Cardiopulmonary exercise testing data processing and storage tools

ABSTRACT Conducting a clinical study is a complex and challenging task, so using technology to facilitate the process is necessary for the medical and scientific community. Data digitization is needed to store and process it. Digitization speeds up data visualization techniques so scientists can analyze it most effectively depending on their goals. In this project, we developed two digital tools related to the cardiopulmonary exercise test (CPET) data, one for digitizing PDF reports and the other for generating databases. The latter was a web-based database visualization dashboard. Users can select parameters of interest and check for differences among subgroups. Basic statistical tests are performed for each variable under analysis, and its results are presented in numerical and graphical formats. The initial statistical tests and the derived recommendation will guide the research team in deeper statistical analysis and robust analysis supporting more decisive conclusions. We ended up with the first version for both tools and validated it using patients' CPET data from the Dante Pazzanese Institute of Cardiology. We conducted a pilot study to verify if the tools served their purpose and observed that both programs worked as planned. The tools can be further tailored to be clinically or research-oriented. On the analysis of CPET's results, the conclusions for our example study were in line with what is presented in the bibliography for cardiorespiratory physiology.

Cardiovascular biomarkers in sickle cell disease− clinical and echocardiographic correlations in a cross-sectional study with 126 patients

INTRODUCTION: Cardiac biomarkers can be useful in understanding the systemic and heart manifestations of sickle cell disease (SCD). Biomarkers reflect various aspects of heart disease (remodeling, injury and myocardial strain), with discriminatory potential for non-cardiac complications. Patients and METHODS: SCD patients (SS/Sß0) in steady state, were studied, correlating clinical manifestations and echo parameters (myocardial work - MW and speckle tracking), pre-MSCD severity score (integrating clinical and echo data), and cardiac biomarkers (high-sensitivity troponins ­ hs-cTn I and T, NT-pro-BNP, ST2s, and galectin-3 - GAL3). Quantitative characteristics were analyzed by Spearman tests, and qualitative characteristics by Mann-Whitney test. Hemolytic Index (HI) was calculated through Principal Component Analysis. Generalized linear Poisson models were generated for hs-cTn, and γ-distribution models were employed for other markers, with final models selected through the Stepwise Backward method. RESULTS: We studied 126 patients (mean age 37.2 ± 11.6 years), 42.1% male, and 80.2% SS. 47% were on hydroxyurea treatment and 30.2% on a chronic transfusion. NT-pro-BNP was elevated in 44% (> 160 ng/mL in 37%), correlated with female gender (p < 0.001), severity score (p = 0.001), uric acid (p = 0.017), HI (p < 0.001), Global Work Index (GWI) (p = 0.003), left atrial (LA) stiffness (p = 0.003), and ventricular mass (VM) (p = 0.02). ST2s were elevated in 11% and correlated with male gender (p > 0.001), HI (p > 0.001), cardiac index (p = 0.015), and LA strain reservoir function (p = 0.034). GAL3 was elevated in 42.8%, correlated with E/e'ratio (p = 0.006), uric acid (p = 0.005), and absence of chronic pain (p = 0.046). hs-cTn correlated with age (c-TnI p = 0.004; c-TnT p > 0.001), HI (p > 0.001), diastolic dysfunction (p > 0.001), left VM (p < 0.001), increased GWI (p < 0.001), and reduced MW efficiency (p < 0.001). hs-cTn I also correlated with increased LA reservoir function (p < 0.001) with reduced conduit function (p < 0.001). hs-cTn T correlated with uric acid (p = 0.001), and in univariate analysis was also correlated with severity score. The values of both hs-cTn correlated with increased GWI (p < 0.001) and reduced MW efficiency (p < 0.001). DISCUSSION: The biomarkers demonstrated various clinical and pathophysiological aspects of SCD. NT-pro-BNP is a routine marker with correlations similar to literature, except for higher values in females, also observed in non-SCD population. ST2 and GAL3 had limited correlations with echo findings, likely due to their production in extracardiac tissues affected by inflammation/vaso-occlusion. Both were linked to the HI, and the decrease in GAL3 in chronic pain can be attributed to chronic opioid use causing reduced synthesis of it. The elevation of hs-cTn was expected due to the analytical characteristics of high-sensitivity assays, but low in terms of the extent of heart involvement. hs-cTnT was more associated with general severity, like in the general population, where it is associated with overall mortality, while hs-cTnI is more connected to heart disease. MW in SCD is optimized to the maximum with a very low Global Work Wasted, and hs-cTn elevation is associated with reduced MW efficiency, indicating mecano-energetic uncoupling and subtle systolic dysfunction. CONCLUSION: Our study demonstrates that cardiac biomarkers can be used for clinical and pathophysiological evaluation, with NT-pro-BNP confirming its role in clinical stratification. ST2s and GAL3 may reveal new pathophysiological pathways in hemolysis and the interaction of opioids and chronic pain. Troponins are promising as prospective tool and may unveil ischemic damage resulting from myocardial mecano-energetic dissociation.

Ferramentas para otimização de digitalização, armazenamento e visualização de dados do teste cardiopulmonar de exercício / Tools for optimizing the digitization, storage and visualization of cardiopulmonary exercise test data

INTRODUÇÃO: O teste cardiopulmonar de exercício (TCPE) é o padrão ouro para função cardiorrespiratória de pacientes. Por conta disso, os dados obtidos por meio deste são extremamente relevantes em pesquisas clínicas. No entanto, há situações em que, por conta de limitações dos equipamentos, a obtenção e armazenamento dos resultados destes exames se torna um processo complicado e dispendioso. OBJETIVO: Desenvolver ferramentas capazes de: 1) Digitalizar os dados nos formatos disponibilizados pelos equipamentos e montar um banco de dados; 2) Permitam uma análise, visualização e comparação otimizadas desses dados. MÉTODOS: Inicialmente, os laudos obtidos diretamente do equipamento são lidos por um programa, os valores dos parâmetros são armazenados durante os tempos mais relevantes do teste (limiares aeróbico e anaeróbico, VO2 máximo etc.) em um arquivo que servirá como banco de dados relacional. Esta base de dados é então utilizada para montar um Dashboard onde se pode selecionar a variável e o intervalo de tempo que se deseja comparar entre os grupos. Nele são mostrados histogramas e gráficos violinos para se comparar visualmente os grupos (figuras 1 e 2) e estatísticas dos dados coletados junto a: 1) Resultado de um teste de normalidade para o parâmetro escolhido; 2) Resultado (p-valor) de um teste de hipótese (paramétrico caso a distribuição seja considerada normal, ou não paramétrico caso contrário); 3) Os tamanhos mínimos que as amostras devem possuir para que os resultados dos testes possam ser considerados estatisticamente significativos. Por fim, é possível analisar, em uma segunda página desta ferramenta os valores individuais de cada paciente, sendo possível filtrar e visualizar apenas os valores dentro de um intervalo desejado. (Figura 4) RESULTADOS: As ferramentas desenvolvidas foram utilizadas na base de dados dos pacientes do Ambulatório de Medicina Cardiovascular Personalizada com os resultados de 102 exames. O Dashboard foi utilizado para realizar um estudo comparativo entre os pacientes homens e mulheres do ambulatório com resultados coerentes com a literatura. CONCLUSÕES: Com estas ferramentas foi possível, rapidamente, digitalizar e armazenar os dados montando um arquivo com as informações de todos pacientes. Ademais, também permitiu realizar um estudo exemplo comparando todos pacientes gerando imagens que auxiliam a visualizar os resultados e os respectivos p-valores para cada comparação.