Iscan ht

Tumor tissues from mice transplanted with cancer cells were fixed in phosphate-buffered 10% formalin and embedded in paraffin. Sections were cut at 5-μm intervals and stained with hematoxylin-eosin according to routine histological protocols. Ultrastructural studies were performed on the cells as previously reported [46 (link)]. Immunohistochemical staining was performed according to the manufacturer's instructions and/or standard protocols, as described previously [47 (link)]. Antibodies used included: anti-CGA (1:500) from Neomarker (California, USA); anti-Ki-67 (1:250), anti-serotonin (1:200), anti-cytokeratin (AE1/AE3 1:50), anti-vimentin (M0725 1:100), and anti-synaptophysin (1:200), all from Dako (California, USA); anti-somatostatin receptor 2A (1:500) and 5 (1:500) from Gramsch (Schwabhausen, Germany); anti-CD56/NCAM (1:100) from Novocastra (Newcastle, England); anti-RB (clone 3H9, 1:300) was from MBL (Nagoya, Japan). VECTASTAIN ABC HRP kit, from Vector Laboratories (California, USA), was used for the analysis. The Ki-67 index was obtained by counting the ratio of Ki-67 positive cells versus total nuclei using VENTANA iScan HT (Arizona, USA).

Hematoxylin and eosin (H&E) stained slides were digitized by scanning using Ventana iScan HT at 40x resolution (0.25 um/pixel resolution). HistoQC^{26 (link)}, an open-source quality control tool for digital pathology slides, was used to identify cases inappropriate for the study due to the presence of large blurry areas, obstructive dotting pen markings, or subcoverslip bubbles. Multinucleation frequency was calculated using a previously described algorithm^{17 (link)}, which generated a multinucleation index (MuNI) characterizing the density of multinucleation events in epithelial regions. Tumor infiltrating lymphocyte (TIL) frequency and intra-tumoral distribution was quantified using a validated machine learning algorithm (OP-TIL).^{18 (link)} The MuNI and OP-TIL risk scores were computed as continuous variables for each patient and a cutoff was then used to stratify patients into low- and high-risk categories as described in previous works.^{17 (link),18 (link)}

Surgical specimens were fixed in 10% formalin, processed, and embedded in paraffin, and then stained with hematoxylin and eosin using the standard protocol. Tumor diameter (or size) was defined as the longest diameter, and tumor width as the maximum width perpendicular to the diameter. Slides with the deepest infiltrated tumor cells were selected and scanned using the Ventana iScan HT slide scanner (Ventana Medical Systems, Tucson, AZ, USA) with a 20× objective lens. Depth of invasion was defined as the depth perpendicular to an imaginary line drawn from the adjacent muscularis mucosae (Figure 2) [29 ]. Estimated tumor volume was calculated using the following equation: 0.5 × diameter × width × depth [9 (link), 10 (link)].

From each block, 5μm sections were cut and stained with PD-L1 (clone SP263, Ventana) on an automated staining platform (Benchmark ULTRA; Ventana). An OptiView DAB IHC Detection Kit (Ventana) and an OptiView Amplification Kit (Ventana) were used according to the manifacturer's recommendations for the visualization of the primary anti PD-L1 antibody.
Stained sections were scanned using Ventana iScan HT and scored based on percentage of positive tumor cells, irrespective of staining intensities, using a four-tiered system: 0=0%, 1=1-4%, 2=5-9%, 3=10-49%, 4=≥50%.
We considered as adequate cores that showed a neoplastic component ≥ 30%; therefore cores with an inferior percentage of neoplastic cells have been excluded.
Macrophages were used as internal control in order to validate the adequacy of PD-L1 staining reaction (Figure 3).

The Ventana SP263 assay (Ventana Medical Systems Inc.) was adopted to assess PD-L1 expression in FFPE tumour samples. Briefly, 5 μm sections were cut and stained for SP263 on the Ventana Benchmark Ultra platform according to the standard protocol. Subsequently, stained sections were scanned using Ventana iScan HT and scored based on the percentage of tumour cells and immune cells showing membranous positivity. PD-L1 evaluation was performed by pathologists who routinely use clone SP263 in their clinical practice. For each case, the highest scoring value across the scores was used. PD-L1 expression on immune cells (IC) and tumour cells (TC), and the combined positive score (CPS) were evaluated and calculated in accordance with previous literature [16 (link)]. SP263 assay has been demonstrated to have similar levels of analytic performance with other FDA-approved PD-L1 IHC assays including 22C3 and 28–8 assays [16 (link), 17 (link)], with advantages of high sensitivity and relatively low cost in China, which contributed to most of the study subjects and where the PD-L1 analysis was performed centrally.

From each block 5 μm sections were cut and stained with anti-PD-L1 (clone SP263, Ventana) on an automated staining platform (Benchmark ULTRA; Ventana). An OptiView DAB IHC Detection Kit (Ventana) and an OptiView Amplification Kit (Ventana) were used according to the manifacturer's recommendations for the visualization of the primary anti PD-L1 antibody.
Stained sections were scanned using Ventana iScan HT and scored based on the percentage of positive tumor cells, irrespective of the staining intensities; a three-tiered system was then applied using the following thresholds: <1%, 1–49% and ≥50%.
Macrophages were used as internal control in order to validate the adequacy of PD-L1 staining reaction.