Functional homologous repair deficiency (HRD), as defined by a RAD51-low score, emerged as an indicator of patient outcome and response to carboplatin in triple negative breast cancer (TNBC) or PARP inhibitors, according to findings from two studies presented at the ESMO Breast Cancer Virtual Congress 2021, held 5 to 8 May.

Although HRD is implicated in TNBC, current genetic and genomic tests measuring HRD show limited predictive value for treatment outcomes, prompting the investigations assessing whether lack of RAD51 foci, due to defective homologous recombination repair (HRR), could be used as an indicator for HRD.

The first analysis was presented by Violeta Serra of the Experimental Therapeutics Group, Vall d’Hebron Institute of Oncology in Barcelona, Spain who noted that HRD is a hallmark of tumours with germline BRCA1/2 mutations; therefore, if lack of RAD51 foci is an indicator of functional HRD it could be a putative biomarker for tumours that can be treated with PARP inhibitors. PARP inhibitors have been approved for the treatment of metastatic breast cancers associated with germline BRCA1/2 mutations (gBRCA), since these tumours are known to have defective double strand DNA break repair.

Dr. Serra explained that germline and somatic alterations within the HRR pathway are found in 60% of TNBC tumours and identifying these patients is an important step to extend the use of PARP inhibitors in the clinic. To this end, this research team evaluated the functional status of HRR in samples from treatment-naive patients with early TNBC by measuring RAD51 foci with an immunofluorescence (IF) tissue-based test and to compare this with genomic HRD tests or treatment activity.

In this post-hoc analysis, they scored the percentage of RAD51 foci during the S/G2-phase of the cell cycle using the RAD51-IF test in baseline and end of treatment (EOT) samples. These paraffin-embedded tumour (FFPE) samples were obtained from 29 patients diagnosed with early TNBC and treated with rucaparib, a PARP inhibitor, in the RIO study (EudraCT 2014-003319-12). The baseline RAD51 score was correlated with the study endpoints, which included HRR alterations, HRDetect status, RAD51 by immunohistochemistry (RAD51-IHC) at EOT, and the change in ctDNA levels.

RAD51-IF in pre-treatment tumours associated with HRD and PARPi activity

The RAD51-IF test was successfully scored in the 17 available baseline tumour samples, and in 27 out of 28 EOT samples. Results from the baseline samples were correlated with genomic HRD for 15 cases and with clinical outcome for 12 cases with available data. The prevalence of functional HRD according to the RAD51-IF test was 47%, compared to 69% of HRDetect or 78% of RAD51-IHC end-of-treatment assay. Rucaparib induced DNA repair by homologous recombination, as seen by the increase in yH2AX-positive cells and the induction of RAD51 foci in paired biopsies.

They also compared RAD51 with HRR alterations, including gene mutations and epigenetic silencing of HRR genes. Six out of eight tumours with known HRR alterations, including 4 tumours with gene mutations and 2 with BRCA1 promoter hypermethylation, had low and concordant RAD51-IF values. Two additional tumours without known HRR alterations, shown in black, had low RAD51 score, suggesting the presence of other underlying HRR alterations beyond those described. They observed that HRDetect identified all tumours with HRR alterations but 3 out of 10 HRDetect-positive tumours showed RAD51 foci and functional HRR.

With rucaparib treatment, ctDNA suppression was higher in 5 RAD51-IF low tumours than in 12 RAD51-high tumours (p = 0.052); 4 out of 5 (80%) RAD51-low tumours demonstrated a decrease in ctDNA levels that was similar to 78% of RAD51-immunohistochemistry EOT samples.

RAD51-IF associated with HRD and response to carboplatin-based neoadjuvant chemotherapy  

The next speaker was Alba Llop-Guevara of the Vall d’Hebron Institute of Oncology in Barcelona, Spain who presented findings from the comparison of the same functional HRD test based on scoring RAD51 foci by IF and genetic/genomic HRD tests. In addition, HRD was assessed as an indicator of patients with primary TNBC who could be sensitive to platinum-based neoadjuvant chemotherapy.

The investigators conducted a retrospective, blinded analysis of samples from patients with TNBC receiving neoadjuvant paclitaxel plus non-pegylated liposomal doxorubicin (Myocet®) and bevacizumab (PM) or PM plus carboplatin in the randomised GeparSixto study (NCT01426880). They quantified the presence of functional biomarkers of HRD, including RAD51, BRCA1, and yH2AX nuclear foci in FFPE tumour samples laid on a tissue microarray format (TMA).

Concordance analyses were performed between the RAD51 score and tumour BRCA (tBRCA) status or genomic HRD score according to myChoice® HRD test. The associations with pathological complete response (pCR) and survival outcomes (disease-free survival (DFS) and overall survival (OS)) were assessed by logistic and Cox regression models, respectively. Functional HRD was predefined as RAD51-low, indicating a RAD51 score ≤10%.

RAD51, BRCA1, and yH2AX were successfully scored in 133 of 200 (67%) TMA cores.

Functional HRD by RAD51-low was detected in 81 of 133 (61%) tumours. The RAD51 test identified 93% of tBRCA-mutated tumours and 45% of the non-tBRCA mutant cases as functional HRD; concordance between RAD51 and genomic HRD was 87% (95% confidence interval [CI] 79-93%).

Regarding clinical outcomes, the pCR rate was similar between treatment arms in patients having RAD51-high tumours; pCR was 31% with PM plus carboplatin compared to 39% with PM (odds ratio [OR] 0.71; 95% CI 0.23-2.24; p = 0.561).

However, significant benefit was obtained from the addition of carboplatin to PM in patients with RAD51-low tumours; pCR was 66% with PM plus carboplatin versus 33% with PM (OR 3.96; 95% CI 1.56-10.05; p = 0.004; interaction test p = 0.023). This benefit maintained statistical significance in the multivariate analysis, supporting the utility of RAD51 as a predictive biomarker for carboplatin-based therapy in TNBC.

Regarding DFS and OS, the trial was not powered to demonstrate benefits in survival outcomes, and no statistically significant association was found from the addition of carboplatin in RAD51-high and RAD51-low tumours.

Prof. Andrew Tutt of the Institute of Cancer Research and Guy’s Hospital King’s College London in London, UK who discussed the studies findings said that RAD51 is a nodal point in the HR pathway, downstream of many HR proteins, potential HRD ‘catch all’ functional biomarker. Clinical context differs between these two abstracts. Assessment of the function of HR by assessment of RAD51 foci is feasible in FFPE material and with baseline not paired biopsies. This appears possible on TMA enabling clinical trial bio-resource analysis. HR DNA repair can return during tumour evolution especially if caused by methylation and so current pre-treatment not archival tissue should be used. HR mutation signature analyses do not change if HR function returns. A sensitive, but more specific predictive approach, may be to combine HR mutation signatures on archival tissue and RAD51 function analysis pre-treatment.


Based on their findings, Dr. Serra and Dr. Llop-Guevara concluded that the RAD51-IF test was feasible and accurate to assess the functional status of HRR and identify PARP inhibitor and platinum sensitive tumours in treatment-naive early TNBC FFPE tumour samples.


Scoring of nuclear RAD51 foci by immunofluorescence (IF) informs about functional HRR status and response to PARP inhibitors or platinum-based drugs in untreated TNBC tumours.

© Alba Llop-Guevara.

They emphasised that the RAD51 test was highly concordant with tumour BRCA mutation, other HRR gene alterations and genomic HRD tests. They noted that RAD51 associated with PARPi activity and independently predicted clinical benefit from adding carboplatin to neoadjuvant chemotherapy in TNBC.

According to the investigators, their results support further development to incorporate RAD51 testing in clinical decision making.

The work in the first study was funded by the Instituto de Salud Carlos III, Asociación Española Contra el Cáncer. Clovis Oncology Inc., Cridlan Foundation, CRUK, Breast Cancer Now, NHS, Wellcome Trust. The second study was funded by ERA-PerMed, Asociación Española Contra el Cáncer and “la Caixa” Foundation and the European Institute of Innovation and Technology/Horizon 2020 (CaixaImpulse programme).


  • 1O – Serra Elizalde V, Llop-Guevara A, Pearson A, et al. Detection of homologous recombination repair deficiency (HRD) in treatment-naive early triple negative breast cancer (TNBC) by RAD51 foci and comparison with DNA-based tests. ESMO Breast Cancer Virtual Congress 2021 (5-8 May).
  • 2O – Llop-Guevara A, Vladimirova V, Schneeweiss A, et al. Association of RAD51 with Homologous Recombination Deficiency (HRD) and clinical outcomes in untreated triple-negative breast cancer (TNBC): analysis of the GeparSixto randomized clinical trial. ESMO Breast Cancer Virtual Congress 2021 (5-8 May).