Justinas Besusparis

A methodology to ensure and improve accuracy of Ki67 labelling index estimation by automated digital image analysis in breast cancer tissue

IntroductionImmunohistochemical Ki67 labelling index (Ki67 LI) reflects proliferative activity and is a potential prognostic/predictive marker of breast cancer. However, its clinical utility is hindered by the lack of standardized measurement methodologies. Besides tissue heterogeneity aspects, the key element of methodology remains accurate estimation of Ki67-stained/counterstained tumour cell profiles. We aimed to develop a methodology to ensure and improve accuracy of the digital image analysis (DIA) approach.MethodsTissue microarrays (one 1-mm spot per patient, n = 164) from invasive ductal breast carcinoma were stained for Ki67 and scanned. Criterion standard (Ki67-Count) was obtained by counting positive and negative tumour cell profiles using a stereology grid overlaid on a spot image. DIA was performed with Aperio Genie/Nuclear algorithms. A bias was estimated by ANOVA, correlation and regression analyses. Calibration steps of the DIA by adjusting the algorithm settings were performed: first, by subjective DIA quality assessment (DIA-1), and second, to compensate the bias established (DIA-2). Visual estimate (Ki67-VE) on the same images was performed by five pathologists independently.ResultsANOVA revealed significant underestimation bias (P < 0.05) for DIA-0, DIA-1 and two pathologists’ VE, while DIA-2, VE-median and three other VEs were within the same range. Regression analyses revealed best accuracy for the DIA-2 (R-square = 0.90) exceeding that of VE-median, individual VEs and other DIA settings. Bidirectional bias for the DIA-2 with overestimation at low, and underestimation at high ends of the scale was detected. Measurement error correction by inverse regression was applied to improve DIA-2-based prediction of the Ki67-Count, in particular for the clinically relevant interval of Ki67-Count < 40%. Potential clinical impact of the prediction was tested by dichotomising the cases at the cut-off values of 10, 15, and 20%. Misclassification rate of 5-7% was achieved, compared to that of 11-18% for the VE-median-based prediction.ConclusionsOur experiments provide methodology to achieve accurate Ki67-LI estimation by DIA, based on proper validation, calibration, and measurement error correction procedures, guided by quantified bias from reference values obtained by stereology grid count. This basic validation step is an important prerequisite for high-throughput automated DIA applications to investigate tissue heterogeneity and clinical utility aspects of Ki67 and other immunohistochemistry (IHC) biomarkers.

A methodology for comprehensive breast cancer Ki67 labeling index with intra-tumor heterogeneity appraisal based on hexagonal tiling of digital image analysis data.

Digital image analysis (DIA) enables higher accuracy, reproducibility, and capacity to enumerate cell populations by immunohistochemistry; however, the most unique benefits may be obtained by evaluating the spatial distribution and intra-tissue variance of markers. The proliferative activity of breast cancer tissue, estimated by the Ki67 labeling index (Ki67 LI), is a prognostic and predictive biomarker requiring robust measurement methodologies. We performed DIA on whole-slide images (WSI) of 302 surgically removed Ki67-stained breast cancer specimens; the tumour classifier algorithm was used to automatically detect tumour tissue but was not trained to distinguish between invasive and non-invasive carcinoma cells. The WSI DIA-generated data were subsampled by hexagonal tiling (HexT). Distribution and texture parameters were compared to conventional WSI DIA and pathology report data. Factor analysis of the data set, including total numbers of tumor cells, the Ki67 LI and Ki67 distribution, and texture indicators, extracted 4 factors, identified as entropy, proliferation, bimodality, and cellularity. The factor scores were further utilized in cluster analysis, outlining subcategories of heterogeneous tumors with predominant entropy, bimodality, or both at different levels of proliferative activity. The methodology also allowed the visualization of Ki67 LI heterogeneity in tumors and the automated detection and quantitative evaluation of Ki67 hotspots, based on the upper quintile of the HexT data, conceptualized as the “Pareto hotspot”. We conclude that systematic subsampling of DIA-generated data into HexT enables comprehensive Ki67 LI analysis that reflects aspects of intra-tumor heterogeneity and may serve as a methodology to improve digital immunohistochemistry in general.

Quantification of myocardial fibrosis by digital image analysis and interactive stereology

BackgroundCardiac fibrosis disrupts the normal myocardial structure and has a direct impact on heart function and survival. Despite already available digital methods, the pathologist’s visual score is still widely considered as ground truth and used as a primary method in histomorphometric evaluations. The aim of this study was to compare the accuracy of digital image analysis tools and the pathologist’s visual scoring for evaluating fibrosis in human myocardial biopsies, based on reference data obtained by point counting performed on the same images.MethodsEndomyocardial biopsy material from 38 patients diagnosed with inflammatory dilated cardiomyopathy was used. The extent of total cardiac fibrosis was assessed by image analysis on Masson’s trichrome-stained tissue specimens using automated Colocalization and Genie software, by Stereology grid count and manually by Pathologist’s visual score.ResultsA total of 116 slides were analyzed. The mean results obtained by the Colocalization software (13.72 ± 12.24%) were closest to the reference value of stereology (RVS), while the Genie software and Pathologist score gave a slight underestimation. RVS values correlated strongly with values obtained using the Colocalization and Genie (r > 0.9, p < 0.001) software as well as the pathologist visual score. Differences in fibrosis quantification by Colocalization and RVS were statistically insignificant. However, significant bias was found in the results obtained by using Genie versus RVS and pathologist score versus RVS with mean difference values of: -1.61% and 2.24%. Bland-Altman plots showed a bidirectional bias dependent on the magnitude of the measurement: Colocalization software overestimated the area fraction of fibrosis in the lower end, and underestimated in the higher end of the RVS values. Meanwhile, Genie software as well as the pathologist score showed more uniform results throughout the values, with a slight underestimation in the mid-range for both.ConclusionBoth applied digital image analysis methods revealed almost perfect correlation with the criterion standard obtained by stereology grid count and, in terms of accuracy, outperformed the pathologist’s visual score. Genie algorithm proved to be the method of choice with the only drawback of a slight underestimation bias, which is considered acceptable for both clinical and research evaluations.Virtual slidesThe virtual slide(s) for this article can be found here: http://www.diagnosticpathology.diagnomx.eu/vs/9857909611227193

Bimodality of intratumor Ki67 expression is an independent prognostic factor of overall survival in patients with invasive breast carcinoma.

Proliferative activity, assessed by Ki67 immunohistochemistry (IHC), is an established prognostic and predictive biomarker of breast cancer (BC). However, it remains under-utilized due to lack of standardized robust measurement methodologies and significant intratumor heterogeneity of expression. A recently proposed methodology for IHC biomarker assessment in whole slide images (WSI), based on systematic subsampling of tissue information extracted by digital image analysis (DIA) into hexagonal tiling arrays, enables computation of a comprehensive set of Ki67 indicators, including intratumor variability. In this study, the tiling methodology was applied to assess Ki67 expression in WSI of 152 surgically removed Ki67-stained (on full-face sections) BC specimens and to test which, if any, Ki67 indicators can predict overall survival (OS). Visual Ki67 IHC estimates and conventional clinico-pathologic parameters were also included in the study. Analysis revealed linearly independent intrinsic factors of the Ki67 IHC variance: proliferation (level of expression), disordered texture (entropy), tumor size and Nottingham Prognostic Index, bimodality, and correlation. All visual and DIA-generated indicators of the level of Ki67 expression provided significant cutoff values as single predictors of OS. However, only bimodality indicators (Ashman’s D, in particular) were independent predictors of OS in the context of hormone receptor and HER2 status. From this, we conclude that spatial heterogeneity of proliferative tumor activity, measured by DIA of Ki67 IHC expression and analyzed by the hexagonal tiling approach, can serve as an independent prognostic indicator of OS in BC patients that outperforms the prognostic power of the level of proliferative activity.

Digital immunohistochemistry: new horizons and practical solutions in breast cancer pathology.

Digital image analysis (DA) brings new opportunities to enhance breast cancer pathology testing by providing tools to read tissue microscopy-based data in a more precise, accurate, and high-throughput manner. The applications vary from very practical solutions assisting in quantification of biomarker expression to more complex efforts to analyse multi-dimensional data. We present our DA experiments to enhance HER2 immunohistochemistry (IHC) and FISH testing and to explore multiple IHC biomarker variation profiles in breast cancer.

Digital immunohistochemistry platform for the staining variation monitoring based on integration of image and statistical analyses with laboratory information system

BackgroundDigital immunohistochemistry (IHC) is one of the most promising applications brought by new generation image analysis (IA). While conventional IHC staining quality is monitored by semi-quantitative visual evaluation of tissue controls, IA may require more sensitive measurement. We designed an automated system to digitally monitor IHC multi-tissue controls, based on SQL-level integration of laboratory information system with image and statistical analysis tools.MethodsConsecutive sections of TMA containing 10 cores of breast cancer tissue were used as tissue controls in routine Ki67 IHC testing. Ventana slide label barcode ID was sent to the LIS to register the serial section sequence. The slides were stained and scanned (Aperio ScanScope XT), IA was performed by the Aperio/Leica Colocalization and Genie Classifier/Nuclear algorithms. SQL-based integration ensured automated statistical analysis of the IA data by the SAS Enterprise Guide project. Factor analysis and plot visualizations were performed to explore slide-to-slide variation of the Ki67 IHC staining results in the control tissue.ResultsSlide-to-slide intra-core IHC staining analysis revealed rather significant variation of the variables reflecting the sample size, while Brown and Blue Intensity were relatively stable. To further investigate this variation, the IA results from the 10 cores were aggregated to minimize tissue-related variance. Factor analysis revealed association between the variables reflecting the sample size detected by IA and Blue Intensity. Since the main feature to be extracted from the tissue controls was staining intensity, we further explored the variation of the intensity variables in the individual cores. MeanBrownBlue Intensity ((Brown+Blue)/2) and DiffBrownBlue Intensity (Brown-Blue) were introduced to better contrast the absolute intensity and the colour balance variation in each core; relevant factor scores were extracted. Finally, tissue-related factors of IHC staining variance were explored in the individual tissue cores.ConclusionsOur solution enabled to monitor staining of IHC multi-tissue controls by the means of IA, followed by automated statistical analysis, integrated into the laboratory workflow. We found that, even in consecutive serial tissue sections, tissue-related factors affected the IHC IA results; meanwhile, less intense blue counterstain was associated with less amount of tissue, detected by the IA tools.

Digital immunohistochemistry wizard: image analysis-assisted stereology tool to produce reference data set for calibration and quality control.

BackgroundDigital image analysis (DIA) enables better reproducibility of immunohistochemistry (IHC) studies. Nevertheless, accuracy of the DIA methods needs to be ensured, demanding production of reference data sets. We have reported on methodology to calibrate DIA for Ki67 IHC in breast cancer tissue based on reference data obtained by stereology grid count. To produce the reference data more efficiently, we propose digital IHC wizard generating initial cell marks to be verified by experts.MethodsDigital images of proliferation marker Ki67 IHC from 158 patients (one tissue microarray spot per patient) with an invasive ductal carcinoma of the breast were used. Manual data (mD) were obtained by marking Ki67-positive and negative tumour cells, using a stereological method for 2D object enumeration. DIA was used as an initial step in stereology grid count to generate the digital data (dD) marks by Aperio Genie and Nuclear algorithms. The dD were collected into XML files from the DIA markup images and overlaid on the original spots along with the stereology grid. The expert correction of the dD marks resulted in corrected data (cD). The percentages of Ki67 positive tumour cells per spot in the mD, dD, and cD sets were compared by single linear regression analysis. Efficiency of cD production was estimated based on manual editing effort.ResultsThe percentage of Ki67-positive tumor cells was in very good agreement in the mD, dD, and cD sets: regression of cD from dD (R2=0.92) reflects the impact of the expert editing the dD as well as accuracy of the DIA used; regression of the cD from the mD (R2=0.94) represents the consistency of the DIA-assisted ground truth (cD) with the manual procedure. Nevertheless, the accuracy of detection of individual tumour cells was much lower: in average, 18 and 219 marks per spot were edited due to the Genie and Nuclear algorithm errors, respectively. The DIA-assisted cD production in our experiment saved approximately 2/3 of manual marking.ConclusionsDigital IHC wizard enabled DIA-assisted stereology to produce reference data in a consistent and efficient way. It can provide quality control measure for appraising accuracy of the DIA steps.

Quantification Accuracy of Liver Fibrosis by In Vivo Elastography and Digital Image Analysis of Liver Biopsy Histochemistry

1Department of Pathology, Forensic Medicine and Pharmacology, Faculty of Medicine, Vilnius University, 21 M. K. Ciurlionio Street, LT-03101 Vilnius, Lithuania 2Université de Caen, Esplanade de la Paix, 14032 Caen, France 3Centre of Hepatology, Gastroenterology and Dietetics, Vilnius University Hospital Santariskiu Clinics, 2 Santariskiu Street, LT-08661 Vilnius, Lithuania 4Clinic of Gastroenterology, Nephrourology and Surgery, Medical Faculty, Vilnius University, 2 Santariskiu Street, LT-08661 Vilnius, Lithuania

Abstract B116: A methodology to ensure and improve accuracy of Ki67 digital immunohistochemistry analysis in breast cancer tissue

Background: Immunohistochemical Ki67 evaluation reflects proliferative activity and is one of most important prognostic/predictive markers of breast cancer. However, standardized and efficient methodologies to accurately and reproducibly measure the Ki67 expression are still to come. Besides tissue processing, sampling, intra-tumour variability, and many other aspects to be considered, key element of the methodology remains accurate enumeration of Ki67-labelling index (LI). We aimed to develop a methodology to estimate and improve accuracy of automated image analysis (IA) approach. Methods: Tissue microarrays (1 mm diameter spot per patient, n=164) from invasive ductal breast carcinoma, stained for Ki67 and digitized by Aperio XT scanner, were used for the study. Reference values (RV) were obtained by counting the LI using stereological frame overlaid on a spot image. To test the degree of inter-observer variation in establishing the RV, the frame counts were performed by 3 observers independently in a subset (n=30) of the TMA images. IA was performed with Aperio Genie/Nuclear algorithms enabling automated selection of tumour tissue. Accuracy of the IA compared to the RV was estimated based on ANOVA, correlation and regression analyses performed with SAS 9.3. Agreement between individual measurements was also estimated based on 95% confidence intervals calculated from the RV according to stereology rules. Several iterations of the IA with adjusted algorithm settings were performed to improve the accuracy. Highly automated calibration cycles were enabled by developing software to integrate processes of the image and statistical analyses. Visual evaluation for the LI on the same images was performed by 3 pathologists (P1, P2, P3). Results: Inter-observer variation between 3 independent frame counts (n=30) was negligible by ANOVA (respectively, mean RV were 28.5, 28.6 and 29.9%) with correlation coefficients 0.97 and above. RV correlated strongly with IA (r=0.95) and P1, P2, P3 (r=0.86, r=0.90, r=0.92, respectively), p Conclusion: Our experiments provide sound and efficient methodology to achieve accurate immunohistochemical Ki67 enumeration by IA, enabled by proper validation and calibration of the measurement against RV obtained by stereological frame counts. Citation Format: Arvydas Laurinavicius, Benoit Plancoulaine, Aida Laurinaviciene, Paulette Herlin, Raimundas Meskauskas, Indra Baltrusaityte, Justinas Besusparis, Nicolas Elie, Philippe Belhomme, Yasir Iqbal, Catherine Bor-Angelier. A methodology to ensure and improve accuracy of Ki67 digital immunohistochemistry analysis in breast cancer tissue. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Breast Cancer Research: Genetics, Biology, and Clinical Applications; Oct 3-6, 2013; San Diego, CA. Philadelphia (PA): AACR; Mol Cancer Res 2013;11(10 Suppl):Abstract nr B116.

A cascade-learning approach for automated segmentation of tumour epithelium in colorectal cancer

Abstract Automated segmentation of tumor epithelial tissue from histological images is a fundamental aspiration of digital pathology to improve biomarker assessment and tissue diagnosis. Accurate tumour segmentation is an important step in many automated digital image analysis applications to be used in clinical practice. In particular, segmentation of tumour, non-tumour epithelium and stromal tissue compartments on immunohistochemistry images presents a challenge. Many artifacts, such as staining and/or illumination variations, can confound image analysis. In this paper, we propose a cascade-learning approach which can diminish the impact of these artifacts. It consists of (a) a set of novel invariant features that encodes meaningful information about the appearance and shape of the region of interest and (b) a novel level set formulation where contour evolution is driven by a probabilistic model of the appearance of the region (based on fuzzy c-means). The merit of our approach is that it exploits both appearance and shape information and combines them in the tissue classification framework. We evaluate the performance of our approach on the segmentation of tumour epithelium in colorectal cancer. The experimental results show that our approach is robust to staining differences, additive noise, intensity inhomogeneities, and can cope with a limited number of training samples, when compared to the state-of-the-art tumour epithelial segmentation methods.