Nanozoomer ht 2

After being killed via CO₂, mice were analyzed macroscopically and weighed (http://eulep.pdn.cam.ac.uk/Necropsy_of_the_Mouse/index.php). All organs (skin, heart, muscle, lung, brain, cerebellum, thymus, spleen, cervical lymph nodes, thyroid, parathyroid, adrenal gland, stomach, intestine, liver, pancreas, kidney, reproductive organs, and urinary bladder) were taken, fixed in 4% buffered formalin and embedded in paraffin for later histological examination. Two-µm-thick sections from each organ sample were generated and stained with hematoxylin and eosin, Periodic acid Schiff stain, Elastica van Gieson (EvG, Weigert’s stain), and/or Movat pentachrome. Thereafter, all sections were scanned using a virtual slide system (NanozoomerHT2.0; Hamamatsu) and evaluated by a board-certified pathologist. In addition, specialized automated image segmentation software (Definiens Tissue Studio (Definiens) and CellProfiler^{70 (link)}) was used to quantify age-related tissue changes via computer-assisted analyses.

Data set A contained 24 whole-slide images of rat liver sections. The data set was divided into four groups of six male Lewis rats each. The individual groups were fed different diets (Ssniff Spezialdiäten GmbH, Soest, Germany) for 3 months:

Ctrl: Normal rat chow

D1: Low methionine-low choline plus high starch diet

D2: Low methionine-low choline plus high fat diet

D3: Methionine-choline-deficient diet

At the end of the feeding periods, a 70% partial hepatectomy was performed and liver tissue was collected for further analysis. Sections of the left lateral lobe and the median lobe were stained with H&E and scanned with a Hamamatsu NanoZoomer HT 2.0 whole-slide scanner at a resolution of 227 nm/pixel.
Steatosis scores were expected to vary significantly between groups because the diets differed in their capacity to induce steatosis. However, only insignificant differences were expected within groups because the respective animals were of the same strain and fed with the same diet for the same time. The groups were sorted according to the average steatosis level induced by the respective diet, as determined by the mean steatosis area fraction within sections.
Rats were obtained from the Central Animal Laboratory, University Hospital Essen, Germany. All procedures were carried out in accordance with German animal welfare legislation.

At 6 weeks, animals were euthanized by an intraperitoneal injection of 140 mg/kg pentobarbital (Euthasol Vet®, Dechra Veterinary products). The heart tissue was collected, weighed after atrial tissue removal and paraffin-embedded for histomorphometry measurements after Hemalun-eosin staining. After slide scanning (Nanozoomer HT 2.0, Hamamatsu photonics K.K., Japan), the LV epicardial and endocardial contours were manually traced on short axis slices obtained in the central third of the heart. LV area (mm²) was determined, and the mean LV thickness (mm) was calculated as a ratio of the LV area over LV external contour length.

As a second dataset, we used scans of 35 H&E-stained slides spaced across a serial section of an entire steatotic mouse liver from a previous study.²¹ (link) In summary, steatosis was induced in a male C57/BL6N mouse (Charles River, Sulzfeld, Germany) by feeding a methionine/choline-deficient high fat diet (E15652-94 EF R/M, high fat MCD mod. low methionine and choline experimental diet; ssniff, Sulzfeld, Germany) for four weeks. All procedures and housing of the animals were strictly carried out according to the German animal welfare legislation (reference number 02-122-12). At the end of the observation period, the animal was sacrificed and the liver explanted, fixed in 5% buffered formalin followed by paraffin embedding. Subsequently, 3 μm sections were prepared using a rotary microtome (Microm HM355S; Thermo Fisher Scientific Microm International GmbH, Walldorf, Germany). A total of 2037 slides were produced and stained in batches of 25 slides. Each batch contained, among other stainings, two subsequent slides stained with H&E and GS, respectively. The slides were digitalized using a whole-slide scanner (NanoZoomer HT 2.0, Hamamatsu Photonics K.K., Hamamatsu City, Japan; at 400-fold optical magnification) at an in-plane image resolution of 227 nm. From these images, 35 undamaged pairs of neighboring H&E and GS WSIs covering the entire liver were selected.

Paraffin hearts sections were incubated with Picrosirius red (VWR) 0.1% in picric acid (Sigma) in a Leica ST5020 automatic strainer during 30 min and dehydrated in ethanol and xylene. Whole sections were observed at a magnification 20x using a Nanozoomer HT 2.0 (Hamamatsu) and fibrosis was quantified using FIBER, Matlab®-based software.

Formalin-fixed paraffin-embedded HES stained slides were digitised with a Nanozoomer HT2.0 (Hamamatsu) at 20× magnification to generate a whole slide imaging (WSI) file in ndpi format. We partitioned the WSI into non-overlapping 220 × 220 pixel tiles at 0.5 mm/pixel resolution (equivalent to 20× magnification) using QuPath v.0.2.3¹⁸ .
In addition, tumour regions of each slide were manually annotated by a pathologist (ALLP). Then, the centroids of each annotation were calculated. A TMA was created based on a circle with a radius of 500 µm from the centre of the centroid of the annotation. The same tiling as described above was kept.