHISTECH was used to visualize images and export the annotated regions, and the exported images were further analyzed in Imaris (Bitplane, Zurich, Switzerland) for 3-dimensional reconstruction, nuclear segmentation, and the quantification and co-localization analysis of dots inside nuclei. Quantification data from Imaris were imported into Excel for statistic analysis.Results: Confocal fluorescence scanning of FISH slides enabled sharper image than wide-field scanning, although it required longer scanning time and larger file storage. More intra-nuclear dots were quantified from multi-layer Z-stack images than single-layer images, and the Z-stack increased scanning time and image file size. Furthermore, there were a reduced in the number of dots and an increased in the number of co-localized dots in extended-focus images compared to Z-stack. Dots in multiple channels were quantified and analyzed automatically, which supports clinical diagnosis of gene amplification, deletion, and translocation. Three-dimensional reconstruction of Z-stack produced precise measurement of spatial distance, which supports molecular research.Conclusion: Confocal provides sharper image than wide-field for FISH slide scanning. Extended focus reduces file size and storage, but could cause inaccurate analysis due to misinterpretation of overlapping information. Z-stack scanning provides high volume image information for spatial analysis. We foresee confocal multi-layer scanning as a digital pathology application tool for FISH imaging in both clinical and research in future.

Multifractal-based nuclei segmentation in fish images

The method for nuclei segmentation in fluorescence in-situ hybridization (FISH) images, based on the inverse multifractal analysis (IMFA) is proposed. From the blue channel of the FISH image in RGB format, the matrix of Holder exponents, with one-by-one correspondence with the image pixels, is determined first. The following semi-automatic procedure is proposed: initial nuclei segmentation is performed automatically from the matrix of Holder exponents by applying predefined hard thresholding; then the user evaluates the result and is able to refine the segmentation by changing the threshold, if necessary. After successful nuclei segmentation, the HER2 (human epidermal growth factor receptor 2) scoring can be determined in usual way: by counting red and green dots within segmented nuclei, and finding their ratio. The IMFA segmentation method is tested over 100 clinical cases, evaluated by skilled pathologist. Testing results show that the new method has advantages compared to already reported methods.