Andrea Peters

Apoptosis as a cellular predictor for histopathologic response to neoadjuvant radiochemotherapy in patients with rectal cancer.

BACKGROUND Tumor shrinkage by preoperative radiochemotherapy (RCT) can markedly improve surgery in locally advanced (T4) rectal cancer with clear resection margins and may enable sphincter preservation in low-lying tumors. However, tumor response varies considerably, even among tumors treated according to the same protocol. If one is able to identify patients with highly radio-responsive tumors at the time of diagnosis, a selective and individualized policy of preoperative RCT might be pursued. METHODS The apoptotic index (AI), Ki-67, p53, and bcl-2 were evaluated by immunohistochemistry on pretreatment biopsies from 44 patients treated uniformly according to a prospective neoadjuvant RCT protocol (CAO/AIO/ARO-94). Treatment response was assessed histopathologically in the resected surgical specimen, using a five-point grading system. Expression of each marker was correlated with tumor response and relapse-free survival after curative surgery. RESULTS Tumors with complete (n = 3) or good (n = 28) response to RCT showed significantly higher pretreatment levels of apoptosis (mean AI: 2.06%) than tumors with moderate (n = 7), minimal (n = 5), or no regression (n = 1) from RCT (AI: 1.44%, p = 0.003). The AI was significantly related to Ki-67 (p = 0.05), but not to p53 and bcl-2 status. Tumor regression and AI best predicted relapse-free survival after combined modality treatment and curative surgery. CONCLUSION Spontaneous apoptosis in rectal cancer may serve as an important predictor of tumor regression from RCT in rectal cancer and as a significant prognosticator of relapse-free survival. Thus, this molecular marker may finally help to tailor therapy with regard to (neo-) adjuvant treatment of rectal cancer.

Microarray versus conventional prediction of lymph node metastasis in colorectal carcinoma

The authors investigated whether microarray‐based gene expression analysis of primary tumor biopsy material could be used to predict lymph node status in patients with colorectal carcinoma (CRC). Lymphatic metastasis strongly determines treatment algorithms in CRC. Currently, postoperative histology results are needed to determine lymph node status. Reliable preoperative information would be useful to advance treatment strategies.

Improving glaucoma diagnosis by the combination of perimetry and HRT measurements

PurposeThe aim of this study was to determine, whether the combination of morphologic data of the optic nerve head and visual field (VF) data would improve diagnosis of glaucoma, on the basis of the measurements alone. Patients and MethodsEighty-eight perimetric glaucomatous and 88 normal optic discs from the Erlangen Glaucoma Registry were matched for age. All normals and patients were examined in a standardized manner (Slitlamp biomicroscopy, gonioscopy, 24 h-applanation tonometry, automated VF testing, 15-degree optic disc stereographs, and Heidelberg Retina Tomograph (HRT)-scanning of the optic disc). The HRT variables were calculated in 4 optic disc sectors. All variables were calculated with the softwares standard reference plane. To gain the same allocation of sectors as provided by the HRT software, the VF responses were averaged within 4 sectors. Classification results of these VF responses were compared with the summarized results within 4 sectors. Six different combinations of morphologic and VF data were used to assess their suitability to diagnose the disease. HRT measurements, and the standard output of the Octopus (HRT/PERI1), HRT measurements and the summarized sectors and their standard deviations (HRT/PERI2), HRT measurements, standard output of the octopus and the summarized sectors and their standard deviations (HRT/PERI1/PERI2), standard output of the Octopus (PERI1), summarized sectors of the Octopus and their standard deviations (PERI2) and HRT measurements. To assess the diagnostic value of the different data sets machine learning classifiers, stabilized linear discriminant analysis, classification trees, bagging, and double-bagging were applied. ResultsCombination of morphologic and VF data improved the automated classification rules. The accuracy to diagnose glaucoma just by VF and HRT indices was maximized for double-bagging using both diagnostic tools. An estimated misclassification probability of less than 0.07 could be achieved for the primary open angle glaucoma patients combining HRT and VF sectors by double bagging. So highest sensitivity was 95% and specificity 91%, achieved by double-bagging and combination of HRT, PERI1, and PERI2. ConclusionsThe combination of optic disc measurements and VF data could not only improve glaucoma diagnosis in future, but could also help to find an objective way to diagnose glaucomatous optic atrophy. The limitation of the topographic relationship between structure and function is the individual variability of the optic disc morphology and the subjective variability of VF testing.

New Glaucoma Classification Method based on Standard Heidelberg Retina Tomograph Parameters by Bagging Classification Trees

PURPOSE In this article we propose and evaluate nonparametric tree classifiers that can handle non-normal data and a large number of possible predictors using the full set of standard Heidelberg Retina Tomograph measurements for classifying glaucoma. METHODS The classifiers were trained and tested using standard Heidelberg Retina Tomograph parameters from examinations of 98 subjects with glaucoma and 98 normal subjects of the Erlangen Glaucoma Registry. All patients and control subjects were evaluated by 15 degrees -optic disc stereographs, Heidelberg Retina Tomograph measurements, standard computerized white-in-white perimetry, and 24-hour-intraocular pressure profiles. The subjects were matched by age and sex. Standard classification trees as well as bagged classification trees were used. The classification outcome of the trees was compared with the classification by two published linear discriminant functions based on Heidelberg Retina Tomograph variables with respect to their cross-validated misclassification error. RESULTS The bagged classification tree had the lowest misclassification error estimate of 14.8% with a sensitivity of 81.6% at a specificity of 88.8%. The cross-validated error rates of the two linear discriminant function procedures were 20.4% (sensitivity 82.6%, specificity 76.7%) and 20.6% (sensitivity 81.4%, specificity 77.3%) for our set of observations. Bagged classification trees were able to reduce the misclassification error of glaucoma classification. CONCLUSIONS Bagged classification trees promise to be a new and efficient approach for glaucoma classification using morphometric 2- and 3-dimensional data derived from the Heidelberg Retina Tomograph, taking into account all given variables.

Comparison of classifiers applied to confocal scanning laser ophthalmoscopy data.

OBJECTIVES Comparison of classification methods using data of one clinical study. The tuning of hyperparameters is assessed as part of the methods by nested-loop cross-validation. METHODS We assess the ability of 18 statistical and machine learning classifiers to detect glaucoma. The training data set is one case-control study consisting of confocal scanning laser ophthalmoscopy measurement values from 98 glaucoma patients and 98 healthy controls. We compare bootstrap estimates of the classification error by the Wilcoxon signed rank test and box-plots of a bootstrap distribution of the estimate. RESULTS The comparison of out-of-bag bootstrap estimators of classification errors is assessed by Spearmans rank correlation, Wilcoxon signed rank tests and box-plots of a bootstrap distribution of the estimate. The classification methods random forests 15.4%, support vector machines 15.9%, bundling 16.3% to 17.8%, and penalized discriminant analysis 16.8% show the best results. CONCLUSIONS Using nested-loop cross-validation we account for the tuning of hyperparameters and demonstrate the assessment of different classifiers. We recommend a block design of the bootstrap simulation to allow a statistical assessment of the bootstrap estimates of the misclassification error. The results depend on the data of the clinical study and the given size of the bootstrap sample.

Influence of alcohol consumption on incidence and severity of open-globe eye injuries in adults

AimTo investigate the influence of alcohol consumption on the occurrence of open-globe injuries in adults.MethodsA retrospective study was made of 100 consecutive patients (81 male, 19 female) with open-globe injuries. Of these patients, 18 exhibited alcohol intoxication (group Ai), and 82 exhibited no alcohol intoxication (group nAi). Investigated parameters were best-corrected visual acuity at day of admission and last examination (logMAR), type of injury according to BETT-classification, extraocular injuries, cause of injury, time and setting of injury, in relation to alcohol consumption and tested for statistical significance with Fisher’s exact test or the Mann–Whitney U test, respectively.ResultsIn group Ai, 83.3% of the patients were male, and in group nAi, 80.5%. Mean logMAR at day of admission was 1.06 ± 0.63 (20/250) in group Ai and 1.08 ± 0.59 (20/250) in group nAi. At last examination, mean logMAR in group Ai was 1.11 ± 0.59 (20/250), in group nAi 0.75 ± 0.60 (20/125). This difference was statistically significant (p = 0.02). In group Ai, significantly more ruptures according to BETT classification occurred (p = 0.05). In group Ai, significantly more additional extraocular injuries occurred compared to group nAi (38.9% versus 6.1%; p = 0.0009). In group Ai, the cause of injury was significantly more often glass (44.4% versus 2.4%; p = 0.0000), in group nAi the injury was more often directly or indirectly caused by tools (74.4% versus 33.3%; p = 0.001). In group Ai, the injury was significantly more often inflicted by others (50.0% versus 9.8%; p = 0.0003). The settings in which the injuries occurred were significantly more often the street in group Ai (44.4% versus 6.1%; p = 0.0002), in group nAi the garden or tool shed (31.7% versus 5.6%; p = 0.02) or the workplace (34.2 % versus 11.1 %; p = 0.04). In group Ai, the injuries occurred significantly more often at night (p = 0.0001) and on weekends (p = 0.0000).ConclusionsOpen-globe eye injuries under alcohol intoxication are more often caused by a third party and have a worse prognosis. Open-globe injuries under alcohol intoxication occur in a different spatio-temporal setting and exhibit a more severe type of injury. Risk behavior combined with alcohol consumption therefore seems to be an independent factor for the incidence of open-globe eye injuries.

Generalised indirect classifiers

Supervised classifiers are usually based on a set of predictors given in the learning sample as well as in later test samples. Especially in the medical field a reduction of the number of examinations is often desired to save patients time and costs. The approach of indirect classification makes use of all available variables of the learning sample, although it classifies based only on a reduced set of variables. A general definition of indirect classification is given and a specific generalised indirect classifier is proposed. This classifier combines an arbitrary number of regression models which predict those variables that are not acquired for future observations. The performance of the generalised indirect classifier is investigated by using a simulation model which mimics different kinds of decision surfaces and by the application to different data sets. Misclassification results of direct and indirect classifiers are compared.

Glaucoma Diagnosis by Indirect Classifiers

Medical decision making is based on various diagnostic measurements and the evaluation of anamnestic information. For example glaucoma diagnosis is based on several direct and indirect assessments of the eye. We discuss possibilities to use a definition of glaucoma to improve supervised glaucoma classification by laser scanning image data. The learning sample consists of laser scanning image data and other diagnostic measurements.