No abstract

Rationale: Transbronchial lung cryobiopsy (TBLC) has emerged as a less invasive method to obtain a tissue diagnosis in patients with interstitial lung disease (ILD). The diagnostic yield of TBLC compared to surgical lung biopsy (SLB) remains uncertain. Objectives: The aim of this study was to determine the diagnostic accuracy of forceps transbronchial lung biopsy (TBLB) and TBLC compared to SLB when making the final diagnosis based on multidisciplinary discussion (MDD). Methods: Patients enrolled in the study underwent sequential TBLB and TBLC followed immediately by SLB. De-identified cases, with blinding of the biopsy method, were reviewed by a blinded pathologist and then discussed at a multidisciplinary conference. Main Results: Between August 2013 and October 2017, we enrolled 16 patients. The raw agreement between TBLC and SLB for the MDD final diagnosis was 68.75% with a Cohen’s kappa of 0.6 (95% CI 0.39, 0.81). Raw agreement and Cohen’s kappa of TBLB versus TBLC and TBLB versus SLB for the MDD final diagnosis were much lower (50%, 0.21 [95% CI 0, 0.42] and 18.75%, 0.08 [95% CI −0.03, 0.19], respectively). TBLC was associated with mild bleeding (grade 1 bleeding requiring suction to clear) in 56.2% of patients. Conclusions: In patients with ILD who have an uncertain type based on clinical and radiographic data and require tissue sampling to obtain a specific diagnosis, TBLC showed moderate correlation with SLB when making the diagnosis with MDD guidance. TBLB showed poor concordance with both TBLC and SLB MDD diagnoses.

Background Immunofluorescent confocal microscopy uses labeled antibodies as probes against specific macromolecules to discriminate between multiple cell types. For images of the developmental mouse lung, these cells are themselves organized into densely packed higher-level anatomical structures. These types of images can be challenging to segment automatically for several reasons, including the relevance of biomedical context, dependence on the specific set of probes used, prohibitive cost of generating labeled training data, as well as the complexity and dense packing of anatomical structures in the image. The use of an application ontology helps surmount these challenges by combining image data with its metadata to provide a meaningful biological context, modeled after how a human expert would make use of contextual information to identify histological structures, that constrains and simplifies the process of segmentation and object identification. Results We propose an innovative approach for the semi-supervised analysis of complex and densely packed anatomical structures from immunofluorescent images that utilizes an application ontology to provide a simplified context for image segmentation and object identification. We describe how the logical organization of biological facts in the form of an ontology can provide useful constraints that facilitate automatic processing of complex images. We demonstrate the results of ontology-guided segmentation and object identification in mouse developmental lung images from the Bioinformatics REsource ATlas for the Healthy lung database of the Molecular Atlas of Lung Development (LungMAP1) program Conclusion We describe a novel ontology-guided approach to segmentation and classification of complex immunofluorescence images of the developing mouse lung. The ontology is used to automatically generate constraints for each image based on its biomedical context, which facilitates image segmentation and classification.

<b>Justificación:</b> La criobiopsia pulmonar transbronquial (TBLC) se ha propuesto como un método menos invasivo para obtener un diagnóstico tisular en pacientes con enfermedad pulmonar intersticial (EPI). El rendimiento diagnóstico de la TBLC en comparación con la biopsia pulmonar quirúrgica (SLB) sigue siendo incierto. <b>Objetivos:</b> Este estudio tuvo como objetivo determinar la precisión diagnóstica de la biopsia pulmonar transbronquial (TBLB) con fórceps y de la TBLC en comparación con la SLB para realizar el diagnóstico final con base en una discusión multidisciplinaria (MDD). <b>Métodos:</b> Los pacientes incluidos en el estudio se sometieron a TBLB y TBLC secuenciales, seguidas inmediatamente de SLB. Los casos, anonimizados y sin información sobre el método de biopsia, fueron revisados por un patólogo, quien trabajó a ciegas con respecto a dicha información, y luego discutidos en una conferencia multidisciplinaria. <b>Resultados:</b> Principales: Entre agosto de 2013 y octubre de 2017 se reclutaron 16 pacientes. La concordancia primaria entre TBLC y SLB para el diagnóstico final de MDD fue de 68.75%, con una kappa de Cohen de 0.6 (IC 95%: 0.39, 0.81). La concordancia primaria y la kappa de Cohen de la TBLC frente a la SLB para el diagnóstico final de MDD fueron mucho menores (50%, 0.21 [IC 95%: 0, 0.42] y 18.75%, 0.08 [IC 95%: -0.03, 0.19], respectivamente). La TBLC se asoció con hemorragia leve (hemorragia de grado 1, que requirió aspiración) en 56.2% de los pacientes. <b>Conclusiones:</b> En los pacientes con EPI de tipo incierto según con los datos clínicos y radiográficos, quienes requieren la toma de muestra de tejido para obtener un diagnóstico específico, la TBLC mostró una correlación moderada con la SLB al realizar el diagnóstico con la guía de una MDD. La TBLB mostró mala concordancia con los diagnósticos de la MDD, tanto mediante TBLC como por SLB.

AbstractBackgroundImmunofluorescent confocal microscopy uses labeled antibodies as probes against specific macromolecules to discriminate between multiple cell types. For images of the developmental mouse lung, these cells are themselves organized into densely packed higher-level anatomical structures. These types of images can be challenging to segment automatically for several reasons, including the relevance of biomedical context, dependence on the specific set of probes used, prohibitive cost of generating labeled training data, as well as the complexity and dense packing of anatomical structures in the image. The use of an application ontology surmounts these challenges by combining image data with its metadata to provide a meaningful biological context, and hence constraining and simplifying the process of segmentation and object identification.ResultsWe propose an innovative approach for the automated analysis of complex and densely packed anatomical structures from immunofluorescent images that utilizes an application ontology to provide a simplified context for image segmentation and object identification. We describe how the logical organization of biological facts in the form of an ontology can provide useful constraints that enhance automatic processing of complex images. We demonstrate the results of ontology-guided segmentation and object identification in mouse developmental lung images from the Bioinformatics REsource ATlas for the Healthy lung (BREATH) database of the Molecular Atlas of Lung Development (LungMAP1) program.ConclusionThe microscopy analysis pipeline library (micap) is available at https://github.com/duke-lungmap-team/microscopy-analysis-pipeline. Code to reproduce our analysis of LungMAP images is also available at https://github.com/duke-lungmap-team/lungmap-pipeline. Finally, the application ontology is available at https://github.com/duke-lungmap-team/lung_ontology and includes example SPARQL queries.ContactAnna Maria Masci email: annamaria.masci@duke.edu