155P - Use of machine learning for the identification of molecular biomarkers to predict response to neoadjuvant chemotherapy in locally advanced breast cancer patients

Background

Breast cancer is a public health issue both in the world and in Mexico, where 70% of patients with the disease are diagnosed in later stages. Because of this, at Instituto Nacional de Cancerología (The National Cancer Institute), neoadjuvant chemotherapy has become the standard for managing locally advanced breast cancer. However, studies show that less than 50% of patients present a complete pathologic response to this regime, which is why is essential to identify molecular biomarkers capable of predicting resistance to this therapy. Thus, the use of machine learning techniques, together with transcriptome analyses could contribute to the management of oncological treatments for this stage of the disease.

Methods

We included 118 samples corresponding to patients classified in the Luminal A, Luminal B, HER2, and Triple Negative subtypes in locally advanced stages who did not present metastasis or had not received treatment at the time of biopsy in the period 2012 - 2015. RNA was extracted from fresh tumor biopsies prior to the administration of neoadjuvant chemotherapy. After obtaining the results of the clinical monitoring of response to the treatment, the patient samples were grouped as "control" if they presented complete pathological response to the treatment, and "case" if they presented resistance. Subsequently, the total RNA of the samples was processed by massive RNA sequencing (RNA-seq) and a differential expression analysis was performed using DESeq2. Finally, machine learning algorithms were used to select a group of genes capable of classifying patients into "resistant" and "non-resistant" and a panel of response biomarkers was proposed.

Results

A set of mRNAs and lncRNAs capable of discriminating the outcome of breast cancer patients to neoadjuvant chemotherapy as well as the enriched pathways associsted with them was identified.

Conclusions

We propose a preliminary biomarker panel, wich will later be tested and validated in large, well-designed, prospective clinical trials. The validation of molecular biomarkers and their eventual incorporation into clinical practice will improve personalized cancer treatments and the management of advanced stages.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

Instituto Nacional de Cancerología INCan (018/055/DII) (CEI/1302/18).

Funding

Instituto Nacional de Cancerología INCan (018/055/DII) (CEI/1302/18).

Disclosure

All authors have declared no conflicts of interest.