The Homologous Recombination Deficiency Scar in Advanced Cancer: Agnostic Targeting of Damaged DNA Repair

Simple Summary Tumor-suppressor genes are involved in DNA break repair through the homologous recombination system and are widely known for their role in hereditary cancer. Beyond breast and ovarian cancer, prostate and pancreatic cancer also have targetable homologous recombination deficiency (HRD) beyond the well-known BRCA1 and BRCA2 with relevance that exceeds diagnostic purposes. In this review, we aim to summarize the roles of HRD across tumor types and the treatment landscape to guide the targeting of damaged DNA repair based on the cancer’s genetic features. Abstract BRCA1 and BRCA2 are the most recognized tumor-suppressor genes involved in double-strand DNA break repair through the homologous recombination (HR) system. Widely known for its role in hereditary cancer, HR deficiency (HRD) has turned out to be critical beyond breast and ovarian cancer: for prostate and pancreatic cancer also. The relevance for the identification of these patients exceeds diagnostic purposes, since results published from clinical trials with poly-ADP ribose polymerase (PARP) inhibitors (PARPi) have shown how this type of targeted therapy can modify the long-term evolution of patients with HRD. Somatic aberrations in other HRD pathway genes, but also indirect genomic instability as a sign of this DNA repair impairment (known as HRD scar), have been reported to be relevant events that lead to more frequently than expected HR loss of function in several tumor types, and should therefore be included in the current diagnostic and therapeutic algorithm. However, the optimal strategy to identify HRD and potential PARPi responders in cancer remains undefined. In this review, we summarize the role and prevalence of HRD across tumor types and the current treatment landscape to guide the agnostic targeting of damaged DNA repair. We also discuss the challenge of testing patients and provide a special insight for new strategies to select patients who benefit from PARPi due to HRD scarring.


Introduction
Targeting homologous recombination deficiency (HRD) has been revealed in the last few years as one of the most promising strategies for various types of tumor. Alteration in the homologous recombination system (HR) genes is prevalent across tumor types and can be mostly found in breast, ovarian, pancreatic and prostate cancer [1,2]. HRD secondary to pathogenic germline and somatic variants in HR-associated genes has been reported as a predictive biomarker to inform about tumor sensitivity to platinum-based regimens and poly-ADP ribose polymerase (PARP) inhibitors (PARPi) [3,4]. Moreover, there are secondary changes to genetic mutations in DNA that can be detected at a structural level as genomic instability and could be associated with HRD response biomarkers that have already been validated. Identifying patients that may respond to direct or indirect therapies against HRD is a need in current clinical practice for cancer management and could optimize the clinical benefits of these therapies.

Homologous Recombination: A Key Pathway in DNA Repair
BRCA1 and BRCA2 are the best-known proteins involved in double-strand DNA break repair by HR. They are two of the main characters in the DNA defect situation, as shown by hundreds of publications reported in recent years [5]. However, genetic and epigenetic inactivation of other HR components can lead to HRD in sporadic cancers, classically termed BRCAness [6]. HR is responsible for repairing DNA before the cell comes into mitosis. It is produced during and immediately after DNA replication in S and G2 phases of the cell cycle, when sister chromatids are available [7]. Double-stranded breaks induced by ionizing radiation or toxic agents as chemotherapy are first sensed by the MRE11-RAD50-NBN (MRN) complex, which loads helicase and exonucleases onto the breaks to start 5 -3 double-stranded DNA resection. ATR then localizes to the ssDNA ends and switches on the ATR-dependent checkpoint, arresting the cell cycle for HR to proceed. Next, BRCA1 is phosphorylated in response to DNA damage by DNA-damage response kinases, such as ATM, ATR and CHK1, which enables the cell to repair DNA before entering mitosis and survive [8][9][10][11]. ATM, ATR, BARD1, RB, p53, p21 and their downstream effectors are involved in induced G1/S arrest [12]. Therefore, BRCA1 loss can result in defective S-G2/M and spindle checkpoints that together with abnormal centrosome duplication and defective DNA damage repair can lead to genetic instability [12]. Furthermore, BRCA2 can help to protect telomere integrity loading RAD51 during S/G2 [7,13]. Proteins involved in the HR system have functions in DNA repair, but also participate in cell cycle regulation, transcriptional activation and chromatin remodeling ( Figure 1).
Cancer genomics often harbor chromosomal aberrations arising from a defective HR pathway. In BRCA mutant cells, chromosomal spreads reveal increased gross chromosomal rearrangements [14]. This leads to the development of assays to evaluate the "genomic scar" left behind by the loss of HR function, irrespective of which component of the pathway was lost.  [7], copyright 2019 the authors, Ivyspring International Publisher under t the Creative Commons Attribution license 4.0 (https://creativecommons.org/licenses/by accessed on 12 May 2022) and Hoppe et al. [6], copyright 2018 by the authors published University Press. The figures have been modified for the purposes of this review. Protein in the HR system have functions in DNA repair, but also participate in cell cycle regul scriptional activation and chromatin remodeling. Genomic scarring is defined by the chromosomal abnormalities related to HRD: (a) telomeric allelic imbalance (TAI) due to ate chromosomal end fusions because of aberrant end joining, (b) loss of heterozygosity lated to inaccurate repair of sister chromatids during the S/G2 cell cycle phase and (c) transitions (LSTs) that are chromosomal breaks of more than 10 Mb.

Prevalence and Prognostic Value of HRD in Cancer
Heritable damage in the DNA repair system can be observed in up to 10% patients. BRCA1 and BRCA2 are the most common of these genetic abnorma breast-ovarian syndrome is the classical phenotype of germline BRCA alteration ever, in the last decade, evidence has shown that somatic events are more freq previously expected and that these aberrations affect tumors beyond breast an cancer. Somatic mutations in BRCA genes are more frequent in ovarian cancer lowed closely by prostate cancer, squamous skin cell carcinoma, breast cance 10%) and pancreatic cancer. (Figure 2). It is noted that frequency varies signifi pending on the population studied, geographic area, type of sample studied or Other relevant HR pathway members include genes such as ATM, PALB2 and RAD51. However, the associations between these genes and an HRD pheno be less consistent than those for BRCA1 and BRCA2 and may vary by the tum  [7], copyright 2019 the authors, Ivyspring International Publisher under the terms of the Creative Commons Attribution license 4.0 (https://creativecommons.org/licenses/by-nc-nd/ 4.0/, accessed on 12 May 2022) and Hoppe et al. [6], copyright 2018 by the authors published by Oxford University Press. The figures have been modified for the purposes of this review. Proteins involved in the HR system have functions in DNA repair, but also participate in cell cycle regulation, transcriptional activation and chromatin remodeling. Genomic scarring is defined by the presence of chromosomal abnormalities related to HRD: (a) telomeric allelic imbalance (TAI) due to inappropriate chromosomal end fusions because of aberrant end joining, (b) loss of heterozygosity (LOH) related to inaccurate repair of sister chromatids during the S/G2 cell cycle phase and (c) large-scale transitions (LSTs) that are chromosomal breaks of more than 10 Mb.

Prevalence and Prognostic Value of HRD in Cancer
Heritable damage in the DNA repair system can be observed in up to 10% of cancer patients. BRCA1 and BRCA2 are the most common of these genetic abnormalities, and breast-ovarian syndrome is the classical phenotype of germline BRCA alteration [1]. However, in the last decade, evidence has shown that somatic events are more frequent than previously expected and that these aberrations affect tumors beyond breast and ovarian cancer. Somatic mutations in BRCA genes are more frequent in ovarian cancer (15%), followed closely by prostate cancer, squamous skin cell carcinoma, breast cancer (around 10%) and pancreatic cancer. (Figure 2). It is noted that frequency varies significantly depending on the population studied, geographic area, type of sample studied or stage.
Other relevant HR pathway members include genes such as ATM, PALB2, CHECK2 and RAD51. However, the associations between these genes and an HRD phenotype may be less consistent than those for BRCA1 and BRCA2 and may vary by the tumor's tissue of origin [15][16][17]. Norquist et al. [17] observed that 6.8% of the ovarian cancer patients included in the GOG 218 trial harbored a non-BRCA somatic HR gene mutation, and the most frequently observed alteration was in ATM. The most frequently altered DNArepair genes in both germline and somatic cells of mCRPC patients are BRCA2, ATM and CHEK2: germline mutations are found in 5%, 2% and 2%, respectively [18,19]. BRCA1/2 homozygous deletions are infrequent except in prostate cancer, where BRCA2 deletions have been reported at 2.6% frequency and accounted for 25% of BRCA1/2-altered cases [20]. of origin [15][16][17]. Norquist et al. [17] observed that 6.8% of the ovarian cancer patients included in the GOG 218 trial harbored a non-BRCA somatic HR gene mutation, and the most frequently observed alteration was in ATM. The most frequently altered DNA-repair genes in both germline and somatic cells of mCRPC patients are BRCA2, ATM and CHEK2: germline mutations are found in 5%, 2% and 2%, respectively [18,19]. BRCA1/2 homozygous deletions are infrequent except in prostate cancer, where BRCA2 deletions have been reported at 2.6% frequency and accounted for 25% of BRCA1/2-altered cases [20]. HRD prevalence based on the measurement of telomeric allelic imbalance (TAI), loss of heterozygosity (LOH) and large-scale state transitions (LST) is the more extended diagnostic method to measure HR status and varies also broadly among different types of tumors, although there is a correlation with BRCA prevalence. TAI, LOH and LST are highly correlated with each other and reflect increasing genomic instability. TAI refers to allelic imbalance extending to the subtelomeric region >11 megabases (Mb) in size. LOH refers to permanent loss of one parent's contributed allele copy at a specific locus, leading to homozygosity at that genomic site. LSTs refer to allelic imbalance > 10 Mb in size between adjacent genomic regions due to translocations or copy gains/losses.
The combination of these three parameters of genomic instability provides HRD scores by measuring LOH, TAI and LST with somatic next-generation DNA sequencing and may estimate the underlying genomic scarring in the context of HR deficiency. HRD scores have been more extensively studied for ovarian cancer, but when the algorithm is applied across different tumor types, it should be considered that some modifications are applied to avoid bias, and there is still no consensus [21]. Thus, this is not a validated method to be used in clinical practice for agnostic tumors [15]. Ovarian cancer, followed by lung adenocarcinoma and breast cancer, are the tumors where these alterations are more frequent, having genomic scores higher than 30 ( Figure 3). In an in silico analysis of 5371 tumors of 15 cancer types available in the TCGA, cancers where platinum constitutes standard first-line therapy showed increased genomic scar scores.
Lotan et al. [15] evaluated HRD scores in prostate cancer and their associations with HR gene mutations, and observed that HRD scores vary significantly between patients harboring BRCA2, ATM and CHEK2. Germline BRCA2-altered prostate cancer patients HRD prevalence based on the measurement of telomeric allelic imbalance (TAI), loss of heterozygosity (LOH) and large-scale state transitions (LST) is the more extended diagnostic method to measure HR status and varies also broadly among different types of tumors, although there is a correlation with BRCA prevalence. TAI, LOH and LST are highly correlated with each other and reflect increasing genomic instability. TAI refers to allelic imbalance extending to the subtelomeric region >11 megabases (Mb) in size. LOH refers to permanent loss of one parent's contributed allele copy at a specific locus, leading to homozygosity at that genomic site. LSTs refer to allelic imbalance > 10 Mb in size between adjacent genomic regions due to translocations or copy gains/losses.
The combination of these three parameters of genomic instability provides HRD scores by measuring LOH, TAI and LST with somatic next-generation DNA sequencing and may estimate the underlying genomic scarring in the context of HR deficiency. HRD scores have been more extensively studied for ovarian cancer, but when the algorithm is applied across different tumor types, it should be considered that some modifications are applied to avoid bias, and there is still no consensus [21]. Thus, this is not a validated method to be used in clinical practice for agnostic tumors [15]. Ovarian cancer, followed by lung adenocarcinoma and breast cancer, are the tumors where these alterations are more frequent, having genomic scores higher than 30 ( Figure 3). In an in silico analysis of 5371 tumors of 15 cancer types available in the TCGA, cancers where platinum constitutes standard first-line therapy showed increased genomic scar scores.
Lotan et al. [15] evaluated HRD scores in prostate cancer and their associations with HR gene mutations, and observed that HRD scores vary significantly between patients harboring BRCA2, ATM and CHEK2. Germline BRCA2-altered prostate cancer patients had the highest HRD scores, germline ATM-altered patients had intermediate scores and germline CHEK2-altered patients had the lowest scores [15].
The most common genomic scar assays reported to date are two commercially available tests that combine tumor BRCA mutation testing with a Genomic Instability Score (GIS) based on quantification of TAI, LOH and LST. These tests are myChoice ® HRD test (Myriad Genetics, Salt Lake City, UT, USA) and Foundation Focus CDxBRCA HRD ® (Foundation Medicine, MA, USA). s 2022, 14, x 5 of 1 had the highest HRD scores, germline ATM-altered patients had intermediate scores and germline CHEK2-altered patients had the lowest scores [15]. The most common genomic scar assays reported to date are two commercially avail able tests that combine tumor BRCA mutation testing with a Genomic Instability Scor (GIS) based on quantification of TAI, LOH and LST. These tests are myChoice ® HRD tes (Myriad Genetics, Salt Lake City, UT, USA) and Foundation Focus CDxBRCA HRD (Foundation Medicine, MA, USA). The figure has been made for the purposes of this review. HRD analysis of TCGA samples acros 15 different cancer types was performed based on the number of Telomeric Allelic Imbalances (TA based on a genomic scar accumulation, the large scale transition (LST) based on a type of genomi scar associated with loss of BRCA1 or BRCA2 and the HRD-LOH based on a scar enriched in high grade serous ovarian cancer patients with a loss of BRCA1 or BRCA2 [21][22][23]. However, the metho originally used for ovarian cancer samples was adapted to avoid bias when the algorithm is applie across different tumors: (1) In the original publication describing TAI [24], all allelic imbalanc events that extended to the telomere were counted, if they did not span the centromere. This result in an overrepresentation of tumors with an uneven copy number among high TAI cases, which ha been corrected in the method used for the present study. (2) The original publication describin HRD-LOH [23] excluded chromosome 17 because LOH on chromosome 17 in the ovarian cance samples is ubiquitous and for this reason did not provide independent information. However, fo this figure, chromosome 17 was not excluded, as chromosome 17 is not ubiquitously lost in all cance types, and therefore may provide independent information in some tumor samples.
Ovarian cancer has the strongest association with HRD, and up to 50% of high-grad serous ovarian carcinoma have a genetic aberration in the HR pathway [25]. In genera somatic aberrations are twice as frequent as germline alterations [20]. In non-endometri oid TP53-mutant endometrial cancer, which is molecularly similar to high-grade serou ovarian carcinoma, a high incidence of HRD genomic scars of up to 48% has been reported [26]. The prognostic significance of HRD in ovarian cancer is controversial. It has bee reported to have more favorable overall survival (OS) compared to non-carriers [27], fo based on a genomic scar accumulation, the large scale transition (LST) based on a type of genomic scar associated with loss of BRCA1 or BRCA2 and the HRD-LOH based on a scar enriched in highgrade serous ovarian cancer patients with a loss of BRCA1 or BRCA2 [21][22][23]. However, the method originally used for ovarian cancer samples was adapted to avoid bias when the algorithm is applied across different tumors: (1) In the original publication describing TAI [24], all allelic imbalance events that extended to the telomere were counted, if they did not span the centromere. This results in an overrepresentation of tumors with an uneven copy number among high TAI cases, which has been corrected in the method used for the present study. (2) The original publication describing HRD-LOH [23] excluded chromosome 17 because LOH on chromosome 17 in the ovarian cancer samples is ubiquitous and for this reason did not provide independent information. However, for this figure, chromosome 17 was not excluded, as chromosome 17 is not ubiquitously lost in all cancer types, and therefore may provide independent information in some tumor samples.
Ovarian cancer has the strongest association with HRD, and up to 50% of high-grade serous ovarian carcinoma have a genetic aberration in the HR pathway [25]. In general, somatic aberrations are twice as frequent as germline alterations [20]. In non-endometrioid TP53-mutant endometrial cancer, which is molecularly similar to high-grade serous ovarian carcinoma, a high incidence of HRD genomic scars of up to 48% has been reported [26]. The prognostic significance of HRD in ovarian cancer is controversial. It has been reported to have more favorable overall survival (OS) compared to non-carriers [27], for both BRCA1 (HR 0.78; 95% CI, 0.68-0.89; p < 0.001) and BRCA2 mutation carriers (HR 0.61; 95% CI, 0.50-0.76; p < 0.001). However, Candido-dos-Reis et al. [28] analyzed the effect of germline BRCA mutations in 4314 ovarian cancer patients with a 10-year follow-up and showed that the better short-term survival observed decreased over time, and patients who harbored a BRCA1 mutation even showed worse OS. Mutations in non-BRCA HR genes, including ATM, CHEK2, PALB2 and RAD51c, have been reported to be predictors of survival in ovarian cancer patients [16].
The prognostic relevance of BRCA in breast cancer is also questionable: some studies demonstrated that patients with a BRCA1/2 mutation had worse OS [29][30][31][32], and other studies showed no significant differences when compared with non-carriers [33][34][35][36]. Patients diagnosed with HRD breast cancer have shown an association with a more aggressive phenotype: BRCA1 is more frequently associated with triple-negative breast cancer, and BRCA2-related breast cancer correlated with a higher histological grade compared to patients who do not have germline mutations [37][38][39] The prognostic significance of HRD in patients with pancreatic adenocarcinoma is currently unknown [40]. Golan et al. [40] analyzed 71 patients with BRCA1/2-asssociated pancreatic cancer and observed an improvement in survival in patients with advanced disease (stage 3 and 4) who had received platinum-based therapy in comparison to those patients who were not treated with these agents. In their study, a more favorable outcome with platinum treatment was suggested, but a statistically significant improvement was not observed [40].
In prostate cancer, germline BRCA2 mutations have been associated with a more aggressive phenotype and poorer outcomes [41]. Castro et al. [42] studied germline DNA repair defects in an unselected cohort of patients with metastatic castration-resistant prostate cancer and observed that gBRCA2 mutation was an independent prognostic factor for cause-specific survival in this setting. The prognostic role of somatic BRCA2 alterations remains unclear.

HRD as an Actionable Target
The treatment landscape has evolved in the last decade, and HRD has been proposed as a predictive biomarker to determine increased sensitivity to platinum chemotherapy and PARPi [43,44].

Platinum
Platinum chemotherapy binds directly to the DNA in order to cause the cytotoxic effect of crosslinking DNA strands and induce double-stranded breaks, which are not repaired in cells that harbor defects in involved DNA repair pathways. In the last decade, BRCA mutations, as lead actors of HRD, have been suggested as predictive biomarkers for response to platinum across different tumor types [45][46][47][48][49][50][51][52] (Table 1). However, despite of high rates of platinum sensitivity in this population, it seems much more is needed beyond BRCA alterations to select candidates for treatment.
In ovarian cancer, the standard first-line chemotherapy regimen includes platinum and taxanes, independently of BRCA status, and response rates are greater than 80%. Platinumfree interval has been identified as a key biomarker to response to subsequent lines. Higher intervals are associated with predominance of HRD in tumor progression, and this will determine the indication for platinum re-treatment.
In breast cancer, an increasing amount of evidence suggests that TNBC patients with BRCA mutations could be more sensitive to platinum-based chemotherapy [46,53]. A recently published meta-analysis by Chai et al. [54] included six trials with HRD in TNBC data and showed that patients with HRD-positive TNBC had higher complete response rates compared to HRD-negative ones after receiving platinum-based neoadjuvant chemotherapy. However, the GeparSixto trial showed that the response to platinum-agents was not dependent on BRCA status [47], and TNBC non-mutated BRCA patients showed increased response rates with carboplatin, meaning that HRD did not predict carboplatin benefit [55].
HRD associated pancreatic cancer was under-identified until recently [56]. A family history of breast, ovarian or pancreatic cancer has been associated with increased sensitivity to platinum drugs as DNA-damaging agents [57], suggesting the presence of DNA repair defects in those patients. However, studies considering only clinical inheritance have failed to demonstrate this clinical biomarker as effective. Thus, Okano et al. [49] evaluated platinum benefit in patients with a family history of ovarian, prostate, pancreatic or breast cancer without analyzing BRCA genes and did not observe a benefit in OS. Patients with HRD pancreatic cancer have shown a clinical benefit and a longer OS due to platinumbased treatments [50,[58][59][60], and data also suggest a notable and significant increase in the response rates: 50-65% [61]. Similar activity of oxaliplatin and cisplatin in patients with germline BRCA and PALB2 mutations has been suggested by retrospective data [50,62], and a survival benefit of platinum-based first line chemotherapy in this subgroup of patients have been observed [50,63].
Metastatic prostate cancer patients harboring DNA repair gene alterations treated with platinum-based chemotherapy showed encouraging antitumor activity [51,52], although the role of HRD in this setting is still controversial [51,52]. Recently, Pokataev et al. [64] published a meta-analysis reporting higher overall survival in patients with HRD, advanced prostate cancer treated with platinum-based chemotherapy. Prospective validation in ongoing randomized clinical trials will be needed to determine the role of platinum treatment in advanced prostate cancer.

PARP Inhibitors
In 2005, two groundbreaking studies observed that tumor cells lacking BRCA1 or BRCA2 were particularly sensitive to PARPi through various mechanisms [66,67]. The main target of PARPi is PARP1, which is involved in the repair of single-strand DNA breaks, so in order to produce cytotoxicity, a defective HR is required [66,67].
PARP1 is a damage sensor that is able to synthesize PAR chains on target proteins near DNA break, and with these PAR chains recruit additional DNA repair effectors [5]. PARPi causes a catalytic inhibition of PARP1 and traps PARP1 by either inhibiting autoPARylation or by causing allosteric changes in its structure [68,69]. Patients' tumors which lack BRCA1 or BRCA2 are not able to repair DNA lesions and try to use error-prone DNA repair pathways that have a cytotoxic effect that kills the cells [5].
In 2015, a basket trial of olaparib in patients with gBRCA1/2 mutations identified responding patients beyond the ovarian or breast cancer population, suggesting that other HR-defective tumors could be suitable for PARPi treatment [70]. The current developments of PARPi in solid tumors are displayed in Table 2.
Three PARPi, olaparib, rucaparib and niraparib, have been approved by the United States Food and Drug Administration (FDA) and the European Medicines Agency (EMA) as maintenance therapy in platinum-sensitive recurrent epithelial ovarian cancer [71][72][73][74]. Rucaparib has been also approved as monotherapy in patients with somatic or germline BRCA1/2 mutations [74]. In 2018, the FDA granted approval to olaparib monotherapy for the first-line maintenance treatment of patients with BRCA-mutated advanced epithelial ovarian, fallopian tube or primary peritoneal cancer who have completely or partially responded to first-line platinum-based chemotherapy based on SOLO-1 trial results [75]. Shortly after, niraparib achieved indication for same setting independently of BRCA status based on a PRIMA trial [76]. Analysis subgroups revealed a benefit for all populations, although less significant for HRD-proficient patients. The PAOLA-1 phase III first-line ovarian cancer maintenance study showed a benefit on progression free survival of the combination of olaparib and bevacizumab compared to bevacizumab in the overall population; however, this benefit was not seen in the subgroup analysis of the HRD-proficient popula-tion [77]. On May 2020, the FDA expanded the approval of olaparib and bevacizumab for first-line maintenance treatment of HRD, advanced ovarian cancer [78].
Metastatic breast cancer patients with a germline BRCA1 or BRCA2 mutation treated with PARPi have better outcomes in terms of PFS compared to standard chemotherapy [79,80]. Recently, the OlympiA trial evaluated PARPi efficacy in early breast cancer after standard adjuvant treatment with chemotherapy and local therapy, achieving in patients with gBRCA1/2 mutations longer survival, free of invasive or distant disease, than the placebo [81].
In metastastic gBRCA-mutated pancreatic adenocarcinoma, olaparib has been approved for the maintenance after at least 16 weeks of first-line platinum-based chemotherapy if the disease has not progressed [82]. The objective response rate was 23% in the olaparib arm; 10% for placebo and 10% of patients from the placebo arm maintained a median duration of response of 24.9 months. On the other hand, cisplatin plus gemcitabine was evaluated in a phase II trial of 50 patients with gBRCA or PALB2-mutated locally advanced or metastastic pancreatic cancer randomly assigned alone or with veliparib [59], and concurrent veliparib did not improve the response rate in this subset of patients.
Several PARPi are currently under development for the treatment of advanced prostate cancer [83][84][85][86][87][88]. Alterations in BRCA2, particularly homozygous deletions, seem to be the best predictor of response to PARP inhibition [89]. In patients with BRCA1/2 alterations, 40-46% radiographic response rates have been reported with the different agents. PSA declines of >50% have been noted in half of the BRCA1/2 patients included in the different trials, despite being heavily pretreated. No differences in efficacy have been reported based on the germline or somatic origin of the alterations [87]. The predictive roles of other HR defects beyond BRCA1/2 remain unclear. Little or no benefit from PARP inhibition has been observed in patients with ATM or CDK12 alterations, whilst the predictive roles of less frequent alterations have not been stablished due to the limited number of patients included in trials. Olaparib has been the only PARPi to be investigated in monotherapy in a phase 3 trial for advanced prostate cancer patients. In the PROfound study, men with metastatic castration-resistant prostate cancer (mCRPC) with alterations in one of the 15 HR genes screened whose disease had already progressed to an AR-targeting inhibitor (ARTi) were randomized to receive treatment with olaparib 300 mg bid or a second ARTi. A benefit in overall survival was observed for patients in cohort A, which included patients with BRCA1, BRCA2 and ATM alterations, whereas no benefit was observed for patients with other alterations included in cohort B. These results led to the EMA approval of olaparib for the treatment of mCRPC patients with BRCA1/2 alterations after disease progression to treatment with an ARTi.
In non-small lung cancer and colorectal cancer, two of the most prevalent tumors, preliminary in vivo studies in cell lines with HRD features have supported the potential use of PARPi [90,91].
Uterine leiomyosarcoma has recently been identified as a sarcoma subtype with characteristic defects in the HR repair pathway and frequent BRCA2 loss [92]. Preclinical data demonstrate marked activity for PARPi in combination with the alkylating agent temozolomide. Ongoing research in order to identify other sarcomas with DNA repair defects is promising, and may offer a new opportunity for the targeted treatment of this rare, aggressive cancer [92].  3 Only two patients had somatic BRCA1/2 mutation.

The Challenge of Testing: Searching for HRD Scar
Next generation sequencing (NGS) techniques, as the current gold standard of genetic diagnosis, have helped to shorten the time to obtain a genetic test result. However, at the same time, an NGS multi-panel may provide data about other HRD genes beyond BRCA where the current evidence as biomarkers to select therapy is limited, hindering decision making in clinical practice. To handle this complexity, an adequate bioinformatic analysis will be key, along with a multidisciplinary approach.
Traditionally, BRCA testing has been conducted in germline DNA triggered by a familial aggregation of cancer. Recent studies have demonstrated that a significant proportion of mutation carriers are undiagnosed due to the lack of a significant family history of cancer [18,42,98], leading to changes in the recommendations for genetic testing. As an example, testing is now recommended for all metastatic prostate cancer patients, regardless of their personal or family history of cancer [99]. Moreover, the advent of therapies that target BRCA alterations and other HRD defects requires the investigation of germline mutations and alterations acquired by the tumor, as described previously.
For that reason, somatic mutation analysis is moving germline testing in various scenarios, such as advanced ovarian and prostate cancer. HRD testing of the tumor directly has the advantage of providing higher rates of positivity compared with germline tests [19]. However, somatic testing is associated with higher rates of failure for sequencing [84]. For that reason, new protocols for improving the conservation and storage of paraffined samples should be implemented in hospitals. In fact, a systematic and consensus protocol for high-quality minimum biomarker testing is being requested by the scientific community [100]. These molecular testing recommendations should be offered to all cancer patients for diagnosis and prevention, detailing the type, the technique and the methods of implementation and ensuring adequate training for clinicians to guide the treatment decisions.
ESMO and NCCN guidelines [98,101,102] recommend pre-test counseling to determinate the most appropriate test for each patient and specific post-test counseling when results are available. Table 3 summarizes the current recommendations for testing based on international clinical guidelines.
A different strategy to identify HRD patients could be to measure the "genomic scarring" associated with loss of function in DNA repair pathways, as genomic instability. In 2012, three SNP-based assays were developed to quantify the extent of chromosomal abnormalities related to HRD: (a) TAI due to inappropriate chromosomal end fusions because of aberrant end joining; (b) LOH-related to inaccurate repair of sister chromatids during the S/G2 cell cycle phase; and (c) LST-chromosomal breaks of more than 10 Mb. These "functional assays" have been proposed as more reliable methods for identifying patients responding to PARPi compared to simply identifying gene mutations, although their role in guiding therapy is pending on validation, and their values and thresholds are heterogeneous across cancer types [103].
In 2020, the phase III PAOLA-1 trial in ovarian cancer was the first one to obtain FDA/EMEA approval for HRD-positive patients to use the PARPi combination with bevacizumab using the Myriad myChoice ® test defined by GIS ≥ 42 [77]. However, the evaluations of the utility of these tests to predict the benefit from PARPi in the BRCAwt populations were preplanned secondary analyses in clinical trials, and there are not definitive results for trials specifically designed for this subpopulation [103,104]. In breast cancer patients, GIS ≥ 42 has been associated with BRCA mutations, and furthermore, TAI has been associated with an improved response to platinum chemotherapy [105,106]. In pancreatic cancer patients, GIS ≥ 42 has shown a sensitivity of 91% and a specificity of 83% for the identification of HRD. Moreover, a higher GIS has been associated with an improved oncological outcome with platinum chemotherapy [107]. Currently, there are different thresholds proposed as best classifiers for HRD score evaluation, but they are still pending validation [108].
In prostate cancer, non-commercial signatures based on genomic instability scores have been explored. An adequate correlation between BRCA-deficient samples and HRDassociated mutational signatures using WGS data was reported [109]. However, there are not clinical trials demonstrating the optimal threshold to assure the role of the HRD score testing as a predictor of treatment with PARPi or platinum therapy.
Another type of functional assay that may have the potential to provide a dynamic readout of HRD scarring is based on the estimation of the amount of nuclear RAD51, a downstream HR protein (a DNA recombinase). RAD51 enables high-fidelity doublestrand DNA repair by facilitating DNA strand invasion into the sister chromatid, a process supported by the BRCA1/PALB2/BRCA2 complex. Reduced, DNA-damaged, induced nuclear RAD51 foci have been associated with BRCA1 or BRCA2 gene defects and PARPi responses [110,111]. This approach is currently under investigation, and functional assays have not yet been validated for pancreatic cancer [112].
Additionally, the HRD profile may change during cancer progression, as reversion mutations of HR genes have been reported to occur in 26% of patients, and this fact may be related to response to previous treatment to generate resistance to platinum or PARPi [113]. In this setting, monitoring the dynamic evaluation of HRD in cancer should be relevant. Moreover, one of the main difficulties in tumors such as pancreatic cancer is obtaining an adequate amount of tissue for genetic testing. Thus, liquid biopsy approaches to identify HRD based on circulating tumoral DNA analysis to assess chromosomal instability or mutational signatures is a promising method under study, but is not ready yet to use in clinical practice (pending validation) [114].

Conclusions
The detection of DNA repair defects related to the HR pathway provide a unique opportunity for the development of treatments in different type of tumors that take advantage of a same tumor feature. Tools and validation trials to identify the optimal HRD test across tumor types are urgently needed.