Chiara Romei

Background: Coronavirus Disease 19 (COVID-19)-associated pneumonia is characterized by a broad spectrum of severity. Endothelial injury is a prominent component of the disease, but the relative contribution of vessels of different sizes to this condition has not been extensively investigated. Quantitative computed tomography (CT) represents a potential tool to evaluate pulmonary vascular involvement. Objectives: To investigate the relative contribution of vessels of different sizes in COVID-19-associated pneumonia of different severity using the Computer-Aided Lung Informatics for Pathology Evaluation and Rating (CALIPER) software. Methods: We conducted a single-center retrospective study, including patients admitted to the Pisa University Hospital between March and April 2020 with COVID-19-associated pneumonia. We excluded patients with severe motion artifacts, inadequate CALIPER segmentation, or inconclusive etiologic confirmation. All CT scans were analyzed using CALIPER, which automatically segmented lung parenchyma and vascular-related structures (VRSs). VRS volumes were stratified by vessel cross-sectional area (20 mm²). Results: We analyzed 258 patients. We observed that an increase in the percentages of VRSs with cross-sectional areas up to 20 mm² was associated with the need for lower intensity of care and with better gas exchange. The opposite trend was observed for the percentages of >20 mm². Conclusions: Our data provide novel information on the involvement of blood vessels of different sizes in the gas exchange impairment of COVID-19 patients. CALIPER represents a promising application for the study of vascular structures of the lung.

Objectives: Lymphoid interstitial pneumonia (LIP) is a rare form of Interstitial Lung Disease (ILD), often associated with Sjögren Disease (SjD). However, the clinical-serologic characteristics of SjD-LIP remain poorly characterized. Our objective was to describe the clinical course and outcome of SjD-associated LIP and to compare this subgroup with other SjD-ILD patterns. Methods: SjD patients with High-resolution Computed Tomography (HRCT)-confirmed ILD followed in Pisa Rheumatology Unit (January 2019–November 2024) were retrospectively enrolled. ILD patterns were classified through multidisciplinary discussion. Clinical and laboratory data were collected according to ESSDAI definitions, along with pulmonary symptoms and function tests (PFTs). Results: Fifty-five SjD-ILD patients were included (M: F = 9:46), of whom 11 were diagnosed with LIP (F: M = 11:0). LIP patients showed thin-walled parenchymal cysts as the predominant HRCT finding, and largely preserved pulmonary function (median forced vital capacity [FVC] 101% [IQR 98–105]; DLCO 76% [IQR 75–81]). After a median 5-years follow-up (IQR 2–7) all LIP patients were alive with stable PFTs. Compared with the remaining 44 non-LIP, LIP patients were younger at SjD diagnosis (P < 0.001) and more frequently presented purpura (P = 0.012), constitutional symptoms (P = 0.001), lymphadenopathy (P = 0.023), hypergammaglobulinemia (P < 0.001), triple anti-Ro60/52/La positivity (P = 0.035) and C3 hypocomplementemia (P = 0.009). ILD preceded SjD diagnosis in 30/44 non-LIP vs. 1/11 LIP patients (P < 0.001), with lower FVC% (P = 0.049) and DLCO% (P = 0.036) in non-LIP. Conclusion: LIP defines a distinct, immunologically active phenotype within the spectrum of SjD-ILD, characterized by greater extrapulmonary systemic involvement and serologic markers of B cell hyperactivity, but limited pulmonary functional impact. These findings support long-term lymphoma surveillance and a potential role for B cell targeted therapies in selected patients.

Background: Interstitial lung disease (ILD) is the most frequent extra-muscular manifestation in patients with idiopathic inflammatory myopathies (IIMs). Although high-resolution chest tomography (HRCT) represents the gold standard for the evaluation of ILD, lung ultrasound (LUS) might be a useful tool for its assessment. The aim of our study was to evaluate the number and distribution of pleural irregularities (PIs) identified by lung US in a cohort of patients with IIMs and to find possible correlations with clinical, serological and HRCT data to verify the potential usefulness of lung US for the study of ILD in patients with IIM. Patients and methods: Fifty-three patients with IIM according to EULAR/ACR classification criteria were enrolled. All patients underwent a clinical evaluation with measurement of disease activity and myositis-specific autoantibodies, pulmonary function tests, HRCT evaluated with the Warrick score, and lung US for the measurement of PIs. Results: The number of PIs was higher in patients with myositis-specific autoantibodies, particularly those with anti-synthetase autoantibodies (p < 0.001) and in patients with high-grade dyspnea (p < 0.03). A negative correlation was identified between PIs and pulmonary function tests, particularly TLC (r = −0.74; p < 0.001) and DLCO (r = −0.56; p < 0.001). Interestingly, PI score was higher in patients with ILD identified with HRCT (p = 0.015) and a positive correlation between PIs and Warrick score (r = 0.542; p < 0.001) was also found. Conclusions: The study of PIs with lung US represents a promising diagnostic tool for the bedside evaluation of patients with IIMs. This can possibly allow for a reduction in unnecessary HRCTs, reducing the exposition of patients to ionizing radiations, optimizing resources and reducing the costs of patients’ management.

Background: Drug-induced interstitial lung disease (DI-ILD) encompasses a heterogeneous spectrum of potentially severe pulmonary toxicities associated with various pharmacological agents. Diagnosis is often delayed due to nonspecific symptoms and imaging findings that can mimic other ILDs or infections, particularly in patients receiving oncological or immunomodulatory therapies. Novel approaches are needed to improve early recognition and management. Methods: We conducted a non-systematic, narrative literature review across major biomedical databases up to June 2025, aimed at describing current and emerging applications of artificial intelligence (AI) and radiomics in the early diagnosis, risk stratification, and personalised management of DI-ILD. Results: Radiomics applied to computed tomography and positron emission tomography/computed tomography enables extraction of high-dimensional quantitative features capturing subclinical alterations undetectable by visual assessment. In DI-ILD (particularly immune checkpoint inhibitor-related pneumonitis), radiomic models show potential diagnostic utility in distinguishing overlapping imaging patterns and predicting fibrotic progression. Integration with clinical, radiological and pharmacogenomic data has improved model performance in several studies. Additionally, AI approaches, including convolutional neural networks and ensemble learning methods, demonstrate promise in enhancing pattern recognition and risk stratification. Conclusions: Radiomics and AI are emerging complementary tools in multidisciplinary management of DI-ILD, offering objective imaging biomarkers and facilitating multimodal data integration to improve diagnostic precision and personalised therapeutic decisions. Nonetheless, reproducibility remains limited by variability in imaging protocols and lack of large-scale prospective multicentre validation. Clinical implementation requires standardised protocols to ensure consistency and reliability. The development of transparent, interpretable models seamlessly integrated into healthcare workflows is essential to fully leverage their real-world potential in proactive patient care.

Purpose The importance of multidisciplinary discussion (MDD) in interstitial lung diseases (ILDs) is well known. The aim of this study was to describe the 10-year experience of the MDD at the Pisa University Hospital in diagnosis and management of ILDs. Material and methods A single-center retrospective observational study was conducted at the Pisa University Hospital. A multidisciplinary group for ILDs diagnosis was instituted in 2012, and all patients referred till August 2022 were included. Results Altogether 1240 patients were evaluated. At baseline assessment, 772 common multidisciplinary diagnoses were achieved, while in 458 cases it was not possible to obtain a definite diagnosis. Most frequent multidisciplinary diagnoses were idiopathic pulmonary fibrosis (IPF), non-specific interstitial pneumonia (NSIP), and smoking-related ILDs. In 9% of cases a histological evaluation was performed, leading to a diagnosis in 75% of cases. One-thousand and fourty-eight cases with at least one year of follow-up after the MDD were revised. Common diagnoses were confirmed or modified, while “diagnostic hypotheses” were changed into “revised diagnoses” or left as hypothesis as they still could not be confirmed. Among common diagnoses, 61% were confirmed, while 36% were not confirmed as patients did not perform further examinations or visits. Eighteen diagnoses were modified (3%). Among the “diagnostic hypotheses,” 57% changed into “revised diagnoses.” Only in 15% of cases a definite diagnosis was not achieved after the revision. Conclusion The role of MDD in ILDs’ assessment and diagnosis was confirmed. A multidisciplinary approach should be considered in the long-term follow-up of ILDs, as diagnosis may change on the basis of clinical course.

Purpose: To differentiate interstitial lung diseases (ILDs) with fibrotic and inflammatory patterns using high-resolution computed tomography (HRCT) and a radiomics-based artificial intelligence (AI) pipeline. Materials and methods: This single-center study included 84 patients: 50 with idiopathic pulmonary fibrosis (IPF)—representative of fibrotic pattern—and 34 with cellular non-specific interstitial pneumonia (NSIP) secondary to connective tissue disease (CTD)—as an example of mostly inflammatory pattern. For a secondary objective, we analyzed 50 additional patients with COVID-19 pneumonia. We performed semi-automatic segmentation of ILD regions using a deep learning model followed by manual review. From each segmented region, 103 radiomic features were extracted. Classification was performed using an XGBoost model with 1000 bootstrap repetitions and SHapley Additive exPlanations (SHAP) were applied to identify the most predictive features. Results: The model accurately distinguished a fibrotic ILD pattern from an inflammatory ILD one, achieving an average test set accuracy of 0.91 and AUROC of 0.98. The classification was driven by radiomic features capturing differences in lung morphology, intensity distribution, and textural heterogeneity between the two disease patterns. In differentiating cellular NSIP from COVID-19, the model achieved an average accuracy of 0.89. Inflammatory ILDs exhibited more uniform imaging patterns compared to the greater variability typically observed in viral pneumonia. Conclusion: Radiomics combined with explainable AI offers promising diagnostic support in distinguishing fibrotic from inflammatory ILD patterns and differentiating inflammatory ILDs from viral pneumonias. This approach could enhance diagnostic precision and provide quantitative support for personalized ILD management.