Biomedical subjects
Publications and source records attributed to J E Weber.
To enable an image understanding system to provide an automated interpretation of diagnostic imagery it must have access to all of the concepts, procedures and methods used by human experts. The paper describes information elicitation from experts of different domains and the construction of a knowledge file. Uncertainty management is based on Bayesian belief network methods.
Bayesian belief networks have a dynamic range and numeric response characteristics that make them uniquely suitable for descriptive classification schemes. Features showing considerable overlap of tolerance regions may be used, in a cumulative manner, to derive unequivocal classification decisions. The numeric response characteristics of Bayesian belief networks are analyzed, and their application as control modules in automated scene segmentation in histopathology is demonstrated.
Uncertainty management for the evaluation of evidence based on linguistic and conceptual data is taking advantage of developments in the Dempster-Shafer (DS) theory of evidence, possibility theory and fuzzy logic. The DS theory offers the capability to assess the uncertainty of different subsets of assertions in a domain and the way in which uncertainty is affected by accumulating evidence. The DS theory goes beyond probability theory in its ability to represent ignorance about certain aspects of a situation. However, the theory is very sensitive to the numerical assessments provided by users and can lead to intuitively unexpected and even undesirable results. Certainty factors are widely used in various expert systems. Their definition and updating may follow either a probabilistic model or fuzzy set theoretic concept.
Expert systems deal with data that are categorical and conceptual, that represent elements of fuzzy sets and that often, by themselves, do not allow an unequivocal decision. The management of uncertainty in expert systems thus becomes a crucial issue. It involves defining measures of uncertainty and procedures for combining accumulating evidence in a manner that properly considers the dependence structure of diagnostic clues. Probability theory offers valuable procedures for uncertainty assessment; however, their practical application in the domain of quantitative histopathology and histopathologic diagnosis can be problematic.
Ten papillary adenocarcinomas of thyroid origin (P-Thy), ten papillary adenocarcinomas of ovarian origin (P-Ov), and eight papillary neoplasms of non-thyroid/non-ovarian origin (P-Other) were morphometrically compared using 19 distinct quantitative nuclear and nucleolar parameters as a database for diagnosis. The selected cases consisted of 16 primary and 12 metastatic neoplasms. It was determined that the P-Thy group had a significantly smaller nucleolar area (NuA) and nucleolar perimeter (NuP), and smaller SDs of nuclear area (NA), NuA, and NuP compared with the P-Ov and P-Other groups (P less than .05). The P-Ov group had a significantly smaller SD of NA compared with the P-Other group (P less than .05). The P-Ov group exhibited the greatest variability among the papillary neoplasms. Linear regression analysis indicated that in the P-Thy group alone there was a significant correlation between mean nuclear form factor (4 pi A/P2) and mean NuA (r = -.82; P less than .01), and mean NP and mean NuA (r = +.77; P less than .01). Linear regression analysis also indicated that in the P-Ov group alone, there was a significant correlation between mean NA and mean NuA (r = +.75; P less than .02). Morphometric domains were established using statistically significant sets of variables that distinguished between the groups. The application of three-dimensional computerized cluster analysis techniques indicated that the P-Thy group consistently had the smallest morphometric domains. It was concluded that ultrastructural morphometric analysis of papillary neoplasms has diagnostic potential and reveals interesting biological relationships among distinct nuclear features in the different groups of neoplasms.
An introduction to, and overview of, expert systems is presented, along with some preliminary comments on their application in diagnostic and analytical histopathology and cytopathology. The terminology common to expert systems is defined, and the nature of expert systems is discussed. In particular, the differences between expert systems and other types of computer programs (e.g., algorithms) or means of solving problems are explored. The rationale for their use and the types of tasks for which they are appropriate are also discussed.
In a previous study, an expert system using visually assessed diagnostic clues for diagnosing colonic tissue as normal, adenoma or adenocarcinoma arrived at diagnoses agreeing with the evaluation by pathologists ("correct diagnoses") for all 49 cases of normal colon, for 49 of 50 cases of adenoma and for 48 of 49 cases of adenocarcinoma. The present study examined the robustness and sensitivity of the expert system to changes in the knowledge base, to changes in criteria specified by the user and to missing information. Alternative rules for combining certainty factors are discussed.
Ectoplasmic specializations (ES) containing packed actin microfilaments are associated with the numerous parallel rows of occluding junctions which form the Sertoli cell (blood-testis) barrier. To determine if ES regulate the structure of the occluding junctions and/or barrier permeability, we experimentally disrupted ES microfilaments in vivo with intratesticularly injected cytochalasin D (CD). Electron microscopic observations of seminiferous tubules from CD-treated (150-500 microM CD; 0.5-12 hr) animals indicated that ES was absent from regions where the Sertoli cell barrier is located. Seminiferous epithelial sheets from uninjected or vehicle-injected animals (1 DMSO: 1 saline) stained with NBD-phallacidin demonstrated the presence of patterned ES actin surrounding the basolateral regions of adjacent Sertoli cells. After exposure to CD, epithelial sheets exhibited increasingly patchy fluorescence indicating progressive F-actin disruption. Freeze-fracture replicas of CD-injected testes revealed numerous focal alterations in the region of occluding junctions which included disorganization of the parallel arrangement of junctional rows, the presence of free-ending rows, clustering of intramembranous particles (IMPs) between rows, reduction in the number of rows, and loss of IMPs on both the P-face and E-face. Tracer experiments, following CD exposure, were conducted to test the integrity of occluding junctions: lanthanum hydroxide, dextrose, or filipin was added, in separate experiments, to the fixative during perfusion-fixation. In another study, serum containing an antibody against adluminal germ cells was injected intratesticularly, and frozen sections were processed for immunofluorescence study. A final study consisted of simultaneous intratesticular infusions of CD and radiolabelled inulin with subsequent intraluminal and peritubular fluid sampling. In animals which were injected with CD, lanthanum was found to enter the adluminal compartment; fixative made hypertonic by addition of dextrose caused germ cells within the adluminal compartment to shrink and produce exaggerated intercellular spaces; filipin-cholesterol perturbations were present between some Sertoli cell junctional rows and on spermatid plasma membranes; and IgG was detected within the adluminal compartment of many seminiferous tubules. None of these adluminal manifestations was noted in control animals or those which received vehicle. Quantitatively, in the in vivo micropuncture experiments, significantly more radiolabelled inulin entered the lumen of seminiferous tubules from CD-treated animals than from those exposed to vehicle.(ABSTRACT TRUNCATED AT 400 WORDS)
The role of expert systems functioning as process controllers in learning image understanding systems is discussed. Numeric learning systems already have found a number of applications in cytologic and histopathologic diagnosis. Depending on the required capabilities, systems of increasing complexity are needed. Expert systems to guide scene segmentation in histopathologic imagery require model-based reasoning. Diagnostic image interpretation with learning capability demands a full model of the human expert's competence, including a considerable variety of knowledge representation schemes and inference strategies, coordinated by a meta-process controller.
Three multivariate methods for structure simplification (cluster analysis, principal component analysis and factor analysis) and three multivariate methods for prediction (discriminant analysis, analysis of variance and regression analysis) are discussed and contrasted, along with basic descriptive statistics for multivariate data. These methods are only a subset of the multivariate methods currently in use, but include those that have thus far been applied in diagnostic cytology. Multivariate analyses provide a basis for describing and analyzing large complex data sets and, with the enormous amounts of data becoming available, their applications in many areas of diagnostic cytology must surely increase.
Corresponding to the rapid increase in the amount of data available for use in clinical diagnoses, there is an increased need for procedures that can provide the diagnostician with meaningful statistical summaries of data and with statements concerning the statistical significance associated with a diagnostic evaluation. It has been demonstrated that multivariate statistical assessment of clinical material can provide consistent, reliable and highly sensitive diagnostic clues, even in instances in which trained personnel are unable to see any change. Several examples of applications of statistical analyses in diagnostic cytology and histopathology are given in this paper. The examples were chosen to be illustrative of the different types of problems for which statistical analyses have been found useful. These problems differ with respect to the extent of the statistical methods thus far developed and the difficulty involved in developing further analyses. For many problems, appropriate statistical analyses are readily available; other problems require definition of custom-made test statistics and, in some cases, also definition of new statistical distributions. The problems discussed here are only a small sample of the existing problems, but they provide at least an indication of the scope of the role that statistics plays in cytopathologic and histopathologic diagnosis.
A computer-based expert system for diagnosing colonic sections as normal, adenoma or adenocarcinoma is described, along with an evaluation of its performance. On the basis of its knowledge base, consisting of the values of diagnostic clues and their associated certainty factors for the possible diagnoses, the system will suggest the diagnosis for new cases presented to it. Using the data provided for 16 diagnostic clues, the system arrived at correct diagnoses for all cases of normal colon, for 49 of 50 cases of adenoma and for 48 of 49 cases of adenocarcinoma. Sample outputs from the expert system are presented and discussed, and the effects of possible alterations in the data base are considered.
Rat Leydig cells were permeabilized and the cytoplasm partially extracted to visualize, describe, and characterize filamentous elements of the cytoskeleton. It was demonstrated by immunofluorescence microscopy that vimentin is abundant within Leydig cells. Ultrastructurally, intermediate filaments in Leydig cells were concentrated at perinuclear sites and comprised bundles that coursed through the cytoplasm. Actin was identified in Leydig cells with the F actin probe, NBD-phallacidin. Fluorescence was strongest at the cortex of the cell. With myosin S-1 subfragments, sparse actin was found positioned almost exclusively in cortical regions of the cell associated with coated pits and in Leydig cell processes.
A morphological evaluation of intercellular bridges was undertaken during rat spermatogenesis. The dimensions and relationships of the bridges were shown to vary during different phases of spermatogenesis. Cellular divisions of spermatogonia and spermatocytes resulted in the partitioning of pre-existing bridges by complex structures termed bridge partitioning complexes, which are described in detail, as is the process whereby new bridges are formed. The structure of premeiotic bridges was generally consistent; however, during spermiogenesis, the structure of bridges and bridge contents were modified at specific phases of their development. The plasma membrane density associated with the cytoplasmic aspect of early step 1 spermatids separated into multiple dense bands that encircled the peripheral aspect of late step 1 spermatid bridges. By step 2 of spermiogenesis, these dense bands became associated with several cisternae of endoplasmic reticulum, which later coalesced into a single saccule that completely encircled the bridge structure by step 4. At steps 10-13 of spermiogenesis, the single saccule of endoplasmic reticulum vesiculated into many smaller cisternae. Also, filament-bounded densities (measuring 10-12 nm in diameter) appeared within the bridge channel. At step 17 of spermiogenesis, the filament-bounded densities were no longer apparent, but an anastomosing network of endoplasmic reticulum, often in the configuration of a sphere, occupied the entire central region of the bridge. In step 19 spermatids, the smooth endoplasmic reticulum within the bridge channel and the multiple cisternae lining the bridge density were gradually displaced. The subsurface density of bridges gradually lost its prominence. Some cytoplasmic lobes were connected by extremely narrow (approximately 22 nm) cytoplasmic channels. Similar-appearing channels were seen on the surface zone of cytoplasmic lobes or residual bodies, this observation suggesting that channels were sites of severence of bridges. Just prior to the separation or disengagement of the spermatid from the cytoplasmic lobe, selected bridges appeared to open to form large masses. After spermiation, residual bodies were not found joined by bridges; but from the size of some of the residual bodies, it was suspected that they were formed by coalescence of more than one cytoplasmic lobe. Freeze-fracture demonstrated few intramembranous particles on either the P or E face of the plasma membrane forming the bridge; this finding suggested bridge structures restricted free lateral movement of membrane constituents across the bridge.(ABSTRACT TRUNCATED AT 400 WORDS)
Filaments about 6-7 nm in diameter were seen associated with germ cell intercellular bridges in detergent-permeabilized cells treated with tannic acid. Approximately 40-50 filaments were present subjacent to the bridge density. Filaments encircled the bridge channel in a manner similar to contractile ring actin filaments of dividing cells. NBD-phallacidin and myosin S-1 subfragments were employed to demonstrate that the filaments observed at intercellular bridges are actin. Intratesticular injection of a single dose of cytochalasin D, a specific inhibitor of actin filaments, caused certain intercellular bridges of spermatids to open within 3 hr after injection, leading to the production of symplasts. During bridge opening, remnants of bridge densities were gradually incorporated into the lateral aspect of the plasma membrane of the symplast. Thus actin, present in bridge structures, appeared to participate in maintaining certain intercellular bridges. A model of intercellular bridge structure is presented.
To better understand, to optimize, and to validate the technique of intratesticular (i.t.) injection, several parameters related to i.t. injection were examined. Volumes exceeding 50 microliters could be injected i.t.; however, testes frequently became excessively turgid and backflow of injected fluids occurred. Thus, a volume of 50 microliters or less was deemed optimal for injection. To determine the rate of distribution of substances throughout the testis, trypan blue was injected i.t. near the caudal pole of the testis, and the movement of dye was monitored. Within 2 min, the dye had spread approximately 1 cm from the site of injection, and in 5 min it had spread twice that distance. In 2 h, the dye had become distributed throughout the testis except at its extreme cranial pole. Seminiferous tubules did not take up dye, indicating that the spread of dye was via peritubular lymphatics. Seminiferous tubule histology appeared virtually unaffected by i.t. injection, even at regions adjacent to the site of injection, when a sterile 26-gauge or smaller bore needle was utilized. To determine disappearance from the testis, radiolabeled inulin was injected i.t. Half time for absorption was achieved at 1.75 h. Potential vehicles were explored in which compounds with a variety of physical properties could be injected. Gum tragacanth, normal saline, ethylene glycol, dimethyl sulfoxide (DMSO) mixed 1:1 with normal saline, sesame oil, and propylene glycol were found to be suitable injection vehicles, whereas ethanol, dissolved in normal saline in concentrations as low as 0.5% was found unsuitable. To assess vehicle efficiency, various vehicles were utilized with a known testicular toxin (taxol) and injected into one testis, and the histology was compared with the contralateral testis injected with vehicle alone. All vehicles, found suitable above, allowed dispersion of taxol to influence areas distant from the site of injection. Intratesticular injection assesses the potential of agents to directly affect the testis, and systemic metabolism is avoided. Their rapid spread throughout the lymphatics of the testes allows seminiferous tubules to be exposed to agents in innocuous vehicles more rapidly and in higher concentration than is often possible when using systemic injections.
Examples are given of the detection of diagnostic clues in quantitative cytology and histopathology by statistical testing, such as may be applied in image analytical procedures. Schematic and other examples are presented of the visual images analyzed by each procedure, whose limits are also discussed. The situations analyzed include increased cellularity, differences in nuclear placement patterns, uniformity of displacement, variance in nuclear diameters and chaincode variance of nuclear placement. A specific model is presented for describing or generating a series of dependent observations representing nuclear placement, based on the Box-Jenkins (ARIMA) models for decomposing a spatial or temporal series into several components. This model describes the statistical observations that are random samples from the series. Finally, one graphic example is given in which visual inspection more readily ascertains an alteration than does statistical analysis of a modest-sized sample.
The use of order statistics to discriminate and classify DNA ploidy patterns is proposed, especially for the classification of additional observations: whether a given sample is more likely to have come from a normal or an abnormal tissue, and with what probability, based on its ploidy pattern. The method involves the order of observations within each of several samples (e.g., euploid and aneuploid DNA patterns) and the use of subsets of the obtained order statistics as independent variables in a linear discriminant analysis. It thus replaces univariate observations by (some of) their order statistics, which are then used as the variables in the discriminant analysis. The procedure does not require normality of distributions or the transformation of nonnormal distributions, as do many discriminant functions; order statistics are usually distribution-free and thus are particularly useful for nonparametric inference. Preliminary simulation studies verified the potential usefulness of the order statistics discriminant function method as applied to DNA ploidy analysis. Its advantages as compared to the usual methods for hypothesis testing, e.g., the use of the chi-square or Kolmogorov-Smirnov tests to as certain "goodness-of-fit," is discussed. The proposed method is easy to implement and easy to interpret; it is also applicable to the study of distributions of other types of measurements.