Effect of mirasol pathogen reduction technology system on immunomodulatory molecules of apheresis platelets.

Pathogen inactivation for platelets by riboflavin system (MIRASOL) efficiently reduces transfusion related pathogen transmission. However little is known about its impact on platelets' immunomodulatory biochemical profile. We aimed was to assess the effects of MIRASOL treatment on platelet quality parameters and immunomodulatory molecules CD62P, RANTES, and CD40L in Single Donor Platelets (SDPs) resuspended in plasma (SDP-P) or T-PAS and additive solution (SDP-A). Twenty nine SDPs (15 SDP-P and 14 SDP-A) were included in the study. Samples were collected before, after MIRASOL treatment and just before transfusion. P-selectin (CD62P), RANTES, and CD40L were tested by ELISA. Platelet products quality assays were also performed. Platelet count/unit decreased after Mirasol treatment by 13 %. The pH of all units decreased over the 5-day storage period but remained above expected limits and the swirling test was positive throughout storage. P-selectin levels were not different between the three different time points in both SDPs-P and SDPs-A while RANTES levels were found to differ statistically significantly at the three different time points in all units and in the SPD-A subgroup. CD40L levels in all SDP products increased slightly during storage but this was not statistically significant. CD62P, RANTES, and CD40L in all time points were elevated in SDPs-A compared to SDPs-P but not at a statistically significant level. In conclusion MIRASOL treatment apart from RANTES increase does not seem to substantially affect platelets associated other cytokines and immunomodulatory molecules namely P-selectin and sCD40L which are implicated in immune transfusion reactions.

HEMOSTASIS DISORDERS IN CORONARY ARTERY DISEASE: A PROSPECTIVE COMPARATIVE STUDY OF 130 PATIENTS.

The laboratory detection of factors that participate in coagulation mechanisms in patients with coronary heart disease may lead to important findings regarding the contribution of endothelial function to atherosclerotic lesions of coronary arteries. The main purpose of this study was to investigate the role of high-sensitivity C-reactive protein (hs-CRP), von Willebrand factor (vWF) activity, thrombomodulin (TM), ADAMTS13 activity and myeloperoxidase (MPO) in patients undergoing coronary angiography due to non-ST-elevation myocardial infarction (NSTEMI), unstable angina (UA) and stable angina pectoris with positive stress testing-induced myocardial ischemia (controls). Furthermore, the measured biomarkers were examined among patients with classical cardiovascular risk factors. 50 NSTEMI patients, 50 UA patients and 30 controls referred to coronary angiography were included in the study. The blood samples were collected before the catheter procedure. MPO, TM and ADAMTS13 activity were measured by enzyme-linked immunosorbent assay (ELISA), while vWF activity was calculated with INNOVANCE vWF Ac. When the laboratory results were compared between the three study groups, hs-CRP was found to be higher in NSTEMI patients compared to UA patients (p=0.0015) and controls (p<0.0001). ADAMTS13 activity was higher in NSTEMI (p=0.0035) and UA patients (p=0.0102) compared to controls and TM was lower in NSTEMI patients compared to UA patients (p=0.0307) and controls (p=0.0002). Moreover, MPO was higher in UA patients compared to the control group (p=0.0227). Finally, each of the aforementioned biomarkers was compared in the presence of the following cardiovascular risk factors: smoking, diabetes mellitus, arterial hypertension, dyslipidemia, chronic kidney disease (CKD) and peripheral artery disease (PAD). The results of this study add more data to the current medical literature concerning the role of coagulation disorders, endothelial damage and immunothrombosis in patients with coronary artery disease and their correlation with traditional risk factors for cardiovascular disease.

Periviable birth: A review of ethical considerations.

BACKGROUND: Advances in perinatology and medical technology have pushed the limits of viability to unprecedented extremes, leading to a growing population of NICU "graduates" with a wide range of health issues. Although survival rates from 22 weeks of gestation onwards have improved over the last 30 years, the incidence of disabilities remains the same. Providing intensive care to a high-risk population with significant mortality and morbidity raises the fundamental conflict between sanctity and quality of life. Potential severe handicap and need for frequent tertiary care inevitably impact the whole family unit and may outweigh the benefit of survival. The aim of this study is to explore and summarize the ethical considerations in neonatal care concerning perivable birth. METHODS: Eligible studies published on PubMed were included after a systematic search using the PICO methodology. RESULTS: Forty-eight studies were systematically reviewed regarding guidelines, withholding or withdrawing treatment, parental involvement, and principles applied in marginal viability. As periviable birth raises an array of complex ethical and legal concerns, strict guidelines are challenging to implement. CONCLUSIONS: Active life-sustaining interventions in neonatology should be balanced against the risk of putting infants through painful and futile procedures and survival with severe sequelae. More evidence is needed on better prediction of long-term outcomes in situations of imminent preterm delivery, while good collaboration between the therapeutic team and the parents for life-and-death decision-making is of utmost importance. HIPPOKRATIA 2021, 25 (1):1-7.

Correlation of single nucleotide polymorphisms in the promoter region of the ANXA5 (annexin A5) gene with recurrent miscarriages in women of Greek origin.

Background: Recent findings show that a number of single nucleotide polymorphisms (SNPs) within the promoter region of the annexin A5-gene (ANXA5) reduce the expression of the reporter gene and so they display a significant association with recurrent pregnancy loss (RPL).Objective: The objective of the present study aimed to address the contribution of ANXA5 M2 haplotype consisting of four minor alleles: (SNP1: (-)467G > A, SNP2: (-)448A > C, SNP3: (-)422T > C, and SNP4: (-)373G > A) in the occurrence of recurrent pregnancy losses in the Greek population, and the role of further two minor alleles: SNP5: (-)302 T > G and SNP6: (-)1C > T as independent risk factors for RPL.Methods: A 752-bp genomic region of ANXA5 promoter was amplified by PCR using specific primers. Genotypic analysis by Sanger sequencing was performed for these six SNPs (minor alleles) in the promoter region of ANXA5 gene, in 100 (100) Greek women with recurrent miscarriages (median =3) and 70 (70) fertile controls. Statistical analysis was done using the SAS 9.3 for Windows (SAS Institute Inc, NC, USA) and SPSS packages for Windows (C.DiMaggio 2013, SAS Institute 2014).Results: This case-control study revealed that there is no significantly increased risk of RPL among the M2/ANXA5 haplotype carriers in the Greek population, as there were no statistical differences between the patients with recurrent pregnancy losses and the fertile controls (11.5% in RPL cases versus 9.29% in controls, p-value: .6364). There was no difference in SNP5 and SNP6 minor carriership between the two groups. In particular, carriers of SNP5 and SNP6 had an increased risk for RPL state with odds ratio: 1.2472 and 1.3846 respectively, however without statistically significant importance.Conclusion: The M2/ANXA5 haplotype does not differ between RPL patients and controls in the Greek population. Also, it is the first time that SNP5 and SNP6 minor alleles were evaluated extensively in women of European origin with recurrent pregnancy losses (RPL), and they do not seem to be independent risk factors in the occurrence of RPL in the Greek population. Though, this has to be confirmed in further and larger clinical trials with women of European origin.

Low molecular weight heparins use in pregnancy: a practice survey from Greece and a review of the literature.

BACKGROUND: Use of LMWH in pregnancy is not only limited to VTE management, but it extends, to the management of vascular gestational complications and the optimization of IVF pregnancies despite the lack of concrete scientific evidence. In this context, we conducted the present study aiming to gain insights regarding the use of LMWH during pregnancy and puerperium. We recorded indication for use, diagnostic work-up as well as the safety and efficacy of the treatment, trying to elucidate the clinical practice in our country. METHODS: We analyzed data regarding 818 pregnant women received LMWH during 2010-2015.Our cohort had a median age of 33.9 years and a BMI of 23.6.There were 4 groups: those with a history of VTE [Group-A: 76], those with pregnancy complications [Group-B: 445], those undergoing IVF [Group-C: 132] and those carrying prothrombotic tendency (thrombophilia, family history of VTE, other) [Group-D: 165]. Mean duration of LMWH administration was 8.6 ± 1.5 months. Out of the total number, 440 received LMWH in fixed prophylactic dose, 272 in higher prophylactic-weight adjusted dose and 106 in therapeutic dose. Moreover, 152 women received in addition low-dose acetylsalicylic acid (ASA). 93.8% of pregnancies were single and 6.2% were multiple ones. Live births occurred in 98.7% of pregnancies. RESULTS: Anticoagulation was efficacious and well tolerated. Seventeen VTE events were recorded; 7 of them antepartum and 10 postpartum. No major bleeding events were observed while 13 clinical relevant non-major bleeding events were recorded. Regarding gestational vascular complications, 28 IUGR events were recorded, as well as 48 cases of preterm labor of which 12 were concomitant with IUGR (25%). Six early pregnancy losses were recorded; there were 3 fetal deaths and 3 cases of pre-eclampsia/eclampsia. CONCLUSIONS: LMWHs are used extensively during pregnancy and puerperium in Greece for VTE treatment and prophylaxis and for a variety of other indications as well. Although the drug has been shown to be both safe and efficacious, its use for some indications has no proven scientific evidence. In order to clearly define the role of LMWHs in pregnancy, beyond thromboprophylaxis, large prospective studies are required, which could be based on the conclusions of this study.

The utility of artificial neural networks and classification and regression trees for the prediction of endometrial cancer in postmenopausal women.

OBJECTIVE: Artificial neural networks (ANNs) and classification and regression trees (CARTs) have been previously used for the prediction of cancer in several fields. In our study, we aim to investigate the diagnostic accuracy of three different methodologies (i.e. logistic regression, ANNs and CARTs) for the prediction of endometrial cancer in postmenopausal women with vaginal bleeding or endometrial thickness ≥5 mm, as determined by ultrasound examination. STUDY DESIGN: We conducted a retrospective case-control study based on data from analysis of pathology reports of curettage specimens in postmenopausal women. METHODS: Classical regression analysis was performed in addition to ANN and CART analysis using the IBM SPSS and Matlab statistical packages. RESULTS: Overall, 178 women were enrolled. Among them, 106 women were diagnosed with carcinoma, whereas the remaining 72 women had normal histology in the final specimen. ANN analysis seems to perform better with a sensitivity of 86.8%, specificity of 83.3%, and overall accuracy (OA) of 85.4%. CART analysis did not perform well with a sensitivity of 78.3%, specificity of 76.4%, and OA of 77.5%. Regression analysis had a poorer predictive accuracy with a sensitivity of 76.4%, a specificity of 66.7%, and an OA of 72.5%. CONCLUSION: Artificial intelligence is a powerful mathematical tool that may significantly promote public health. It may be used as a non-invasive screening tool to guide clinicians involved in primary care decision making when endometrial pathology is suspected.

Fine needle aspiration cytology of nodular thyroid lesions: a 2-year experience of the Bethesda system for reporting thyroid cytopathology in a large regional and a university hospital, with histological correlation.

OBJECTIVE: Thyroid fine needle aspiration (FNA) contributes to the appropriate management of nodular thyroid lesions. The introduced categories in the Bethesda system for reporting thyroid cytopathology (TBSRTC) are associated with an implied cancer risk, providing a clinical management guideline. This study aims to evaluate the reproducibility of this implied risk and to compare the results from two different cytopathology departments. METHODS: Five hundred histologically confirmed FNAs, studied since the introduction of TBSRTC, were obtained from 4208 and 3587 FNAs performed in a large regional hospital in Herakleion, Crete (group A) and a university hospital in Athens (group B), respectively. Reports were issued according to TBSRTC. Aspirates were prepared with ThinPrep(®) and evaluated by two experienced cytopathologists. The reproducibility and accuracy were evaluated. RESULTS: The proportion test for suspicious for malignancy (SFM) and malignant (M) cytology reports (P < 0.0001), and the number of malignancies on histology (P < 0.0001), were significantly higher in group A than in group B, consistent with a higher incidence of thyroid carcinomas in southern Greece. Although the malignancy rates were higher in group A than in group B for all categories, except M (A, 99.3%; B, 100%), the difference was only significant for benign aspirates (P = 0.0303). Malignancy rates for all categories in group A were above the TBSRTC recommended range, but were consistent with an increased prevalence of malignancy in that centre, differences in reporting practice and the variable ranges reported in the literature. There was lower sensitivity (P = 0.019) and overall accuracy (P = 0.003) in group A relative to group B, but no difference in specificity. CONCLUSIONS: TBSRTC provides valuable information for the appropriate management of nodular thyroid lesions, both in a university and a large regional hospital.

The use of an immunocytochemical double-labeling staining can display the distribution of Bcl-2/Ki-67 cells in endometrial adenocarcinomas as well as in normal endometrium.

BACKGROUND: Endometrial adenocarcinoma is the most common gynecologic malignancy and is the fifth most frequently diagnosed cancer in women. The objective of the present study was to determine the expression of a proliferation marker, Ki-67 and an apoptosis inhibitor, Bcl-2, by double-label staining in endometrial adenocarcinomas and in normal endometrium samples, to evaluate the differences in the immunocytochemical expression of Ki-67 and Bcl-2, and finally to correlate the results with tumor grade and stage. METHODS: This study was carried on 270 endometrial samples, collected during a 27 month period, freshly resected from women who underwent total abdominal hysterectomy. RESULTS: Ki-67 and Bcl-2 expressions were studied by immunocytochemistry. Bcl-2 expression was strong and homogeneous in normal (proliferative, secretory and atrophic) endometrium and more frequent in low-grade endometrioid carcinomas. Completely negative staining of Bcl-2 was found to be strictly related to high-grade endometrioid carcinomas. Ki-67 expression was higher in patients with high-grade endometrioid carcinomas. Proliferative endometrium showed inconclusive Ki-67 expression levels and in the secretory endometrium Ki-67 positive cells were remarkably diminished and even disappeared. Completely negative staining of Ki-67 was found to be strictly related to atrophic endometrium. CONCLUSIONS: Immunocytochemical double-label staining can be used to display the distribution of Bcl-2 and Ki-67 cells in endometrioid carcinomas as well as normal endometrium, and, in addition to cytomorphologic features, contributes to its accurate diagnosis.

Neural network application in the discrimination of benign from malignant gastric cells.

OBJECTIVE: To investigate the potential value of morphometry and neural networks for the discrimination of benign from malignant gastric lesions. STUDY DESIGN: One thousand cells from 19 cases of cancer, 19 cases of gastritis and 56 cases of ulcer were selected as a training set, and an additional 4,000 cells from the same cases of cancer, gastritis and ulcer were used as a test set. Images of routinely processed gastric smears stained by the Papanicolaou technique were analyzed by a custom-made image analysis system. RESULTS: Application of the neural network gave correct classification in 96% of benign cells and 89% of malignant cells. CONCLUSION: The results indicate that the use of neural networks and image morphometry may offer useful information concerning the potential of malignancy in gastric cells.

Learning vector quantizer in the investigation of thyroid lesions.

OBJECTIVE: To investigate the capability of the learning vector quantizer (LVQ) in the discrimination of benign from malignant thyroid lesions. STUDY DESIGN: The study was performed on May-Grünwald-Giemsa-stained smears taken by fine needle aspiration (FNA). Using a custom image analysis system, 25 features that describe the size, shape and texture of approximately 100 nuclei were measured from each case. Statistical features were extracted from each case, and a linear regression analysis was performed to detect the statistically significant features. The cases were distributed by category, as follows: 100 cases of goiter and follicular adenomas, 11 cases of follicular carcinoma, 35 cases of papillary carcinoma, 24 cases of oncocytic adenoma, 8 cases of oncocytic carcinoma and 20 cases of Hashimoto thyroiditis. About 30% of the cases from each class were used for training two LVQ classifiers. The remaining 139 cases, out of a total of 198, were used as the test set. A classifier was used to discriminate into four classes and a second into two classes. RESULTS: The application of LVQ neural networks allows good discrimination between benign and malignant lesions (O.A. = 97.8). However, reliable discrimination of the cytologic types of the lesions was not obtained. CONCLUSION: These results indicate that the use of neural networks combined with image morphometry may offer useful information on the potential for malignancy of thyroid lesions and may improve the diagnostic accuracy of FNA of the thyroid gland, especially in cases of follicular neoplasms classified as suspicious for malignancy and in cases of oncocytic tumors.

Static cytometry and neural networks in the discrimination of lower urinary system lesions.

OBJECTIVES: To investigate the potential of morphometry and artificial intelligence tools for the discrimination of benign and malignant lower urinary system lesions. METHODS: The study group included 50 cases of lithiasis, 61 cases of inflammation, 99 cases of benign prostatic hyperplasia, 5 cases of in situ carcinoma, 71 cases of grade I transitional cell carcinoma of the bladder (TCCB), and 184 cases of grade II and grade III TCCB. Images of voided urine smears stained by the Giemsa technique were analyzed by a custom image analysis system. The analysis gave a data set of features from 45,452 cells. A learning vector quantizer (LVQ)-type neural network (NN) was used to discriminate benign from malignant cells on the basis of the extracted morphometric and textural features. The data from 13,636 randomly selected cells were used as a training set and the data from the remaining 31,816 cells made up the test set. Similarly, in an attempt to discriminate at the patient level, 30% of the cases randomly selected were used to train an LVQ NN and the remaining 329 cases were used for the test. RESULTS: The application of the LVQ NN enabled the correct classification of 95.42% of the benign cells and 86.75% of the malignant cells, giving an overall accuracy rate of 90.63%. At the patient level, the LVQ NN enabled the correct classification of 100% of benign cases and 95.6% of malignant cases, giving an overall accuracy rate of 97.57%. CONCLUSIONS: NNs combined with image analysis offer useful information in the discrimination of benign and malignant cells and lesions of the lower urinary system.

Comparing neural networks in the discrimination of benign from malignant lower urinary tract lesions.

OBJECTIVE: To compare the performance of two different neural networks (NNs) in the discrimination of benign and malignant lower urinary tract lesions. MATERIALS AND METHODS: A group of patients was evaluated, comprising 50 cases of lithiasis, 61 of inflammation, 99 of benign prostatic hyperplasia (BPH), five of in situ carcinoma, 71 of grade I transitional cell carcinoma of the bladder (TCCB), and 184 of grade II and grade III TCCB. Images of routinely processed voided urine smears were stained using the Giemsa technique and analysed using an image-analysis system, providing a dataset of 45452 cells. Two NN models of the back propagation (BP) and learning vector quantizer (LVQ) type were used to discriminate benign from malignant cells and lesions, based on morphometric and textural features. The data from 13636 randomly selected cells (30% of the total data) were used as a training set and data from the remaining 31816 cells comprised the test set. Similarly, in an attempt to discriminate patients, 30% of the cases, selected randomly, were used to train a BP and an LVQ NN, with the remaining 329 cases used for the test set. The data used for training and testing were the same for the two kinds of classifiers. RESULTS: The two NNs gave similar results, with an overall accuracy of discrimination of approximately 90.5% at the cellular level and of approximately 97% for individual patients. There were no statistically significant differences between the two NNs at the cellular or patient level. CONCLUSIONS: The use of NNs and image morphometry could increase the diagnostic accuracy of voided urine cytology; despite the different nature of the two classifiers, the results obtained were very similar.

A comparative study of three variations of the learning vector quantizer in the discrimination of benign from malignant gastric cells.

A prospective study was undertaken to investigate the potential value of morphometry and artificial neural networks (ANN) for the discrimination of benign and malignant gastric lesions. Two thousand five hundred cells from 23 cases of cancer, 19 cases of gastritis and 58 cases of ulcer were selected as a training set, and an additional 8524 cells from an equal number of cases of cancer, gastritis and ulcer were used as a test set. Images of routine processed gastric smears stained by the Papanicolaou technique were processed by a custom image analysis system. The application of the learning vector quantization (LVQ) classifier enabled correct classification of > 97% of benign cells and > 95% of malignant cells, obtaining an overall accuracy of > 97%. This study presents the capabilities of ANN, and also indicates that ANN and image morphometry may offer useful information on the potential of malignancy in gastric cells.

Back propagation neural network in the discrimination of benign from malignant lower urinary tract lesions.

PURPOSE: We investigated the potential value of morphometry and artificial intelligence tools to discriminate between benign and malignant lower urinary tract lesions. MATERIALS AND METHODS: The lesions included lithiasis in 50 cases, inflammation in 61, benign prostatic hyperplasia in 99, carcinoma in situ in 5, and grade I and grades II and III transitional cell carcinoma of the bladder in 71 and 184, respectively. Images of routine processed voided urine smears stained by the Giemsa technique were analyzed using a custom image analysis system, providing a data set of 45,452 cells. A neural net model of the back propagation type was used to discriminate benign from malignant cells based on the extracted morphometric and textural features. Data from 13,636 randomly selected cells (30% of the total data) were used as a training set and the data from the remaining 31,816 cells comprised the test set. In a similar attempt to discriminate at the patient level data on 30% of those randomly selected were used to train a back propagation neural net and data on the remaining 329 were used for testing. RESULTS: Application of the back propagation neural net enabled the correct classification of 95.34% of benign and 86.71% of malignant cells with overall 90.57% accuracy. At the patient level the back propagation neural net enabled the correct classification of 100% of those with benign and 94.51% of those with malignant disease with overall 96.96% accuracy. CONCLUSIONS: The use of neural nets and image morphometry may increase the speed of cytological diagnosis and the diagnostic accuracy of voided urine cytology.

Application of the learning vector quantizer to the classification of breast lesions.

OBJECTIVE: To investigate the potential of the learning vector quantization (LVQ) neural network for the discrimination of benign from malignant breast lesions. STUDY DESIGN: Using a custom image analysis system on Giemsa-stained smears, 25 parameters describing the size, shape and texture of the cell nucleus were measured. Three thousand nuclei from a total of 9,356 were selected as a training set for the neural network, and the whole data set was used for testing. An additional 238 cells from 16 cases without final cytologic diagnoses were evaluated by the system. The total number of cells (9,594) was collected from 100 patients (68 carcinomas and 32 benign lesions). RESULTS: Cytologic examination of the cases gave two false positive and two false negative results. However, in eight cases of ductal breast carcinoma and in eight cases of benign lesions, histologic confirmation was necessary in order to confirm the cytologic diagnosis. Application of the LVQ permitted correct classification of 87.41% of the cells. Classification at the patient level by using a hypothesis test for proportion with a hypothesis value equal to 50% permitted the correct diagnosis in 98% of patients. CONCLUSION: These results indicate that the use of neural networks combined with image morphometry and statistical techniques may offer useful information about the potential for malignancy, improving the diagnostic accuracy of fine needle aspiration of breast lesions.

Comparative study of artificial neural networks in the discrimination between benign from malignant gastric cells.

OBJECTIVE: To compare the accuracy of two different artificial neural networks (ANNs) for the discrimination of benign and malignant gastric lesions using morphometric and textural data on the nucleus. STUDY DESIGN: Three thousand cells from 30 cancer cases, 26 cases of gastritis and 64 cases of ulcer were selected as a training set, and an additional 10,300 cells from equal cases of cancer, gastritis and ulcer were used as a test set using two different neural net architectures: back propagation (BP) and learning vector quantizer (LVQ). Images of routinely processed gastric smears stained by the Papanicolaou technique were processed by a custom image analysis system. RESULTS: Application of the BP and three variations of the LVQ established correct classification of more than 97% of the benign cells and more than 95% of the malignant cells, obtaining an overall accuracy of more than 97%. CONCLUSION: This study not only presents a comparative study of the abilities of ANNs but also indicates that the use of ANNs and image morphometry may offer useful information on the potential of malignancy of gastric cells.

Potential of the back propagation neural network in the morphologic examination of thyroid lesions.

OBJECTIVE: To investigate the potential of back propagation (BP) neural networks (NNs) in the discrimination of benign from malignant thyroid lesions. STUDY DESIGN: The study was performed on May-Grünwald-Giemsa-stained smears obtained by fine needle aspiration (FNA). Using a custom image analysis system, 26 features that describe the size, shape and texture of the nucleus were measured from each cell. The cases were distributed according to categories, as follows: 25 cases of goiter and follicular adenomas, 1 case of follicular carcinoma, 12 cases of papillary carcinoma, 6 cases of oncocytic adenoma, 3 cases of oncocytic carcinoma and 4 cases of Hashimoto thyroiditis. From each case about 100 nuclei were measured; they formed a pool of 13,850 feature vectors. Out of this pool, 2,770 vectors were randomly selected to form the training set, and the remaining 11,080 vectors formed the test set. RESULTS: The application of a BP NN on the nuclear measurements permitted correct classification of 90.61% nuclei. Classification at the patient level was performed using a hypothesis test for proportion and two different hypothesis values, one equal to the overall accuracy of the NN and one equal to 50%. The second method permitted correct classification of 98% of patients. CONCLUSION: These results indicate that the use of NNs combined with image morphometry and statistical techniques may offer useful information on the potential malignancy of thyroid cells and may improve the diagnostic accuracy of FNA of the thyroid gland, especially in cases classified as suspicious for malignancy.

Potential of the back propagation neural network in the discrimination of benign from malignant gastric cells.

OBJECTIVE: To investigate the potential value of morphometry and the back propagation neural network for the discrimination of benign and malignant gastric lesions. STUDY DESIGN: The study group consisted of 23 cases of cancer, 19 of gastritis and 58 of ulcer. Images of routinely processed gastric smears stained by the Papanicolaou technique were processed by a custom image analysis system. Analysis of the images gave a data set of 11,024 cells. Two different neural net architectures were used to classify benign from malignant cells based on the extracted morphometric and textural features. The data from 2,500 randomly selected cells were used as a training set, and the data from the remaining 8,524 cells were applied as a test set. RESULTS: Application of the back propagation neural network permitted the correct classification of 97.6% of benign cells and 95% of malignant cells with overall accuracy 97.3%. CONCLUSION: These results indicate that neural networks and image morphometry may offer useful information about the potential for malignancy in gastric cells.