CUPCAKE is a randomized, non-comparative, multicenter, proof-of-concept phase II trial, using the Trials within Cohorts concept(1) to assess the clinical utility of ctDNA monitoring combined with 68Ga-FAPI-46-PET-CT imaging upon ctDNA detection for the surveillance of patients with a non-metastatic TNBC at high risk of relapse.

The study has two steps. In Step 1, patients who have completed the treatments for a localized TNBC will undergo ctDNA monitoring every ~4 months (± 2 weeks). In Step 2, patients for whom ctDNA will be detected will then be randomized between an observation arm, in which monitoring will continue until the detection of a clinical relapse, and an experimental arm, in which the ctDNA detection will be revealed to both the patient and the clinician: patients will then undergo a 18F-FDG PET-CT and a 68Ga-FAPI-46-PET-CT, in addition to whatever workup the investigator will deem necessary.

The CUPCAKE trial will follow the Trials within Cohorts (TwiCs) approach. Non-metastatic TNBC patients at high risk of relapse will be included, after having signed a written informed consent, in a cohort allowing them to be followed by ctDNA monitoring every 4 months.

For each patient included, a ctDNA detection assay will be performed in blood samples every 4 months for a maximum of 24 months, while extra-plasma will be banked. ctDNA results will be available with a turnaround time of less than 3 weeks. When negative, ctDNA detection results will not be disclosed to patients nor clinicians.

First line therapy will not be started until a metastatic relapse has been found by imagining: no treatment will be started in the sole basis of a positive ctDNA test.

If, at any timepoint, ctDNA is detected (molecular relapse), patients will be randomized in a 1:1 ratio.

• In the experimental arm, patients and their treating physician will be made aware of the molecular relapse (positive ctDNA detection results). To locate metastatic deposits, patients will be offered to undergo a whole-body imaging with 18F-FDG PET-CT and 68Ga-FAPI-46-PET-CT, in addition to any other workup considered as relevant by their treating physician. If/when a clinical/radiological relapse is observed, the patient performance status will be registered (secondary objective) and systemic or local treatments will be decided by physicians. These treatments could be informed by the genetic landscape of the relapse, assessed by ctDNA.
• In the control arm, patients and their treating physician will not be made aware of the molecular relapse and will continue the standard surveillance with repeated ctDNA test every 4 months (blinded) for a maximum of 24 months from enrolement in the study. At the time of the clinical/radiological diagnosis of relapse, similar procedures will be performed (18F-FDG PET-CT, 68Ga-FAPI-46-PET-CT, and tumor genetic landscape assessment by ctDNA analysis).

Inclusion Criteria:

1. Patients must have signed a written informed consent before inclusion

2. Patients must be female ≥ 18 years old

3. Patients diagnosed with a non-metastatic TNBC (ER & PR <10%, HER2- per ASCO/CAP guidelines). Patients must have been previously evaluated by a 18F-FDG PET-CT or a bone scintigraphy combined with a thorax, abdomen and pelvis CT scan with contrast

4. Patients who have undergone surgery with curative intent for their non-metastatic TNBC. Surgery must have been performed between 3 to 9 months before inclusion. Patients must have initiated their adjuvant therapy, whenever indicated, since at least 12 weeks. For patients receiving an experimental adjuvant treatment in a clinical trial, any intervention planned as part of this trial must be completed before inclusion.

5. High-risk primary tumor, defined as:

   1. Lack of pathological complete response after neoadjuvant chemotherapy (RCB I, II or III; RCB I being capped to a maximum of 30% of included patients) OR, in the absence of neoadjuvant chemotherapy,
   2. Stage IIB-III (i.e., T2N1, any T3-T4, any N2-3) OR
   3. Any loco-regional relapse occurring after a prior ipsilateral, curatively treated TNBC

6. No sign of local or distant relapse, as per investigator assessment

7. Performance status < 2

8. Available FFPE tumor block with > 10% cellularity or 11 tumor sections with >10% cellularity

9. Patient able to comply with protocol requirements

10. Patients covered by a health insurance

Exclusion Criteria:

1. Any uncontrolled disease, metabolic dysfunction, physical examination finding, or clinical laboratory finding or any other medical condition that, in the opinion of the investigator, interferes with the trial procedures

2. Male participants

3. Patients with altered mental status or psychiatric disorder that, in the opinion of the investigator, would preclude a valid patient informed consent.

4. Patients who have difficulty undergoing trial procedures for geographic, social or psychological reasons

5. Person deprived of liberty or under guardianship

6. History of another primary malignancy except for the following :

   1. Basal cell carcinoma or any in situ carcinoma treated with curative intent
   2. Any stage I-II malignancy treated with curative intent with no evidence of active disease in the last five years

7. For step #2 (randomization after ctDNA detection): clinical/radiological metastatic relapse before the detection of the molecular relapse.

Investigators will share de-identified data sets with interested researchers, educators or clinicians. Materials generated under the project will be disseminated in accrodance with Institut Curie policies.

Data requests can be submitted starting 9 months after article publication and will be made accessible for up to 12 months.

Access to trial individual participant data can be requested by qualified researchers engaging in independent scientific reserach, and will be provided following review and approval of a research proposal and Statistical Analysis Plan (SAP) and execution of a data sharing agreement (DSA).

Upon detection of ctDNA during Step1, patients will be centrally assigned to the control or the experimental arm (1:1) using the eCRF.

Randomization will be stratified by RCB score (I/II vs. III/NA because no neoadjuvant chemotherapy) and by time to molecular relapse (< 12 months vs. ≥ 12 months between surgery and molecular relapse).