Patients and Samples
We searched pharmacy records of Oulu University Hospital and identified 54 patients who had received intravenous trastuzumab for metastatic breast cancer at least once during 2009–2014. Median age of the patients at the time of diagnosis was 58 years and 37 (68.5%) of the patients were oestrogen receptor positive tumours (Table 1). Median overall survival and median survival in metastatic disease were 58 and 39 months, respectively. Of all the patients 22 (40.7%) had primary metastatic disease at the time of diagnosis and 32 (59.3%) had relapsed disease.
Table 1. Patient demographics (n=54)
|
n (%)
|
Estrogen receptor positivity
|
37 (68.5)
|
Primary metastatic disease
|
22 (40.7)
|
Liver metastasis
|
20 (37.0)
|
Primary liver metastasis
|
12 (22.2)
|
Pulmonary metastasis
|
20 (37.0)
|
Primary pulmonary metastasis
|
17 (31.5)
|
Brain metastasis
|
19 (35.2)
|
Primary brain metastasis
|
2 (3.7)
|
Bone metastasis
|
29 (53.7)
|
Primary bone metastasis
|
25 (46.3)
|
Median age at diagnosis, years (range)
|
58 (28-82)
|
Median survival, months (range)
|
58 (4-242)
|
Median survival in metastatic disease, months (range)
|
39 (0-217)
|
The most common site of metastasis was bone, which was observed in 29 patients (53.7%) and 25 of whom (46.3%) had bone as a primary site of metastasis. Liver and pulmonary metastasis were both seen in 20 patients (37.0%). Of the patients with liver metastasis, 12 (22.2%) had liver as the primary site of metastasis, whereas 17 (31.5%) of the patients with pulmonary metastasis had lungs as the primary site of metastasis. Brain metastasis was observed in 19 patients (35.2%), but only 2 (3.7%) had brain as the primary site of metastasis (Table 1).
In our previous studies with metastatic HER2+ breast cancer[14, 15], we found that high infiltration of both CD8+ T-cells and M1-like macrophages in the centre of the tumour (CT) were associated with improved survival. We updated the follow-up time and status of the patients and reanalysed the survival estimates. High density of CD8+ T-cells in CT (HR 0.321, 95% CI 0.163–0.634), high density of M1-like macrophages in CT (HR 0.339, 95% CI 0.161–0.715) and the combination of CD8+ T-cells and M1-like macrophages in CT (high density of CD8 or M1 in CT; HR 0.246, 95% CI 0.094–0.641), and high density of CD8 and M1 in CT (HR 0.082, 95% CI 0.025–0.272) predicted improved survival. With this updated follow-up, the survival differences were more pronounced than in our previous studies (Supplementary figure is provided in the Online Resource 1).
Prognosis of the patients according to the metastatic sites
Bone, liver, lungs, and brain were the most common sites of metastasis in this study, and these were selected for further analysis. The impact of the metastatic sites to the overall survival of the patients during the whole disease course were investigated with Kaplan-Meier analysis. Liver, bone, and brain metastases were linked with worse prognosis; however, only liver metastasis was statistically significant (HR 1.809, 95% CI 1.004–3.262, Fig. 1a).Interestingly, pulmonary metastasis showed tendency towards improved survival (HR 0.662, 95% CI 0.361–1.216) but this, however, was not statistically significant (Fig. 1a).
Next, we evaluated whether the primary site of metastasis would also be linked to survival differences. We observed that liver (HR 4.171, 95% CI 2.009–8.656), bone (HR 2.047, 95% CI 1.115–3.757), and brain (HR 5.317, 95% CI 1.202–23.528) as a primary site of metastases were linked to poor prognosis, whereas primary pulmonary metastasis again showed tendency towards improved survival (HR 0.655, 95% CI 0.346–1.239) (Fig. 1b).
Correlation of the immune profiles and metastasis sites
Next, we wanted to question, whether different metastatic sites were associated with variable primary tumour immune profiles. Since the survival of patients with liver and pulmonary metastasis varied most considerably, we chose these two sites for further analysis on immune profiles. Even though the sites of primary metastasis showed more significant survival differences, the number of primary cases was too low for reliable statistical analysis and, we therefore chose to analyze patients with either liver or pulmonary metastasis during the whole follow-up. We included numerous immunology related cell types (markers) in the analysis: T-cells (CD3), cytotoxic T-cells (CD8), regulatory T-cells (FoxP3), NK-cells (CD56), M1-like (iNOS) and M2-like (CD163) macrophages, immunosuppressive STAB1+ macrophages and IDO1+ immune cells. We also evaluated IDO1 and CD47 expression in tumour cells. The scoring of the immunological cells was carried out using computer assisted counting of the positively stained cells/mm2, while tumour cell markers were scored according to staining intensity or counted histoscore (based on the proportion of positive tumour cells and staining intensity).
T-cells and M2-like macrophages were the most abundant cell types both in the invasive margin (IM) and center of the tumour (CT) (Table 2). Mann-Whitney test was used to compare the median immune cell densities of the tumours, both in the IM and CT, in patient with or without liver and with or without pulmonary metastasis. Patients with liver metastasis had lower amount of CD3+ T-cells in the CT (p=0.030) and M1-like macrophages in the CT (p=0.025) than patients without liver metastasis (Fig. 2a). In addition, patients with pulmonary metastasis had lower amount of STAB1+ macrophages in the IM (p=0.032) and CT (p=0.006) (Fig. 2b). The other studied immunological markers did not show any statistically significant associations with liver or pulmonary metastasis (data not shown).
Table 2. Median densities of the studied immune cells
|
All cases, n=54 (range) (cells/mm2)
|
Liver metastasis, n=20 (range) (cells/mm2)
|
Pulmonary metastasis, n=20 (range) (cells/mm2)
|
CD3 IM
|
573 (20 – 2520)
|
479 (71 – 1120)
|
659 (20 – 2459)
|
CD3 CT
|
242 (19 – 2051)
|
158 (19 – 537)
|
170 (19 – 1641)
|
CD8 IM
|
207 (20 – 1307)
|
176 (20 – 554)
|
240 (27 – 1307)
|
CD8 CT
|
120 (4 – 1201)
|
103 (4 – 223)
|
97 (4 – 1201)
|
iNOS IM
|
42 (4 – 145)
|
32 (4 – 78)
|
38 (10 – 99)
|
iNOS CT
|
37 (4 – 173)
|
20 (4 – 53)
|
36 (6 – 172)
|
CD163 IM
|
338 (24 – 957)
|
421 (219 – 653)
|
372 (24 – 653)
|
CD163 CT
|
295 (15 – 983)
|
267 (37 – 695)
|
321 (15 – 678)
|
STAB1 IM
|
82 (13 – 172)
|
70 (25 – 172)
|
36 (18 – 145)
|
STAB1 CT
|
30 (7 – 118)
|
33 (12 – 72)
|
18 (7 – 36)
|