Mitotic Figure Recognition: Agreement among Pathologists and Computerized Detector

Despite the prognostic importance of mitotic count as one of the components of the Bloom – Richardson grade [3], several studies ([2, 9, 10]) have found that pathologists’ agreement on the mitotic grade is fairly modest. Collecting a set of more than 4,200 candidate mitotic figures, we evaluate pathologists' agreement on individual figures, and train a computerized system for mitosis detection, comparing its performance to the classifications of three pathologists. The system’s and the pathologists’ classifications are based on evaluation of digital micrographs of hematoxylin and eosin stained breast tissue. On figures where the majority of pathologists agree on a classification, we compare the performance of the trained system to that of the individual pathologists. We find that the level of agreement of the pathologists ranges from slight to moderate, with strong biases, and that the system performs competitively in rating the ground truth set. This study is a step towards automatic mitosis count to accelerate a pathologist's work and improve reproducibility.

Download Full-text

Deep learning algorithms out-perform veterinary pathologists in detecting the mitotically most active tumor region

Scientific Reports ◽

10.1038/s41598-020-73246-2 ◽

2020 ◽

Vol 10 (1) ◽

Cited By ~ 1

Author(s):

Marc Aubreville ◽

Christof A. Bertram ◽

Christian Marzahl ◽

Corinne Gurtner ◽

Martina Dettwiler ◽

...

Keyword(s):

Deep Learning ◽

Ground Truth ◽

Mitotic Count ◽

Mitotic Figure ◽

Two Stage ◽

Data Set ◽

Area Selection ◽

Tumor Region ◽

Cutaneous Mast Cell ◽

Mitotic Figures

Abstract Manual count of mitotic figures, which is determined in the tumor region with the highest mitotic activity, is a key parameter of most tumor grading schemes. It can be, however, strongly dependent on the area selection due to uneven mitotic figure distribution in the tumor section. We aimed to assess the question, how significantly the area selection could impact the mitotic count, which has a known high inter-rater disagreement. On a data set of 32 whole slide images of H&E-stained canine cutaneous mast cell tumor, fully annotated for mitotic figures, we asked eight veterinary pathologists (five board-certified, three in training) to select a field of interest for the mitotic count. To assess the potential difference on the mitotic count, we compared the mitotic count of the selected regions to the overall distribution on the slide. Additionally, we evaluated three deep learning-based methods for the assessment of highest mitotic density: In one approach, the model would directly try to predict the mitotic count for the presented image patches as a regression task. The second method aims at deriving a segmentation mask for mitotic figures, which is then used to obtain a mitotic density. Finally, we evaluated a two-stage object-detection pipeline based on state-of-the-art architectures to identify individual mitotic figures. We found that the predictions by all models were, on average, better than those of the experts. The two-stage object detector performed best and outperformed most of the human pathologists on the majority of tumor cases. The correlation between the predicted and the ground truth mitotic count was also best for this approach (0.963–0.979). Further, we found considerable differences in position selection between pathologists, which could partially explain the high variance that has been reported for the manual mitotic count. To achieve better inter-rater agreement, we propose to use a computer-based area selection for support of the pathologist in the manual mitotic count.

Download Full-text

A detailed analysis of anatomical plausibility of crossed and uncrossed streamline rendition of the dentato-rubro-thalamic tract (DRT(T)) in a commercial stereotactic planning system

Acta Neurochirurgica ◽

10.1007/s00701-021-04890-4 ◽

2021 ◽

Author(s):

Volker A. Coenen ◽

Bastian E. Sajonz ◽

Peter C. Reinacher ◽

Christoph P. Kaller ◽

Horst Urbach ◽

...

Keyword(s):

Surgical Planning ◽

Ground Truth ◽

Planning System ◽

Single Subject ◽

Fiber Structure ◽

Native Space ◽

Volume Of Interest ◽

Mri Scans ◽

Magnetic Resonance Imaging Mri ◽

The Individual

Abstract Background An increasing number of neurosurgeons use display of the dentato-rubro-thalamic tract (DRT) based on diffusion weighted imaging (dMRI) as basis for their routine planning of stimulation or lesioning approaches in stereotactic tremor surgery. An evaluation of the anatomical validity of the display of the DRT with respect to modern stereotactic planning systems and across different tracking environments has not been performed. Methods Distinct dMRI and anatomical magnetic resonance imaging (MRI) data of high and low quality from 9 subjects were used. Six subjects had repeated MRI scans and therefore entered the analysis twice. Standardized DICOM structure templates for volume of interest definition were applied in native space for all investigations. For tracking BrainLab Elements (BrainLab, Munich, Germany), two tensor deterministic tracking (FT2), MRtrix IFOD2 (https://www.mrtrix.org), and a global tracking (GT) approach were used to compare the display of the uncrossed (DRTu) and crossed (DRTx) fiber structure after transformation into MNI space. The resulting streamlines were investigated for congruence, reproducibility, anatomical validity, and penetration of anatomical way point structures. Results In general, the DRTu can be depicted with good quality (as judged by waypoints). FT2 (surgical) and GT (neuroscientific) show high congruence. While GT shows partly reproducible results for DRTx, the crossed pathway cannot be reliably reconstructed with the other (iFOD2 and FT2) algorithms. Conclusion Since a direct anatomical comparison is difficult in the individual subjects, we chose a comparison with two research tracking environments as the best possible “ground truth.” FT2 is useful especially because of its manual editing possibilities of cutting erroneous fibers on the single subject level. An uncertainty of 2 mm as mean displacement of DRTu is expectable and should be respected when using this approach for surgical planning. Tractographic renditions of the DRTx on the single subject level seem to be still illusive.

Download Full-text

Fully Data-Driven Pseudohealthy Synthesis for Planning Valve-Sparing Aortic Root Reconstruction using Conditional Variational Autoencoders

Current Directions in Biomedical Engineering ◽

10.1515/cdbme-2020-3072 ◽

2020 ◽

Vol 6 (3) ◽

pp. 284-287

Author(s):

Jannis Hagenah ◽

Mohamad Mehdi ◽

Floris Ernst

Keyword(s):

Aortic Root ◽

Similarity Index ◽

Ground Truth ◽

Representation Learning ◽

Patient Specific ◽

Ultrasound Images ◽

Specific Geometry ◽

The Individual ◽

Native Root ◽

Original Information

AbstractAortic root aneurysm is treated by replacing the dilated root by a grafted prosthesis which mimics the native root morphology of the individual patient. The challenge in predicting the optimal prosthesis size rises from the highly patient-specific geometry as well as the absence of the original information on the healthy root. Therefore, the estimation is only possible based on the available pathological data. In this paper, we show that representation learning with Conditional Variational Autoencoders is capable of turning the distorted geometry of the aortic root into smoother shapes while the information on the individual anatomy is preserved. We evaluated this method using ultrasound images of the porcine aortic root alongside their labels. The observed results show highly realistic resemblance in shape and size to the ground truth images. Furthermore, the similarity index has noticeably improved compared to the pathological images. This provides a promising technique in planning individual aortic root replacement.

Download Full-text

WattScale

ACM/IMS Transactions on Data Science ◽

10.1145/3406961 ◽

2021 ◽

Vol 2 (1) ◽

pp. 1-25

Author(s):

Srinivasan Iyengar ◽

Stephen Lee ◽

David Irwin ◽

Prashant Shenoy ◽

Benjamin Weil

Keyword(s):

Cooling System ◽

Large Population ◽

Ground Truth ◽

Detection Algorithm ◽

Building Envelope ◽

Ground Truth Data ◽

Heating And Cooling ◽

Data Driven Approach ◽

Execution Mode ◽

The Individual

Buildings consume over 40% of the total energy in modern societies, and improving their energy efficiency can significantly reduce our energy footprint. In this article, we present WattScale, a data-driven approach to identify the least energy-efficient buildings from a large population of buildings in a city or a region. Unlike previous methods such as least-squares that use point estimates, WattScale uses Bayesian inference to capture the stochasticity in the daily energy usage by estimating the distribution of parameters that affect a building. Further, it compares them with similar homes in a given population. WattScale also incorporates a fault detection algorithm to identify the underlying causes of energy inefficiency. We validate our approach using ground truth data from different geographical locations, which showcases its applicability in various settings. WattScale has two execution modes—(i) individual and (ii) region-based, which we highlight using two case studies. For the individual execution mode, we present results from a city containing >10,000 buildings and show that more than half of the buildings are inefficient in one way or another indicating a significant potential from energy improvement measures. Additionally, we provide probable cause of inefficiency and find that 41%, 23.73%, and 0.51% homes have poor building envelope, heating, and cooling system faults, respectively. For the region-based execution mode, we show that WattScale can be extended to millions of homes in the U.S. due to the recent availability of representative energy datasets.

Download Full-text

Autonomous feature type selection based on environment using expectation maximization in self-localization

International Journal of Advanced Robotic Systems ◽

10.1177/1729881418814701 ◽

2018 ◽

Vol 15 (6) ◽

pp. 172988141881470

Author(s):

Nezih Ergin Özkucur ◽

H Levent Akın

Keyword(s):

Expectation Maximization ◽

Autonomous Robots ◽

Expectation Maximization Algorithm ◽

Sensory Information ◽

Ground Truth ◽

Local Environment ◽

Localization Algorithm ◽

Data Set ◽

Public Data ◽

The Individual

Self-localization in autonomous robots is one of the fundamental issues in the development of intelligent robots, and processing of raw sensory information into useful features is an integral part of this problem. In a typical scenario, there are several choices for the feature extraction algorithm, and each has its weaknesses and strengths depending on the characteristics of the environment. In this work, we introduce a localization algorithm that is capable of capturing the quality of a feature type based on the local environment and makes soft selection of feature types throughout different regions. A batch expectation–maximization algorithm is developed for both discrete and Monte Carlo localization models, exploiting the probabilistic pose estimations of the robot without requiring ground truth poses and also considering different observation types as blackbox algorithms. We tested our method in simulations, data collected from an indoor environment with a custom robot platform and a public data set. The results are compared with the individual feature types as well as naive fusion strategy.

Download Full-text

Multiple Noisy Label Distribution Propagation for Crowdsourcing

Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2019/204 ◽

2019 ◽

Cited By ~ 1

Author(s):

Hao Zhang ◽

Liangxiao Jiang ◽

Wenqiang Xu

Keyword(s):

Supervised Learning ◽

Real World ◽

Effective Means ◽

Ground Truth ◽

Cost Effective ◽

Nearest Neighbors ◽

True Label ◽

Real World Datasets ◽

The Individual ◽

Label Distribution

Crowdsourcing services provide a fast, efficient, and cost-effective means of obtaining large labeled data for supervised learning. Ground truth inference, also called label integration, designs proper aggregation strategies to infer the unknown true label of each instance from the multiple noisy label set provided by ordinary crowd workers. However, to the best of our knowledge, nearly all existing label integration methods focus solely on the multiple noisy label set itself of the individual instance while totally ignoring the intercorrelation among multiple noisy label sets of different instances. To solve this problem, a multiple noisy label distribution propagation (MNLDP) method is proposed in this study. MNLDP first transforms the multiple noisy label set of each instance into its multiple noisy label distribution and then propagates its multiple noisy label distribution to its nearest neighbors. Consequently, each instance absorbs a fraction of the multiple noisy label distributions from its nearest neighbors and yet simultaneously maintains a fraction of its own original multiple noisy label distribution. Promising experimental results on simulated and real-world datasets validate the effectiveness of our proposed method.

Download Full-text

Image Processing for Localization and Parameterization of the Glandular Ducts of Colon in Inflammatory Bowel Diseases

Image Processing ◽

10.4018/978-1-4666-3994-2.ch036 ◽

2013 ◽

pp. 688-708

Author(s):

Stanislaw Osowski ◽

Michal Kruk ◽

Robert Koktysz ◽

Jaroslaw Kurek

Keyword(s):

Inflammatory Bowel Diseases ◽

Svm Classifier ◽

Diagnostic Features ◽

Human Expert ◽

Stage Of Development ◽

Colon Biopsy ◽

Bowel Diseases ◽

Inflammatory Bowel ◽

The Individual ◽

Computerized System

This chapter presents the computerized system for automatic analysis of the medical image of the colon biopsy, able to extract the important diagnostic knowledge useful for supporting the medical diagnosis of the inflammatory bowel diseases. Application of the artificial intelligence methods included in the developed automatic system allowed the authors to obtain the unique numerical results, impossible for achieving at the visual inspection of the image by the human expert. The developed system enabled the authors to perform all steps in an automatic way, including the segmentation of the image, leading to the extraction of all glandular ducts, parameterization of the individual ducts and creation of the diagnostic features, as well as characterizing the recognition problem. These features put to the input of SVM classifier enable to associate them with the stage of development of the inflammation. The numerical experiments have shown that the system is able to process successfully the images at different stages of development of the inflammation. Its important advantage is automation of this very difficult work, not possible to be done manually, even by a human expert.

Download Full-text

Ego-zones: non-symmetric dependencies reveal network groups with large and dense overlaps

Applied Network Science ◽

10.1007/s41109-019-0192-6 ◽

2019 ◽

Vol 4 (1) ◽

Cited By ~ 1

Author(s):

Milos Kudelka ◽

Eliska Ochodkova ◽

Sarka Zehnalova ◽

Jakub Plesnik

Keyword(s):

Network Structure ◽

Real World ◽

Ground Truth ◽

Structural Similarity ◽

Global Network ◽

Overlapping Communities ◽

Detection Algorithms ◽

First Case ◽

New Perspective ◽

The Individual

Abstract The existence of groups of nodes with common characteristics and the relationships between these groups are important factors influencing the structures of social, technological, biological, and other networks. Uncovering such groups and the relationships between them is, therefore, necessary for understanding these structures. Groups can either be found by detection algorithms based solely on structural analysis or identified on the basis of more in-depth knowledge of the processes taking place in networks. In the first case, these are mainly algorithms detecting non-overlapping communities or communities with small overlaps. The latter case is about identifying ground-truth communities, also on the basis of characteristics other than only network structure. Recent research into ground-truth communities shows that in real-world networks, there are nested communities or communities with large and dense overlaps which we are not yet able to detect satisfactorily only on the basis of structural network properties.In our approach, we present a new perspective on the problem of group detection using only the structural properties of networks. Its main contribution is pointing out the existence of large and dense overlaps of detected groups. We use the non-symmetric structural similarity between pairs of nodes, which we refer to as dependency, to detect groups that we call zones. Unlike other approaches, we are able, thanks to non-symmetry, accurately to describe the prominent nodes in the zones which are responsible for large zone overlaps and the reasons why overlaps occur. The individual zones that are detected provide new information associated in particular with the non-symmetric relationships within the group and the roles that individual nodes play in the zone. From the perspective of global network structure, because of the non-symmetric node-to-node relationships, we explore new properties of real-world networks that describe the differences between various types of networks.

Download Full-text

Testing G-CSF responsiveness predicts the individual susceptibility to infection and consecutive treatment in recipients of high-dose chemotherapy

Blood ◽

10.1182/blood-2010-06-290080 ◽

2011 ◽

Vol 117 (7) ◽

pp. 2121-2128 ◽

Cited By ~ 13

Author(s):

Christian Straka ◽

Michael Sandherr ◽

Hans Salwender ◽

Hannes Wandt ◽

Bernd Metzner ◽

...

Keyword(s):

Unknown Origin ◽

High Dose Chemotherapy ◽

Individual Risk ◽

High Dose ◽

Open Label ◽

Open Label Study ◽

Susceptibility To Infection ◽

Intravenous Antibiotics ◽

Grade 3 ◽

The Individual

Abstract The individual risk of infection and requirements for medical treatment after high-dose chemotherapy have been unpredictable. In this prospective, multicenter, open-label study we investigated the potential of granulocyte colony-stimulating factor (G-CSF) responsiveness as a predictor. A total of 168 patients with multiple myeloma or lymphoma received a single dose of subcutaneous G-CSF (lenograstim, 263 μg) after high-dose chemotherapy. Highly variable leukocyte peaks were measured and grouped as low (quartile 1; leukocytes 100-10 100/μL), medium (quartile 2; leukocytes > 10 100-18 300/μL), and high (quartiles 3/4; leukocytes > 18 300-44 800/μL). G-CSF responsiveness (low vs medium vs high) was inversely correlated with febrile neutropenia (77% vs 60% vs 48%; P = .0037); the rate of infection, including fever of unknown origin (91% vs 67% vs 54%; P < .0001); days with intravenous antibiotics (9 vs 6 vs 5; P < .0001); and antifungal therapy (P = .042). In multivariate analysis, G-CSF responsiveness remained the only factor significantly associated with infection (P = .016). In addition, G-CSF responsiveness was inversely correlated with grade 3/4 oral mucositis (67% vs 33% vs 23%; P < .0001). G-CSF responsiveness appears as a signature of the myeloid marrow reserve predicting defense against neutropenic infection after intensive chemotherapy. This study is registered at http://www.clinicaltrials.gov as NCT01085058.

Download Full-text

Low-dose dexamethasone and thalidomide with higher frequency zoledronic acid (dtZ) for newly diagnosed multiple myeloma

Journal of Clinical Oncology ◽

10.1200/jco.2007.25.18_suppl.18506 ◽

2007 ◽

Vol 25 (18_suppl) ◽

pp. 18506-18506

Author(s):

G. Teoh ◽

D. Tan ◽

C. Chuah ◽

W. Hwang ◽

R. Yiu ◽

...

Keyword(s):

Multiple Myeloma ◽

Zoledronic Acid ◽

Complete Remission ◽

Median Time ◽

Low Dose ◽

Dental Treatment ◽

Osteonecrosis Of The Jaw ◽

Newly Diagnosed ◽

Grade 3 ◽

The Individual

18506 Background: Although dexamethasone (Dex), thalidomide (Thal) and zoledronic acid (Zol) have frequently been combined for the treatment of multiple myeloma (MM), the ideal dosing schedule is unknown. We previously reported that lower doses of Dex and Thal can be effectively combined with high-frequency dosing of Zol (Haematologica 2005). Methods: This “dtZ” regimen - which comprises weekly Dex 20 mg OM for 4 days, Thal 50 mg ON, and 3-weekly Zol 4 mg - resulted in an impressive response rate (RR) of 61.6% and near complete remission (nCR)/complete remission (CR) rate of 7.7% in 26 patients with relapsed/refractory MM. Results: In this present study, we treated 22 newly diagnosed MM patients with “dtZ” and report an even more impressive RR of 100.0% and nCR/CR rate of 20–35%. The median time to response was 1.8 months and median time to maximum response was 2.2 months. The median time to progression (TTP) has not been achieved yet. As expected, low-dose Dex/Thal resulted in lower (18.1%) grade 3 or 4 toxicities. These were all infections; which lead to further dose-reduction of Dex. There were no thromboembolic events, despite the fact that aspirin was not routinely given. Of particular interest, 3- weekly Zol was not associated with any significant decrease in renal function, and none of our patients developed osteonecrosis of the jaw (ONJ). In fact, at the time of writing of this abstract, more than 1,000 doses of Zol had been administered in a 3-weekly fashion to these as well as other patients, and only 1 patient developed ONJ. This patient who had already received greater than 20 doses of Zol healed uneventfully after receiving appropriate outpatient dental treatment, and subsequently received another 8 doses of Zol with no recurrence of ONJ. Conclusion: In conclusion, the Zol-based “dtZ” regimen is potentially a highly-effective and safe frontline regimen for MM. Using Zol every 3 weeks with lower doses of Dex and Thal does not appear to increase the rate or severity of nephrotoxicity or ONJ. Although we do not know exactly why every patient responded to “dtZ”, we speculate that this could be due to a critical balance that has been achieved between the anti-MM, anti-osteoclastic and immunostimulatory effects of the individual drugs of the combination. No significant financial relationships to disclose.

Download Full-text