The use of the decision tree technique and image cytometry to characterize aggressiveness in World Health Organization (WHO) grade II superficial transitional cell carcinomas of the bladder

Automatic Prediction of Meningioma Grade Image Based on Data Amplification and Improved Convolutional Neural Network

Meningioma is the second most commonly encountered tumor type in the brain. There are three grades of meningioma by the standards of the World Health Organization. Preoperative grade prediction of meningioma is extraordinarily important for clinical treatment planning and prognosis evaluation. In this paper, we present a new deep learning model for assisting automatic prediction of meningioma grades to reduce the recurrence of meningioma. Our model is based on an improved LeNet-5 model of convolutional neural network (CNN) and does not require the extraction of the diseased tissue, which can greatly enhance the efficiency. To address the issue of insufficient and unbalanced clinical data of meningioma images, we use an oversampling technique which allows us to considerably improve the accuracy of classification. Experiments on large clinical datasets show that our model can achieve quite high accuracy (i.e., as high as 83.33%) for the classification of meningioma images.

Prediction of radiosensitivity in human bladder cell lines using nuclear chromatin phenotype

BACKGROUND

Nuclear texture analysis measures phenotypic changes in chromatin distribution within a cell nucleus, while the alkaline Comet assay is a sensitive method for measuring the extent of DNA breakage in individual cells. The authors aim to use both methods to provide information about the sensitivity of cells to ionizing radiation.

METHODS

The alkaline Comet assay was performed on six human bladder carcinoma cell lines and one human urothelial cell line exposed to gamma-radiation doses from 0 to 10 Gy. Nuclear chromatin texture analysis of 40 features was then performed in the same cell lines exposed to 0, 2, and 6 Gy to explore if nuclear phenotype was related to radiation sensitivity.

RESULTS

Comet assay results demonstrated that the cell lines exhibited different levels of radiosensitivity and could be divided into a radiosensitive and a radioresistant group at >6 Gy. Using stepwise discriminant analysis, a subset of important nuclear texture features that best discriminated between sensitive and resistant cell lines were identified A classification function, defined using these features, correctly classified 81.75% of all cells into their radiosensitive or radioresistant groups based on their pretreatment chromatin phenotype. Posttreatment chromatin changes also varied between cell lines, with sensitive cell lines showing a relaxed chromatin conformation following radiation, whereas resistant cell lines exhibited chromatin condensation.

CONCLUSIONS

The authors conclude that the alkaline Comet assay and nuclear texture methodologies may prove to be valuable aids in predicting the response of tumor cells to radiotherapy.

Automated detection of immunofluorescently labeled cytomegalovirus-infected cells in isolated peripheral blood leukocytes using decision tree analysis

BACKGROUND

Cytomegalovirus (CMV) infection continues to be a major problem for immunocompromised patients. Detection of viral antigens in leukocytes (antigenemia assay) is widely used for the diagnosis of CMV infection and for guiding antiviral therapy. The antigenemia technique, contingent upon the manual microscopic analysis of rare cells, is a laborious task that is subject to human error. In this study, we combine automated microscopy with artificial intelligence for reliable detection of fluorescently labeled CMV-infected cells.

METHODS

Cytospin preparations of peripheral blood leukocytes were immunofluorescently labeled for the CMV lower matrix phosphoprotein (pp65) and scanned in the Rare Event Imaging System (REIS), a fully automated image cytometer. The REIS detected potential positive objects and digitally recorded 49 measured cellular features for each identified case. The measurement data of these objects were analyzed by the See5 decision tree (DT) algorithm to ascertain whether they were true-positive detections.

RESULTS

The DT was built from the measurement data of 2,047 true- and 2,028 false-positive detections, collected from 32 patient samples. By designating misclassifications of false-negatives three times more costly, the 10-fold cross-validation sensitivity, specificity, and misclassification error of the assay was 94.3%, 56.2%, and 25%, respectively. The method was also validated using an independent test set of 21 patient samples, in which similar results were obtained.

CONCLUSIONS

To our knowledge, this study represents the first attempt to improve the accuracy of rare event image cytometry through the implementation of artificial intelligence methodology. Results suggest that cost-sensitive decision tree analysis of digitally measured cellular features vastly improves the performance of rare event image cytometry.

Prognostic values of galectin-3 and the macrophage migration inhibitory factor (MIF) in human colorectal cancers

This study aims to investigate whether the immunohistochemical levels of expression of galectin-3 and the macrophage migration inhibitory factor (MIF) are associated with prognostic values in human colorectal tumors. This was performed on 99 specimens including 69 colorectal tumors (17 Dukes A, 19 Dukes B, 15 Dukes C and 18 metastatic tumors that we labeled as D), 10 hepatic metastases from colorectal cancers and 20 normal specimens (biopsies). The immunohistochemical levels of expression of MIF and galectin-3 were quantified on routine histological slides by means of computer-assisted microscopy. Separate analyses were performed on epithelial and connective tissue. The levels of expression of both MIF and galectin-3 were very significantly higher in epithelial tumor tissue when compared with normal epithelial specimens. A positive and significant correlation between MIF and galectin-3 expression was evidenced in connective tumor tissue, and in particular in the cases associated with short survival periods (less than 5 years). In the case of the Dukes A or B tumors, we established two new prognostic groups (labeled I and II) on the basis of the levels of galectin-3 expression measured in the tumor epithelium. In the case of the Dukes C or D tumors, we established two other prognostic groups (labeled III and IV) on the basis of the levels of MIF expression measured in the connective tissue. Kaplan-Meyer analyses confirmed the additional prognostic values (as compared with conventional clinical staging) given by this new classification (groups I to IV). They show that the Dukes A or B tumors characterized by low levels of galectin-3 expression in the tumor epithelium are associated with significantly better prognoses than those characterized by high levels. In addition, the Dukes C or D tumors characterized by high levels of MIF expression in the connective tumor tissue are associated with significantly better prognoses than those characterized by low levels. In conclusions, MIF and galectin-3 expression levels in colorectal tumors are related to their levels of biological aggressiveness. These markers could be used to identify patients at risk, for whom more aggressive adjuvant therapy seems to be indicated.

Multicriteria fuzzy assignment method: a useful tool to assist medical diagnosis

The aim of this paper is to provide a concise portrayal of medical applications of a new fuzzy classification method called PROAFTN, which uses a multicriteria decision aid approach. The review summarises and discusses medical applications of the proposed method in acute leukemia, astrocytic and bladder tumours. Although still an investigative method, the preliminary results are very encouraging and demonstrate the potential performances of this procedure for solving medical classification problems.

Identification by quantitative chromatin pattern analysis of patients at risk for recurrence of superficial transitional bladder carcinoma

PURPOSE

Based on the actual clinical outcomes of 132 fully documented patients with superficial transitional cell carcinoma of the bladder, we characterize the risk of recurrence and/or progression by computer assisted image microscopy applied to Feulgen stained nuclei.

MATERIALS AND METHODS

Each tumor was characterized by the conventional grading and staging systems as well as by cytometry generated variables describing nuclear DNA content, nuclear morphometry and chromatin patterns. These data were submitted to discriminant analysis to establish a model distinguishing between 2 groups of patients. Group 1 included cases with remission for more than 60 months and group 2 cases presented with recurrence with or without progression within 12 months of transurethral bladder resection. This latter model was then validated by Kaplan-Meyer analysis of the full data set.

RESULTS

As evidenced by Kaplan-Meier analysis, the discriminant factor generated by discriminant analysis of cytometry generated variables provided a cutoff value for distinguishing between low and high risks of recurrence (p <0.00001). In contrast, conventional grading and staging systems were not able to make such efficient distinction.

CONCLUSIONS

These 2 groups can be used as references with which new cases can be compared to prognosticate disease behavior independently of histopathological grading and/or clinical staging.

Improving the prognostic value of histopathological grading and clinical staging in renal cell carcinomas by means of computer-assisted microscopy

The present work aims to refine prognosis in cases of renal cell carcinoma (RCC) by integrating a variety of parameters with the prognostic information provided by histopathological grading and clinical staging, carried out on a series of 97 RCCs. To this end, Feulgen-stained RCC cell nuclei were characterized by means of 38 variables describing nuclear DNA ploidy levels and morphology. All of these data were subjected to a principal components analysis. On the basis of this multivariate analysis, Fuhrman grade II was subdivided into grades II- and II+, and Fuhrman grade III into grade III- and III+. The same kind of subcategorization was performed in the case of the T2 and T3 clinical stages. The results show that the classification into grade II- and III- RCCs correspond to a more favourable prognosis than grade II+ and III+, to which shorter survival periods were attributable. Similar results were obtained for the subcategorization of the T2 and T3 clinical stages. Very simple biological characterizations of these grade- or stage-related RCC groups were obtained by means of a decision tree approach applied to the cytometry-generated variables. The resulting classification rules were validated on a new series of 18 patients and enabled very accurate predictions of survival.

Image cytometry as a discriminatory tool for cytologic specimens obtained by endoscopic retrograde cholangiopancreatography

BACKGROUND

Routine brush cytology is relatively insensitive for the diagnosis of biliary and pancreatic malignancy. Sensitivity can be improved by measuring DNA and proliferation. The goal of this study was to assess the discriminatory capacity of image cytometry using pancreaticobiliary brush cytology specimens obtained during endoscopic retrograde cholangiopancreatography (ERCP). Analysis included morphometry, DNA quantification, and characterization of nuclear chromatin distribution and condensation.

METHODS

Brush cytology specimens were obtained during ERCP from 22 chronic pancreatitis specimens, 11 pancreatic adenocarcinoma specimens, 13 primary sclerosing cholangitis specimens, and 11 cholangiocarcinoma specimens and contrasted with 25 normal epithelia specimens. A SAMBA 2005 image processor was used to analyze Feulgen stained chromatin density and distribution. Discriminant analysis of 37 morphonuclear variables was performed to characterize differences between: 1) chronic pancreatitis and pancreatic adenocarcinoma, and 2) primary sclerosing cholangitis and cholangiocarcinoma.

RESULTS

Chronic pancreatitis was distinguished from pancreatic adenocarcinoma (P < or = 0.001); sensitivity and specificity were both estimated to be 82%. Primary sclerosing cholangitis was distinguished from cholangiocarcinoma (P < or = 0.01); sensitivity and specificity were estimated to be 82% and 85%, respectively.

CONCLUSIONS

Multiparameter image cytometry has potential as an adjuvant diagnostic technique in patients with pancreaticobiliary malignancy.

Classification strategies for the grading of renal cell carcinomas, based on nuclear morphometry and densitometry

The various grading systems proposed for renal cell carcinomas all suffer from problems related to inter-observer variability. Some of these grading systems are based, either partially or wholly, on morphonuclear criteria, such as nuclear size and shape, anisonucleosis, and chromatin pattern. These criteria can be quantitatively (and thus objectively) evaluated by means of the computer-assisted microscopic analysis of Feulgen-stained nuclei. In the present work, 39 quantitative variables, including two morphometric, 28 chromatin pattern-related, and nine DNA ploidy level-related, were computed for 65 renal cell carcinomas. The actual diagnostic information contributed by each variable was determined by means of multifactorial statistical analysis (discriminant analysis) and two artificial intelligence-related methods of data classification (the decision tree and production rule methods). The results show that quantitative information, as provided by the computer-assisted microscopy of Feulgen-stained nuclei and analysed by means of artificial intelligence-related methods of data classification, contributes significant diagnostic information for the grading of renal cell carcinoma, thus reducing the problem of inter-observer reproducibility.