Abstract text (incl. figure legends and references)

Introduction: Paraffin samples still remain the primary source of pathological diagnosis and for understanding disease development and progression. Due to the vast amount of paraffin samples available to scientific laboratories and universities around the world [1], we need to rethink modalities for imaging targeted structures that go beyond the limits of standard light microscopy (LM) and that achieve higher resolution. Objectives: Our goal is to develop a correlated workflow for targeting volumes of interest within a standard paraffin block. Starting with X-ray based imaging to create a reference volume, we then combine LM and SEM imaging to address specific subvolumes. Materials & Methods: Samples - Tissue from human lungs affected by COVID-19 was fixed in formalin and embedded in paraffin (FFPE) according to standard protocols. X-ray microscopy - Entire paraffin blocks were μCT-imaged (Phase Contrast CT) using a synchrotron radiation-based system (SRCT, Synchrotron Elettra, Trieste), initially with an isometric voxel size of 4μm. Based on the 3D scan of the entire block, targeted punch biopsies (cylinders with a diameter of 4 mm) were taken and scanned again with a voxel size of 2μm. Histology - Serial sections (5μm thick) were cut from a selected volume, stained (AB-PAS) and imaged using digital LM (slide scanner). Electron microscopy - Sections were stained with 3% uranyl acetate and lead citrate or/and sputtered with 3nm of Pt/Pd (80/20%). SEM (Zeiss Ultra 55) analysis was performed on regions pre-defined in the LM scan. Integration – To adjust possible deformations of the sections due to LM/SEM preparation, the data was registered with ,,Fuxlastix'' [2] Results: Comparison of a virtual slice from an X-ray volume (fig 1A) with a physical section cut from the corresponding paraffin block and stained for glycoproteins (fig 1B) shows perivascular fibrotic tissue formation in both modalities. SEM imaging (fig 1C, D) of regions defined in the LM scan illustrates the amount of detail obtainable at high resolution. By integrating the LM and SEM data into the X-ray volume, it is possible to identify and segment extended extracellular matrix elements such as the elastin cables (yellow in fig 2B & C) within the volume. Combining a cross-section of an alveolar tip (insets fig 2A) highlighted in purple in the X-ray volume (fig 2 B top) with the segmented elastin cables (fig 2B bottom) indicates where the elastin cables "penetrate" the alveolar tip. Conclusion: The use of X-ray tomography for non-destructive targeting is a goal-oriented method. We have developed a consistent correlative workflow for imaging paraffin samples in order to go beyond the restricted spatial resolution offered by traditional LM analysis. Furthermore, by using SRCT-based segmentations, we are able to make interpretations of their 3D context and represent them ultrastructurally by imaging the slice in the SEM. [1] L Cheng, DG Bostwick, (Eds.) - Essentials of Anatomic Pathology, Humana Press, 2002 [2] J Albers et al. - Elastic transformation of histological slices allows precise co‐registration with microCT data sets for a refined virtual histology approach, Scientific Reports, 2021