Using data mining techniques to support breast cancer diagnosis

More than ever, there is an increase of the number of decision support methods and computer aided diagnostic systems applied to various areas of medicine. In breast cancer research, many works have been done in order to reduce false-positives when used as a double reading method. In this study, we aimed to present a set of data mining techniques that were applied to approach a decision support system in the area of breast cancer diagnosis. This method is geared to assist clinical practice in identifying mammographic findings such as microcalcifications, masses and even normal tissues, in order to avoid misdiagnosis. In this work a reliable database was used, with 410 images from about 115 patients, containing previous reviews performed by radiologists as microcalcifications, masses and also normal tissue findings. Throughout this work, two feature extraction techniques were used: the gray level co-occurrence matrix and the gray level run length matrix. For classification purposes, we considered various scenarios according to different distinct patterns of injuries and several classifiers in order to distinguish the best performance in each case described. The many classifiers used were Naïve Bayes, Support Vector Machines, k-nearest Neighbors and Decision Trees (J48 and Random Forests). The results in distinguishing mammographic findings revealed great percentages of PPV and very good accuracy values. Furthermore, it also presented other related results of classification of breast density and BI-RADS® scale. The best predictive method found for all tested groups was the Random Forest classifier, and the best performance has been achieved through the distinction of microcalcifications. The conclusions based on the several tested scenarios represent a new perspective in breast cancer diagnosis using data mining techniques.

A novel antibody-like material for breast cancer antigen CA15-3, used to track breast cancer by potentiometric transduction

This work presents the development of a low cost sensor device for the diagnosis of breast cancer in point-of-care, made with new synthetic biomimetic materials inside plasticized poly(vinyl chloride), PVC, membranes, for subsequent potentiometric detection. This concept was applied to target a conventional biomarker in breast cancer: Breast Cancer Antigen (CA15-3). The new biomimetic material was obtained by molecularly-imprinted technology. In this, a plastic antibody was obtained by polymerizing around the biomarker that acted as an obstacle to the growth of the polymeric matrix. The imprinted polymer was specifically synthetized by electropolymerization on an FTO conductive glass, by using cyclic voltammetry, including 40 cycles within -0.2 and 1.0 V. The reaction used for the polymerization included monomer (pyrrol, 5.0×10-3 mol/L) and protein (CA15-3, 100U/mL), all prepared in phosphate buffer saline (PBS), with a pH of 7.2 and 1% of ethylene glycol. The biomarker was removed from the imprinted sites by proteolytic action of proteinase K. The biomimetic material was employed in the construction of potentiometric sensors and tested with regard to its affinity and selectivity for binding CA15-3, by checking the analytical performance of the obtained electrodes. For this purpose, the biomimetic material was dispersed in plasticized PVC membranes, including or not a lipophilic ionic additive, and applied on a solid conductive support of graphite. The analytical behaviour was evaluated in buffer and in synthetic serum, with regard to linear range, limit of detection, repeatability, and reproducibility. This antibody-like material was tested in synthetic serum, and good results were obtained. The best devices were able to detect 5 times less CA15-3 than that required in clinical use. Selectivity assays were also performed, showing that the various serum components did not interfere with this biomarker. Overall, the potentiometric-based methods showed several advantages compared to other methods reported in the literature. The analytical process was simple, providing fast responses for a reduced amount of analyte, with low cost and feasible miniaturization. It also allowed the detection of a wide range of concentrations, diminishing the required efforts in previous sample pre-treating stages.

Pulmonary thromboembolism in AIDS patient with chronic venous insufficiency, pulmonary tuberculosis and breast cancer: a case report and pathophysiology review

Recent literature reports thrombotic episodes occurring in patients with HIV infection associated with other abnormalities including neoplasms and infections predisposing to a hypercoagulable state. We report a 47-year-old woman who developed pulmonary thromboembolism in association with HIV infection, pulmonary tuberculosis and breast cancer. She was treated with rifampin, isoniazid, pyrazinamide; heparin, phenprocoumon, zidovudine, lamivudine and efavirenz. Acid fast bacilli were visualized in a sputum smear and three months after, Mycobacterium tuberculosis was isolated from lymph node biopsy during a episode of immune reconstitution. The isolated mycobacteria showed sensitivity to all first-line drugs. HIV infection, breast cancer and pulmonary tuberculosis have several mechanisms that induce hypercoagulable state and can lead to thromboembolic complications. Pulmonary thromboembolism in this patient was a diagnostic challenge because of all the other severe diseases that she experienced at the same time.

Effects of Environmental Organochlorine Pesticides on Human Breast Cancer: Putative Involvement on Invasive Cell Ability

Human exposure to persistent organic pollutants (POPs) is a certainty, even to long banned pesticides like o,p′-dichlorodiphenyltrichloroethane (o,p′-DDT), and its metabolites p,p′-dichlorodiphenyldichloroethylene (p,p′-DDE), and p,p′-dichlorodiphenyldichloroethane (p,p′-DDD). POPs are known to be particularly toxic and have been associated with endocrine-disrupting effects in several mammals, including humans even at very low doses. As environmental estrogens, they could play a critical role in carcinogenesis, such as in breast cancer. With the purpose of evaluating their effect on breast cancer biology, o,p′-DDT, p,p′-DDE, and p,p′-DDD (50–1000 nM) were tested on two human breast adenocarcinoma cell lines: MCF-7 expressing estrogen receptor (ER) α and MDA-MB-231 negative for ERα, regarding cell proliferation and viability in addition to their invasive potential. Cell proliferation and viability were not equally affected by these compounds. In MCF-7 cells, the compounds were able to decrease cell proliferation and viability. On the other hand, no evident response was observed in treated MDA-MB-231 cells. Concerning the invasive potential, the less invasive cell line, MCF-7, had its invasion potential significantly induced, while the more invasive cell line MDA-MB-231, had its invasion potential dramatically reduced in the presence of the tested compounds. Altogether, the results showed that these compounds were able to modulate several cancer-related processes, namely in breast cancer cell lines, and underline the relevance of POP exposure to the risk of cancer development and progression, unraveling distinct pathways of action of these compounds on tumor cell biology.

Longitudinal analysis of tumor marker CEA of breast cancer patients from Braga's hospital

Allied to an epidemiological study of population of the Senology Unit of Braga’s Hospital that have been diagnosed with malignant breast cancer, we describe the progression in time of repeated measurements of tumor marker Carcinoembryonic antigen (CEA). Our main purpose is to describe the progression of this tumor marker as a function of possible risk factors and, hence, to understand how these risk factors influences that progression. The response variable, values of CEA, was analyzed making use of longitudinal models, testing for different correlation structures. The same covariates used in a previous survival analysis were considered in the longitudinal model. The reference time used was time from diagnose until death from breast cancer. For diagnostic of the models fitted we have used empirical and theoretical variograms. To evaluate the fixed term of the longitudinal model we have tested for a changing point on the effect of time on the tumor marker progression. A longitudinal model was also fitted only to the subset of patients that died from breast cancer, using the reference time as time from date of death until blood test.

High-Content Analysis of Breast Cancer Using Single-Cell Deep Transfer Learning

High-content analysis has revolutionized cancer drug discovery by identifying substances that alter the phenotype of a cell, which prevents tumor growth and metastasis. The high-resolution biofluorescence images from assays allow precise quantitative measures enabling the distinction of small molecules of a host cell from a tumor. In this work, we are particularly interested in the application of deep neural networks (DNNs), a cutting-edge machine learning method, to the classification of compounds in chemical mechanisms of action (MOAs). Compound classification has been performed using image-based profiling methods sometimes combined with feature reduction methods such as principal component analysis or factor analysis. In this article, we map the input features of each cell to a particular MOA class without using any treatment-level profiles or feature reduction methods. To the best of our knowledge, this is the first application of DNN in this domain, leveraging single-cell information. Furthermore, we use deep transfer learning (DTL) to alleviate the intensive and computational demanding effort of searching the huge parameter's space of a DNN. Results show that using this approach, we obtain a 30% speedup and a 2% accuracy improvement.

Breast Cancer Presents With a Paraneoplastic Neurologic Syndrome

BACKGROUND: Paraneoplastic neurologic syndromes (PNS) pose quite an uncommon neurological complication, affecting less than 1% of patients with breast cancer. Nearly one third of these patients lack detectable onconeural antibodies (ONAs), and improvement in neurologic deficits with concomitant cancer treatments is achieved in less than 30% of cases. CASE PRESENTATION: A 42-year-old, premenopausal woman presented with facial paralysis on the central left side accompanied by a left tongue deviation, an upward vertical nystagmus, moderate spastic paraparesis, dystonic posturing of the left foot, lower limb hyperreflexia and bilateral extensor plantar reflex. After ruling out all other potential neurologic causes, PNS was suspected but no ONAs were found. A PET-CT scan detected increased metabolism in the right breast, as well as an ipsilateral thoracic interpectoral adenopathy. Core biopsy confirmed the presence of an infiltrating duct carcinoma. After breast surgery, the neurologic symptoms disappeared. One week later, the patient was readmitted to the hospital with a bilateral fatigable eyelid ptosis, and two weeks later, there was a noticeable improvement in eyelid ptosis, accompanied by a rapid and progressive development of lower spastic paraparesis. She started adjuvant treatment with chemotherapy with marked clinical and neurological improvement, and by the end of radiotherapy, there were no signs of neurologic impairment. CONCLUSION: This case study highlights the importance of a high level of vigilance for the detection of PNS, even when ONAs are not detected, as the rapid identification and treatment of the underlying tumor offers the best chance for a full recovery.

A comparison between digital breast tomosynthesis and mammography: dosimetric and image quality characterization of clinical system

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica

Pseudoneoplastic lesion of the breast caused by Schistosoma mansoni

A case of a pseudoneoplastic lesion of the breast clinically and sonographically suggestive of a fibroadenoma is reported. Excisional biopsy revealed the nodule was an inflammatory process consequent to infection by Schistosoma mansoni.

Optimization of breast tomosynthesis image reconstruction using parallel computing

Breast cancer is the most common cancer among women, being a major public health problem. Worldwide, X-ray mammography is the current gold-standard for medical imaging of breast cancer. However, it has associated some well-known limitations. The false-negative rates, up to 66% in symptomatic women, and the false-positive rates, up to 60%, are a continued source of concern and debate. These drawbacks prompt the development of other imaging techniques for breast cancer detection, in which Digital Breast Tomosynthesis (DBT) is included. DBT is a 3D radiographic technique that reduces the obscuring effect of tissue overlap and appears to address both issues of false-negative and false-positive rates. The 3D images in DBT are only achieved through image reconstruction methods. These methods play an important role in a clinical setting since there is a need to implement a reconstruction process that is both accurate and fast. This dissertation deals with the optimization of iterative algorithms, with parallel computing through an implementation on Graphics Processing Units (GPUs) to make the 3D reconstruction faster using Compute Unified Device Architecture (CUDA). Iterative algorithms have shown to produce the highest quality DBT images, but since they are computationally intensive, their clinical use is currently rejected. These algorithms have the potential to reduce patient dose in DBT scans. A method of integrating CUDA in Interactive Data Language (IDL) is proposed in order to accelerate the DBT image reconstructions. This method has never been attempted before for DBT. In this work the system matrix calculation, the most computationally expensive part of iterative algorithms, is accelerated. A speedup of 1.6 is achieved proving the fact that GPUs can accelerate the IDL implementation.

Transimpedance amplifier for early detection of breast cancer

Breast cancer is the most common type of cancer worldwide. The effectiveness of its treatment depends on early stage detection, as well as on the accuracy of its diagnosis. Recently, diagnosis techniques have been submitted to relevant breakthroughs with the upcoming of Magnetic Resonance Imaging, Ultrasound Sonograms and Positron Emission Tomography (PET) scans, among others. The work presented here is focused on studying the application of a PET system to a Positron Emission Mammography (PEM) system. A PET/PEM system works under the principle that a scintillating crystal will detect a gamma-ray pulse, originated at the cancerous cells, converting it into a correspondent visible light pulse. The latter must then be converted into an electrical current pulse by means of a Photo- -Sensitive Device (PSD). After the PSD there must be a Transimpedance Amplifier (TIA) in order to convert the current pulse into a suitable output voltage, in a time period lower than 40 ns. In this Thesis, the PSD considered is a Silicon Photo-Multiplier (SiPM). The usage of this recently developed type of PSD is impracticable with the conventional TIA topologies, as it will be proven. Therefore, the usage of the Regulated Common-Gate (RCG) topology will be studied in the design of the amplifier. There will be also presented two RCG variations, comprising a noise response improvement and differential operation of the circuit. The mentioned topology will also be tested in a Radio-Frequency front-end, showing the versatility of the RCG. A study comprising a low-voltage self-biasing feedback TIA will also be shown. The proposed circuits will be simulated with standard CMOS technology (UMC 130 nm), using a 1.2 V power supply. A power consumption of 0.34 mW with a signal-to-noise ratio of 43 dB was achieved.

Assessment of Notch1 ligands production - key protein targets in breast cancer

Cell-to-cell communication is required for many biological processes in development and adult life. One of the most common systems utilized by a wide range of eukaryotes is the Notch signalling pathway. Four Notch receptors and five ligands have been identified in mammals that interact via their extracellular domains leading to transcription activation. Studies have shown that the Notch ligands expression is undetectable in normal breast tissues, but moderate to high expression has been detected in breast cancer. Thus, any of the Notch1 ligands can be studied as possible therapeutic targets for breast cancer. To study Notch pathway proteins there is the need to obtain stable protein solutions. E. coli is the host of excellence for recombinant proteins for the ease of use, fast growth and high cell densities. However, the expression of mammalian proteins in such systems may overwhelm the bacterial cellular machinery, which does not possess the ability for post-translational modifications, or dedicated compartments for protein synthesis. Mammalian cells are therefore preferred, despite their technical and financial increased demands. We aim to determine the best expression and purification conditions for the different ligand protein constructs, to develop specific function-blocking antibodies using the Phage Display technology. Moreover, we propose to crystallize the Notch1 ligands alone and in complex with the phage display selected antibodies, unveiling molecular details. hJag2DE3 and hDll1DE6 proteins were purified from refolded inclusion bodies or mammalian cell culture supernatants, respectively, and purity was confirmed by SDS-PAGE (>95%). Protein produced in mammalian cells showed to be more stable, apparently with the physiological disulfide pattern, contrary to what was observed in the refolded protein. Several nano-scale crystallization experiments were set up in 96-well plates, but no positive result was obtained. We will continue to pursue for the best expression for the Notch ligand constructs in both expression systems.

Microwave imaging of the axilla to aid breast cancer diagnosis

Breast cancer is the most common type of cancer among women all over the world. An important issue that is not commonly addressed in breast cancer imaging literature is the importance of imaging the underarm region—where up to 80% of breast cancer cells can metastasise to. The first axillary lymph nodes to receive drainage from the primary tumour in the breast are called Sentinel Node. If cancer cells are found in the Sentinel Node, there is an increased risk of metastatic breast cancer which makes this evaluation crucial to decide what follow-up exams and therapy to follow. However, non-invasive detection of cancer cells in the lymph nodes is often inconclusive, leading to the surgical removal of too many nodes which causes adverse side-effects for patients. Microwave Imaging is one of the most promising non-invasive imaging modalities for breast cancer early screening and monitoring. This novel study tests the feasibility of imaging the axilla region by means of the simulation of an Ultra-Wideband Microwave Imaging system. Simulations of such system are completed in several 2D underarm models that mimic the axilla. Initial imaging results are obtained by means of processing the simulated backscattered signals by eliminating artefacts caused by the skin and beamforming the processed signals in order to time-align all the signals recorded at each antenna. In this dissertation several image formation algorithms are implemented and compared by visual inspection of the resulting images and through a range of performance metrics, such as Signal-to-Clutter Ratio and FullWidth Half Maximum calculations. The results in this study showed that Microwave Imaging is a promising technique that might allow to identify the presence and location of metastasised cancer cells in axillary lymph nodes, enabling the non-invasive evaluation of breast cancer staging.

Phage display as a tool for development of novel therapeutics for breast cancer

Phage display technology is a powerful platform for the generation of highly specific human monoclonal antibodies (Abs) with potential use in clinical applications. Moreover, this technique has also proven to be a reliable approach in identifying and validating new cancer-related targets. For scientific or medical applications, different types of Ab libraries can be constructed. The use of Fab Immune libraries allows the production of high quality and affinity antigen-specific Abs. In this work, two immune human phage display IgG Fab libraries were generated from the Ab repertoire of 16 breast cancer patients, in order to obtain a tool for the development of new therapeutic Abs for breast cancer, a condition that has great impact worldwide. The generated libraries are estimated to contain more than 108 independent clones and a diversity over 90%. Libraries validation was pursued by selection against BSA, a foreign and highly immunogenic protein, and HER2, a well established cancer target. Preliminary results suggested that phage pools with affinity for these antigens were selected and enriched. Individual clones were isolated, however, it was not possible to obtain enough data to further characterize them. Selection against the DLL1 protein was also performed, once it is a known ligand of the Notch pathway, whose deregulation is associated to breast cancer, making it an interesting target for the generation of function-blocking Abs. Selection resulted in the isolation of a clone with low affinity and Fab expression levels. The validation process was not completed and further effort will have to be put in this task in the future. Although immune libraries concept implies limited applicability, the library reported here has a wide range of use possibilities, since it was not restrained to a single antigen but instead thought to be used against any breast cancer associated target, thus being a valuable tool.