RESUMEN
Radiomics and artificial intelligence have the potential to become a valuable tool in clinical applications. Frequently, radiomic analyses through machine learning methods present issues caused by high dimensionality and multicollinearity, and redundant radiomic features are usually removed based on correlation analysis. We assessed the effect of preprocessing-in terms of voxel size resampling, discretization, and filtering-on correlation-based dimensionality reduction in radiomic features from cardiac T1 and T2 maps of patients with hypertrophic cardiomyopathy. For different combinations of preprocessing parameters, we performed a dimensionality reduction of radiomic features based on either Pearson's or Spearman's correlation coefficient, followed by the computation of the stability index. With varying resampling voxel size and discretization bin width, for both T1 and T2 maps, Pearson's and Spearman's dimensionality reduction produced a slightly different percentage of remaining radiomic features, with a relatively high stability index. For different filters, the remaining features' stability was instead relatively low. Overall, the percentage of eliminated radiomic features through correlation-based dimensionality reduction was more dependent on resampling voxel size and discretization bin width for textural features than for shape or first-order features. Notably, correlation-based dimensionality reduction was less sensitive to preprocessing when considering radiomic features from T2 compared with T1 maps.
RESUMEN
INTRODUÇÃO: Síncope inexplicada é considerada um fator de risco de morte súbita na cardiomiopatia hipertrófica (CMH). Em sua patogênese estão envolvidos mecanismos diversos, incluindo a dificuldade de adaptação da resistência vascular sistêmica ao exercício e ao estresse ortostático, que pode ser influenciada por uma disfunção do sistema nervoso autônomo. Os objetivos deste estudo foram comparar a função nervosa autonômica em portadores de CMH com e sem síncope, bem como avaliar o valor diagnóstico do teste de inclinação (TI) na investigação de síncope nessa população. MÉTODOS: Foram incluídos 37 pacientes, 16 com síncope inexplicada à avaliação rotineira e 21 sem síncope. A função nervosa autonômica foi medida pela sensibilidade do barorreflexo (BR) espontâneo e do induzido por fenilefrina e pela variabilidade da freqüência cardíaca (VFC). As variáveis da VFC consideradas no domínio do tempo foram: desvio-padrão de todos os intervalos RR normais (SDNN); raiz quadrada da média do quadrado das diferenças entre intervalos RR normais adjacentes (RMSSD); e percentagem de intervalos RR adjacentes com diferença superior a 50 ms (pNN50), durante o eletrocardiograma de 24 horas. No domínio da freqüência, foram considerados os componentes de alta, baixa e muito baixa freqüência e a densidade total do espectro, tanto em valores absolutos como em unidades normalizadas, em repouso e aos 60 graus de inclinação. As medidas da pressão arterial sistólica e diastólica, batimento a batimento, e as medidas do índice sistólico, do índice cardíaco e da resistência vascular sistêmica, obtidas pela cardiografia por impedância, foram comparadas, entre os grupos, a 0, 30 e 60 graus de inclinação. O TI consistiu na exposição dos pacientes a 60º de inclinação por 40 minutos, ou até uma resposta positiva. RESULTADOS: A sensibilidade do BR, tanto espontâneo (16,46±12,99 vs 18,31±9,88 ms/mmHg, p=0,464) como induzido por fenilefrina (18,33±9,31 vs 15,83±15,48 ms/mmHg, p=0,521)...
BACKGROUND: Unexplained syncope is considered a risk factor for sudden death in hypertrophic cardiomyopathy (HCM). Several mechanisms are involved in its pathogenesis, including the difficulty in adaptation of the systemic vascular resistance to exertion and to orthostatic stress, which may be influenced by a dysfunction of the autonomic nervous system. The purposes of this study were to compare the autonomic nervous function in patients with HCM with and without syncope and to assess diagnostic value of the head-up tilting test (HUT) in this population. METHODS: Thirty seven patients were included: 16 with unexplained syncope at routine evaluation and 21 without syncope. The autonomic nervous function was assessed by spontaneous and phenylephrine-induced baroreflex sensitivity (BRS) and by heart rate variability (HRV). Considered HRV variables in time domain were: standard deviation of normal RR intervals (SDNN), root mean square of successive differences (RMSSD), and percentage of adjacent normal RR intervals which differ by at least 50 ms (pNN50), during 24 hours electrocardiogram recording. In frequency domain, high, low and very low frequency bands and the spectrum total power density were considered, both in absolute values and in normalized units, at rest and at 60-degree tilting. Measures of systolic and diastolic blood pressure, beat to beat, and measures of stroke index, cardiac index, and systemic vascular resistance, obtained by impedance cardiography, were compared between the groups, at 0-, 30- and 60-degree tilting. The HUT consisted in exposure to 60º for 40 minutes, or until a positive response. RESULTS: Spontaneous BRS measures were similar between the syncope and non-syncope groups (16.46±12.99 vs 18.31±9.88 ms/mmHg, p=0.464), as well as phenylephrine induced BRS (18.33±9.31 vs 15.83±15.48 ms/mmHg, p=0.521). No differences were found between SDNN values (137.69±36.62 vs 145.95±38.07 ms, p=0.389). The syncope group presented lower values...