RESUMEN
OBJECTIVES: Although classification systems and scores for capillaroscopy interpretation have been published, there is a lack of homogenization for the procedure, especially in the way and place the images are taken, the counting of the capillaries and the measuring of their size. Our objective is to provide a deep learning-based software to obtain objective and exhaustive data for the whole nailfold without increasing the time or effort needed to do the examination, or requiring expensive equipment. METHODS: An automated software to count nailfold capillaries has been designed, through an exploratory image dataset of 2,713 images with 18,000 measurements of 3 different types. Subsequently, application rules have been created to detect the morphology of nailfold videocapillaroscopy images, through a training set of images. The software reliability has been evaluated with standard metrics used in the machine learning field for object detection tasks, comparing automatic and manual counting on the same NVC images. RESULTS: A mean average precision (mAP) of 0.473 is achieved for detecting and classifying capillaries and haemorrhages by their shape, and a mAP of 0.515 is achieved for detecting and classifying capillaries by their size. A precision of 83.84% and a recall of 92.44% in the identification of capillaries was estimated. CONCLUSIONS: Deep learning is a useful tool in nailfold videocapillaroscopy that allows to analyse objectively and homogeneously images taken with multiple devices. It should make the assessment of the capillary morphology in nailfold video capillaroscopy easier, quicker, more complete and accessible to everyone.
Asunto(s)
Angioscopía Microscópica , Uñas , Humanos , Angioscopía Microscópica/métodos , Reproducibilidad de los Resultados , Uñas/diagnóstico por imagen , Uñas/irrigación sanguínea , Capilares/diagnóstico por imagen , Programas InformáticosRESUMEN
INTRODUCTION: Our objective was to evaluate the pulmonary hypertension (PH) data for Spanish patients with systemic sclerosis (SSc), define the PH types and determine the associated factors. METHOD: Descriptive study of PH-related data from the multicentre RESCLE registry. Estimated systolic pulmonary artery pressure (esPAP), measured via echocardiogram was considered elevated if ≥ 35 mmHg. Left heart disease (LHD) and interstitial lung disease (ILD) were identified. When performed, data from right heart catheterisation (RHC) were collected. RESULTS: esPAP was elevated in 350 of 808 patients (43.3%). One hundred and forty-four patients (17.8%) were considered to have PH (88 via RHC and the rest due to elevated esPAP along with evidence of significant LHD or ILD): PAH 3.7%, secondary to ILD 8.3%, secondary to LHD 2.8% and unclassified 3%. Prevalence of elevated esPAP was greater in diffuse SSc (dSSc) than in limited scleroderma (lSSc) (50.5 vs. 42.2%, p 0.046). In the group with elevated esPAP, a lower prevalence of anti-centromere antibodies (41.9% vs. 52.3%, p 0.006) and a greater prevalence of anti-topoisomerase-1 antibodies (ATA) (25.1% vs. 18.6%, p 0.04) were observed compared to the group with normal esPAP. Patients with elevated esPAP had a lower rate of digital ulcers (50.6% vs. 60.2%, p 0.007) and esophageal involvement (83.6% vs. 88.7%, p 0.07) and higher rate of renal crisis (4.6% vs. 1.8%, p 0.066). CONCLUSIONS: Prevalence of PAH was lower than expected (3.7%). Probability of having elevated esPAP was higher among patients with dSSc and among those with ATA.
Asunto(s)
Hipertensión Pulmonar/epidemiología , Esclerodermia Sistémica/epidemiología , Adulto , Anciano , Anticuerpos Antinucleares , Centrómero/inmunología , Comorbilidad , Femenino , Humanos , Hipertensión Pulmonar/inmunología , Masculino , Persona de Mediana Edad , Prevalencia , Sistema de Registros , Esclerodermia Sistémica/inmunología , España/epidemiologíaRESUMEN
BACKGROUND AND OBJECTIVE: To describe the uveitis pattern in our geographic area. Recent demographic, environmental and scientific changes can determine uveitis pattern changes, which we aim to investigate. PATIENTS AND METHODS: A cross-sectional study between 1(st) January 2009 and 30(th) June 2010 was done. All uveitis patients visited at our institution were included. Anatomical (IUSG & SUN) and aetiological patterns (by "tailored" tests), age, sex, origin and laterality were analyzed. Results from first visits and controls were compared statistically, by chi-square and ANOVA tests. RESULTS: From 416 patients included, 150 (36%) were anterior uveitis, 128 (31%) posterior, 101 (24%) panuveitis and 37 (9%) intermediate uveitis (IU). Fifty-six percent (236) were bilateral, mean age was 46 years (6-87) and 58% were women. Fourteen percent (58) were from non-Spanish origin. By aetiology, 20% were unclassifiable, 31% infectious, 26% associated with systemic immune diseases and 23% were ocular specific syndromes. Among classified causes, herpes virus (11%), Toxoplasma (8%), Behçet (8%) and tuberculosis (6%) were the most common. IU were diagnosed more frequently at first time in foreign patients (P=.018); they affected younger patients (P=.008), were more chronic (P=.003) and unclassifiable (P=.024). CONCLUSION: In our area, 80% of uveitis can be correctly classified. IU have an increased frequency among foreigners, and they use to be unclassifiable and chronic. Tuberculosis (6%) and Birdshot chorio-retinopathy (4%) must be kept in mind as uveitis causes.
Asunto(s)
Uveítis , Adolescente , Adulto , Anciano , Anciano de 80 o más Años , Niño , Estudios Transversales , Femenino , Humanos , Masculino , Persona de Mediana Edad , Estudios Retrospectivos , España/epidemiología , Uveítis/diagnóstico , Uveítis/epidemiología , Uveítis/etiología , Adulto JovenRESUMEN
OBJECTIVE: To investigate the incidence of clinical and immunological characteristics of a large cohort of Spanish patients with scleroderma (SSc) and identifying factors associated with particular organ manifestations assessed by a nationwide cross-sectional analysis. METHODS: We classified SSc patients in 4 subsets using a modification of LeRoy and Medsger classification that included: "prescleroderma" (pre-SSc), limited cutaneous SSc (lcSSc), diffuse cutaneous SSc (dcSSc), and SSc sine scleroderma (ssSSc). Fourteen Spanish centers participated in patient recruitment. On January 2008, the database included 916 consecutive Spanish SSc patients, 801 women (87.4%) and 115 men (12.6%), all of whom fulfilled the classification criteria proposed by LeRoy and Medsger. Epidemiological, clinical, and laboratory data were collected according to a standard protocol. Mean age at diagnosis was 51.2 ± 15.1 years and mean age at disease onset was 44.9.0 ± 15.8 years. lcSSc was the most frequent subset (61.8%) followed by dcSSc (26.5%), ssSSc (7.5%), and preSSc (4%) subsets. Gender ratios were as follows: dcSSc subset, 200 women and 43 men (4.7:1); lcSSc subset, 503 women and 63 men (ratio 7.9:1), and ssSSc subset, 62 women and 7 men (ratio 8.9:1). Digital ulcers, interstitial lung disease (ILD), musculoeskeletal and esophageal involvement, and scleroderma renal crisis were more frequent in dcSSc than lcSSc and ssSSc subsets. The incidence of pulmonary arterial hypertension assessed by echocardiography was similar in all subsets but mean estimated systolic pulmonary arterial pressure was higher in ssSSc than in lcSSc subset (47.3 ± 23.9 mm Hg vs 39.6 ± 19.2 mm Hg; P < 0.03). Cardiac involvement was identified more frequently in ssSSc than in dcSSc and lcSSc subsets (49.3% vs 32.5% and 31.1%, respectively; P = 0.015 and P = 0.004 for both comparisons). Acro-osteolysis (8.2% vs 2.4%, P = 0.049), calcinosis (19.8% vs 7.2%, P < 0.05), and sicca syndrome (37.5% vs 14.5%, P < 0.0001) were more frequent in lcSSc than in ssSSc subsets. The frequency of clinical manifestations related to the presence of anticentromere antibodies or antitopoisomerase I antibodies was very similar to that identified in patients categorized to lcSSc and dcSSc, respectively. However, in multivariate studies, the ranking of the variables according to their overall explanatory effect on the model showed that the contributory effect of the antibody status was not greater than that of the clinical categorization into lcSSc and dcSSc for the majority of disease manifestations, but, in important manifestations, as ILD, absence of anticentromere antibodies was an independent predictor factor. CONCLUSIONS: The classification of SSc into dcSSc, lcSSc, and ssSSc subsets is the one that most closely reflects the natural history of the disease, as they presented clear clinical differences. The immunological profile helps to define important visceral alteration as ILD. Finally, to improve early diagnosis of SSc, patients with preSSc should be considered both to trace the true evolution of the disease and to define which patients could benefit from therapeutic measures able to prevent the appearance of visceral involvements.