Atezolizumab Plus Anthracycline-Based Chemotherapy in Metastatic Triple-Negative Breast Cancer: The Randomized, Double-Blind Phase 2b ALICE Trial

nature medicine
Article https://doi.org/10.1038/s41591-022-02126-1
Atezolizumab plus anthracycline-based

chemotherapy in metastatic triple-negative
breast cancer: the randomized, double-blind
phase 2b ALICE trial
A list of authors and their affiliations appears at the end of the paper
Received: 17 October 2022
Accepted: 8 November 2022 Immune checkpoint inhibitors have shown efficacy against metastatic
Published online: 8 December 2022 triple-negative breast cancer (mTNBC) but only for PD-L1positive disease.
The randomized, placebo-controlled ALICE trial (NCT03164993)
Check for updates evaluated the addition of atezolizumab (anti-PD-L1) to immune-stimulating
chemotherapy in mTNBC. Patients received pegylated liposomal
doxorubicin (PLD) and low-dose cyclophosphamide in combination with
atezolizumab (atezo-chemo; n = 40) or placebo (placebo-chemo; n = 28).
Primary endpoints were descriptive assessment of progression-free survival
in the per-protocol population (>3 atezolizumab and >2 PLD doses; n = 59)
and safety in the full analysis set (FAS; all patients starting therapy; n = 68).
Adverse events leading to drug discontinuation occurred in 18% of patients
in the atezo-chemo arm (7/40) and in 7% of patients in the placebo-chemo
arm (2/28). Improvement in progression-free survival was indicated in the
atezo-chemo arm in the per-protocol population (median 4.3 months versus
3.5 months; hazard ratio (HR) = 0.57; 95% confidence interval (CI) 0.33–0.99;
log-rank P = 0.047) and in the FAS (HR = 0.56; 95% CI 0.33–0.95; P = 0.033).
A numerical advantage was observed for both the PD-L1positive (n = 27;
HR = 0.65; 95% CI 0.27–1.54) and PD-L1negative subgroups (n = 31; HR = 0.57,
95% CI 0.27–1.21). The progression-free proportion after 15 months was
14.7% (5/34; 95% CI 6.4–30.1%) in the atezo-chemo arm versus 0% in the
placebo-chemo arm. The addition of atezolizumab to PLD/cyclophos
phamide was tolerable with an indication of clinical benefit, and the findings
warrant further investigation of PD1/PD-L1 blockers in combination with
immunomodulatory chemotherapy.
The prognosis for patients with metastatic triple-negative breast can- as monotherapy2. In combination with chemotherapy, atezolizumab
cer (mTNBC) is poor, with a median survival of approximately 1 year, (anti PD-L1) and pembrolizumab (anti-PD1) have shown activity against
and the therapeutic options are limited1. Immune checkpoint inhibi- PD-L1positive mTNBC3,4. Atezolizumab in combination with nab-paclitaxel
tors (ICIs) targeting PD1/PD-L1 are effective against metastatic dis- and pembrolizumab in combination with nab-paclitaxel, paclitaxel or
ease in many cancer forms but have limited efficacy against mTNBC gemcitabine/carboplatin are currently approved for this indication by
e-mail: jonky@ous-hf.no
Nature Medicine | Volume 28 | December 2022 | 2573–2583 2573

the European Medicines Agency, whereas only the pembrolizumab The secondary study objectives included assessment of objective
combinations are approved by the US Food & Drug Administration. The response rate (ORR), duration of response (DoR), durable response
approval for atezolizumab in the United States was withdrawn in August rate (DRR), clinical benefit rate (CBR), overall survival (OS), quality of
2021 after the publication of negative data from IMpassion131 (ref. 5) and life, candidate biomarkers and changes in immune cell populations.
positive data from KEYNOTE-355 (ref. 4). IMpassion130 was the first trial The protocol power estimates focused on durable clinical benefit,
to demonstrate an effect of adding immunotherapy to chemotherapy as measured by 15-month PFS. This milestone was chosen because
in mTNBC. This trial indicated that atezolizumab produced a survival chemotherapy responses rarely last that long in mTNBC, whereas
benefit when combined with nab-paclitaxel and that the effect applied durable responses are a hallmark of ICI activity1,24.
only to PD-L1positive disease3. Intriguingly, atezolizumab did not show any
effect against mTNBC in IMpassion131 (ref. 5), where the chemotherapy Results
backbone was paclitaxel. These contrasting findings have highlighted Patient characteristics
questions about how ICI activity is influenced by the chemotherapy The study recruited patients with mTNBC who were 18 years of age
backbone and by the use of steroids. Importantly, the addition of PD1/ or older, with an Eastern Cooperative Oncology Group (ECOG) per-
PD-L1 inhibitors to chemotherapy has not shown any efficacy against formance status of 0 or 1, who had received no more than one prior
PD-L1negative mTNBC, which represents a large proportion of patients3,5,6. line of chemotherapy in the metastatic setting and had measurable
Doxorubicin and cyclophosphamide are described as potent disease according to Immunotherapy Response Evaluation Criteria in
inducers of immunogenic cell death7–9, and there is evidence suggest- Solid Tumors (iRECIST)25. For patients who had received (neo)adjuvant
ing that their pro-survival effect in patients depends on the immune treatment with anthracyclines or cyclophosphamide, a disease-free
response7,10. There is yet no robust evidence showing that these or interval of at least 12 months was required. Key exclusion criteria were
other chemotherapies perceived to be immunogenic yield clinically central nervous system (CNS) disease (except asymptomatic lesions)
relevant synergies with ICIs, as has been hypothesized11. A few inter- and autoimmune disease requiring systemic treatment. A full list of
esting observations have, however, been reported. In the TONIC trial, inclusion and exclusion criteria is presented in the Methods section.
induction therapy with doxorubicin yielded the highest response rate Patients were enrolled at five centers in Norway and Den-
to nivolumab, and there was some evidence of immune activation in the mark. Between 24 August 2017 and 21 December 2021, 70 patients
tumor12. In the neoadjuvant setting, the KEYNOTE-522, GeparNuevo and were randomly assigned to receive atezolizumab and chemother-
IMpassion031 trials all demonstrated significantly increased response apy (atezo-chemo; n = 42 (60%)) or placebo and chemotherapy
rates from adding anti-PD1/PD-L1 to chemotherapy13–15, whereas the (placebo-chemo; n = 28 (40%)). A CONSORT flow diagram is shown in
NEOTRIP trial did not16. GeparNuevo indicated a benefit only for those Fig. 1. The 68 patients who received at least one dose of the allocated
starting anti-PD-L1 therapy 2 weeks before chemotherapy. Intrigu- treatment were included in the FAS (40 in the atezo-chemo group and
ingly, the chemotherapy backbone in KEYNOTE-522, GeparNuevo 28 in the placebo-chemo group), and 59 patients were included in the PP
and IMpassion031 contained anthracyclines and cyclophosphamide, population receiving >3 atezolizumab/placebo doses and >2 PLD doses
whereas NeoTRIP employed only taxanes and carboplatin before sur- (36 in the atezo-chemo group and 23 in the placebo-chemo group).
gery. However, it is not known if these observations are causally related Patient characteristics at baseline were mostly well-balanced
to the chemotherapy. The pathological complete response rate was between the treatment arms (Table 1). The atezo-chemo group had a
relatively high (60%) in the one-armed NeoPACT trial, which combined lower proportion of patients with ECOG performance status 0, a higher
pembrolizumab with an anthracycline-free regimen17. median age, a lower proportion with liver metastases, a lower propor-
Here we report the results of the randomized, double-blind, tion with more than two metastatic sites and a higher proportion with
placebo-controlled phase 2b study ALICE (NCT03164993), which PD-L1positive disease.
evaluated the safety and efficacy of adding atezolizumab to pegylated
liposomal doxorubicin (PLD) combined with low-dose metronomic Primary efficacy assessment
cyclophosphamide in patients with mTNBC18. This chemotherapy was At the data cutoff on 5 July 2022, the median follow-up time was
selected based on the hypothesis that it would trigger ICI sensitivity in 32.2 months (interquartile range (IQR) 27.4–40.9 months). A PFS event
patients who are otherwise non-responsive. Low-dose cyclophospha- had occurred in 57 patients (96.6%) in the PP population (23/23 in the
mide is reported to decrease the levels of regulatory T cells (Tregs)19. placebo-chemo group and 34/36 in the atezo-chemo group). All PFS
This has led to an interest in using low-dose cyclophosphamide in events were caused by disease progression. PFS in the PP population
cancer vaccine trials, yielding partially contradictory findings20–23. The was improved in the atezo-chemo arm compared to the placebo-chemo
pegylated liposomal form of doxorubicin was selected to obviate the arm (hazard ratio (HR) = 0.57; 95% confidence interval (CI) 0.33–0.99;
need for steroids, minimize the adverse effects (AEs) on the heart and P = 0.047; Fig. 2a), as hypothesized for the descriptive primary efficacy
allow for continued treatment beyond otherwise mandatory anthracy- endpoint. Median PFS was 4.3 months in the atezo-chemo arm versus
cline limits. To improve toxicity control and limit deep lymphopenia, 3.5 months in the placebo-chemo arm. The proportion without pro-
which may preclude ICI activity, PLD was administered every 2nd week gression or death 15 months after randomization was 14.7% (95% CI
instead of every 4th week, with an option for dose reduction. 6.4–30.1%) in the atezo-chemo group and 0% in the placebo-chemo
Inclusion of patients both with and without PD-L1 expression group. PFS was numerically improved in the PD-L1positive population
in tumors was allowed in this trial, to investigate both populationsʼ (co-primary endpoint; HR = 0.65, 95% CI 0.27–1.54; Fig. 2b).
response to PD-L1 blockade when added to the selected chemotherapy.
Stratification for PD-L1 was not performed at inclusion, because no Safety assessment and treatment exposure
PD-L1 test was in use for TNBC at the study sites at the time the study was Table 2 gives a summary of the safety data. A list of all AEs is available in
initiated. The primary study objectives were the descriptive assessment Supplementary Table 1. The median treatment period was 5.2 months
of safety and efficacy, as measured by progression-free survival (PFS). (IQR 2.3–7.2) in the atezo-chemo arm and 3.2 months (IQR 2.0–5.7)
The primary hypotheses were that the atezolizumab-chemotherapy in the placebo-chemo arm. The median number of PLD doses in the
(atezo-chemo) arm would have prolonged PFS compared to the control atezo-chemo arm was 11 (IQR 6–16) compared to seven (IQR 5–13) in
arm and that the atezo-chemo combination had acceptable safety and the placebo-chemo arm. An AE of any grade was recorded in 100% of
tolerability. PFS was measured in the per-protocol (PP) population and patients in both arms. Grade 3–4 AEs occurred in 62% of patients in the
the PD-L1positive PP population as primary analyses and in the full analy- atezo-chemo group and in 43% of patients in the placebo-chemo group.
sis set (FAS) and the PD-L1negative PP population as secondary analyses. The most common grade 3–4 AEs were decreased lymphocyte count (15%

Assessed for eligibility (n = 79)
Excluded (n = 9)
• Not meeting inclusion criteria (n = 9)
Randomized (n = 70)
Assigned to chemo + placebo (n = 28) Allocation Assigned to chemo + atezo (n = 42)

• Received allocated intervention (n = 28) • Received allocated intervention (n = 40)
• Did not receive allocated intervention (n = 2)
- Inclusion criteria no longer met
Discontinued intervention (n = 28) Follow-up Discontinued intervention (n = 38)

• Disease progression (n = 26) • Disease progression (n = 33)
• Adverse event (n = 1) • Adverse event (n = 2)
• Investigator’s decision (n = 1) • Investigator’s decision (n = 1)
• Completed 2-year intervention (n = 2)
Still on study treatment (n = 2)
Analyzed populations: Analysis Analyzed populations:

Full analysis set (n = 28) Full analysis set (n = 40)
Per-protocol population (n = 23) Per-protocol population (n = 36)

Excluded: Excluded:
• Insufficient cumulative IMP dose (n = 4) • Insufficient cumulative IMP dose (n = 4)
• Not evaluable by iRECIST (n = 1)
Discontinued the trial (n = 24) Discontinued the trial (n = 28)

• Death (n = 23) • Death (n = 28)
• Lost to follow-up (n = 1)
- Censored for survival Still on study treatment (n = 2)
In survival follow-up (n = 4) In survival follow-up (n = 10)
Fig. 1 | Patient flow diagram. The FAS included all patients who had received any IMP. The PP population comprised all patients who had received >3 doses of
atezolizumab/placebo and >2 doses of PLD and could be evaluated for tumor response.
in atezo-chemo and 18% in placebo-chemo) and rash (18% in atezo-chemo in Extended Data Fig. 1. In the PD-L1negative subgroup, a numerical PFS
and 0% in placebo-chemo). Serious adverse events (SAEs) occurred in improvement was observed in the PP population (HR = 0.57, 95% CI
48% of patients in the atezo-chemo group and in 29% of patients in 0.27–1.21; Fig. 2c) and in the FAS (HR = 0.55, 95% CI 0.27–1.14; Extended
the placebo-chemo group. There were no deaths related to AEs. AEs Data Fig. 1b). Three patients, all in the atezo-chemo arm, had PFS beyond
leading to permanent discontinuation of any study drug occurred in the scheduled treatment period of 2 years (Extended Data Fig. 2). These
18% of patients in the atezo-chemo group and in 7% of patients in the three long-term responders had PD-L1negative disease. One patient in the
placebo-chemo group, whereas atezo/placebo was discontinued due to atezo-chemo arm had a PFS of 4.7 months by iRECIST25 and 1.8 months
an AE in 12% and 4%, respectively. Immune-related adverse events (irAEs) by RECIST version 1.1 (ref. 26), whereas the estimates were the same by
of any grade were recorded in 28% and 18% of patients in the atezo-chemo both criteria for all other patients (Extended Data Fig. 2). There was
and placebo-chemo arms, respectively. Grade 3–4 irAEs occurred in 10% no significant difference in OS between the treatment groups (FAS:
of patients in the atezo-chemo arm compared to 4% of patients in the HR = 0.75, 95% CI 0.43–1.30; Fig. 2e). The time to deterioration of global
placebo-chemo arm. The most common irAEs in the atezo-chemo arm health status/quality of life was improved in the atezo-chemo group
were hypothyroidism (12%), pneumonitis (10%) and rash (8%) (Table 2). (FAS: HR = 0.25; 95% CI 0.09–0.67; Fig. 2f).
The treatment was considered to be tolerable. The PFS advantage for the atezo-chemo arm seemed consistent
across the protocol-specified subgroups defined by clinical parameters
Secondary efficacy and biomarker endpoints (treatment line, disease-free interval, de novo metastatic disease and
PFS in the FAS (n = 68) was improved in the atezo-chemo arm compared metastatic site) (Fig. 3). The only subgroup with an unfavorable HR
to the placebo-chemo arm (HR = 0.56; 95% CI 0.33–0.95, P = 0.033; in the stratified analyses was patients with more than two metastatic
Fig. 2d). OS and PFS for the PD-L1positive and PD-L1negative FAS are shown sites, which was not a pre-defined subgroup.

Table 1 | Demographics and baseline characteristics of the study cohort
FAS PP population
Placebo-chemo Atezo-chemo P value Placebo-chemo Atezo-chemo P value
(n = 28) (n = 40) (n = 23) (n = 36)
Age 0.16 0.17
Median (range) 52.5 (28.0–74.0) 58.5 (31.0–77.0) 52.0 (28.0–74.0) 59.0 (31.0–77.0)
Gender 0.40 0.42
Female 28 (100%) 39 (97.5%) 23 (100%) 35 (97.2%)
Male 0 (0%) 1 (2.5%) 0 (0%) 1 (2.8%)
ECOG performance status 0.50 0.99
0 21 (75.0%) 27 (67.5%) 16 (69.6%) 25 (69.4%)
1 7 (25.0%) 13 (32.5%) 7 (30.4%) 11 (30.6%)
De novo metastatic disease 0.74 0.93
Yes 8 (28.6%) 10 (25.0%) 6 (26.1%) 9 (25.0%)
No 20 (71.4%) 30 (75.0%) 17 (73.9%) 27 (75.0%)
Bone metastases 0.23 0.18
Yes 16 (57.1%) 17 (42.5%) 13 (56.5%) 14 (38.9%)
No 12 (42.9%) 23 (57.5%) 10 (43.5%) 22 (61.1%)
Liver metastases 0.17 0.36
Yes 13 (46.4%) 12 (30.0%) 9 (39.1%) 10 (27.8%)
No 15 (53.6%) 28 (70.0%) 14 (60.9%) 26 (72.2%)
Lung metastases 0.44 0.69
Yes 10 (35.7%) 18 (45.0%) 9 (39.1%) 16 (44.4%)
No 18 (64.3%) 22 (55.0%) 14 (60.9%) 20 (55.6%)
Lymph node metastases 0.49 0.94
Yes 13 (46.4%) 22 (55.0%) 13 (56.5%) 20 (55.6%)
No 15 (53.6%) 18 (45.0%) 10 (43.5%) 16 (44.4%)
CNS metastases 0.80 0.75
Yes 1 (3.6%) 1 (2.5%) 1 (4.3%) 1 (2.8%)
No 27 (96.4%) 39 (97.5%) 22 (95.7%) 35 (97.2%)
Number of metastatic sites 0.24 0.12
≤2 15 (53.6%) 27 (67.5%) 12 (52.2%) 26 (72.2%)
>2 13 (46.4%) 13 (32.5%) 11 (47.8%) 10 (27.8%)
>3 3 (10.7%) 7 (17.5%) 3 (13.0%) 5 (13.9%)
Line of chemotherapy 0.81 0.37
1st 16 (57.1%) 24 (60.0%) 12 (52.2%) 23 (63.9%)
2nd 12 (42.9%) 16 (40.0%) 11 (47.8%) 13 (36.1%)
Previous anthracycline treatment 0.90 0.71
Yes 20 (71.4%) 28 (70.0%) 17 (73.9%) 25 (69.4%)
No 8 (28.6%) 12 (30.0%) 6 (26.1%) 11 (30.6%)
PD-L1 status 0.22 0.22
Negative 17 (60.7%) 19 (47.5%) 14 (60.9%) 17 (47.2%)

Table 1 (continued) | Demographics and baseline characteristics of the study cohort
FAS PP population
Placebo-chemo Atezo-chemo P value Placebo-chemo Atezo-chemo P value
(n = 28) (n = 40) (n = 23) (n = 36)
Positive 10 (35.7%) 21 (52.5%) 8 (34.8%) 19 (52.8%)
Missing 1 (3.6%) 0 (0%) 1 (4.3%) 0 (0%)
Intrinsic breast cancer subtype 0.44 0.48
Luminal A 2 (7.1%) 0 (0%) 2 (8.7%) 0 (0%)
Luminal B 1 (3.6%) 1 (2.5%) 1 (4.3%) 1 (2.8%)
HER2-enriched 2 (7.1%) 4 (10.0%) 2 (8.7%) 4 (11.1%)
Basal 12 (42.9%) 22 (55.0%) 11 (47.8%) 20 (55.6%)
Missing 11 (39.3%) 13 (32.5%) 7 (30.4%) 11 (30.6%)
BRCA mutation status 0.54 0.54
BRCA1 mutation 1 (3.6%) 2 (5.0%) 1 (4.3%) 2 (5.6%)
Normal variant 12 (42.9%) 22 (55.0%) 11 (47.8%) 22 (61.1%)
Missing 15 (53.6%) 16 (40.0%) 11 (47.8%) 12 (33.3%)
Two-sided P values were calculated using the Wilcoxon rank-sum test to compare age distributions and the chi-square test for all other variables. In comparison of the number of metastatic
sites, only the first two groups were included in the chi-square test owing to the overlap between the two latter groups.
a Progression–free survival
b Progression–free survival
c Progression–free survival
PP population PD–L1positive PP population PD–L1negative PP population
1.0 1.0 1.0
0.9 0.9 0.9
HR 95% CI HR 95% CI HR 95% CI
0.8 0.8 0.8
Survival probability
0.7
0.57 0.33–0.99 0.7
0.65 0.27–1.54 0.7
0.57 0.27–1.21
0.6 0.6 0.6
0.5 Placebo–chemo 0.5 Placebo–chemo 0.5 Placebo–chemo
0.4 Atezo–chemo 0.4 Atezo–chemo 0.4 Atezo–chemo
0.3 0.3 0.3
0.2 0.2 0.2
0.1 0.1 0.1
P = 0.047 P = 0.33
0 0 0
0 3 6 9 12 15 18 21 24 27 30 33 36 0 3 6 9 12 15 18 21 24 27 30 33 36 0 3 6 9 12 15 18 21 24 27 30 33 36
Time (months) Time (months) Time (months)
Number at risk Number at risk Number at risk
Placebo–chemo 23 13 4 2 0 0 0 0 0 0 0 0 0 Placebo–chemo 8 6 2 1 0 0 0 0 0 0 0 0 0 Placebo–chemo 14 7 2 1 0 0 0 0 0 0 0 0 0
Atezo–chemo 36 25 11 7 5 5 4 3 3 2 2 1 0 Atezo–chemo 19 12 8 4 2 2 1 0 0 0 0 0 0 Atezo–chemo 17 13 3 3 3 3 3 3 3 2 2 1 0
d e f
Progression–free survival Overall survival Global health status/quality of life
FAS FAS time to deterioration, FAS
1.0 1.0 1.0
Proportion without deterioration
0.9 0.9 0.9

Placebo–chemo
HR 95% CI HR 95% CI Atezo–chemo
0.8 0.8 0.8
0.7
0.56 0.33–0.95 0.7
0.75 0.43–1.30 0.7
0.6 0.6 0.6
0.5 Placebo–chemo 0.5 Placebo–chemo 0.5
0.4 Atezo–chemo 0.4 Atezo–chemo 0.4
0.3 0.3 0.3
0.2 0.2 0.2
HR 95% CI
0.1 0.1 0.1 0.25 0.09–0.67
P = 0.033 P = 0.0029
0 0 0
0 3 6 9 12 15 18 21 24 27 30 33 36 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Time (months) Time (months) Time (months)
Number at risk Number at risk Number at risk
Placebo–chemo 28 13 4 2 0 0 0 0 0 0 0 0 0 Placebo–chemo 28 25 20 15 14 12 9 7 5 4 2 1 1 1 0 0 0 Placebo–chemo 25 22 17 14 7 6 4 1 1 0 0 0 0 0
Atezo–chemo 40 26 11 7 5 5 4 3 3 2 2 1 0 Atezo–chemo 40 39 36 27 23 17 11 10 9 8 6 4 4 3 1 1 1 Atezo–chemo 37 32 26 23 19 18 10 8 6 6 6 5 1 0
Fig. 2 | Kaplan–Meier plots of survival outcomes and quality of life. a–d, the 62 patients (91%) in the FAS for whom a baseline value was available. The
PFS assessed according to iRECIST in the PP population (a), the PD-L1positive (b) proportions of patients who completed this item at each of the subsequent
and PD-L1negative (c) PP population subsets and the FAS (d). e, OS in the FAS. f, timepoints are available in Extended Data Table 4. HRs with CIs were estimated
Time to deterioration (reduction ≥20 points) of the global health status/quality using the Cox proportional hazards method. P values were calculated by the
of life score in the EORTC QLQ-C15-PAL questionnaire. The analysis includes log-rank method.
The ORR, CBR, DRR and DoR were numerically higher in the (7.9–35.6), and the CBR was 50% (35.2–64.8) versus 35.7% (20.7–54.2).
atezo-chemo arm, in both the PP population and the FAS (Extended The DRR was 15.0% (7.1–29.1) in the atezo-chemo arm compared to
Data Table 1). In the FAS, the ORR was 27.5% (6.1–42.8%) versus 17.9% 3.6% (0.2–17.7) in the placebo-chemo arm, whereas the median DoR

Table 2 | Summary of AEs (FAS) Exploratory assessment of effects on lymphocyte subsets

We hypothesized that the applied chemotherapy would cause a reduc-
Placebo-chemo Atezo-chemo tion in Treg levels, as reported for low-dose cyclophosphamide19. To
n = 28 n = 40 evaluate how the chemotherapy with or without atezolizumab affected
Any grade Grade 3–4 Any grade Grade 3–4 circulating immune cell subsets, paired samples of peripheral blood
mononuclear cells (PBMCs) from two timepoints (pre-treatment and
Any AE 100% (28) 43% (12) 100% (40) 62% (25)
week 8) from 47 patients (all patients in the FAS with paired samples
Any SAE 29% (8) 18% (5) 48% (19) 38% (15) available) were assessed by an exploratory flow cytometry analysis.
Any IMP discontinued 7% (2) 4% (1) 18% (7) 10% (4) As expected, the absolute cell counts for all measured immune cell
due to AE subsets decreased after treatment, with B cells and Tregs being most
Atezo/placebo 4% (1) 0 12% (5) 10% (4) affected (Fig. 4a). Furthermore, we found that the percentage of Tregs
discontinued due to AE as a fraction of CD4+ T cells decreased in both arms after therapy
irAE (Fig. 4b), with a larger decrease in mean value in the atezo-chemo arm
(from 3.28% to 2.33%; P < 0.0001) than in the placebo-chemo arm (from
Any 18% (5) 4% (1) 28% (11) 10% (4)
2.82% to 2.20%; P = 0.054).
Hypothyroidism 7% (2) 0 12% (5) 0
Pneumonitis 4% (1) 0 10% (4) 5% (2) Exploratory biomarker analyses
In the analyses of pre-defined biomarkers, we found that only patients
Rash 4% (1) 0 8% (3) 5% (2)
with above-median TIS appeared to benefit from the addition of
Hyperthyroidism 7% (2) 0 2% (1) 0 atezolizumab (Fig. 3). Previous studies indicated that patients with
Pancreatic enzymes 4% (1) 4% (1) 0 0 values within the highest tertile (above the 66th percentile; TIShigh)
increased have improved responses to PD-1 blockers28,29. Applying the same
Pancreatitis 0 0 2% (1) 2% (1) cutoff (66th percentile) for an exploratory analysis in our dataset,
Pyrexia 0 0 2% (1) 2% (1)
we found that the TIShigh group appeared to have improved PFS in the
atezo-chemo arm, with an HR of 0.42 (95% CI 0.17–1.03), and comprised
AEs occurring in ≥25% in either group
all atezo-chemo patients with a PFS >12 months (Extended Data Fig. 3).
Rash 39% (11) 0 65% (26) 18% (7) Exploratory analyses suggested that patients with low Treg lev-
Nausea 54% (15) 0 57% (23) 5% (2) els at baseline had a substantial PFS benefit in the atezo-chemo arm
(HR = 42, 95% CI 0.17–1.00, P = 0.043; Fig. 4c), whereas those with high
Palmar-plantar 11% (3) 4% (1) 52% (21) 8% (3)
erythrodysaesthesia Treg levels had no benefit (Fig. 4d). The level of tumor-infiltrating
syndrome lymphocytes (TILs) was also evaluated in an explorative biomarker
Fatigue 43% (12) 0 50% (20) 5% (2) analysis, by assessment of biopsies obtained at study entry, and scored
as TILlow (n = 47) or TILhigh (n = 16). As shown in Fig. 3, the observed PFS
Mucosal inflammation 21% (6) 0 48% (19) 0
benefit for the atezo-chemo arm was relatively similar for the TILlow
Constipation 43% (12) 0 45% (18) 0 (HR = 0.62) and TILhigh (HR = 0.71) subgroups.
Lymphocyte count 36% (10) 18% (5) 45% (18) 15% (6) We further explored how the TIS correlated with the PD-L1 status
decreased and TIL score. The TIS and PD-L1 analyses were performed on the same
Musculoskeletal pain 21% (6) 0 30% (12) 5% (2) biopsy in all cases where enough material was available (28/44 biop-
AEs were graded according to CTCAE version 4.0.
sies), whereas the TIL scoring was performed on a different biopsy.
An overview of the biopsy sites is given in Extended Data Table 3. The
results showed that the TIS correlated with PD-L1 status (P = 0.0082),
with a median scaled TIS value of 0.56 (IQR 0.18–1.00) for PD-L1positive
was 7.3 months (IQR 2.1–19.6) versus 3.7 months (IQR 1.6–4.8). Among versus −0.54 for PD-L1negative (IQR −0.93 to 0.39) (Extended Data Fig. 4).
patients with PD-L1negative disease, none recorded a durable response A proportion of 25% in the PD-L1negative group had a TIS above the median
(>6 months) after placebo-chemo compared to 15.8% (5.5–37.6) in the value in the PD-L1positive group. The median scaled TIS was 0.55 and −0.21
atezo-chemo group (Extended Data Table 1). in the TILhigh and TILlow groups, respectively. A proportion of 32% in
We further investigated the potential impact of imbal- the TILlow group had a TIS above the median value in the TILhigh group.
ances between the treatment arms by exploratory multivariable
analyses performed in the FAS. For this purpose, we selected fac- Discussion
tors expected to be of particular clinical relevance and adjusted There is no consensus on the optimal chemotherapy regimen for com-
for one factor in each analysis (Extended Data Table 2). The unad- bination with ICI, and there is a need to investigate whether immuno-
justed HR for PFS in the atezo-chemo-arm was 0.56, with a P value logically cold tumors can be made susceptible to ICI therapy. This is a
of 0.033. The adjusted HRs were 0.50, 0.60, 0.57 and 0.58 after clinically important knowledge gap for mTNBC, which has a poor prog-
inclusion of ECOG status, age, treatment line and PD-L1 status, nosis and where ICI benefits have been observed only for PD-L1positive
respectively. disease. The ALICE trial is, to our knowledge, the first randomized study
The pre-specified biomarker analyses comprised PD-L1 expression in mTNBC investigating the addition of any ICI to anthracyclines, even
(SP142 immunohistochemistry (IHC) assay), tumor mutational bur- though anthracyclines are widely used for this disease. There was a
den (TMB) and the NanoString tumor inflammation signature (TIS)27 plausible rationale for exploring the selected combination, based on
(immune gene expression). The median score was used as a cutoff for the perceived immunogenic properties of anthracyclines, the avoid-
TIS and TMB. Whereas the PFS benefit from atezolizumab appeared ance of steroids with PLD and the reported effects of low-dose cyclo-
not to depend on PD-L1 status or TMB, the HR was numerically better phosphamide on Tregs. The results indicate improved therapeutic
for patients with a high TIS score (Fig. 3). The TMB was generally low, efficacy for the atezo-chemo arm, as measured by the primary endpoint
with a median value of 2.14 mutations per megabase (mut/Mb; IQR (descriptive comparison of PFS). The PFS HR benefit was numerically
1.4–2.8). Only one patient, in the placebo-chemo group, had a TMB large and reached statistical significance both in the PP population
>10 mut/Mb. and the FAS, even though the patient number was relatively small and

Subgroup Atezo− Placebo− HR (95 % CI)

chemo chemo
All 40 28 0.56 (0.33 − 0.95)

Previous chemotherapy in metastatic setting
Yes 16 12 0.64 (0.29 − 1.43)
No 24 16 0.53 (0.26 − 1.07)
Disease−free interval
<24 months 10 8 0.58 (0.20 − 1.71)
≥ 24 months 20 12 0.55 (0.25 − 1.22)
ECOG performance status
0 27 21 0.70 (0.37 − 1.30)
1 13 7 0.21 (0.06 − 0.74)
De novo metastatic
Yes 10 8 0.57 (0.21 − 1.53)
No 30 20 0.57 (0.30 − 1.06)
Bone metastases
Yes 17 16 0.51 (0.24 − 1.08)
No 23 12 0.63 (0.30 − 1.36)
Liver metastases
Yes 12 13 0.59 (0.25 − 1.41)
No 28 15 0.59 (0.30 − 1.16)
Lung metastases
Yes 18 10 0.63 (0.28 − 1.44)
No 22 18 0.47 (0.23 − 0.97)
Lymph node metastases
Yes 22 13 0.87 (0.43 − 1.79)
No 18 15 0.35 (0.16 − 0.78)
Number of metastatic sites
≤2 27 15 0.41 (0.20 − 0.82)
>2 13 13 1.31 (0.55 − 3.09)
Intrinsic subtype
Basal−like 22 12 0.77 (0.36 − 1.64)
Other 5 5 0.44 (0.10 − 1.86)
PD−L1 status
Positive 21 10 0.63 (0.28 − 1.43)
Negative 19 17 0.55 (0.27 − 1.14)
Tumor−infiltrating lymphocytes
High 10 6 0.71 (0.24 − 2.06)
Low 27 20 0.62 (0.33 − 1.16)
Tumor inflammation signature
< median 11 11 1.04 (0.42 − 2.58)
≥ median 16 6 0.67 (0.24 − 1.90)
Tumor mutational burden
< median 13 13 0.85 (0.37 − 1.97)
≥ median 21 6 0.86 (0.34 − 2.20)
0.06 0.13 0.25 0.50 1.00 2.00
HR (PFS)
Favors atezo−chemo Favors placebo−chemo
Fig. 3 | Subgroup analyses of PFS (FAS). HRs (center square) with 95% CIs (error The number of patients in each subgroup/arm is indicated. Intrinsic subtype,
bars) for PFS in the atezo-chemo group versus the placebo-chemo group in the PD-L1 status, TIS, tumor lymphocyte infiltration and TMB were not available for
FAS. HRs with CIs were calculated using the Cox proportional hazards method. all patients.
lower than originally planned. The same applied to the proportion of findings need to be confirmed in an independent study. There was no
patients without disease progression after 15 months. The trial was not difference in the median OS, whereas the secondary efficacy endpoints
powered to test the superiority hypothesis of the addition of atezoli- related to tumor response (ORR, CBR, DRR and DoR) were in favor of
zumab. Accordingly, the statistical analyses were descriptive, and the the atezo-chemo arm, mostly with overlapping CIs. The development

a b
P = 0.0066
P = 0.0737
P = 0.0539
P = 0.0004
P = 0.0034
P = 0.0047
P = 0.4426
P < 0.0001
P < 0.0001
P < 0.0001
P < 0.0001
P < 0.0001
P = 0.0003
P < 0.0001
2 2
P < 0.0001
P = 0.5226
P < 0.0001
P = 0.2396
P = 0.1169
P < 0.0001
0 1
log2 fold change

log2 fold change
–2 0
–4 –1
–6 –2
CD4 CD8 Treg
eg
lls
lls
lls
lls
8
4
D
D
ce
ce
ce
ce
Tr
C
C
B Placebo-chemo
KT
-T
N
gd
N
Atezo-chemo
c d
Treglow patients Treghigh patients
1.0 1.0
0.9 0.9
HR 95% CI HR 95% CI
Progression–free survival
Progression–free survival
0.8 0.8
0.42 0.17–1.00 1.57 0.55–4.44
0.7 0.7
0.6 0.6
0.5 Placebo–chemo 0.5 Placebo–chemo
0.4 Atezo–chemo 0.4 Atezo–chemo
0.3 0.3
0.2 0.2
0.1 0.1
P = 0.043
0 0
0 3 6 9 12 15 18 21 24 27 30 33 0 3 6 9 12 15 18 21 24 27 30 33
Time (months) Time (months)
Number at risk Number at risk
Placebo–chemo 11 6 2 1 0 0 0 0 0 0 0 0 Placebo–chemo 7 4 1 1 0 0 0 0 0 0 0 0
Atezo–chemo 14 10 6 5 4 4 4 3 3 2 2 1 Atezo–chemo 15 8 0 0 0 0 0 0 0 0 0 0
Fig. 4 | Effect of therapy on immune cell subsets and association of Tregs cell subsets are defined as follows: CD4 T cells (CD3+CD4+CD8−), CD8 T cells
with PFS. PBMCs collected at pre-trial screening and after 8 weeks of trial (CD3+CD4−CD8+), Tregs (CD3+CD4+Foxp3+CD25Hi), B cells (CD3−CD19+),
treatment were assessed for different immune cell subsets by flow cytometry. NK cells (CD3−CD56+), NKT cells (CD3+CD56+) and gd-T cells (CD3+gd-TCR+).
All patients in the FAS for whom paired samples were available were included in b, Percentage of T cell subsets. CD4 and CD8 subsets are shown as a percentage of
the analysis (placebo-chemo n = 18, atezo-chemo n = 29). The gating strategy is total lymphocytes, and Tregs are shown as a percentage of total CD4+ T cells.
shown in Supplementary Fig. 1. Cell counts were estimated by multiplying the Two-tailed P values were calculated using the Wilcoxon matched-pairs signed-
percentage of each subset within the lymphocyte gate by the clinical lymphocyte rank test. c,d, Kaplan–Meier analysis of PFS in the placebo-chemo group
differential cell count obtained for each patient at each timepoint. Fold changes compared to the atezo-chemo group in patients with low (≤ median; c) and high
from screening were calculated and log2-transformed. The baseline level is (> median; d) levels of Tregs as percentage of total CD4+ T cells. Two-tailed P
indicated with a dotted line. Data are presented as box and whisker plots, with the values were calculated using the log-rank test and HRs with CIs using the Cox
center line showing the median and the hinges showing the IQR. Whiskers show proportional hazards method.
minimum and maximum values. a, Immune cell absolute cell counts. Immune
in the global health status score indicated a favorable effect of atezoli- melanoma and other cancer forms where ICI has a proven benefit24.
zumab on the quality of life. At the timepoint of 15 months, all patients without disease progression
The nature, frequency and severity of AEs were as expected based belonged to the atezolizumab arm. The number of patients remaining
on the properties of the study drugs and the advanced disease stage. for these analyses was small, and the data should be interpreted with
We recorded a higher frequency of SAEs and high-grade AEs in the caution. The PFS signal of long-term efficacy is, however, worth not-
atezo-chemo arm. This may be due to toxicity from atezolizumab but ing as there is an unmet medical need in mTNBC for agents providing
may also reflect longer duration of treatment in the atezo-chemo group. sustained benefit. On the other hand, even though the survival data
A role of higher treatment exposure is supported by the increased are not mature, it appears evident that the median survival will remain
frequency of non-immune-related AEs as well as the immune-related similar between the arms. Further follow-up is ongoing to assess if a
events. High-grade irAEs were observed in only 10% of the atezo-chemo proportion of patients experience long-term survival.
group and were generally manageable. The proportion of patients who ALICE was a small randomized study, which means that known
discontinued therapy due to AEs was in line with IMpassion130 (ref. 3) and unknown imbalances between the arms represent a major limi-
and IMpassion131 (ref. 5). tation. We performed multivariable analyses for selected factors of
Whereas the improvement in median PFS was only modest, the perceived clinical significance. The results suggested that imbalances
indicated atezolizumab benefit was evident for the tail of the PFS in these factors could not explain the indicated atezolizumab PFS ben-
curve, as reflected by the DRRs. This is in line with the experience from efit. The HR adjusted for ECOG indicated a larger PFS benefit for the

atezo-chemo-arm than the parent analysis, whereas the other adjusted approved diagnostic biomarker assays. This observation would be
HRs were similar to the unadjusted value. In the subgroup analyses, the worth investigating in independent and larger cohorts, using TIS and
numerical HR advantage for the atezo-chemo group was conserved in all other gene expression signatures. Whether or not the concept of immu-
groups defined by clinical parameters, apart from those with more than nogenic cell death applies, it is possible that some level of pre-existing
two metastatic sites, which had a sample size of only 21 patients. The immune activation in the tumor is required for the benefit of PD1/PD-L1
CIs were generally wide, due to the small sample size, but still seemed blockers. This form of immune activation may not necessarily lead to
to suggest that ECOG >0, fewer than three metastatic sites and lack of expression of PD-L1 but may be captured by gene expression assays.
lymph node involvement were associated with increased benefit from It will be important to investigate whether gene expression signa-
atezolizumab. There is no clear explanation for these observations, tures vary between primary tumors and different metastatic lesions
and we cannot exclude that they are incidental due to multiple testing. and how this compares to variations in PD-L1 status. A considerable
Interestingly, the PFS advantage appears to apply even to the discrepancy has been reported between biopsies obtained at differ-
PD-L1negative population, and the three patients with >24-month PFS in ent timepoints31,32. For clinical practice, there is a need to determine
the PD-L1negative group had all been randomized to the atezo-chemo arm. whether more than one biopsy is required for biomarker assessment
Atezolizumab was administered in the same dose as in IMpassion130 and if more reproducible and informative assays can be established.
(ref. 3) and IMpassion131 (ref. 5). Furthermore, the enrollment criteria In our dataset, the benefit from atezolizumab was restricted to
and patient characteristics in ALICE, IMpassion130 and IMpassion131 those with a below-median level of Tregs in blood. This was an explora-
were mostly similar, apart from the allowance in the ALICE trial for one tory biomarker test and needs further investigation and validation.
previous line of chemotherapy in the metastatic setting. The subgroup A possible explanation may be that different suppressive mecha-
analyses from ALICE did not suggest that the treatment line affected nisms may be involved and that high Treg levels compensate for PD1/
the PFS advantage for the atezo-chemo arm. The ALICE data are, thus, PD-L1-mediated suppression. There has been a large interest in coun-
consistent with the hypothesis that the applied chemotherapy may tering Tregs in cancer therapy. We investigated the use of metronomic
trigger ICI sensitivity in PD-L1negative patients with mTNBC who are oth- cyclophosphamide for this purpose. The previously reported data
erwise non-responsive. We have, as yet, no mechanistic evidence for on cyclophosphamide and Tregs are sparse and partially contradic-
this from ALICE patients and plan to investigate if the therapy modified tory19–23. It is, therefore, interesting that we observed a decrease in
the tumor microenvironment by analyzing study biopsies. Tregs. Low-dose cyclophosphamide has only mild side effects and
It should be emphasized that cross-trial comparisons of PD-L1positive/ may be widely applicable across cancer forms. We do not know if the
negative
populations are associated with uncertainty. In the ALICE trial, the changes in peripheral blood are reflected in the tumor microenviron-
PD-L1 assay, interpretation method and choice of biopsy were defined in ment and if the therapy affects Treg function. Further exploration
the protocol and statistical analysis plan. We employed the same assay of this is needed. It may also be noted that a decrease in Tregs is not
(SP142) and cutoff (≥1%) as used in IMpassion130 (ref. 3) and IMpas- necessarily beneficial as it may enhance autoimmunity.
sion131 (ref. 5). If the patient had positive and negative PD-L1 status in dif- The radiological response was assessed locally, without a cen-
ferent biopsies, the most recent biopsy obtained before study therapy tral blinded review. This setup may have affected the consistency in
was used. The SP142 assay was chosen because it was recommended by response measurement across study sites but should not lead to a
Roche and reported to be more predictive for the effect of atezolizumab systematic bias in favor of any arm, as the study was placebo-controlled.
in mTNBC than other assays, which identify more patients as PD-L1positive Only one patient had a different timepoint for progression between
(ref. 6). The decision to use the last biopsy was based on the assumption iRECIST and RECIST version 1.1. The data, therefore, suggest that
that this is most relevant for the metastatic disease and was in accord- pseudo-progression33 was not a common feature.
ance with the national guidelines for PD-L1 testing for mTNBC in Norway. Several chemotherapeutic agents have been hypothesized to
However, the choice of biopsy varies between trials. In IMpassion130, induce immunogenic cell death. However, anthracyclines trigger
it was not defined whether the biopsy should be from a primary tumor release of four major damage‐associated molecular patterns9 and are
or metastasis3. In the SAFIR-02 trial, the SP142 assay was employed on a considered to be particularly immunogenic8,9,11. As anthracyclines are
metastatic biopsy obtained <1 year before enrollment30. Data from the also potent drugs against TNBC, it is important to establish whether
IMpassion130 cohort suggest that PD-L1 assessment on both primary synergy between anthracyclines and ICIs can be achieved and if this
and metastatic tumor is informative for atezolizumab benefit31. It has would make a larger proportion of patients with TNBC responsive
previously been reported that the PD-L1positive proportion is higher in to ICIs. Moreover, some patients are resistant to taxanes and, there-
primary tumors, compared to metastatic biopsies32. Accordingly, it is fore, not candidates for therapy with the atezolizumab/nab-paclitaxel
possible that some of the patients classified as PD-L1negative in ALICE and regimen. The discrepancy between the results from IMpassion130 and
other trials could have had a PD-L1positive primary tumor. The proportion IMpassion131 have highlighted the need to document the efficacy of
of patients with PD-L1positive disease in ALICE (47%) is in line with other each ICI/chemotherapy combination. Pembrolizumab has been tested
mTNBC studies employing the SP142 assay, including IMpassion130 (41% with other agents but not with anthracyclines, which is employed in
PD-L1positive) and IMpassion131 (45% PD-L1positive)3,5,30–32. Accordingly, the many countries as first-line therapy for mTNBC.
chosen assay and observed PD-L1positive percentage appear to suggest PLD is more expensive than several other anthracyclines, and this is
that the PD-L1negative ALICE population is comparable to the PD-L1negative a hurdle for the implementation of the findings from the ALICE trial. On
groups in other mTNBC trials. The PD-L1 classification issue is, however, the other hand, continued PLD administration is feasible in long-term
complicated, and further investigation of PD-L1 status across clinical ICI responders, whereas other anthracyclines have mandatory restric-
trial cohorts would be of interest. tions on accumulative dose. Steroids are needed to control side effects
The TIS is based on the expression of 18 immune-related genes of many agents, and their use is difficult to limit in a real-world setting,
and was developed to predict response to anti-PD-1 therapy27. In our but they are rarely required for PLD therapy. The PLD dosing schedule
dataset, a high TIS score appeared to be associated with clinical benefit used in ALICE (every 2nd week) is used for Kaposi sarcoma and has
from atezolizumab, whereas a high TMB or PD-L1 positivity did not. been explored in some breast cancer studies, with or without low-dose
Interestingly, all patients with a PFS of 12 months belonged to the top cyclophosphamide34,35, but is not regularly used for TNBC. This may
tertile of TIS, which is a cutoff reported by others29. Furthermore, the represent a challenge for implementing the ALICE regimen, even if
lack of association with TMB may reflect that the TMB was generally low. the cumulative dose over 4 weeks corresponds to common practice.
However, the data support the notion that gene expression signatures In conclusion, this study indicates that the addition of atezoli-
may identify a responsive subgroup that is not detected by regulatory zumab to PLD and low-dose cyclophosphamide improved PFS in

patients with mTNBC. A benefit was also indicated in patients with 14. Mittendorf, E. A. et al. Neoadjuvant atezolizumab in combination
PD-L1negative disease. The combination therapy was tolerable, and no with sequential nab-paclitaxel and anthracycline-based
new safety signals were identified. Subgroup analyses suggest that a chemotherapy versus placebo and chemotherapy in patients
high immune gene expression in the tumor may predict benefit from with early-stage triple-negative breast cancer (IMpassion031): a
atezolizumab. There was no difference in median survival, and it will randomised, double-blind, phase 3 trial. Lancet 396, 1090–1100
be important to establish if there is a survival benefit for a proportion (2020).
of the patients. The flow cytometry data support the hypothesis that 15. Schmid, P. et al. Pembrolizumab for early triple-negative breast
low-dose metronomic cyclophosphamide leads to decreased Treg cancer. N. Engl. J. Med. 382, 810–821 (2020).
levels. The findings in this small randomized trial are consistent with 16. Gianni, L. et al. Pathologic complete response (pCR) to
the concept that ICI may synergize with selected immune-stimulating neoadjuvant treatment with or without atezolizumab in
chemotherapy and provide a basis for further investigations in larger triple-negative, early high-risk and locally advanced breast
mTNBC cohorts. The results also suggest that similar immunomodu- cancer: NeoTRIP Michelangelo randomized study. Ann. Oncol. 33,
latory chemotherapies may be considered in combination with ICI in 534–543 (2022).
other cancer forms. 17. Sharma, P. et al. Clinical and biomarker results of neoadjuvant
phase II study of pembrolizumab and carboplatin plus docetaxel
Online content in triple-negative breast cancer (TNBC) (NeoPACT). J. Clin. Oncol.
Any methods, additional references, Nature Portfolio reporting sum- 40, 513–513 (2022).
maries, source data, extended data, supplementary information, 18. Kyte, J. A., Røssevold, A., Falk, R. S. & Naume, B. ALICE: a
acknowledgements, peer review information; details of author con- randomized placebo-controlled phase II study evaluating
tributions and competing interests; and statements of data and code atezolizumab combined with immunogenic chemotherapy in
availability are available at https://doi.org/10.1038/s41591-022-02126-1. patients with metastatic triple-negative breast cancer. J. Transl.
Med 18, 252 (2020).
References 19. Ghiringhelli, F. et al. Metronomic cyclophosphamide regimen
1. Waks, A. G. & Winer, E. P. Breast cancer treatment: a review. JAMA selectively depletes CD4+CD25+ regulatory T cells and restores
321, 288–300 (2019). T and NK effector functions in end stage cancer patients. Cancer
2. Winer, E. P. et al. Pembrolizumab versus investigator-choice Immunol. Immunother. 56, 641–648 (2007).
chemotherapy for metastatic triple-negative breast cancer 20. Walter, S., Weinschenk, T., Reinhardt, C. & Singh-Jasuja, H.
(KEYNOTE-119): a randomised, open-label, phase 3 trial. Lancet Single-dose cyclophosphamide synergizes with immune
Oncol. 22, 499–511 (2021). responses to the renal cell cancer vaccine IMA901.
3. Schmid, P. et al. Atezolizumab and nab-paclitaxel in advanced Oncoimmunology 2, e22246 (2013).
triple-negative breast cancer. N. Engl. J. Med. 379, 2108–2121 (2018). 21. Slingluff, C. L. Jr. et al. Randomized multicenter trial of the effects
4. Cortes, J. et al. Pembrolizumab plus chemotherapy versus of melanoma-associated helper peptides and cyclophosphamide
placebo plus chemotherapy for previously untreated locally on the immunogenicity of a multipeptide melanoma vaccine.
recurrent inoperable or metastatic triple-negative breast cancer J. Clin. Oncol. 29, 2924–2932 (2011).
(KEYNOTE-355): a randomised, placebo-controlled, double-blind, 22. Emens, L. A. et al. Timed sequential treatment with
phase 3 clinical trial. Lancet 396, 1817–1828 (2020). cyclophosphamide, doxorubicin, and an allogeneic
5. Miles, D. et al. Primary results from IMpassion131, a double-blind, granulocyte-macrophage colony-stimulating factor-secreting
placebo-controlled, randomised phase III trial of first-line breast tumor vaccine: a chemotherapy dose-ranging factorial
paclitaxel with or without atezolizumab for unresectable locally study of safety and immune activation. J. Clin. Oncol. 27,
advanced/metastatic triple-negative breast cancer. Ann. Oncol. 5911–5918 (2009).
32, 994–1004 (2021). 23. Anders, C. K. et al. Evaluating the efficacy of a priming dose of
6. Rugo, H.S. et al. PD-L1 immunohistochemistry assay comparison cyclophosphamide prior to pembrolizumab to treat metastatic
in atezolizumab plus nab-paclitaxel-treated advanced triple- triple negative breast cancer. J. Immunother. Cancer 10,
negative breast cancer. J. Natl Cancer Inst. 113, 1733−1743 (2021). e003427 (2022).
7. Sistigu, A. et al. Cancer cell-autonomous contribution of type I 24. Hodi, F. S. et al. Nivolumab plus ipilimumab or nivolumab alone
interferon signaling to the efficacy of chemotherapy. Nat. Med. versus ipilimumab alone in advanced melanoma (CheckMate
20, 1301–1309 (2014). 067): 4-year outcomes of a multicentre, randomised, phase 3
8. Kroemer, G., Senovilla, L., Galluzzi, L., Andre, F. & Zitvogel, L. trial. Lancet Oncol. 19, 1480–1492 (2018).
Natural and therapy-induced immunosurveillance in breast 25. Seymour, L. et al. iRECIST: guidelines for response criteria for use
cancer. Nat. Med. 21, 1128–1138 (2015). in trials testing immunotherapeutics. Lancet Oncol. 18, e143–e152
9. Bezu, L. et al. Combinatorial strategies for the induction of (2017).
immunogenic cell death. Front. Immunol. 6, 187 (2015). 26. Schwartz, L. H. et al. RECIST 1.1—update and clarification: from
10. Apetoh, L. et al. Toll-like receptor 4-dependent contribution of the the RECIST committee. Eur. J. Cancer 62, 132–137 (2016).
immune system to anticancer chemotherapy and radiotherapy. 27. Ayers, M. et al. IFN-γ-related mRNA profile predicts clinical
Nat. Med. 13, 1050–1059 (2007). response to PD-1 blockade. J. Clin. Invest. 127, 2930–2940 (2017).
11. Pfirschke, C. et al. Immunogenic chemotherapy sensitizes tumors 28. Danaher, P. et al. Pan-cancer adaptive immune resistance as
to checkpoint blockade therapy. Immunity 44, 343–354 (2016). defined by the tumor inflammation signature (TIS): results from The
12. Voorwerk, L. et al. Immune induction strategies in metastatic Cancer Genome Atlas (TCGA). J. Immunother. Cancer 6, 63 (2018).
triple-negative breast cancer to enhance the sensitivity to PD-1 29. Damotte, D. et al. The tumor inflammation signature (TIS) is
blockade: the TONIC trial. Nat. Med. 25, 920–928 (2019). associated with anti-PD-1 treatment benefit in the CERTIM
13. Loibl, S. et al. A randomised phase II study investigating pan-cancer cohort. J. Transl. Med. 17, 357 (2019).
durvalumab in addition to an anthracycline taxane-based 30. Bachelot, T. et al. Durvalumab compared to maintenance
neoadjuvant therapy in early triple-negative breast cancer: chemotherapy in metastatic breast cancer: the randomized
clinical results and biomarker analysis of GeparNuevo study. phase II SAFIR02-BREAST IMMUNO trial. Nat. Med. 27, 250–255
Ann. Oncol. 30, 1279–1288 (2019). (2021).

31. Emens, L. A. et al. Atezolizumab and nab-paclitaxel in Publisher’s note Springer Nature remains neutral with regard to
advanced triple-negative breast cancer: biomarker evaluation jurisdictional claims in published maps and institutional affiliations.
of the IMpassion130 study. J. Natl Cancer Inst. 113, 1005–1016
(2021). Open Access This article is licensed under a Creative Commons
32. Li, Y. et al. Prevalence study of PD-L1 SP142 assay in metastatic Attribution 4.0 International License, which permits use, sharing,
triple-negative breast cancer. Appl. Immunohistochem. Mol. adaptation, distribution and reproduction in any medium or format,
Morphol. 29, 258–264 (2021). as long as you give appropriate credit to the original author(s) and the
33. Hodi, F.S. et al. Evaluation of immune-related response source, provide a link to the Creative Commons license, and indicate
criteria and RECIST v1.1 in patients with advanced melanoma if changes were made. The images or other third party material in this
treated with pembrolizumab. J. Clin. Oncol. 34, 1510–1517 article are included in the article’s Creative Commons license, unless
(2016). indicated otherwise in a credit line to the material. If material is not
34. Jehn, C. F. et al. Biweekly pegylated liposomal doxorubicin included in the article’s Creative Commons license and your intended
(Caelyx) in heavily pretreated metastatic breast cancer: a phase 2 use is not permitted by statutory regulation or exceeds the permitted
study. Clin. Breast Cancer 16, 514–519 (2016). use, you will need to obtain permission directly from the copyright
35. Dellapasqua, S. et al. Pegylated liposomal doxorubicin in holder. To view a copy of this license, visit http://creativecommons.
combination with low-dose metronomic cyclophosphamide as org/licenses/by/4.0/.
preoperative treatment for patients with locally advanced breast
cancer. Breast 20, 319–323 (2011). © The Author(s) 2022
Andreas Hagen Røssevold 1,2, Nikolai Kragøe Andresen 1,2, Christina Annette Bjerre 3, Bjørnar Gilje4,
Erik Hugger Jakobsen5, Sunil Xavier Raj6, Ragnhild Sørum Falk7, Hege Giercksky Russnes8, Thea Jahr9,
Randi Ruud Mathiesen10, Jon Lømo8, Øystein Garred8, Sudhir Kumar Chauhan2, Ragnhild Reehorst Lereim2, Claire Dunn2,
Bjørn Naume10,11 & Jon Amund Kyte 1,2
Department of Clinical Cancer Research, Oslo University Hospital, Oslo, Norway. 2Department of Cancer Immunology, Institute for Cancer Research,
1
Oslo University Hospital, Oslo, Norway. 3Department of Oncology, Rigshospitalet, Copenhagen, Denmark. 4Department of Hematology and Oncology,
Stavanger University Hospital, Stavanger, Norway. 5Department of Oncology, Sygehuset Lillebælt, Vejle, Denmark. 6Department of Oncology,
St. Olav University Hospital, Trondheim, Norway. 7Oslo Centre for Biostatistics and Epidemiology, Oslo University Hospital, Oslo, Norway. 8Department of
Pathology, Oslo University Hospital, Oslo, Norway. 9Department of Radiology and Nuclear medicine, Oslo University Hospital, Oslo, Norway. 10Department
of Oncology, Oslo University Hospital, Oslo, Norway. 11Institute of Clinical Medicine, University of Oslo, Oslo, Norway. e-mail: jonky@ous-hf.no

Methods negativity are defined as <1% and <10%, respectively, of cells

Study design and approvals expressing hormonal receptors via IHC analysis.
The ALICE trial was an investigator-initiated, multi-center, randomized, 2. Adequate newly obtained core or excisional biopsy of a tumor
double-blind, parallel-group, placebo-controlled phase 2b trial evalu- lesion not previously irradiated. No anti-tumor treatment is
ating the safety and efficacy of the combination of atezolizumab and allowed between the timepoint for biopsy and study entry. If a
chemotherapy in mTNBC. Oslo University Hospital was the sponsor. patient has undergone chemotherapy in the metastatic setting,
Protocol approval was obtained from the Regional Committee for Medi- a new biopsy must be obtained after this therapy.
cal Research Ethics South-East Norway (EC ID: 14195), the Research Eth- 3. Measurable disease according to iRECIST.
ics Committee in Denmark (EC ID: H-18018750), the Norwegian Medical 4. Signed informed consent form.
Agency (ID: 16/11993), the Danish Medicines Agency (ID: 2018051636) 5. Women or men aged ≥18 years.
and institutional review boards. The trial was conducted according to 6. ECOG performance status of 0 or 1.
the guidelines of Good Clinical Practice, the principles of the World 7. In patients who have received (neo)adjuvant treatment with
Medical Association’s Declaration of Helsinki and the CONSORT 2010 anthracyclines or cyclophosphamide, a minimum of 12 months
guidelines. All patients provided written informed consent. Patients from treatment with anthracyclines or cyclophosphamide until
did not receive any financial compensation. The main contents of the relapse of disease is required.
protocol were published previously18. The last version of the protocol is 8. A maximum of one previous line with chemotherapy in the
available as Supplementary Material. Trial registration: NCT03164993; metastatic setting.
EudraCT: 2016-003570-40. 9. Female patients of childbearing potential should have a nega-
According to the initial protocol, the plan was to enroll 75 patients tive urine or serum pregnancy test within 7 days before receiv-
in the intention-to-treat (ITT) population and use this for all efficacy ing the first dose of study medication. If the urine test is positive
analyses as well as for the safety assessment. Owing to the small sample or cannot be confirmed as negative, a serum pregnancy test will
size and concerns that the scientific output of the study would be be required.
affected by patients leaving the trial before any therapeutic effect could 10. Female patients of childbearing potential should agree to
be expected, or without response evaluation, the protocol was amended remain abstinent (refrain from heterosexual intercourse) or
to allow the enrollment of 75 patients in a PP population. This decision use contraceptive methods that result in a failure rate of <1%
was based on the consideration that the sample size of 75 was marginal, per year, during the treatment period and for at least 5 months
compared to the statistical estimates and the aim of comparing between after the last dose of study therapy. A woman is considered to
the arms the proportion of long-term responders (>15-month PFS), be of childbearing potential if she is postmenarcheal, has not
which were expected to represent only a small number of patients (see reached a postmenopausal state (≥12 continuous months of
‘Statistical analysis’ below). Furthermore, we expected that patients amenorrhea with no identified cause other than menopause)
leaving the trial very early would be less informative for the efficacy and has not undergone surgical sterilization (removal of ovaries
and toxicity of atezolizumab, and we wanted to decide up-front (in this and/or uterus). Examples of contraceptive methods with a
double-blind trial) a definition of a more informative population. The failure rate of <1% per year include bilateral tubal ligation,
amendment was implemented in March 2018, after enrollment of nine male sterilization, proper use of hormonal contraceptives that
patients. The PP population comprised all patients who had received inhibit ovulation and hormone-releasing intrauterine devices.
>3 doses of atezolizumab/placebo and >2 doses of PLD and could be Periodic abstinence (for example, calendar, ovulation, symp-
evaluated for tumor response. PFS in the PD-L1positive PP population tothermal or postovulation methods) and withdrawal are not
was added as a co-primary outcome in a later amendment (protocol acceptable methods of contraception.
version 3.0, implemented in April 2021). Enrollment was stopped on 31 11. Male patients should agree to use an adequate method of con-
December 2021 owing to slow patient accrual, after the introduction of traception starting with the first dose of study therapy through
atezolizumab/nab-paclitaxel as standard therapy for PD-L1positive mTNBC. 3 months after the last dose of study therapy.
Data lock and unblinding were performed 6 months after enrollment 12. Able to swallow orally administrated medication.
of the last patients to allow for a reasonable observation time. All pro- 13. Adequate organ function, defined as absolute neutrophil count
tocol amendments were implemented before data lock and unblind- ≥1.20 × 109/L, lymphocyte count ≥0.50 × 109/L, thrombocytes
ing. No interim analyses were performed. Data lock was performed ≥80 × 109/L, hemoglobin ≥9 g/dl (≥ 5.6 mmol/L), creatinine
on 5 July 2022. ≤1.5× the upper limit of normal (ULN) or glomerular filtration
rate/creatinine clearance ≥40 ml/min, total bilirubin ≤1.5× ULN,
Patients AST and ALT ≤2.5× ULN (≤5× ULN for patients with liver metas-
Patients were enrolled in five hospitals in Norway and Denmark tases), albumin ≥2.5 g/dl and International Normalized Ratio
(Oslo University Hospital, Stavanger University Hospital, St. Olavs (INR) or prothrombin time ≤1.5× ULN for patients not receiving
University Hospital (Trondheim), Vejle Hospital and Rigshospitalet anticoagulant therapy.
(Copenhagen)).
The full inclusion and exclusion criteria are listed below. Exclusion criteria
1. Malignancies other than breast cancer within 5 years before
Inclusion criteria randomization, with the exception of those with a negligible
1. Metastatic or incurable locally advanced, histologically docu- risk of metastasis or death and treated with expected curative
mented TNBC (negative for HER2, ER and PR). HER2 negativity outcome (such as adequately treated carcinoma in situ of the
is defined as either of the following by local laboratory assess- cervix or basal or squamous cell skin cancer).
ment: in situ hybridization (ISH) non-amplified (ratio of ERBB2 2. Patients with known PD-L1positive TNBC, as assessed by the
to CEP17 <2.0 or single probe average ERBB2 gene copy number Ventana SP142 assay (IC ≥1%) and no previous chemotherapy
<4 signals per cell) or IHC 0 or IHC 1+ (if more than one test in the metastatic setting, should be offered standard therapy
result is available and not all results meet the inclusion criterion with nab-paclitaxel/atezolizumab outside of the trial, if they
definition, all results should be discussed with the principal had a disease-free interval of >12 months after previous (neo)
investigator to establish eligibility of the patient). ER and PR adjuvant chemotherapy, unless the patient for other reasons
Nature Medicine
should not receive nab-paclitaxel, according to own prefer- 13. Major surgical procedure within 14 days before randomization
ences, drug availability or recommendations by the treating or anticipation of the need for a major surgical procedure dur-
physician. A history of progression on taxanes in the neoadju- ing the course of the study other than for diagnosis. Placement
vant setting, or severe side effects from taxane therapy, may of central venous access catheter(s) is not considered a major
represent sufficient reason to offer the patient inclusion into the surgical procedure and is, therefore, permitted.
ALICE trial, if the physician considers that the patient should 14. A history of severe allergic, anaphylactic or other hypersensi-
receive anthracyclines rather than taxanes as first-line therapy tivity reactions to chimeric or humanized antibodies or fusion
for metastatic disease. If more than one TNBC biopsy has been proteins.
evaluated for PD-L1 by the SP142 assay, and the results differ, the 15. Known hypersensitivity or allergy to biopharmaceuticals pro-
patient’s PD-L1 status determination will be based on best clini- duced in Chinese hamster ovary cells or any component of the
cal judgment. atezolizumab formulation.
3. Spinal cord compression not definitively treated with surgery 16. Known hypersensitivity to doxorubicin or cyclophosphamide
and/or radiation or previously diagnosed and treated spinal or any of their excipients.
cord compression without evidence that disease has been clini- 17. A history of autoimmune disease that has required sys-
cally stable for >2 weeks before randomization. temic treatment in the past 2 years (that is, with use of
4. Known CNS disease, except for asymptomatic CNS metastases, disease-modifying agents, corticosteroids or immunosup-
provided that all of the following criteria are met: pressive drugs). Replacement therapy (for example, thyroxin,
a. Measurable disease outside the CNS. insulin or physiologic corticosteroid replacement therapy for
b. No metastases to mesencephalon, pons, medulla oblongata adrenal or pituitary insufficiency) is not considered a form of
or spinal cord. systemic treatment. Patients with eczema, psoriasis, lichen
c. No ongoing requirement for dexamethasone as therapy simplex chronicus or vitiligo with dermatologic manifesta-
for CNS disease. tions only (for example, no psoriatic arthritis) are permitted
d. No radiation of brain lesions within 7 days before provided that they meet all of the following conditions:
randomization. a. Rash must cover less than 10% of body surface area.
e. No leptomeningeal disease. b. Disease is well-controlled at baseline and only requiring
f. Patients with symptomatic CNS metastases must receive low-potency topical steroids.
radiation therapy and/or surgery for CNS metastases. After c. No acute exacerbations of underlying condition within the
treatment, these patients may be eligible, if all other criteria last 12 months (not requiring psoralen plus ultraviolet A radi-
are met. ation (PUVA), methotrexate, retinoids, biologic agents, oral
5. Uncontrolled pleural effusion, pericardial effusion or ascites. calcineurin inhibitors or high-potency or oral steroids).
Patients with indwelling catheters (for example, PleurX) are
allowed. 14. Undergone allogeneic stem cell or solid organ transplantation.
6. Uncontrolled tumor-related pain. Patients requiring narcotic 15. A history of idiopathic pulmonary fibrosis (including pneumo-
pain medication must be on a stable regimen at study en- nitis) drug-induced pneumonitis, organizing pneumonia (that
try. Symptomatic lesions (for example, bone metastases or is, bronchiolitis obliterans, cryptogenic organizing pneumonia)
metastases causing nerve impingement) amenable to pallia- or evidence of active pneumonitis on screening chest CT scan.
tive radiotherapy should be treated before randomization. History of radiation pneumonitis in the radiation field (fibrosis)
Asymptomatic metastatic lesions whose further growth would is permitted.
likely cause functional deficits or intractable pain (for example, 16. A positive test for HIV.
epidural metastasis that is not presently associated with spinal 17. Active hepatitis B (defined as having a positive hepatitis B sur-
cord compression) should be considered for locoregional face antigen (HBsAg) test at screening) or hepatitis C. Patients
therapy if appropriate before randomization. with past hepatitis B virus (HBV) infection or resolved HBV
7. Ionized calcium >1.2× ULN. The use of bisphosphonates is infection (defined as having a negative HBsAg test and a posi-
allowed. tive antibody to hepatitis B core antigen (anti-HBc) antibody
8. Pregnant or breastfeeding. test) are eligible. Patients positive for hepatitis C virus (HCV)
9. Evidence of significant uncontrolled concomitant disease that antibody are eligible only if polymerase chain reaction (PCR) is
could affect compliance with the protocol or interpretation of negative for HCV RNA.
results, including significant liver disease (such as cirrhosis, 18. Active tuberculosis.
uncontrolled major seizure disorder or superior vena cava 19. Currently receiving study therapy or has participated in a study of
syndrome). an investigational agent and received study therapy or used an in-
10. Significant cardiovascular disease, such as New York Heart As- vestigational device within 4 weeks of the first dose of treatment.
sociation (NYHA) cardiac disease (class II or higher), myocardial 20. Received treatment with immune checkpoint modulators,
infarction within 3 months before randomization, unstable including anti-CTLA-4, anti-PD-1 or anti-PD-L1 therapeutic
arrhythmias or unstable angina. Patients with a known left ven- antibodies.
tricular ejection fraction (LVEF) <40% will be excluded. Patients 21. Received treatment with systemic immunostimulatory agents
with known coronary artery disease, congestive heart failure (including, but not limited, to interferons or IL-2) within 4 weeks
not meeting the above criteria or LVEF <50% must be on a stable or five half-lives of the drug (whichever is shorter) before
medical regimen that is optimized in the opinion of the treating randomization.
physician, in consultation with a cardiologist if appropriate. 22. Received treatment with systemic corticosteroids or other
11. Severe infection within 14 days before randomization, requiring systemic immunosuppressive medications (including, but not
hospitalization. limited to, prednisone, dexamethasone, cyclophosphamide,
12. Received oral or intravenous (i.v.) antibiotics within 1 week be- azathioprine, methotrexate, thalidomide and anti-tumor necro-
fore cycle 1, day 1. Patients receiving routine antibiotic prophy- sis factor (TNF) agents) within 2 weeks before randomization
laxis (for example, to prevent chronic obstructive pulmonary or anticipated requirement for systemic immunosuppressive
disease exacerbation or for dental extraction) are eligible. medications during the trial.
Nature Medicine
a. Patients who have received acute, low-dose, systemic immu- Tumor evaluation by CT of the chest, abdomen and pelvis and a
nosuppressant medications (for example, a one-time dose of bone scan (MRI, PET or scintigraphy) was performed within 21 days of
dexamethasone for nausea) may be enrolled in the study. randomization and repeated every 8 weeks from the first day of treat-
b. Patients with a history of allergic reaction to i.v. contrast re- ment for the first 12 months and every 12 weeks thereafter. Response
quiring steroid pre-treatment should have baseline and sub- evaluation was done using iRECIST, a version of the RECIST modified
sequent tumor assessments performed using MRI. to capture the response patterns of immunotherapies25,26,36. In patients
c. The use of inhaled corticosteroids for chronic obstructive with unconfirmed disease progression by iRECIST (iUPD), continued
pulmonary disease, mineralocorticoids (for example, fludro- treatment was allowed until progression was confirmed (iCPD) in
cortisone) for patients with orthostatic hypotension and patients considered clinically stable according to the criteria in the
low-dose supplemental corticosteroids for adrenocortical iRECIST guideline.
insufficiency are allowed. Safety was monitored continuously throughout the study. All AEs
23. Received anti-cancer therapy (medical agents or radiation) were recorded and classified according to the Common Terminology
within 1 week before study cycle 1, day 1. Criteria for Adverse Events (CTCAE) version 4.0, regardless of relation
24. A history or current evidence of any condition, therapy or labo- to the study drugs. An independent safety monitoring committee
ratory abnormality that might confound the results of the trial, periodically reviewed the study safety data.
interfere with the patient’s participation for the full duration of
the trial or is not in the best interest of the patient to partici- Endpoints
pate, in the opinion of the treating investigator. The co-primary endpoints were safety and a descriptive comparison of
25. Known psychiatric or substance abuse disorders that would the PFS in the two arms. Safety was evaluated as the incidence, nature
interfere with cooperation and the requirements of the trial. and severity of AEs. Efficacy was evaluated as time-to-event and by
26. Any reason why, in the opinion of the investigator, the patient the proportion of patients without a PFS event 15 months after rand-
should not participate. This includes a careful evaluation of omization. The secondary efficacy outcomes included OS, ORR, DoR,
whether standard therapy is preferable to the study therapy, DRR and CBR. Owing to the limited sample size of this phase 2 trial,
for the individual patient. all comparisons of survival and response rates were of a descriptive
nature. The exploratory outcomes included analysis of PFS in prede-
fined subgroups defined by PD-L1 status, disease-free interval, line of
Randomization and masking treatment, metastatic sites, TMB, intrinsic breast cancer subtype and
Participants were randomly assigned (2:3) to receive either placebo TIS27, as well as the assessment of immunological response, evaluation
+ chemotherapy (placebo-chemo) or atezolizumab + chemotherapy of potential biomarkers for clinical response, toxicity and immune
(atezo-chemo). Randomization was done by the investigators using an response and developments in PROs (FQ, NRS pain intensity and EORTC
unstratified permuted block design through an interactive web response QLQ-C15-PAL37). PD-L1 expression, mutation load and immune gene
system implemented in the electronic case report form Viedoc v4 (Viedoc expression were predefined potential biomarkers. PFS was defined
Technologies). Randomization listings were generated by an independ- as the time from randomization to disease progression according to
ent statistician at Clinical Trial Unit Research Support Services, Oslo iRECIST, as assessed by the investigator, or death from any cause. OS
University Hospital, using Stata 14 software (StataCorp). Atezolizumab was defined as the time from randomization to death from any cause.
and matching placebo had identical packaging, labeling, appearance and ORR was defined as the proportion of patients with a partial or com-
administration schedules. Participants, investigators and study site per- plete response by iRECIST. DoR was defined as the time from the first
sonnel (other than pharmacy personnel involved in placebo/drug prepa- documentation of an objective response to the time of progression or
ration) were blinded to the treatment assignment until database lock. death, and DRR was defined as the proportion of patients with a DoR
of ≥6 months. CBR was defined as the proportion of patients who had
Procedures either an objective response or stable disease lasting at least until the
The chemotherapy regimen was the same for both treatment groups radiological evaluation at 24 weeks ± 7 days.
and consisted of PLD (20 mg/m2 i.v. on day 1 of each 14-day cycle)
and cyclophosphamide (50 mg by mouth (p.o.) daily in every other Statistical analysis
14-day cycle). Atezolizumab (840 mg) or placebo was given i.v. on The primary efficacy hypothesis was that atezo-chemo would lead
day 1 of each cycle. Treatment continued until disease progression, to prolonged PFS compared to placebo-chemo in mTNBC. The initial
unacceptable toxicity, withdrawal of consent, investigator’s decision sample size was calculated based on a two-sided test with an alpha
or up to 24 months. Treatment beyond 24 months was allowed for level of 10% and a power of 80% to detect an absolute reduction of
selected patients based on a risk/benefit analysis performed by the 15% in the proportion without progression or death at 15 months
study management. in the atezo-chemo arm compared to the placebo-chemo arm. The
The baseline assessment included a medical history, a full clini- estimated sample size was 75 patients (45 in atezo-chemo and 30 in
cal examination, a complete blood count (CBC), a comprehensive placebo-chemo).
blood chemistry panel and a cardiac assessment by electrocardiogram Safety was evaluated in the FAS, which included all patients who
and echocardiography. CBC and blood chemistry were repeated on received one dose of any of the investigational medical products (IMPs).
day 1 of each treatment cycle, and a full clinical examination, electro- The primary efficacy analysis was done in the PP population and in
cardiogram and echocardiography were repeated every 4th cycle. the PD-L1positive PP subpopulation. Secondary efficacy analyses were
Patient-reported outcomes (PROs) were recorded on day 1 of cycles performed in the FAS, PP, PD-L1positive PP and PD-L1negative PP populations.
1, 5, 9, 13 and 25 and at the time of treatment discontinuation, using For patients who discontinued the trial without progression or death,
the Chalder fatigue questionnaire (FQ), the European Organization PFS was censored at the date of the last tumor assessment. Patients
for Research and Treatment of Cancer Quality of Life Questionnaire who could not be evaluated for tumor response were censored 1 day
Core 15 Palliative Care (EORTC QLQ-C15-PAL) and the NRS pain scale. after randomization. Patients who were alive at the time of data lock
After consent, tumor tissue, blood, urine and feces samples were col- were censored at the last timepoint at which they were confirmed
lected for a research biobank at several timepoints before, during and to be alive.
after trial treatment. Extraction of PBMCs was only done at the three Changes in the EORTC QLQ-C15 global health status/quality of life
Norwegian study sites. score were analyzed by time-to-deterioration and mean change from
Nature Medicine
baseline with 95% CI. Deterioration was defined as a score reduction of MarkDuplicatesSpark, followed by calibration of base qualities by
≥20 points from baseline. Patients with missing baseline values were GATK BaseRecalibrator and GATK ApplyBQSR. The depth of sequenc-
not included in these analyses, and patients with no follow-up values ing coverage was estimated on the recalibrated .bam files by GATK
were censored for deterioration on day 1. Patients who missed two DepthOfCoverage 4.2.5.0 (ref. 42). Structural variants were detected
consecutive assessments were censored for deterioration at the last by Manta43. Somatic variant calling was performed by Mutect2 (ref. 42)
assessment before this. and Strelka 2 (ref. 44) to detect single-nucleotide variations and small
Comparisons of time-to-event outcomes between groups were insertions and deletions. Somatic variants detected in both variant call-
made using the Kaplan–Meier method, with P values calculated using ers were selected by BCFTools 1.9 (ref. 45) and used in further analysis.
the log-rank method. The median follow-up time was calculated using
the reverse Kaplan–Meier method, with censoring for OS as the event. TMB. TMB was calculated following the Uniform TMB estimation
HRs with 95% CIs were estimated using the Cox proportional haz- method proposed by Merino et al.46. In brief, the median sequenc-
ards model. The Efron method was used for handling of tied event ing depth of the tumor samples was >300×. Mutations were filtered
times. CIs for proportions were estimated by the Wilson score method. before TMB calculation to include only coding, non-synonymous
Circulating immune cell populations across different timepoints were single-nucleotide polymorphisms and insertions and deletions. In
compared by fold change, with P values calculated using Wilcoxon addition, mutations with a tumor allele frequency ≥5%, tumor read
matched-pairs signed-rank test. All reported P values are two-sided. depth ≥25 and a minimum of three supporting reads were included
Statistical analyses were done in R version 4.1.3 (ref. 38). in the calculation. Samples were excluded from further analysis if
Time-to-event analyses were performed using R packages survival more than 50% of variants were removed by the quality filters. Data
(version 3.4) and survminer (version 0.4.9). from 53/68 patients in the FAS were available after quality control.
The approximate size of the human exome (35.01 Mb) was used as a
PD-L1 scoring of tumors denominator in the TMB calculation. For patients with more than one
The PD-L1 status of tumor samples was assessed by experienced pathol- sample, the highest TMB was considered representative.
ogists who were blinded for treatment group and clinical outcome,
according to VENTANA PD-L1 (SP142) Assay Interpretation Guide for Gene expression analysis (NanoString)
Triple-Negative Breast Carcinoma39. Results were reported as the per- The Research-use-only Human nCounter Breast Cancer 360 (BC360)
centage of the tumor area made up of antibody-stained immune cells, panel identifies 776 genes of interest using optical barcodes consisting
and tumors were considered PD-L1positive if this value was ≥1%. of capture and reporter probes that hybridize with mRNA targets. This
enables digital counting of individual mRNA in each sample. The BC360
Assessment of TILs panel and kit were provided by NanoString Technologies.
Assessment of TILs was done by examination of hematoxylin and eosin Tumor mRNA isolation from 45 patients was performed using
(H&E)-stained sections of tumor biopsies collected during the study 10-µm FFPE tissue sections, mounted on histology slides and deparaffi-
screening period by two experienced breast cancer pathologists, who nized using (R)-(+)-Limonene (Merck, 183164) and rehydrated absolute
were blinded for treatment group and clinical outcome. Scoring was ethanol and air-dried for a minimum of 15 minutes. Macrodissection
performed based on the presence and abundance of lymphocytes was done with an H&E guide slide, collecting only tumor tissue from
within the borders of the invasive tumor. The scoring system included the sections. RNA isolation of the collected tumor tissue was done
two categories for low infiltration (score 0–1) and two categories for using the High Pure FFPET RNA Isolation Kit (Roche, 06650775001)
high infiltration (score 2–3). and protocol, with the 55 °C incubation time increased to 2 hours.
Hybridization of mRNA and capture and reporter probes was done
Assessment of TMB using the nCounter XT Assay protocol (NanoString Technologies)
DNA extraction from tumor biopsies and blood genomic DNA. and incubated at 65 °C for 20 hours. Post-hybridization processing
Fresh frozen tumor biopsies were quickly disrupted and homogenized was performed on the nCounter Prep Station (NanoString Technolo-
using TissueLyzer (Qiagen). Subsequently, DNA/RNA/proteins were gies) at the ‘high sensitivity’ setting. The nCounter Digital Analyzer
isolated from the tumor lysate using the AllPrep DNA/RNA/Protein (NanoString Technologies) was used for digital counting and data col-
Mini Kit (80004, Qiagen). Genomic DNA from blood was isolated using lection with field of view (FOV) set to 555. Best practice for the BC360
FlexiGene DNA Kit (51206, Qiagen). Each step was carried out as per the assays was followed. Preliminary quality control was performed using
manufacturer’s instructions. DNA quality was measured by a NanoDrop the RCCCollector (NanoString Technologies) tool before further data
spectrophotometer (Thermo Fisher Scientific), and the DNA concentra- processing. The determination of intrinsic breast cancer subtypes
tion was determined by a Qubit fluorometer (Thermo Fisher Scientific) and calculation of TIS were done by NanoString Technologies. As the
before further processing and analysis. distribution and clinically relevant cutoff values are not established
for TIS, the values were scaled to a mean of 0 and a standard deviation
Whole-exome sequencing. The samples were further processed of 1 before presentation, to achieve numerical values that reflect the
and sequenced by the Genomics Core Facility, Institute for Cancer distribution of TIS in the study population.
Research, Oslo University Hospital. In brief, 50 ng of DNA was processed
following the manufacturer’s recommendations for library preparation Flow cytometry analysis of peripheral blood immune cell
and target enrichment using Twist Comprehensive Human Exome and subsets
Twist Library Preparation EF Kit (Twist Bioscience). Subsequently, the PBMCs collected at screening and day 1 of the fifth treatment cycle
samples were sequenced by paired-end sequencing 2×151 bp on the (8 weeks after start of therapy) were isolated from whole blood using
Illumina NovaSeq 6000 system. LymphoPrep Cell Separation Media (Abbott Rapid Diagnostics), frozen
and stored in liquid nitrogen until assessed for immune cell populations
Variant calling. Germline and somatic variant calling was performed by flow cytometry. Paired samples from each patient were stained
by the Bioinformatics Core Facility, Institute for Cancer Research, and analyzed on the same day. The experiments were not replicated.
Oslo University Hospital, and analyzed by the nf-core/Sarek 2.7 pipe- PBMCs were initially incubated with antibodies for the surface markers
line40. In brief, fastq files containing the sequencing raw data were CD3-BUV395, CD8-FITC, CD4-BV510, γδ-TCR-BUV737, CD19-BUV563,
aligned to the human GRCh38 reference genome by BWA-mem41. CD56-Alexa Fluor 647 (BD Biosciences), CD25-BV605 (BioLegend)
After sequence alignment, the duplicate reads were marked by GATK and Fixable Viability Dye eFluor780 (Thermo Fisher Scientific). After
Nature Medicine
fixation and permeabilization using eBioscience Foxp3/Transcription Acknowledgements

Factor Staining Buffer Set (Thermo Fisher Scientific), PBMCs were incu- The sponsor of this trial was Oslo University Hospital (OUH). The authors
bated with an antibody for the intracellular antigen Foxp3-PE (Thermo thank the patients, their families and the user representative. We also
Fisher Scientific). Samples were acquired using BD FACSymphony A5 thank the personnel at the study hospitals for their contributions and
flow cytometer (BD Biosciences), and the data were analyzed with excellent patient care. Special thanks to E. Borgen, E.-J. Sande, T. Guren,
FlowJo (Tree Star) and GraphPad Prism (GraphPad Software). I. Abdi, M. Stensland, G. Sæter, M. Nguyen, T. Arntsen, G. Skorstad, N.
Moharrami, L. Julsrud and G. Hagen. We are also grateful to T. Ahlquist
Role of the funding source and the team at Roche, to the OUH Bioinformatics Core Facility, the OUH
This study was supported by grants from the Norwegian Health Region Genomics Core Facility, the OUH Flow Cytometry Core Facility and to
South-East and the Norwegian Cancer Society/Norwegian Breast Can- the contributing units at the OUH Department of Pathology, headed by
cer Society. Roche supported the study with free drug (atezolizumab), M. Førsund, I. J. Ryen and A. Renolen. Roche supported the study with
free SP142 kits and a funding contribution. NanoString supported the free drug (atezolizumab), free SP142 assays and a funding contribution.
study with free assays and analyses. Roche provided critical reviews of NanoString supported the study with free assays and analyses. The
the protocol. None of the funders had any role in the study design, data study was also supported by grants from the Norwegian Health Region
collection, data analysis, data interpretation or writing of the report. South-East (grants 2017100 to J.A.K. and 2017122 to B.N.) and grants
Roche and NanoString reviewed the first version of this manuscript. from the Norwegian Cancer Society/Norwegian Breast Cancer Society
(grants 182632 and 214972 to J.A.K).
Reporting summary
Further information on research design is available in the Nature Port- Author contributions
folio Reporting Summary linked to this article. J.A.K. was the coordinating investigator of the study and was
responsible for study conception and design and acquisition of
Data availability funding and approvals. He also contributed as an investigator and
Any request for raw or analyzed data will be reviewed by the study team, medical monitor and to patient recruitment, data collection and data
and a response can be expected within 14 days. The data generated in interpretation. The manuscript was written by J.A.K and A.H.R., with
this study are subject to patient confidentiality, and the transfer of contributions from all authors. A.H.R. and N.K.A. were investigators
data or materials will require approval from the Data Privacy Officer at OUH and medical monitors for the other sites and contributed
and the institutional review board at Oslo University Hospital and to patient recruitment and data acquisition, curation and analysis.
from the Regional Committee for Medical and Health Research Ethics C.A.B., B.G., E.H.J. and S.X.R were the principal investigators at their
South-East Norway and the Research Ethics Committee in Denmark. respective study sites. C.A.B. was coordinating investigator for
Any shared data will be de-identified. Requests should be made to the Denmark. R.S.F. was the study statistician. B.N. contributed to the
corresponding author ( jonky@ous-hf.no). study conception and design and to patient recruitment and data
interpretation. T.J. performed radiological assessments. R.R.M. was
References an investigator and medical monitor. H.G.R., J.L. and Ø.G. were study
36. Eisenhauer, E. A. et al. New response evaluation criteria in solid pathologists. S.C., R.R.L. and C.D. performed translational laboratory
tumours: revised RECIST guideline (version 1.1). Eur. J. Cancer 45, analyses. All authors approved the final version of this manuscript.
228–247 (2009).
37. Groenvold, M. et al. The development of the EORTC Competing interests
QLQ-C15-PAL: a shortened questionnaire for cancer patients in J.A.K. has, in the last 3 years, received research support from
palliative care. Eur. J. Cancer 42, 55–64 (2006). Bristol Myers Squibb, F. Hoffmann-La Roche, NanoString and NEC
38. R Core Team. R: a language and environment for statistical OncoImmunity and has previously received advisory board/lecture
computing (R Foundation for Statistical Computing, 2022). honoraria from pharmaceutical companies, including Roche. B.G. has,
39. Vennapusa, B. et al. Development of a PD-L1 complementary in the last 3 years, received honoraria for advisory boards from Eli Lilly,
diagnostic immunohistochemistry assay (SP142) for atezolizumab. Gilead Sciences, Daiichi Sankyo, Roche and Pierre Fabre. S.X.R. and J.L.
Appl. Immunohistochem. Mol. Morphol. 27, 92–100 (2019). have received lecture honoraria from Pfizer, Novartis and AstraZeneca.
40. Garcia, M. et al. Sarek: a portable workflow for whole-genome C.B. has received travel grants or advisory board/consultant honoraria
sequencing analysis of germline and somatic variants. F1000Res. from Eli Lilly, Pfizer, AstraZeneca, Gilead Sciences, MSD and Daiichi
9, 63 (2020). Sankyo. All other authors declare no competing interests.
41. Li, H. & Durbin, R. Fast and accurate long-read alignment
with Burrows–Wheeler transform. Bioinformatics 26, 589–595 Additional information
(2010). Extended data is available for this paper at https://doi.org/10.1038/
42. Van der Auwera, G. A. & O’Connor, B. D. Genomics in the Cloud: s41591-022-02126-1.
Using Docker, GATK, and WDL in Terra 1st ed (O’Reilly Media,
2020). Supplementary information The online version contains supplementary
43. Chen, X. et al. Manta: rapid detection of structural variants material available at https://doi.org/10.1038/s41591-022-02126-1.
and indels for germline and cancer sequencing applications.
Bioinformatics 32, 1220–1222 (2016). Correspondence and requests for materials should be addressed to
44. Kim, S. et al. Strelka2: fast and accurate calling of germline and Jon Amund Kyte.
somatic variants. Nat. Methods 15, 591–594 (2018).
45. Danecek, P. et al. Twelve years of SAMtools and BCFtools. Peer review information Nature Medicine thanks the anonymous
Gigascience 10, giab008 (2021). reviewers for their contribution to the peer review of this work.
46. Merino, D. M. et al. Establishing guidelines to harmonize tumor Primary handling editor: Saheli Sadanand, in collaboration with the
mutational burden (TMB): in silico assessment of variation in TMB Nature Medicine team.
quantification across diagnostic platforms: phase I of the Friends
of Cancer Research TMB Harmonization Project. J. Immunother. Reprints and permissions information is available at
Cancer 8, e000147 (2020). www.nature.com/reprints.
Nature Medicine
Extended Data Fig. 1 | Kaplan-Meier plots of progression-free and overall with confidence intervals were calculated using the Cox proportional hazards
survival by PD-L1 status (FAS). Progression-free survival by iRECIST in the method. PD-L1, Programmed death-ligand 1; FAS, full analysis set; HR, hazard
PD-L1 positive (a) and negative (b) subgroups of the full analysis set. Overall ratio; CI, confidence interval.
survival in the PD-L1 positive (c) and negative (d) full analysis set. Hazard ratios
Nature Medicine
Extended Data Fig. 2 | Swimmer plot. Duration of treatment and treatment responses in the PP population. The length of the bars represent the time from
randomization to disease progression (iRECIST) or censoring. Symbols denote the first time a treatment response (other than progression) is documented, patients
censored for progression, and the last date of study treatment.
Nature Medicine
Extended Data Fig. 3 | Kaplan-Meier plots of progression-free survival with analysis of pre-treatment biopsies on the NanoString BC360 assay. All patients in
high and low Tumor Inflammation Signature. Progression-free survival the full analysis set for whom samples were available were included in the analysis
by iRECIST in the subpopulation with high (> 66th percentile) and low (≤ 66th (placebo-chemo n = 17, atezo-chemo n = 27). Hazard ratios (HR) with confidence
percentile) tumor inflammation signature (TIS) in the atezo-chemo (a) and intervals (CI) were calculated using the Cox proportional hazards method.
placebo-chemo (b) group. Tumor Inflammation Signature was assessed by
Nature Medicine
Extended Data Fig. 4 | Correlation between Tumor Inflammation Signature pre-treatment biopsies on the NanoString BC360 assay. All patients for whom
and other biomarkers. a: Tumor Inflammation Signature (TIS) in patients a suitable archival biopsy was available were included in the analysis (n = 44).
with PD-L1 negative and positive tumors. Both analyses were done on archival PD-L1 expression (SP142 ASSAY) and TIL were assessed once for each biopsy, by
pre-study biopsies. For 28 of the 44 patients, the analyses were done in the two experienced breast cancer pathologists that were blinded for treatment arm
same tumor sample, while 16 of the tumors analyzed for PD-L1 did not contain and clinical outcome. Two-sided p-values were calculated using the Wilcoxon
enough material to be analyzed for TIS. For these biopsies, the Nanostring assay rank sum test. In the box plots, the center line represents the median value and
was performed on a different archival biopsy. b: TIS in patients with low and the box limits represent the upper and lower quartiles. The whiskers extends to
high infiltration of lymphocytes in tumor. Tumor-infiltrating lymphocyte (TIL) the extreme values, omitting outliers extending > 1.5 x IQR from the box limits,
abundance was assessed in study-specific pre-treatment biopsies, whereas TIS these are shown in the dot plot. TIS, Tumor Inflammation Signature; PD-L1,
was analyzed in archival biopsies, due to the limited volume of the screening programmed death-ligand 1; TIL, tumor-infiltrating lymphocytes.
biopsies. Tumor Inflammation Signature was assessed by a single analysis of
Nature Medicine
Extended Data Table 1 | Summary of response variables. Proportions of patients with objective response, clinical benefit,
and durable response (> 6 months) as assessed by iRECIST. Confidence intervals for proportions were calculated using the
Wilson score method. ORR, objective response rate; CBR, clinical benefit rate; DRR, durable response rate; PP, per-protocol
population; FAS, full analysis set; PD-L1, programmed death-ligand 1
Nature Medicine
Extended Data Table 2 | Multivariable Cox regression. The hazard ratio (HR) and associated p-value for progression-free
survival in the atezo-chemo arm versus the placebo-chemo arm is shown unadjusted and after inclusion of selected
baseline characteristics. Age is treated as a continuous variable and all other covariates as categorical. Hazard ratios and
two-sided p-values were calculated using the Cox proportional hazards method. ECOG, Eastern Cooperative Oncology
Group; PD-L1, programmed death-ligand 1
Nature Medicine
Extended Data Table 3 | Biopsy sites. Biopsies used in the analysis of Tumor Inflammation Signature (TIS), programmed
death-ligand 1 (PD-L1) expression and tumor-infiltrating lymphocytes (TIL), sorted by organ / tissue biopsied
Nature Medicine
Extended Data Table 4 | Proportion of patients responding to EORTC QLQ-C15-PAL The table shows the proportion of
patients that responded to the quality of life item in the EORTC QLQ-C15-PAL questionnaire at each of the protocol-defined
time points, as a percentage of all patients who remained in the study until that time point
Nature Medicine
≥ ≥
γδ
≥
γδ

Atezolizumab Plus Anthracycline-Based Chemotherapy in Metastatic Triple-Negative Breast Cancer: The Randomized, Double-Blind Phase 2b ALICE Trial

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Atezolizumab Plus Anthracycline-Based Chemotherapy in Metastatic Triple-Negative Breast Cancer: The Randomized, Double-Blind Phase 2b ALICE Trial

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Atezolizumab Plus Anthracycline-Based Chemotherapy in Metastatic Triple-Negative Breast Cancer: The Randomized, Double-Blind Phase 2b ALICE Trial

Uploaded by

Copyright:

Available Formats

nature medicine

Atezolizumab plus anthracycline-based

Nature Medicine | Volume 28 | December 2022 | 2573–2583 2573

Nature Medicine | Volume 28 | December 2022 | 2573–2583 2574

Assessed for eligibility (n = 79)

Assigned to chemo + placebo (n = 28) Allocation Assigned to chemo + atezo (n = 42)

Discontinued intervention (n = 28) Follow-up Discontinued intervention (n = 38)

Still on study treatment (n = 2)

Analyzed populations: Analysis Analyzed populations:

Per-protocol population (n = 23) Per-protocol population (n = 36)

Discontinued the trial (n = 24) Discontinued the trial (n = 28)

In survival follow-up (n = 4) In survival follow-up (n = 10)

Nature Medicine | Volume 28 | December 2022 | 2573–2583 2575

Table 1 | Demographics and baseline characteristics of the study cohort

Age 0.16 0.17

Gender 0.40 0.42

Female 28 (100%) 39 (97.5%) 23 (100%) 35 (97.2%)

Male 0 (0%) 1 (2.5%) 0 (0%) 1 (2.8%)

ECOG performance status 0.50 0.99

0 21 (75.0%) 27 (67.5%) 16 (69.6%) 25 (69.4%)

1 7 (25.0%) 13 (32.5%) 7 (30.4%) 11 (30.6%)

De novo metastatic disease 0.74 0.93

Yes 8 (28.6%) 10 (25.0%) 6 (26.1%) 9 (25.0%)

No 20 (71.4%) 30 (75.0%) 17 (73.9%) 27 (75.0%)

Bone metastases 0.23 0.18

Yes 16 (57.1%) 17 (42.5%) 13 (56.5%) 14 (38.9%)

No 12 (42.9%) 23 (57.5%) 10 (43.5%) 22 (61.1%)

Liver metastases 0.17 0.36

Yes 13 (46.4%) 12 (30.0%) 9 (39.1%) 10 (27.8%)

No 15 (53.6%) 28 (70.0%) 14 (60.9%) 26 (72.2%)

Lung metastases 0.44 0.69

Yes 10 (35.7%) 18 (45.0%) 9 (39.1%) 16 (44.4%)

No 18 (64.3%) 22 (55.0%) 14 (60.9%) 20 (55.6%)

Lymph node metastases 0.49 0.94

Yes 13 (46.4%) 22 (55.0%) 13 (56.5%) 20 (55.6%)

No 15 (53.6%) 18 (45.0%) 10 (43.5%) 16 (44.4%)

CNS metastases 0.80 0.75

Yes 1 (3.6%) 1 (2.5%) 1 (4.3%) 1 (2.8%)

No 27 (96.4%) 39 (97.5%) 22 (95.7%) 35 (97.2%)

Number of metastatic sites 0.24 0.12

≤2 15 (53.6%) 27 (67.5%) 12 (52.2%) 26 (72.2%)

>2 13 (46.4%) 13 (32.5%) 11 (47.8%) 10 (27.8%)

>3 3 (10.7%) 7 (17.5%) 3 (13.0%) 5 (13.9%)

Line of chemotherapy 0.81 0.37

1st 16 (57.1%) 24 (60.0%) 12 (52.2%) 23 (63.9%)

2nd 12 (42.9%) 16 (40.0%) 11 (47.8%) 13 (36.1%)

Previous anthracycline treatment 0.90 0.71

Yes 20 (71.4%) 28 (70.0%) 17 (73.9%) 25 (69.4%)

No 8 (28.6%) 12 (30.0%) 6 (26.1%) 11 (30.6%)

PD-L1 status 0.22 0.22

Negative 17 (60.7%) 19 (47.5%) 14 (60.9%) 17 (47.2%)

Nature Medicine | Volume 28 | December 2022 | 2573–2583 2576

Table 1 (continued) | Demographics and baseline characteristics of the study cohort

Positive 10 (35.7%) 21 (52.5%) 8 (34.8%) 19 (52.8%)

Missing 1 (3.6%) 0 (0%) 1 (4.3%) 0 (0%)

Intrinsic breast cancer subtype 0.44 0.48

Luminal A 2 (7.1%) 0 (0%) 2 (8.7%) 0 (0%)