Selection of specimens and collection of tissue fragments

This is a prospective study that includes 30 consecutive thyroid surgical specimens received for observation at the Pathology Laboratory of Institute of Molecular Pathology and Immunology of University of Porto, from February 2023 to May 2023, that accomplished to inclusion criteria described next. In this series, the thyroid surgical specimens should contain at least one well-limited follicular derived nodule with 10 mm in largest diameter or more. This(ese) well-limited nodule(s) should be suitable to be sampled in a 10-mm(width) × 10-mm(long) × 3-mm(thickness) tissue fragment that also included the radial margin. A single fragment was collected per nodule. One to four nodules were sampled per specimen. A total of 40 fragments, corresponding to the 30 surgical specimens from 30 different patients, were collected. The radial margin represented in each fragment was selectively inked with a brush. Of the 40 fragments, 10 were inked in red, 10 were inked in black, 10 were inked in blue, and 10 were inked in green (Epredia©, Kalamazoo, USA) (Fig. 1). If two nodules from the same lobe were to be sampled and included in the study, the used ink was the same to prevent colour contamination. Afterwards, all inked surgical margins were sprayed with an inking fixative constituted by a solution of absolute ethanol with glacial acetic acid.

Fig.1

Hematoxylin and eosin (H&E) whole slide images of four cases with their surgical margins inked with different colours: A red, B black, C blue, D green

Each of the 40 selected tissue fragments was processed (Donatello™ series 2 tissue processor, Diapath™, Martinengo, Italy) in a single cassette, together with the remaining cassettes of the corresponding case.

Glass slide preparation

After manual embedding, each of the 40 paraffin blocks was sectioned in a microtome HistoCore BIOCUT® (Leica Biosystems, Melbourne, UK) according to the following:

First section: 2 µm section to be attached to a regular non-adhesive slide (Süsse Labortechnik GmbH, Gudensberg, Germany) (all nodules)

Second section: 3 µm section to be attached to a regular non-adhesive slide (only black inked nodules)

Third section: 4 µm section to be attached to a regular non-adhesive slide (only black inked nodules)

Forth section: 2 µm section to be attached to an adhesive slide (TOMO slide, Matsunami Glass Ind., Ltd., Tokyo, Japan) (only black inked nodules)

All slides were stained by the hematoxylin and eosin (H&E) technique (Tissue-Tek Prisma® Plus automatic stainer, Sakura™, Tokyo, Japan) and coverslipped with film (Tissue-Tek Film®, Sakura™, Tokyo, Japan).

Scanning, observation, and annotation procedures

All slides were scanned on the Pannoramic 1000® Scanner (3DHISTECH Ltd, Budapest, Hungary) at 20 × (0.25 µm/pixel), using the scanning protocol clinically validated for histological WSIs currently used in our laboratory and managed by SlideCenter software (3DHISTECH Ltd., Budapest, Hungary). The time for scanning and the size file for each slide were recorded.

The WSIs were included in the virtual tray of the respective case for observation by one pathologist (CE) using the SlideViewer software (3DHISTECH, Ltd., Budapest, Hungary) (Fig. 2). On top of the WSI obtained from the first Sect. (2 µm, non-adhesive slide)m a digital measurement of the tissue was obtained. At 0.4 × magnification view and using the tool “draw a closed polygon,” we calculated the area and perimeter of each fragment. Within this area were included: the larger tissue fragment and very small tissue fragments detached, that could be identified as thyroid tissue. Small tissue fragments that were aligned with the largest fragment were included in a single measurement while tissue fragments located far from the largest fragment were measured separately and summed to those of the corresponding largest fragment (Fig. 3).

Fig.2

Aspect of the 4 different type of section of the same nodule disclosing the diagnosis of benign nodule in the setting of follicular nodular disease. A Hematoxylin and eosin (H&E), 40 × , 2 µm, non-adhesive slide; B H&E, 40 × , 3 µm, non-adhesive slide; C H&E, 40 × , 4 µm, non-adhesive slide; D H&E, 40 × , 2 µm, adhesive slide

Fig. 3

Measurement of the perimeter and area of the thyroid tissue fragments using digital tools on the whole slide image. A Measurement of the perimeter of a fragment inked red (arrow) that was representative of a follicular nodular disease with predominant macrofollicular growth pattern, H&E, 0.4 × , inset 41 × ; B measurement of the area and perimeter of a fragment inked green including a small fragment aligned with the largest fragment, this fragment was representative of a solid pattern papillary carcinoma, H&E, 0.4 × , inset 41 × ; C measurement of the area and perimeter of a fragment inked blue including 3 small fragments detached, this fragment was representative of a follicular nodular disease with dominant microfollicular pattern, H&E, 0.4 × 

All 40 nodules were observed and classified according to the 5th edition of the WHO book for Endocrine and Neuroendocrine tumours. For each nodule, the following data was recorded: age (years) and gender of the patient, largest size of the nodule (mm), location of the nodule, diagnosis, dominant histological pattern represented in the WSI, area of the fragment (mm2), perimeter of the fragment (mm), scanning time per each section (seconds), file size of each WSI per section (Mb), scanning time per area per section (seconds/mm2), and file size per area per section (Mb/mm2).

Statistical analysis

Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 29.0 for Windows. Fisher’s exact test was used for comparison of qualitative variables. The pair t test, the one-way ANOVA, and the Kruskal–Wallis test (KW) were used for comparison of quantitative variables, after testing for normal distribution through the Kolmogorov–Smirnov test. The level of significance was set at p < 0.05.