SUPPLEMENTARY MATERIAL
Material and Methods
Pathology
For those patients who underwent preoperative systemic treatment, tumor regression grade (TRG) was scored for each CLM according to the Rubbia-Brandt system26. The percentage of CLM surface with necrosis, acellular mucin, fibrosis and residual tumor was also recorded. Patients with multiple CLM showing different TRG were categorized according to the worst metastasis score (highest TRG).
Immunohistochemistry
Antigen retrieval was performed by using EDTA (Biocare Medical). Antigen detection was performed by using 3,3-diaminobenzidine (DAB chromogen) and counter-staining of nuclei was performed with hemotoxylin. Images were obtained using optical microscopy (Olympus BX53). The immune-reactivity to CD3+ T cell and NKp46+ NK cells was evaluated in three different areas for each case of CLM: a) inside the tumor (intra-tumor), b) at the tumor border/peri-tumoral border (border), and c) inside the liver parenchyma free of tumor surrounding the lesion (peri-tumor). For each area (a,b,c), we determined the frequency of CD3+ T cells and NKp46+ NK cells in 10 consecutive 40x microscopic fields (HPF) after having pre-analytically evaluated the homogenous distribution of all immune reactive cells in all areas. The choice of the first field was done arbitrarily. 10 HPF correspond approximately to 1 cm2 that is the size of the lesions that were sampled by surgeons for pathology examination. Hence, the number of examined HPF fields was sufficient to examine the entire sampled tumor area of the majority of CLM specimens. Each HPF correspond to an area of 1 mm2 detected by the Olympus BX53 microscope equipped with an eyepiece field of 22 mm and a 40x objective. Therefore, we used the cell/mm2 parameter in the text. Given the high numbers of tumor-infiltrating CD3pos T cells, we determined their percentages instead of absolute numbers for mm2 by using the automatic count with the software Image-Pro Premier Version 9.1 (Media Cybernetics). The percentages of CD3+ cells were evaluated as a result of the ratio given by the number of immune-reactive cells and the number of cells counted in the field under examination. The automatic count of the percentages of CD3+ T cells was then also conformed with manual count. The number of NK cells for mm2 was assessed by manual count. Two pathologists evaluated CD3+ T cells as a percentage and NKp46+ NK cells as absolute number per field blinded to any patient clinical data.
Table 1-supplementary. Pathological response
Variable / N (%)Surgical minimal margin (mm)
Median, range / 0, 0-10
Tumor component
0-50%
>50% / 50 (41)
71 (59)
Fibrotic changes
0-50%
>50% / 107 (88)
14 (12)
Mucous substance
0-50%
>50% / 116 (96)
5 (4)
Necrosis
0-50%
>50% / 112 (92.5)
9 (7.5)
Pathological response (TRG)
TRG 5 (no regression, only tumor cells)
TRG 4 (dominant tumor)
TRG 3 (dominant fibrotic changes with few tumor cells)
TRG 2 (very few tumor cells in fibrotic tissue)
TRG 1 (no tumor cells) / 13 (14)
53 (55)
21 (22)
7 (7)
2 (2)
Table 2-supplementary. Immunological response
Variable / N (%)Intratumoral compartment
CD3+ cells
% cell/mm2 ± SD
>1%
Range %
NKp46+ cells
cell/mm2 ± SD
>1
Range / 0.79 ± 1.27
28 (23)
0-9%
0.55 ± 1.53
18 (15)
7 (6)
0-11
Border between the tumor and the non-tumoral tissue
CD3+ cells
% cell/mm2 ± SD
>1%
Range %
NKp46+ cells
cell/mm2 ± SD
>1
Range / 2.88 ± 2.08
104 (86)
76 (63)
0-12%
1.28 ± 1.58
46 (38)
0-8
1