Nailfold videocapillaroscopy (NVC) is the gold standard for diagnosing systemic sclerosis (SSc) and differentiating primary from secondary Raynaud's phenomenon. The CAPI-Score algorithm, designed for simplicity, classifies capillaroscopy scleroderma patterns (CSPs) using a limited number of capillary variables. This study aims to develop a more advanced machine learning (ML) model to improve CSP identification by integrating a broader range of statistical variables while minimising examiner-related bias. A total of 1,780 capillaroscopies were randomly and blindly analysed by 3-4 trained observers. Consensus was defined as agreement among all but one observer (partial consensus) or unanimous agreement (full consensus). Capillaroscopies with at least partial consensus were used to train ML-based classification models using CatBoost software, incorporating 24 capillary architecture-related variables extracted via automated NVC analysis. Validation sets were employed to assess model performance. Of the 1,490 capillaroscopies classified with consensus, 515 achieved full consensus. The model, evaluated on partial and full consensus datasets, achieved 0.912, 0.812, and 0.746 accuracy for distinguishing SSc from non-SSc, among SSc patterns, and between normal and non-specific patterns, respectively. When evaluated on full consensus only, accuracy improved to 0.910, 0.925, and 0.933. CAPI-Detect outperformed CAPI-Score, revealing novel capillary variables critical to ML-based classification. CAPI-Detect, an ML-based model, provides an unbiased, quantitative analysis of capillary structure, shape, size, and density, significantly improving capillaroscopic pattern identification.
Systemic sclerosis (SSc) can virtually affect any organ system (such as lungs, kidneys, gastrointestinal tract, and heart). However, it is the pulmonary manifestations that account for the majority of deaths, especially interstitial lung disease (ILD) and pulmonary arterial hypertension (PAH).
Objectives
Our aim was to assess the differences between severe and mild-to-moderate ILD in SSc.
Methods
A descriptive study was performed, using the available data from the Spanish Scleroderma Study Group (RESCLE). ILD was deemed as serious when forced vital capacity (FVC) was <50%. Patients were classified attending the modified classification criteria proposed by LeRoy and Megdsger.
Results
Fourteen referral centers for SSc participated in the registry. By April 2014, 1374 patients with SSc had been enrolled, 541 of whom (39.4%) had ILD, which was severe in 72 of them (13.2%). There were no significant differences as far as sex and age at onset is concerned. Patients with diffuse SSc presented with severe ILD more frequently than those with limited SSc (57% vs. 35%, p=0.002), as well as those who had tested positive for ATA (51% vs. 33%, p=0.005). Aditionally, prevalence of FVC<50% was higher in patients with myopathy (32% vs. 15%, p=0.002). Mean FVC was 40.2±6.4 in the severe ILD group, whilst it was 80.3±18.9 in the mild-to-moderate one (p<0.001), and mean DLco was 36.7±15.2 and 62.9±34.5, respectively (p<0.001). Likewise, DLco<70% was also more frequent among patients with severe ILD (100% vs. 69%, p<0.001), as well as mean DLco/VA (56.2±24.2 vs. 74.2±42.0, p=0.002). PAPs was equally higher when FVC<50% (42.2±18.2 vs. 35.1±13.4, p=0.034), and so was the frequency of PAPs>40mmHg (66% vs. 29%, p<0.001) and PAH by right heart catheterism (19% vs. 11%, p=0.050). Finally, by means of a multiple logistic regression, both ATA positivity [OR 0.17 (0.05–0.58), p=0.005] and low DLco [0.93 (0.91–0.95), p=0.000] were found to be related with FVC<50%
Conclusions
Patients with ACA positivity and with a limited variant of SSc seem to be at lower risk of severe interstitial lung involvement. Furthermore, the presence of myopathy may contribute to explain the decrease of FVC in SSc patients.
References
Steen VD, Medsger TA. Changes in causes of death in systemic sclerosis, 1972–2002. Ann Rheum Dis 2007; 66: 940–4. LeRoy EC, Medsger TA, Jr. Criteria for the classification of early systemic sclerosis. J Rheumatol 2001; 28: 1573–6.
Systemic lupus erythematosus (SLE) is a genetically complex rheumatic disease with heterogeneous clinical manifestations. Recent studies have suggested the existence of a genetic basis for the diverse SLE clinical phenotypes. Also, there is increasing evidence indicating that a substantial part of the genetic variation associated with complex diseases is explained by small-effect genes from common genetic pathways.
Objectives
The objective of the present study was to identify new genetic variation associated with SLE phenotypes using a genome-wide association study at the pathway level.
Methods
A total of 598,258 SNPs were genotyped in a discovery cohort of n=482 SLE patients of southern European ancestry using the Illumina platform Quad610. After quality control analysis, including ancestry estimation using principal-component analysis, genome-wide pathway analysis was performed. A total of 14 clinically relevant SLE phenotypes were tested for association with n>700 reference genetic pathways. Significantly associated pathways (corrected P-value <0.05) were subsequently tested for validation in an independent cohort of n=425 SLE patients from the same ancestry. Both discovery and validation cohort patients were Caucasian European and from Spanish origin, and were recruited by rheumatology departments from n=15 Spanish university hospitals. The validated genetic pathways were functionally characterized using in silico analysis on cell types of relevance in SLE pathogenesis.
Results
In the discovery stage, two genetic pathways were significantly associated with the presence of oral ulcers and antinuclear antibodies in SLE (PFDR<0.05). In the replication stage, we validated the association between oral ulcers and vascular endothelial growth factor (VEGF) genetic pathway (P=1.3e-2). Analyzing the transcriptional effect of the topical immunotherapies used for the treatment of oral ulcers in SLE, we found a significant differential expression of VEGF pathway genes (P<0.05).
Conclusions
In this work we have performed the first genome-wide association study for clinically relevant SLE phenotype using a pathway-based approach. With this new approach, we have identified and validated the association of VEGF genetic pathway with oral ulcers in SLE. These findings represent an important step towards the characterization of the genetic basis of phenotype heterogeneity in SLE.
Abstract To assess severity markers and outcomes of patients with systemic sclerosis (SSc) with or without pulmonary arterial hypertension (PAH-SSc/non-PAH-SSc), and the impact of interstitial lung disease (ILD) on PAH-SSc. Non-PAH-SSc patients from the Spanish SSc registry and PAH-SSc patients from the Spanish PAH registry were included. A total of 364 PAH-SSc and 1589 non-PAH-SSc patients were included. PAH-SSc patients had worse NYHA-functional class (NYHA-FC), worse forced vital capacity (FVC) (81.2 ± 20.6% vs 93.6 ± 20.6%, P < 0.001), worse tricuspid annular plane systolic excursion (TAPSE) (17.4 ± 5.2 mm vs 19.9 ± 6.7 mm, P < 0.001), higher incidence of pericardial effusion (30% vs 5.2%, P < 0.001) and similar prevalence of ILD (41.8% vs. 44.9%). In individuals with PAH-SSc, ILD was associated with worse hemodynamics and pulmonary function tests (PFT). Up-front combination therapy was used in 59.8% and 61.7% of patients with and without ILD, respectively. Five-year transplant-free survival rate was 41.1% in PAH-SSc patients and 93.9% in non-PAH-SSc patients (P < 0.001). Global survival of PAH-SSc patients was not affected by ILD regardless its severity. The multivariate survival analysis in PAH-SSc patients confirmed age at diagnosis, worse NYHA-FC, increased PVR, reduced DLCO, and lower management with up-front combination therapy as major risk factors. In conclusion, in PAH-SSc cohort risk of death was greatly increased by clinical, PFT, and hemodynamic factors, whereas it was decreased by up-front combination therapy. Concomitant ILD worsened hemodynamics and PFT in PAH-SSc but not survival regardless of FVC impairment.