Toward human-level concept learning: Pattern benchmarking for AI algorithms

Artificial intelligence (AI) today is very successful at standard pattern-recognition tasks due to the availability of large amounts of data and advances in statistical data-driven machine learning. However, there is still a large gap between AI pattern recognition and human-level concept learning. Humans can learn amazingly well even under uncertainty from just a few examples and are capable of generalizing these concepts to solve new conceptual problems. The growing interest in explainable machine intelligence requires experimental environments and diagnostic/benchmark datasets to analyze existing approaches and drive progress in pattern analysis and machine intelligence. In this paper, we provide an overview of current AI solutions for benchmarking concept learning, reasoning, and generalization; discuss the state-of-the-art of existing diagnostic/benchmark datasets (such as CLEVR, CLEVRER, CLOSURE, CURI, Bongard-LOGO, V-PROM, RAVEN, Kandinsky Patterns, CLEVR-Humans, CLEVRER-Humans, and their extension containing human language); and provide an outlook of some future research directions in this exciting research domain.

ManuKnowVis: How to Support Different User Groups in Contextualizing and Leveraging Knowledge Repositories

We present ManuKnowVis, the result of a design study, in which we contextualize data from multiple knowledge repositories of a manufacturing process for battery modules used in electric vehicles. In data-driven analyses of manufacturing data, we observed a discrepancy between two stakeholder groups involved in serial manufacturing processes: Knowledge providers (e.g., engineers) have domain knowledge about the manufacturing process but have difficulties in implementing data-driven analyses. Knowledge consumers (e.g., data scientists) have no first-hand domain knowledge but are highly skilled in performing data-driven analyses. ManuKnowVis bridges the gap between providers and consumers and enables the creation and completion of manufacturing knowledge. We contribute a multi-stakeholder design study, where we developed ManuKnowVis in three main iterations with consumers and providers from an automotive company. The iterative development led us to a multiple linked view tool, in which, on the one hand, providers can describe and connect individual entities (e.g., stations or produced parts) of the manufacturing process based on their domain knowledge. On the other hand, consumers can leverage this enhanced data to better understand complex domain problems, thus, performing data analyses more efficiently. As such, our approach directly impacts the success of data-driven analyses from manufacturing data. To demonstrate the usefulness of our approach, we carried out a case study with seven domain experts, which demonstrates how providers can externalize their knowledge and consumers can implement data-driven analyses more efficiently.

What is Missing in XAI So Far?

With the perspective on applications of AI-technology, especially data intensive deep learning approaches, the need for methods to control and understand such models has been recognized and gave rise to a new research domain labeled explainable artificial intelligence (XAI). In this overview paper we give an interim appraisal of what has been achieved so far and where there are still gaps in the research. We take an interdisciplinary perspective to identify challenges on XAI research and point to open questions with respect to the quality of the explanations regarding faithfulness and consistency of explanations. On the other hand we see a need regarding the interaction between XAI and user to allow for adaptability to specific information needs and explanatory dialog for informed decision making as well as the possibility to correct models and explanations by interaction. This endeavor requires an integrated interdisciplinary perspective and rigorous approaches to empirical evaluation based on psychological, linguistic and even sociological theories.

Explanatory machine learning for justified trust in human-AI collaboration: Experiments on file deletion recommendations

In the digital age, saving and accumulating large amounts of digital data is a common phenomenon. However, saving does not only consume energy, but may also cause information overload and prevent people from staying focused and working effectively. We present and systematically examine an explanatory AI system (Dare2Del), which supports individuals to delete irrelevant digital objects. To give recommendations for the optimization of related human-computer interactions, we vary different design features (explanations, familiarity, verifiability) within and across three experiments (N ₁ = 61, N ₂ = 33, N ₃= 73). Moreover, building on the concept of distributed cognition, we check possible cross-connections between external (digital) and internal (human) memory. Specifically, we examine whether deleting external files also contributes to human forgetting of the related mental representations. Multilevel modeling results show the importance of presenting explanations for the acceptance of deleting suggestions in all three experiments, but also point to the need of their verifiability to generate trust in the system. However, we did not find clear evidence that deleting computer files contributes to human forgetting of the related memories. Based on our findings, we provide basic recommendations for the design of AI systems that can help to reduce the burden on people and the digital environment, and suggest directions for future research.

Explaining and Evaluating Deep Tissue Classification by Visualizing Activations of Most Relevant Intermediate Layers

Deep Learning-based tissue classification may support pathologists in analyzing digitized whole slide images. However, in such critical tasks, only approaches that can be validated by medical experts in advance to deployment, are suitable. We present an approach that contributes to making automated tissue classification more transparent. We step beyond broadly used visualizations for last layers of a convolutional neural network by identifying most relevant intermediate layers applying Grad-CAM. A visual evaluation by a pathologist shows that these layers assign relevance, where important morphological structures are present in case of correct class decisions. We introduce a tool that can be easily used by medical experts for such validation purposes for any convolutional neural network and any layer. Visual explanations for intermediate layers provide insights into a neural network’s decision for histopathological tissue classification. In future research also the context of the input data must be considered.

IRVINE: A Design Study on Analyzing Correlation Patterns of Electrical Engines

In this design study, we present IRVINE, a Visual Analytics (VA) system, which facilitates the analysis of acoustic data to detect and understand previously unknown errors in the manufacturing of electrical engines. In serial manufacturing processes, signatures from acoustic data provide valuable information on how the relationship between multiple produced engines serves to detect and understand previously unknown errors. To analyze such signatures, IRVINE leverages interactive clustering and data labeling techniques, allowing users to analyze clusters of engines with similar signatures, drill down to groups of engines, and select an engine of interest. Furthermore, IRVINE allows to assign labels to engines and clusters and annotate the cause of an error in the acoustic raw measurement of an engine. Since labels and annotations represent valuable knowledge, they are conserved in a knowledge database to be available for other stakeholders. We contribute a design study, where we developed IRVINE in four main iterations with engineers from a company in the automotive sector. To validate IRVINE, we conducted a field study with six domain experts. Our results suggest a high usability and usefulness of IRVINE as part of the improvement of a real-world manufacturing process. Specifically, with IRVINE domain experts were able to label and annotate produced electrical engines more than 30% faster.

Generating contrastive explanations for inductive logic programming based on a near miss approach

In recent research, human-understandable explanations of machine learning models have received a lot of attention. Often explanations are given in form of model simplifications or visualizations. However, as shown in cognitive science as well as in early AI research, concept understanding can also be improved by the alignment of a given instance for a concept with a similar counterexample. Contrasting a given instance with a structurally similar example which does not belong to the concept highlights what characteristics are necessary for concept membership. Such near misses have been proposed by Winston (Learning structural descriptions from examples, 1970) as efficient guidance for learning in relational domains. We introduce an explanation generation algorithm for relational concepts learned with Inductive Logic Programming (GeNME). The algorithm identifies near miss examples from a given set of instances and ranks these examples by their degree of closeness to a specific positive instance. A modified rule which covers the near miss but not the original instance is given as an explanation. We illustrate GeNME with the well-known family domain consisting of kinship relations, the visual relational Winston arches domain, and a real-world domain dealing with file management. We also present a psychological experiment comparing human preferences of rule-based, example-based, and near miss explanations in the family and the arches domains.

Automatic Detection of Pain from Facial Expressions: A Survey

Pain sensation is essential for survival, since it draws attention to physical threat to the body. Pain assessment is usually done through self-reports. However, self-assessment of pain is not available in the case of noncommunicative patients, and therefore, observer reports should be relied upon. Observer reports of pain could be prone to errors due to subjective biases of observers. Moreover, continuous monitoring by humans is impractical. Therefore, automatic pain detection technology could be deployed to assist human caregivers and complement their service, thereby improving the quality of pain management, especially for noncommunicative patients. Facial expressions are a reliable indicator of pain, and are used in all observer-based pain assessment tools. Following the advancements in automatic facial expression analysis, computer vision researchers have tried to use this technology for developing approaches for automatically detecting pain from facial expressions. This paper surveys the literature published in this field over the past decade, categorizes it, and identifies future research directions. The survey covers the pain datasets used in the reviewed literature, the learning tasks targeted by the approaches, the features extracted from images and image sequences to represent pain-related information, and finally, the machine learning methods used.

Beneficial and harmful explanatory machine learning

Given the recent successes of Deep Learning in AI there has been increased interest in the role and need for explanations in machine learned theories. A distinct notion in this context is that of Michie’s definition of ultra-strong machine learning (USML). USML is demonstrated by a measurable increase in human performance of a task following provision to the human of a symbolic machine learned theory for task performance. A recent paper demonstrates the beneficial effect of a machine learned logic theory for a classification task, yet no existing work to our knowledge has examined the potential harmfulness of machine’s involvement for human comprehension during learning. This paper investigates the explanatory effects of a machine learned theory in the context of simple two person games and proposes a framework for identifying the harmfulness of machine explanations based on the Cognitive Science literature. The approach involves a cognitive window consisting of two quantifiable bounds and it is supported by empirical evidence collected from human trials. Our quantitative and qualitative results indicate that human learning aided by a symbolic machine learned theory which satisfies a cognitive window has achieved significantly higher performance than human self learning. Results also demonstrate that human learning aided by a symbolic machine learned theory that fails to satisfy this window leads to significantly worse performance than unaided human learning.

The Next Generation of Medical Decision Support: A Roadmap Toward Transparent Expert Companions

Increasing quality and performance of artificial intelligence (AI) in general and machine learning (ML) in particular is followed by a wider use of these approaches in everyday life. As part of this development, ML classifiers have also gained more importance for diagnosing diseases within biomedical engineering and medical sciences. However, many of those ubiquitous high-performing ML algorithms reveal a black-box-nature, leading to opaque and incomprehensible systems that complicate human interpretations of single predictions or the whole prediction process. This puts up a serious challenge on human decision makers to develop trust, which is much needed in life-changing decision tasks. This paper is designed to answer the question how expert companion systems for decision support can be designed to be interpretable and therefore transparent and comprehensible for humans. On the other hand, an approach for interactive ML as well as human-in-the-loop-learning is demonstrated in order to integrate human expert knowledge into ML models so that humans and machines act as companions within a critical decision task. We especially address the problem of Semantic Alignment between ML classifiers and its human users as a prerequisite for semantically relevant and useful explanations as well as interactions. Our roadmap paper presents and discusses an interdisciplinary yet integrated Comprehensible Artificial Intelligence (cAI)-transition-framework with regard to the task of medical diagnosis. We explain and integrate relevant concepts and research areas to provide the reader with a hands-on-cookbook for achieving the transition from opaque black-box models to interactive, transparent, comprehensible and trustworthy systems. To make our approach tangible, we present suitable state of the art methods with regard to the medical domain and include a realization concept of our framework. The emphasis is on the concept of Mutual Explanations (ME) that we introduce as a dialog-based, incremental process in order to provide human ML users with trust, but also with stronger participation within the learning process.

Stop Thinking: An Experience Sampling Study on Suppressing Distractive Thoughts at Work

In modern work environments, it can be difficult for workers to avoid becoming distracted from their current task. This study investigates person-situation interactions to predict thought control activities (kind of self-control), which aim to stop distracting thoughts that enter the mind. Specifically, it was examined (1) how challenging work demands (time pressure, task complexity) activate workers' thought control to stop distractive thoughts (n level 2 = 143) and relate to the effort to do so (n level 2 = 91) in daily working life and (2) how these relationships differ according workers' general cognitive ability to suppress unwanted thoughts. To understand these person-situation interactions, an experience sampling study was combined with a laboratory task assessing the ability to suppress unwanted thoughts (think/no-think task). Multilevel modeling revealed that workers' engage more often and more intensively in thought control activities at a moderate level of time pressure but only when they had a higher general ability to suppress unwanted thoughts. For workers with a lower ability to suppress unwanted thoughts, increasing time pressure was negatively related to thought control activities, even at very low levels of time pressure. Thus, whether time pressure activates or hinders thought control depends on individuals' ability to suppress distractive thoughts.

Time to Forget

Abstract. Emerging technologies at work encourage the collection and storage of large amounts of data. However, these vast quantities of data are likely to impair efficient work decisions by employees over time, with negative consequences for the organization. As human attention increasingly represents the scarce resource at work, the present paper focuses on a mechanism of attentional control at work – namely, intentionally forgetting unwanted and outdated internal (e. g., knowledge) and external (e. g., digital objects) information. The purpose of this paper is threefold. Based on a short review of the research on intentional forgetting, a prototypical conceptualization of an interactive assistive system (Dare2Del, cognitive companion) is provided, which should support employees in temporally ignoring or permanently deleting outdated information. Then, we completed a critical incident study to examine why and when employees might want to forget information at work, and to identify in which working situations an assistive system should be particularly helpful.

Problems of video-based pain detection in patients with dementia: a road map to an interdisciplinary solution

BACKGROUND

Given the unreliable self-report in patients with dementia, pain assessment should also rely on the observation of pain behaviors, such as facial expressions. Ideal observers should be well trained and should observe the patient continuously in order to pick up any pain-indicative behavior; which are requisitions beyond realistic possibilities of pain care. Therefore, the need for video-based pain detection systems has been repeatedly voiced. Such systems would allow for constant monitoring of pain behaviors and thereby allow for a timely adjustment of pain management in these fragile patients, who are often undertreated for pain.

METHODS

In this road map paper we describe an interdisciplinary approach to develop such a video-based pain detection system. The development starts with the selection of appropriate video material of people in pain as well as the development of technical methods to capture their faces. Furthermore, single facial motions are automatically extracted according to an international coding system. Computer algorithms are trained to detect the combination and timing of those motions, which are pain-indicative.

RESULTS/CONCLUSION

We hope to encourage colleagues to join forces and to inform end-users about an imminent solution of a pressing pain-care problem. For the near future, implementation of such systems can be foreseen to monitor immobile patients in intensive and postoperative care situations.

Circular test structure for the determination of piezoelectric constants of Sc x Al1-x N thin films applying Laser Doppler Vibrometry and FEM simulations

Piezoelectric scandium aluminium nitride (Sc _x Al_1-x N) thin films offer a large potential for the application in micro electromechanical systems, as advantageous properties of pure AlN thin films are maintained, but combined with an increased piezoelectric actuation and sensing potential. Sc _x Al_1-x N thin films with x = 27% have been prepared by DC reactive magnetron sputtering to find optimized deposition parameters to maximize the piezoelectric constants d₃₃ and d₃₁. For the accurate and simultaneous measurement of these constants Laser Doppler Vibrometry has been applied and compared to finite element (FEM) simulations. The electrode design has been optimized to rotational symmetric structures enabling a 180° phase shifted excitation, so that a straight-forward comparison of experimental displacement curves with those obtained from FEM is feasible.

Acetabular bone reconstruction in revision arthroplasty: a comparison of freeze-dried, irradiated and chemically-treated allograft vitalised with autologous marrow versus frozen non-irradiated allograft

Deficiencies of acetabular bone stock at revision hip replacement were reconstructed with two different types of allograft using impaction bone grafting and a Burch-Schneider reinforcement ring. We compared a standard frozen non-irradiated bone bank allograft (group A) with a freeze-dried irradiated bone allograft, vitalised with autologous marrow (group B). We studied 78 patients (79 hips), of whom 87% (69 hips) had type III acetabular defects according to the American Academy of Orthopaedic Surgeons classification at a mean of 31.4 months (14 to 51) after surgery. At the latest follow-up, the mean Harris hip score was 69.9 points (13.5 to 97.1) in group A and 71.0 points (11.5 to 96.5) in group B. Each hip showed evidence of trabeculation and incorporation of the allograft with no acetabular loosening. These results suggest that the use of an acetabular reinforcement ring and a living composite of sterile allograft and autologous marrow appears to be a method of reconstructing acetabular deficiencies which gives comparable results to current forms of treatment.

[The internet as an information source for family caregivers of dementia patients]

OBJECTIVE

What significance has the Internet as a source of information for family caregivers of dementia patients?

METHODS

In Middle Franconia (Germany), 391 family caregivers were requested to participate in a questionnaire-based postal survey about the significance of the Internet and other sources of information on dementia. The family caregivers in question were the relatives of patients of the Memory Clinic at Erlangen University Psychiatric Hospital, members of the Alzheimer's Society of Middle Franconia or the Nuremberg Family Counselling Society.

RESULTS

Younger and better-educated family caregivers more often own a computer with Internet access than older ones. The Internet is in 4th place on their list of sources of information. Although doctors are by far the most important source, counselling centres and literature are rated only just before the Internet.

CONCLUSIONS

The Internet is particularly significant for the younger better-educated family caregivers, independent of gender, as a source of information on the diagnosis and treatment of dementia.

MAGE-C1/CT-7 expression in plasma cell myeloma: sub-cellular localization impacts on clinical outcome

Plasma cell myelomas (PMs) have a poor prognosis. Cancer-testis (CT) antigens are immunogenic proteins, representing potential targets for tumor vaccination strategies. The expression of the CT antigens GAGE, MAGE-A4, MAGE-C1/CT-7, and NY-ESO-1 was investigated on paraffin-embedded bone marrow biopsies from 219 PM and 8 monoclonal gammopathy of undetermined significance (MGUS) patients. The frequency and prognostic impact of these CT antigens were compared with known morphological prognostic markers (i.e. Mib1 labeling index) and the presence of the translocations t(4;14)(p16.3; q32) and t(11;14)(q13;q32). We show that MAGE-C1/CT-7 is the most prevalent CT antigen, expressed in 57% of PMs in a high percentage of tumor cells. While MAGE-C1/CT-7 was absent in non-malignant plasma cells, plasma cells of patients with MGUS did express MAGE-C1/CT-7, but no other CT antigens. MAGE-C1/CT-7 was more frequently expressed in PMs with an elevated proliferation rate (Mib1 >10%) compared to PMs with a low proliferation rate (Mib1 <or=10%, 71%versus 29%, P < 0.001) and correlated with overall survival, depending on its subcellular distribution. PMs with pure cytoplasmic MAGE-C1/CT-7 expression showed a better prognosis (48 months versus 33 months, P < 0.05) than PMs with combined nuclear-cytoplasmic or nuclear expression only. Thus, expression of MAGE-C1/CT-7 in patients with monoclonal gammopathies represents a predictor of outcome and overt malignant transformation.

Changes in acral blood flux under local application of ropivacaine and lidocaine with and without an adrenaline additive: a double-blind, randomized, placebo-controlled study

Vascular effects of local anesthetics are especially important in dermatological surgery. In particular, adequate perfusion must be ensured in order to offset surgical manipulations during surgical interventions at the acra. However, the use of adrenaline additives appears fraught with problems when anesthesia affects the terminal vascular system, particularly during interventions at the fingers, toes, penis, outer ears, and tip of the nose. We studied skin blood flux at the fingerpads via laser Doppler flowmetry over the course of 24 hours in a prospective, double-blind, randomized, placebo-controlled study with 20 vascularly healthy test persons following Oberst's-method anesthetic blocks. In each case, 6 ml ropivacaine (7.5 mg/ml) (A), lidocaine 1% without an additive (B), and lidocaine 1% with an adrenaline additive (1:200,000) (C) was used respectively as a verum. Isotonic saline solution was injected as a placebo (D). Measurements were carried out with the aid of a computer simultaneously at D II and D IV on both hands. Administration of (A) led to increased blood flux (+155.2%); of (B) initially to a decrease of 27%; of (C) to a reduction of 55% which was reversible after 40 minutes and of (D) to no change.(A) resulted in sustained vasodilatation which was still demonstrable after 24 h. (B) had notably less vasodilative effect, although comparison with (D) clearly showed that (B) is indeed vasodilative. (C) resulted in only a passing decrease in perfusion; this was no longer measurable when checked after 6 and 24 h. This transient inadequacy of blood flux also appeared after administration of (D). These tests show that adrenaline additive in local anesthesia does not decrease blood flow more than 55% for a period of 16 min. Following these results an adrenaline additive can be safely used for anesthetic blocks at the acra in healthy persons.

[A novel therapeutic approach to bone replacement: vitalisation of industrial processed allogenic bone graft with autologous bone marrow]

AIM

Osseous defects of the human skeleton are a common problem in reconstructive orthopaedic and trauma surgery. In spite of a long-standing history of research on different organic and inorganic bone replacement materials, a clinically applicable alternative to autogenous bone transplants has not been found.

METHODS

The current paper describes a novel bone replacement material, which is vitalised through injection of autologous bone marrow. Industrially processed bone allografts were used for bone replacement; the cancellous bone graft was gained from femoral heads from living human donors and processed by a French tissue bank (Banque de Tissus France, TBF). Sterile packed allogenic bone grafts can be stored at room temperature, are easy to handle and can be adapted to almost every bone defect due to their stability and versatility.

RESULTS

From July 2003 to February 2006 125 patients,70 men and 55 women, underwent bone grafting with composite transplants composed of autogenous marrow cells and industrial processed bone allografts. The average age at surgery was 62 (14-84) years; the follow-up period was 11 months. The indication to use the allograft included the whole spectrum of bony defects conventionally treated with autologous cancellous bone grafts such as primary and revision surgery in THA (n = 64) and the treatment of pseudarthrosis particularly of the lower limb (n = 30).

CONCLUSION

In comparison to current bone replacement materials applied in clinical use, this approach promises to replicate the biomechanical and biological properties of autogenous cancellous bone. This is an important advantage over conventional approaches to ensure successful healing and regeneration of bone defects. Hence application of vitalised bone allograft represents an important alternative to conventional autogenous cancellous bone which avoids certain complications associated with harvesting and transplantation of autogenous bone. This applies in particular in situations in which harvesting of autogenous cancellous bone is not or no longer possible.

The Influence of Fluoroquinolone Drugs on the Bacterial Growth of S. epidermidis Utilizing the Unique Potential of Vibrational Spectroscopy

Increasing resistance of many antibiotics has made the design of new drugs necessary. To assist a target-oriented search for new structures and for the elucidation of the mode of action of existing drugs, powerful analytical techniques are required. In this work, vibrational spectroscopy is used to shed more light on the as-yet elusive interaction of gyrase inhibitors of the fluoroquinolone type with their biological target inside the Gram-positive bacterium Staphylococcus epidermidis by investigating whole-cell changes that occur as a result of the presence of the drug moxifloxacin. IR absorption and Raman spectra with excitation off resonance (lambda exc = 532 nm) and in resonance with the biological targets DNA and the aromatic amino acids of gyrase (lambda exc = 244 nm) were recorded for unperturbed bacteria and bacteria in varying drug concentrations (0.08, 0.16, 0.27, and 0.62 microg moxifloxacin/mL bacterial culture). The spectral changes caused by the action of the drug were analyzed with the help of statistical methods, such as hierarchical cluster analysis (HCA), principal component analysis (PCA), and Fisher's linear discriminant analysis (LDA) combined with variable selection. The wavenumbers mostly affected by the action of the drug could be assigned to protein and DNA moieties, supporting the proposed mechanisms of a tertiary complex of the fluoroquinolone, the enzyme gyrase, and DNA.

Transformational and derivational strategies in analogical problem solving

Analogical problem solving is mostly described as transfer of a source solution to a target problem based on the structural correspondences (mapping) between source and target. Derivational analogy (Carbonell, Machine learning: an artificial intelligence approach Los Altos. Morgan Kaufmann, 1986) proposes an alternative view: a target problem is solved by replaying a remembered problem-solving episode. Thus, the experience with the source problem is used to guide the search for the target solution by applying the same solution technique rather than by transferring the complete solution. We report an empirical study using the path finding problems presented in Novick and Hmelo (J Exp Psychol Learn Mem Cogn 20:1296-1321, 1994) as material. We show that both transformational and derivational analogy are problem-solving strategies realized by human problem solvers. Which strategy is evoked in a given problem-solving context depends on the constraints guiding object-to-object mapping between source and target problem. Specifically, if constraints facilitating mapping are available, subjects are more likely to employ a transformational strategy, otherwise they are more likely to use a derivational strategy.

Characterization of bacterial growth and the influence of antibiotics by means of UV resonance Raman spectroscopy

In this work we monitor the bacterial growth of a Bacillus pumilus batch culture by means of UV resonance Raman spectroscopy. Excitation with a wavelength of 244 nm especially enhances the Raman scattering of the aromatic amino acids and the nucleic acid bases and therefore is a good method to track the metabolic changes that occur during bacterial growth. Furthermore, a drug from the fluoroquinolone group is added to the bacterial suspension at the beginning of the exponential growth phase. With the help of chemometrical methods such as hierarchical cluster analysis (HCA) and principal component analysis (PCA) it is possible to visualize the small changes that occur in the UV resonance Raman spectra due to the interaction of the drug with its biological targets DNA and the enzyme gyrase within the bacterial cell.

T-zone lymphoma with cutaneous involvement: a case report and review of the literature

T-zone lymphoma (TZL) is a rare subtype of nodal peripheral T-cell lymphoma characterized by a clonal expansion of T-zone lymphocytes accompanied by a proliferation of other T-zone constituents. Non-specific cutaneous alterations are seen in about one-third of all cases, but specific cutaneous involvement is extremely rare. We present a case of TZL with secondary skin infiltration, review the literature on cutaneous manifestations of TZL and discuss the differential diagnosis of TZL.

Microarrays of bladder cancer tissue are highly representative of proliferation index and histological grade

The number of genes suggested to play a role in cancer biology is rapidly increasing. To be able to test a large number of molecular parameters in sufficiently large series of primary tumours, a tissue microarray (TMA) approach has been developed where samples from up to 1000 tumours can be simultaneously analysed on one glass slide. Because of the small size of the individual arrayed tissue samples (diameter 0.6 mm), the question arises of whether these specimens are representative of their donor tumours. To investigate how representative are the results obtained on TMAs, a set of 2317 bladder tumours that had been previously analysed for histological grade and Ki67 labelling index (LI) was used to construct four replica TMAs from different areas of each tumour. Clinical follow-up information was available from 1092 patients. The histological grade and the Ki67 LI were determined for every arrayed tumour sample (4x2317 analyses each). Despite discrepancies in individual cases, the grade and Ki67 information obtained on minute arrayed samples were highly similar to the data obtained on large sections (p<0.0001). Most importantly, every individual association between grade or Ki67 LI and tumour stage or prognosis (recurrence, progression, tumour-specific survival) that was observed in large section analysis could be fully reproduced on all four replica TMAs. These results show that intra-tumour heterogeneity does not significantly affect the ability to detect clinico-pathological correlations on TMAs, probably because of the large number of tumours that can be included in TMA studies. TMAs are a powerful tool for rapid identification of the biological or clinical significance of molecular alterations in bladder cancer and other tumour types.

Microarrays of bladder cancer tissue are highly representative of proliferation index and histological grade

The number of genes suggested to play a role in cancer biology is rapidly increasing. To be able to test a large number of molecular parameters in sufficiently large series of primary tumours, a tissue microarray (TMA) approach has been developed where samples from up to 1000 tumours can be simultaneously analysed on one glass slide. Because of the small size of the individual arrayed tissue samples (diameter 0.6 mm), the question arises of whether these specimens are representative of their donor tumours. To investigate how representative are the results obtained on TMAs, a set of 2317 bladder tumours that had been previously analysed for histological grade and Ki67 labelling index (LI) was used to construct four replica TMAs from different areas of each tumour. Clinical follow-up information was available from 1092 patients. The histological grade and the Ki67 LI were determined for every arrayed tumour sample (4x2317 analyses each). Despite discrepancies in individual cases, the grade and Ki67 information obtained on minute arrayed samples were highly similar to the data obtained on large sections (p<0.0001). Most importantly, every individual association between grade or Ki67 LI and tumour stage or prognosis (recurrence, progression, tumour-specific survival) that was observed in large section analysis could be fully reproduced on all four replica TMAs. These results show that intra-tumour heterogeneity does not significantly affect the ability to detect clinico-pathological correlations on TMAs, probably because of the large number of tumours that can be included in TMA studies. TMAs are a powerful tool for rapid identification of the biological or clinical significance of molecular alterations in bladder cancer and other tumour types.

High-throughput tissue microarray analysis of 3p25 (RAF1) and 8p12 (FGFR1) copy number alterations in urinary bladder cancer

Studies by comparative genomic hybridization revealed that the chromosomal regions 3p25 and 8p11-p12 are recurrently amplified in bladder cancer. To investigate the prevalence of DNA copy number alterations in these chromosomal regions and study their clinical significance, we used probes for the RAF1 (3p25) and FGFR1 (8p12) genes for fluorescence in situ hybridization. A tissue microarray containing 2317 tumors was analyzed. The analysis revealed RAF1 amplification in 4.0% and FGFR1 amplification in 3.4% of interpretable tumors. In addition, deletions were found at the 3p25 locus in 2.2% and at the 8p11-12 locus in 9.9% of interpretable tumors. Both amplifications and deletions of RAF1 and FGFR1 were significantly associated with high tumor grade (P < 0.0001), advanced stage (P < 0.0001), and poor survival (P < 0.05) if tumors of all of the stages where analyzed together. RAF1 amplifications were associated with subsequent tumor progression in pT1 carcinomas (P < 0.05). The marked differences in the frequency of all of the analyzed changes between pTa grade 1/grade 2 and pT1-4 carcinomas support the concept of these tumor groups representing different tumor entities.