Inese Polaka

Pepsinogen testing for evaluation of the success of Helicobacter pylori eradication at 4 weeks after completion of therapy

BACKGROUND AND OBJECTIVE Pepsinogen levels in plasma are increased by inflammation in the gastric mucosa, including inflammation resulting from Helicobacter pylori infection. A decrease in pepsinogen II level has been suggested as a reliable marker to confirm the successful eradication of infection. The aim of our study was to evaluate the potential role of pepsinogens I and II, gastrin-17 and H. pylori antibodies in confirming successful eradication. MATERIAL AND METHODS Altogether 42 patients (25 women, 17 men), mean age 45 years (range 23-74), were enrolled. Pepsinogens I and II, gastrin-17 and H. pylori IgG antibodies were measured in plasma samples using an ELISA test (Biohit, Oyj., Finland) before the eradication and 4 weeks after completing the treatment. The success of eradication was determined by a urea breath test. RESULTS Eradication was successful in 31 patients (74%) and unsuccessful in 11 patients (26%). Pepsinogen II decreased significantly in both the successful (P=0.029) and unsuccessful (P=0.042) eradication groups. Pepsinogen I decreased significantly in the successful (P=0.025) but not the unsuccessful (P=0.29) eradication group. The pepsinogen I/II ratio increased in the successful eradication group (P=0.0018) but not in the group in which treatment failed (P=0.12). There were no differences in gastrin-17 or H. pylori antibody values. CONCLUSIONS A decrease in pepsinogen II levels cannot be used as a reliable marker for the successful eradication of H. pylori 4 weeks after the completion of treatment. The increase in pepsinogen I/II ratio reflects differences in pepsinogen production following the eradication irrespective of improvement in atrophy.

Detection of gastric atrophy by circulating pepsinogens: A comparison of three assays

Circulating levels of pepsinogens have been used in high gastric cancer‐risk Asian and European populations to triage endoscopic evaluation for more severe pathology. There are different analytic methods with uncertain correlations. We therefore compared diagnostic performance of three commonly used pepsinogen assays to detect histologically confirmed gastric atrophy.

Multicentric randomised study of Helicobacter pylori eradication and pepsinogen testing for prevention of gastric cancer mortality: the GISTAR study

Introduction Population-based eradication of Helicobacter pylori has been suggested to be cost-effective and is recommended by international guidelines. However, the potential adverse effects of widespread antibiotic use that this would entail have not been sufficiently studied. An alternative way to decrease gastric cancer mortality is by non-invasive search for precancerous lesions, in particular gastric atrophy; pepsinogen tests are the best currently available alternative. The primary objective of GISTAR is to determine whether H pylori eradication combined with pepsinogen testing reduces mortality from gastric cancer among 40–64-year-old individuals. The secondary objectives include evaluation of H pylori eradication effectiveness in gastric cancer prevention in patients with precancerous lesions and evaluation of the potential adverse events, including effects on microbiome. Methods and analysis Individuals are recruited from general population (50% men) in areas with high gastric cancer risk in Europe and undergo detailed lifestyle and medical history questionnaire before being randomly allocated to intervention or control groups. The intervention group undergoes H pylori testing and is offered eradication therapy if positive; in addition, pepsinogen levels are detected in plasma and those with decreased levels are referred for upper endoscopy. All participants are offered faecal occult blood testing as an incentive for study participation. Effectiveness of eradication and the spectrum of adverse events are evaluated in study subpopulations. A 35% difference in gastric cancer mortality between the groups is expected to be detectable at 90% power after 15 years if 30 000 individuals are recruited. Biological materials are biobanked for the main and ancillary studies. The study procedure and assumptions will be tested during the pilot phase. Ethics and dissemination The study was approved by the respective ethics committees. An independent Data Safety and Monitoring Board has been established. The findings will be published in peer-reviewed journals and presented at scientific meetings. Trial registration number NCT02047994

Decision Tree Classifiers in Bioinformatics

Decision Tree Classifiers in Bioinformatics This paper presents a literature review of articles related to the use of decision tree classifiers in gene microarray data analysis published in the last ten years. The main focus is on researches solving the cancer classification problem using single decision tree classifiers (algorithms C4.5 and CART) and decision tree forests (e.g. random forests) showing strengths and weaknesses of the proposed methodologies when compared to other popular classification methods. The article also touches the use of decision tree classifiers in gene selection. Lēmumu koku klasifikatori bioinformātikā Rakstā piedāvāts literatūras apskats, analizējot zinātniskos rakstus, kas apskata klasifikācijas koku un to ansambļu metožu izmantošanu klasifikācijas uzdevuma risināšanai bioinformātikā. Apskatīts vēža klasifikācijas uzdevums, kurā nosaka vēža tipu vai pacienta diagnozi (slims vai vesels) pēc gēnu ekspresijas datiem (mikrorežga formāta dati). Apskatīti vairāki raksti, kas analizē dažādu klasifikācijas metožu pielietošanas iespējas šādu bioinformātikas uzdevumu risināšanā un salīdzina to veiktspēju, izmantojot dažādas datu kopas un pirmapstrādes pieejas. Klasifikatoru salīdzināšanā ņemts vērā arī īpatnējais datu raksturs - dati satur vairākus tūkstošus atribūtu (gēnu) un salīdzinoši maz ierakstu (daži desmiti vai simti), kas apgrūtina klasisko datu ieguves metožu darbību. Apskatītajos rakstos aprakstītās lēmumu koku metodes šajā rakstā tiek salīdzinātas pēc to efektivitātes (klasifikācijas kļūda/precizitāte), kas uzrādīta vairākās populārās gēnu mikrorežga datu kopās (leikēmijas, limfomas u.c. datu kopas). Rakstā arī apskatītas uz lēmumu koku izmantošanu balstītas metodes, kas izmantotas gēnu atlasei. Šādas metodes ir, piemēram, gēnu lietderības noteikšana pēc lēmumu koku klasifikatoru konstruēšanā izmantotās atribūtu informatīvuma novērtēšanas pieejas (Information Gain u.c.) un gadījuma lēmumu koku mežu generēšana, nosakot visbiežāk izmantotos gēnus, kas tiek atlasīti tālākajam darbam. Kopumā lēmumu koku klasifikatoru veiktspēja ir līdzvērtīga vai pārspēj citas klasiskās metodes, veicot pareizu datu pirmapstrādi. Lēmumu koku klasifikatoru ansambļu veiktspēja lielākoties pārspēj vienkāršu lēmumu koku klasifikatoru veiktspēju, ņemot vērā šādu klasifikatoru nestabilitāti. Lēmumu koku priekšrocība ir arī to vieglā interpretējamība un to spēja atklāt sakarības datos, kas var palīdzēt atklāt gēnu lomu slimības diagnostikā un ārstēšanā. Деревья решений в биоинформатике В статье предложен обзор литературы, анализ научных статей, которые рассматривают применение методов деревьев решений и их ансамблей для решения задач классификации в биоинформатике. Рассматривается задача классификации рака, которая определяет тип рака или диагноз пациента (больной или здоровый) по данным экспрессии генов (данные формата микрочипов). Рассматриваются статьи, в которых анализируются возможности применения различных методов классификации в области биоинформатики при решении подобных задач и сравнивается их производительность с помощью различных наборов данных и подходов предобработки. При сравнении классификаторов также принимается во внимание особый характер данных - данные содержат несколько тысяч признаков (генов) и относительно небольшое число записей (несколько десятков или сотен), что осложняет работу классических методов добычи данных. Методы деревьев решений, рассматриваемые в статьях, сравниваются в данной статье по их эффективности (ошибка /точность классификации), показанной в экспериментах с популярными наборами данных генных микрочипов (наборами данных о лейкемии, лимфоме и другими). В статье также обсуждается использование методов на основе деревьев решений для отбора генов. Такие методы включают в себя, например, использование подходов к оценке информативности атрибутов (Information Gain и т. д.), которые используются при построении классификаторов деревьев решений, и генерацию случайных лесов деревьев решений для определения наиболее часто используемых генов, которые отбираются для дальнейшей работы. В целом, классификаторы деревьев решений по производительности равны или превосходят другие традиционные методы, производя правильную предварительную обработку данных. Ансамбли классификаторов деревьев решений в значительной степени превосходят простые классификаторы деревьев решений по производительности с учетом нестабильности классификаторов. Преимущество методов деревьев решений заключается в том, что их легко интерпретировать, и они способны обнаруживать взаимосвязи в данных, которые могут помочь определить роль гена в диагностике и лечении заболеваний.

Comparison of the yield from two faecal immunochemical tests at identical cutoff concentrations - a randomized trial in Latvia.

Objective We have compared the performance of two faecal immunochemical tests (FIT) in an average-risk population. Materials and methods Altogether, 10 000 individuals aged 50–74 were selected randomly from the population of Latvia in 2011 and assigned randomly either to OC-Sensor or to FOB Gold single-time testing. Positivity of the test, frequency of colonic lesions, number needed to screen (NNscreen) and scope for the detection of an advanced neoplasm (cancer and advanced adenoma) were compared between the tests using the same cutoff concentrations in µg/g faeces. Confidence intervals (CIs) at 95% were calculated. Results Positivity with the cutoff set at 10 µg/g faeces was 12.8% (95% CI: 11.4–14.2) for FOB Gold and 8.3% (95% CI: 7.2–9.4) for OC-Sensor (P<0.001). Positivity was higher in men and the older age groups. Colonoscopy compliance was 55.5%. There was no significant difference between the two tests at comparable cutoff concentrations in µg/g, colonoscopy attendance rate or colonoscopy results. For advanced neoplasm detection, there was no significant difference in number needed to scope and NNscreen at a cutoff of 10 µg/g faeces; however, lower NNscreen was required to detect advanced neoplasms with the FOB Gold test at increased cutoff concentrations. Conclusion Different quantitative FIT systems may report different positivity rate at identical cutoff concentrations, which has to be considered when implementing the use of FIT in national screening programmes.

Method for decomposition of monolithic enterprise applications

Enterprise applications are used for managing operational data and are aimed at improving business efficiency. Many enterprise applications have been developed over the past three decades and are often referred to as legacy systems. Usually they are monolith, inflexible, poorly documented and hard to maintain. These issues can be addressed by improving their design using decomposition. The purpose of this paper is to introduce a method for decomposition of large scale enterprise applications and to present initial practical evaluation results. The paper systematically defines the general decomposition process including its tailoring options. The general process is adapted and applied for application in the industrial case study investigated as part of the university-industry collaboration research project.

Source Code Driven Decomposition of Object-Oriented Legacy Systems

Many enterprise applications have been developed over the last three decades therefore known as legacy systems. Usually they are monolith, inflexible, poorly documented and hard to maintain, however they are important to enterprises. The evolution of these systems depends on their decomposability. The purpose of this paper is to summarize existing knowledge, requirements and limitations for object-oriented legacy system decomposition based on systematic literature review. The investigation is performed as a part of the university-industry collaboration research project.

Decomposition of Enterprise Application: A Systematic Literature Review and Research Outlook

Abstract Enterprise applications are aimed at managing enterprise operational data and improving business efficiency. Many enterprise applications have been developed over the past three decades, therefore, known as legacy systems. Usually, they are monolith, inflexible, poorly documented and hard to maintain. The purpose of this paper is to describe best practices and limitations for enterprise application decomposition based on the results of the systematic literature review in order to introduce an approach for enterprise application decomposition. The paper focuses on decomposition of large-scale systems using clustering methods. The investigation is performed as part of the university-industry collaboration research project.

The Application of Class Structure to Classification Tasks

Abstract This article presents an approach in bioinformatics data analysis and exploration that improves classification accuracy by learning the inner structure of the data. The diseases studied in bioinformatics (diagnostic, prognostic etc. studies) often have the known or yet undiscovered subtypes that can be used while solving bioinformatics tasks providing more information and knowledge. This study deals with the problem above by studying inner class structures (probable disease subtypes) using a cluster analysis to find classification subclasses and applying it in classification tasks. The study also analyses possible cluster merges that would best describe classes. Evaluation is carried out using four classification methods that can be successfully used in bioinformatics: Naïve Bayes classifiers, C4.5, Random Forests and Support Vector Machines.

Feature Ranking by Classification Accuracy Estimation of Multiple Data Samples

Abstract This article considers the gene ranking algorithm for the microarray data. The rank vector is estimated by classifications of the random data samples. At each iteration, the ranks of genes participating in the successful classification become higher. Unlike other methods of feature selection, the proposed algorithm allows increasing the generality of the classification models by construction of the balanced training samples and taking into account the descriptiveness of the gene combinations by the subset estimation.