RESUMO
Background: Interstitial lung disease (ILD) is a common manifestation of idiopathic inflammatory myopathies (IIM) and a substantial contributor to hospitalisation, increased morbidity, and mortality. In-vivo evidence of ongoing tissue remodelling in IIM-ILD is scarce. We aimed to evaluate fibroblast activation in lungs of IIM-patients and control individuals using 68Ga-labelled inhibitor of Fibroblast-Activation-Protein (FAPi) based positronic emission tomography and computed tomography imaging (PET/CT). Methods: In this prospective observational pilot study, consecutive patients with IIM and participants without rheumatic conditions or ILD serving as a control group were recruited at the Medical University of Vienna, Austria, and underwent FAPi PET/CT imaging. Standard-of-care procedures including clinical examination, assessment of severity of dyspnoea, high-resolution computed tomography (HR-CT), and pulmonary function testing (PFT) were performed on all patients with IIM at baseline and for patients with IIM-ILD at follow-up of 12 months. Baseline pulmonary FAPi-uptake was assessed by the maximum (SUVmax) and mean (SUVmean) standardized uptake values (SUV) over the whole lung (wl). SUV was corrected for blood pool background activity and target-to-background ratios (TBR) were calculated. We compared pulmonary FAPi-uptake between patients with IIM-ILD and those without ILD, as well as controls, and correlated baseline FAP-uptake with standard diagnostic tools such as HR-CT and PFT. For predictive implications, we investigated whether patients with IIM and progressive ILD exhibited higher baseline FAPi-uptake compared to those with stable ILD. Metrics are reported as mean with standard deviation (±SD). Findings: Between November 16, 2021 and October 10, 2022, a total of 32 patients were enrolled in the study. Three participants from the control group were excluded due to cardiopulmonary disease. In individuals with IIM-ILD (n = 14), wlTBRmax and wlTBRmean were significantly increased as compared with both non-ILD-IIM patients (n = 5) and the control group (n = 16): wlTBRmax: 2.06 ± 1.04 vs. 1.04 ± 0.22 (p = 0.019) and 1.08 ± 0.19 (p = 0.0012) and wlTBRmean: 0.45 ± 0.19 vs. 0.26 ± 0.06 (p = 0.025) and 0.27 ± 0.07 (p = 0.0024). Similar values were observed in wlTBRmax or wlTBRmean between non-ILD IIM patients and the control group. Patients with progressive ILD displayed significantly enhanced wlTBRmax and wlTBRmean values at baseline compared to patients with stable ILD: wlTBRmax: 1.30 ± 0.31 vs. 2.63 ± 1.04 (p = 0.0084) and wlTBRmean: 0.32 ± 0.08 vs. 0.55 ± 0.19 (p = 0.021). Strong correlations were found between FAPi-uptake and disease extent on HR-CT (wlTBRmax: R = 0.42, p = 0.07; wlTBRmean: R = 0.56, p = 0.013) and severity of respiratory symptoms determined by the New York Heart Association (NYHA) classification tool (wlTBRmax: R = 0.52, p = 0.022; wlTBRmean: R = 0.59, p = 0.0073). Further, pulmonary FAPi-uptake showed inverse correlation with forced vital capacity (FVC) (wlTBRmax: R = -0.56, p = 0.012; wlTBRmean: R = -0.64, p = 0.0033) and diffusing capacity of the lungs for carbon monoxide (DLCO) (wlTBRmax: R = -0.52, p = 0.028; wlTBRmean: R = -0.68, p = 0.0017). Interpretation: Our study demonstrates higher fibroblast activation in patients with IIM-ILD compared to non-ILD patients and controls. Intensity of pulmonary FAPi accumulation was associated with progression of ILD. Considering that this study was carried out on a small population, FAPi PET/CT may serve as a useful non-invasive tool for risk stratification of lung disease in IIM. Funding: The Austrian Research Fund.
RESUMO
Background: This study investigated the performance of ensemble learning holomic models for the detection of breast cancer, receptor status, proliferation rate, and molecular subtypes from [18F]FDG-PET/CT images with and without incorporating data pre-processing algorithms. Additionally, machine learning (ML) models were compared with conventional data analysis using standard uptake value lesion classification. Methods: A cohort of 170 patients with 173 breast cancer tumors (132 malignant, 38 benign) was examined with [18F]FDG-PET/CT. Breast tumors were segmented and radiomic features were extracted following the imaging biomarker standardization initiative (IBSI) guidelines combined with optimized feature extraction. Ensemble learning including five supervised ML algorithms was utilized in a 100-fold Monte Carlo (MC) cross-validation scheme. Data pre-processing methods were incorporated prior to machine learning, including outlier and borderline noisy sample detection, feature selection, and class imbalance correction. Feature importance in each model was assessed by calculating feature occurrence by the R-squared method across MC folds. Results: Cross validation demonstrated high performance of the cancer detection model (80% sensitivity, 78% specificity, 80% accuracy, 0.81 area under the curve (AUC)), and of the triple negative tumor identification model (85% sensitivity, 78% specificity, 82% accuracy, 0.82 AUC). The individual receptor status and luminal A/B subtype models yielded low performance (0.46-0.68 AUC). SUVmax model yielded 0.76 AUC in cancer detection and 0.70 AUC in predicting triple negative subtype. Conclusions: Predictive models based on [18F]FDG-PET/CT images in combination with advanced data pre-processing steps aid in breast cancer diagnosis and in ML-based prediction of the aggressive triple negative breast cancer subtype.