Radiological Prediction of Response to Neoadjuvant Chemoimmunotherapy for Triple-negative Breast Cancer

Triple-negative breast cancers (TNBC) account for approximately 15% of breast cancers, representing about 8,800 cases per year. The therapeutic management of localized TNBC is based on a multimodal strategy: neoadjuvant or adjuvant chemotherapy (depending on tumor stage), surgery, and radiotherapy. The advantage of a neoadjuvant approach lies in the ability to assess each patient's response to treatment. This response is defined by the Residual Cancer Burden (RCB) classification, which includes four stages, from 0 to 3. A high stage (RCB 3) indicates resistance to neoadjuvant treatment and is associated with a poorer prognosis. Conversely, RCB 0 (also known as a complete pathological response or pCR) corresponds to the absence of any remaining tumor cells and is associated with an excellent prognosis.

Since the results of the KEYNOTE-522 trial, neoadjuvant treatments have been based on sequential chemo-immunotherapy with pembrolizumab every three weeks, combined in sequence 1) with 12 weekly cycles of carboplatin and paclitaxel, and in sequence 2) with four cycles of epirubicin and cyclophosphamide every three weeks. Pembrolizumab is continued after surgery for a total of one year of exposure (neoadjuvant + adjuvant treatment). The addition of pembrolizumab to chemotherapy in this study increased the pCR rate (defined as the absence of invasive residuals: ypT0N0 or ypTisN0) to 64.8% compared to 51.2% in the chemotherapy/placebo arm. However, this clinical benefit is associated with an increase in side effects related to immunotherapy exposure.

A standard radiological assessment before neoadjuvant treatment is essential to define the optimal therapeutic strategy. This typically includes mammography and ultrasound, often combined with either MRI or contrast-enhanced mammography. Before surgery, imaging is repeated to guide the surgical procedure. Additionally, an 18F-FDG PET/CT scan is performed before treatment initiation to confirm the absence of distant metastases.

Recent data have questioned the role of surgery in patients with TNBC, given the high pCR rates (over 50%). Considering the chemosensitivity of these tumors and the consequences of surgery (aesthetic/psychological impact, functional issues such as pain or limited joint mobility), the de-escalation of surgery is being debated. To explore this, it is necessary to identify a strategy that can reliably predict histological response to treatment, particularly complete pathological response (defined as the absence of invasive residuals: ypT0N0 or ypTis).

Beyond response prediction, several questions have arisen following the introduction of immunotherapy in TNBC neoadjuvant treatment. In terms of efficacy, no reliable biomarker currently exists in routine practice to select patients who would benefit the most from this chemo-immunotherapy. Paradoxically, although PD-1/PD-L1 status has been shown to predict pembrolizumab response in metastatic TNBC, it has not demonstrated its utility in the neoadjuvant setting. Tumor mutational burden (TMB) and the number of tumor-infiltrating lymphocytes (TILs) are associated with better prognoses but do not allow for precise patient stratification. Notably, TILs correlate with higher pCR rates (in TNBC and HER2-amplified breast cancers receiving neoadjuvant therapy) and are partially predictive of overall survival. While extensive research has examined the tumor and its microenvironment to predict treatment efficacy, no marker is currently reliable enough for routine clinical use.

Immunotherapy use is also associated with frequent toxicities, which may lead to treatment discontinuation. Again, no marker can predict the occurrence of immune-related adverse events. A better understanding of the biological mechanisms underlying treatment response and immunotherapy-related toxicities could help identify new predictive markers.

The objective of our study is to determine which combination of radiological factors most accurately assesses tumor response to neoadjuvant treatment, by comparing radiological findings with objective histological response data (current gold standard - ypT0N0 or ypTisN0). Regarding biological data, we aim to identify transcriptomic markers in circulating PBMCs at baseline, as well as profile variations during treatment, associated with better histological outcomes and the occurrence of toxicities. The goal is to identify initial profiles or transcriptomic changes indicating the immune system's activation, leading to an anti-tumor effect or immune-related toxicities.