Search

The initial stage of a lung examination involves the segmentation of a CT image, a process that has been put under a lot of pressure with the high demand for chest scans and accurate segmentations. Current automatic segmentation algorithms are either non-robust for different datasets, not easily accessible, or time-consuming. Furthermore, classification of lung diseases such as IPF and NSIP is a difficult task often requiring decision-making between pathologists, radiologists and clinicians to make an accurate prognosis. Therefore, this thesis aims to create two algorithms easily accessible through a common medical software, 3D Slicer, with simple user interfaces for more efficient lung analysis. The first one is a fully automatic segmentation algorithm with a manual adjustment option. It is robust and developed on a diverse dataset, demonstrating a high accuracy with a median Dice score of0,967. The second one is a lung vessel tree morphometry algorithm which computes various parameters correlated to the vessel tree and its structure, providing insight into morphological changes. It shows great usability but has certain limitations, making it not entirely finished for clinical research but acts as an excellent starting point for a future project. The segmentation algorithm was developed using classical image processing techniques making it comprehensible. The distinctive feature of this algorithm is the entropy map used, enabling an effective way in distinguishing between the fibrotic regions of the lungs with surrounding soft tissue and therefore increasing its applicability on lungs with various diseases. The lung vessel tree morphometry algorithm utilized a segmentation of the lung vessels to organize them into a tree-like structure. The structure was divided into branches where each branch was used to calculate different parameters such as its level within the tree hierarchy, the length of the branch and more. These parameters were displayed and color-coded fo, Det inledande skedet av en lungundersökning involverar segmenteringen av en CT-bild, en process som har satts under mycket press på grund utav den höga efterfrågan på bröstskanningar och noggrann segmentering. Aktuella automatiska segmenteringsalgoritmer är antingen icke-robusta för olika dataset, ej lättillgängliga eller tidskrävande. Dessutom är klassificering av lungsjukdomar som IPF och NSIP en svår uppgift som ofta kräver beslutsfattande mellan patologer, radiologer och kliniker för att göra en korrekt prognos. Därför syftar denna rapport till att skapa två lättillgängliga algoritmer genom en ofta användmedicinsk programvara, 3D Slicer, bestående utav enkla användargränssnitt för en effektivare analys av lungorna. Den första är en helautomatisk segmenteringsalgoritm med ett manuellt justeringsalternativ. Den är robust och utvecklad på ett mångsidigt dataset som har demonstrerat en hög noggrannhet med en median Dice-score på 0,967. Den andra är en morfometri algoritm för lungkärlsträd som beräknar olika parametrar korrelerade till kärlträdet och dess struktur, vilket ger insikt i morfologiska förändringar. Den visar stor användbarhet men innehåller begränsningar, vilket gör den ej helt färdig för klinisk forskning utan fungerar som en utmärkt utgångspunkt för framtida arbete. Segmenteringsalgoritmen utvecklades med hjälp av klassiska bildbehandlingsmetoder vilket gör den mer lättförstådd. Det utmärkande för denna algoritm är entropikartan som används, vilket möjliggör ett effektivt sätt att skilja mellan de fibrotiska regionerna i lungorna med omgivande mjukdelar, detta gör den mer användbar på lungor med olika sjukdomar. Algoritmen för lungkärlsträdets morfometri använde en segmentering av lungkärlen för att sedanorganiseras i en trädliknande struktur. Strukturen var uppdelad i grenar där varje gren användes för att beräkna olika parametrar såsom dess nivå inom trädhierarkin, grenens längd med mera. Dessutom uppvisades dessa parametrar och färgkodades för vidar

The initial stage of a lung examination involves the segmentation of a CT image, a process that has been put under a lot of pressure with the high demand for chest scans and accurate segmentations. Current automatic segmentation algorithms are either non-robust for different datasets, not easily accessible, or time-consuming. Furthermore, classification of lung diseases such as IPF and NSIP is a difficult task often requiring decision-making between pathologists, radiologists and clinicians to make an accurate prognosis. Therefore, this thesis aims to create two algorithms easily accessible through a common medical software, 3D Slicer, with simple user interfaces for more efficient lung analysis. The first one is a fully automatic segmentation algorithm with a manual adjustment option. It is robust and developed on a diverse dataset, demonstrating a high accuracy with a median Dice score of0,967. The second one is a lung vessel tree morphometry algorithm which computes various parameters correlated to the vessel tree and its structure, providing insight into morphological changes. It shows great usability but has certain limitations, making it not entirely finished for clinical research but acts as an excellent starting point for a future project. The segmentation algorithm was developed using classical image processing techniques making it comprehensible. The distinctive feature of this algorithm is the entropy map used, enabling an effective way in distinguishing between the fibrotic regions of the lungs with surrounding soft tissue and therefore increasing its applicability on lungs with various diseases. The lung vessel tree morphometry algorithm utilized a segmentation of the lung vessels to organize them into a tree-like structure. The structure was divided into branches where each branch was used to calculate different parameters such as its level within the tree hierarchy, the length of the branch and more. These parameters were displayed and color-coded fo, Det inledande skedet av en lungundersökning involverar segmenteringen av en CT-bild, en process som har satts under mycket press på grund utav den höga efterfrågan på bröstskanningar och noggrann segmentering. Aktuella automatiska segmenteringsalgoritmer är antingen icke-robusta för olika dataset, ej lättillgängliga eller tidskrävande. Dessutom är klassificering av lungsjukdomar som IPF och NSIP en svår uppgift som ofta kräver beslutsfattande mellan patologer, radiologer och kliniker för att göra en korrekt prognos. Därför syftar denna rapport till att skapa två lättillgängliga algoritmer genom en ofta användmedicinsk programvara, 3D Slicer, bestående utav enkla användargränssnitt för en effektivare analys av lungorna. Den första är en helautomatisk segmenteringsalgoritm med ett manuellt justeringsalternativ. Den är robust och utvecklad på ett mångsidigt dataset som har demonstrerat en hög noggrannhet med en median Dice-score på 0,967. Den andra är en morfometri algoritm för lungkärlsträd som beräknar olika parametrar korrelerade till kärlträdet och dess struktur, vilket ger insikt i morfologiska förändringar. Den visar stor användbarhet men innehåller begränsningar, vilket gör den ej helt färdig för klinisk forskning utan fungerar som en utmärkt utgångspunkt för framtida arbete. Segmenteringsalgoritmen utvecklades med hjälp av klassiska bildbehandlingsmetoder vilket gör den mer lättförstådd. Det utmärkande för denna algoritm är entropikartan som används, vilket möjliggör ett effektivt sätt att skilja mellan de fibrotiska regionerna i lungorna med omgivande mjukdelar, detta gör den mer användbar på lungor med olika sjukdomar. Algoritmen för lungkärlsträdets morfometri använde en segmentering av lungkärlen för att sedanorganiseras i en trädliknande struktur. Strukturen var uppdelad i grenar där varje gren användes för att beräkna olika parametrar såsom dess nivå inom trädhierarkin, grenens längd med mera. Dessutom uppvisades dessa parametrar och färgkodades för vidar

Background: Although increasing data supports the use of transbronchial lung cryobiopsies (TBLCs) for the diagnosis of diffuse parenchymal lung diseases (DPLDs), its role as an alternative to surgical lung biopsy (SLB) is still under debate. The aim of this study was to assess the benefit of additional SLBs performed in selected patients after TBLCs. Method: We conducted a multicentric Belgian prospective trial in which SLBs were performed after TBLCs when the pathological diagnosis was uncertain or if a nonspecific interstitial pneumonia (NSIP) pattern was observed hypothesizing that SLB could provide additional information and that a co-existent UIP pattern could be missed. Results: Eighty-one patients with TBLCs performed for a DPLD were included in the study between April 2015 and December 2019. A specific histological diagnosis was obtained in 52 patients (64%) whereas no pathological diagnosis following TBLCs was obtained in 13 patients (16%) and a pattern suggestive of a NSIP was observed in 16 patients (20%). Fourteen out of these 29 patients had SLBs after TBLCs. SLBs showed a UIP pattern in 11 (79%), a pattern suggestive of a hypersensitivity pneumonitis in two (14%) and a NSIP pattern in one patient (7%). Among the 16 patients with pathological NSIP following TBLCs, six underwent a SLBs showing a UIP in five and confirming a NSIP in one patient only. A retrospective pathological analysis of patients having both procedures showed a lower diagnostic confidence and agreement among pathologists for TBLCs compared to SLBs. Major factors underlying the added value of SLBs were the bigger size of the sample as well as the subpleural localization of the biopsies. Conclusions: TBLCs are useful in the setting of DPLDs with a good diagnostic yield. However, our study suggests that SLB provides critical additional information in case TBLCs are inconclusive or show a pattern suggestive of a NSIP, questioning the accuracy of TBLC to adequately identify this histologica, SCOPUS: ar.j, info:eu-repo/semantics/published

type:Original, Background:Inflammatory cytokines have been reported to play important roles in the pathogenesis ofinterstitial lung diseases. However, their individual roles in idiopathic interstitial pneumonitis (IIP) and inthe other types of interstitial pneumonitis (IP), including collagen vascular disease associated interstitialpneumonitis (CVD-IP), remain unknown. In this study, we measured the bronchoalveolar lavage (BAL)fluid levels of several cytokines in patients with IIP and CVD-IP.Methods:Cell subpopulations in BAL fluid were counted, and BAL fluid levels of interleukin( IL)-2, -6,-7, -8, -17, interferon (IFN)-g , tumor necrosis factor (TNF)-a , and transforming growth factor (TGF)-b 1 were measured using a bead suspension array or an enzyme-linked immunosorbent assay( ELISA) kitin 16 patients (8 men, 8 women;mean age, 60.0±9.9 years) with idiopathic nonspecific interstitial pneumonitis(NSIP), 5 patients (3 men, 2 women;mean age, 69.0±4.8 years) with idiopathic usual interstitialpneumonitis (UIP), 5 patients (3 men, 2 woman;mean age, 66.3±5.5 years) with rheumatoid arthritis inCVD-IP (RA), and 5 patients (3 man, 2 women;mean age, 52.3±14.5 years) with dermatomyositis inCVD-IP( DM), and 13 patients( 3 men, 10 women;mean age, 51.8±17.2 years) with sarcoidosis.Results:BAL cell subpopulations had high amounts of lymphocytes in NSIP and sarcoidosis, and neutrophilsin RA. Levels of IL-7 were significantly (P<0.05) higher in DM (9.2±2.2 pg/ml) than in RA (4.5±1.7 pg/ml). IL-7 in DM was significantly( P<0.05) correlated with lymphocytes. The levels of TNF-a werehighest for RA( 18.6±29.5 pg/ml), compared with other IPs( 1.6±0.7 pg/ml in UIP, 1.8±2.7 pg/ml in NSIP,4.4±4.5 pg/ml in DM), and sarcoidosis( 5.8±6.1 pg/ml). In addition, the levels of IL-17 were highly detectablein RA( 2.5±2.5 pg/ml), but not in NSIP, UIP, or sarcoidosis.Conclusions:Differences in the cell types of BAL fluid and the level of each cytokine between patientswith IIP and CVD-IP might reflect pathogenesis an, identifier:2, identifier:KJ00008178439

Background: Interstitial lung diseases (ILDs) are a heterogeneous group of disorders characterized by inflammation and/or fibrosis of the lungs, varying from idiopathic interstitial pneumonias to secondary variants, including the ILDs associated to connective tissue diseases (CTDs). In addition, a number of patients are recognized as unclassifiable ILD (U-ILD), because of the inability to reach a definite diagnosis; some of them show autoimmune manifestations not fulfilling the classification criteria of a given CTD. The term interstitial pneumonia with autoimmune features (IPAF) has been recently proposed for this particular ILD subset. Methods: Here, we report our experience resulting from the integrated - pneumology/rheumatology - approach to patients with suspected ILDs or CTDs referred to our university-based Center for the Rare Pulmonary Diseases and Rheumatology Unit, from January 2009 to June 2015, with particular attention to the above-mentioned U-ILD, IPAF, and undifferentiated connective tissue disease (UCTD). The comparative analysis of these clinical variants was carried out; moreover, the observed findings were compared with the results of the updated review of the literature. Results: After the first clinical assessment, the U-ILD were identified in 50 patients; afterwards, on the basis of clinico-serological and radiological findings U-ILD group was subdivided into 2 subgroups, namely U-ILD without any clinical extra-thoracic manifestations and/or immunological alterations (15 pts) and IPAF according to the above-mentioned classification criteria (35 pts). Patients with either IPAF or U-ILD were compared with a series of 52 stable UCTD (disease duration ≥. 3. years), followed at our Rheumatology Unit. Some important differences were evidenced among the 3 series of U-ILD, IPAF, and UCTD: firstly, female gender was more frequent in patients with UCTD (86%) or IPAF (69%) compared with U-ILD (60%) or idiopathic pulmonary fibrosis (24%; p = 0.001). In add