Landscape of homologous recombination deficiencies in solid tumours: analyses of two independent genomic datasets

Prediction of homologous recombination deficiency from routine histology with attention-based multiple instance learning in nine different tumor types

BACKGROUND

Homologous recombination deficiency (HRD) is recognized as a pan-cancer predictive biomarker that potentially indicates who could benefit from treatment with PARP inhibitors (PARPi). Despite its clinical significance, HRD testing is highly complex. Here, we investigated in a proof-of-concept study whether Deep Learning (DL) can predict HRD status solely based on routine hematoxylin & eosin (H&E) histology images across nine different cancer types.

METHODS

We developed a deep learning pipeline with attention-weighted multiple instance learning (attMIL) to predict HRD status from histology images. As part of our approach, we calculated a genomic scar HRD score by combining loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large-scale state transitions (LST) from whole genome sequencing (WGS) data of n = 5209 patients across two independent cohorts. The model's effectiveness was evaluated using the area under the receiver operating characteristic curve (AUROC), focusing on its accuracy in predicting genomic HRD against a clinically recognized cutoff value.

RESULTS

Our study demonstrated the predictability of genomic HRD status in endometrial, pancreatic, and lung cancers reaching cross-validated AUROCs of 0.79, 0.58, and 0.66, respectively. These predictions generalized well to an external cohort, with AUROCs of 0.93, 0.81, and 0.73. Moreover, a breast cancer-trained image-based HRD classifier yielded an AUROC of 0.78 in the internal validation cohort and was able to predict HRD in endometrial, prostate, and pancreatic cancer with AUROCs of 0.87, 0.84, and 0.67, indicating that a shared HRD-like phenotype occurs across these tumor entities.

CONCLUSIONS

This study establishes that HRD can be directly predicted from H&E slides using attMIL, demonstrating its applicability across nine different tumor types.

Germline and somatic testing for homologous repair deficiency in patients with prostate cancer (part 1 of 2)

BACKGROUND/OBJECTIVES

Unfortunately, not all metastatic castration resistant prostate cancer (mCRPC) patients receive available life-prolonging systemic therapies, emphasizing the need to optimize mCRPC treatment selections. Better guidelines are necessary to determine genetic testing in prostate cancer.

SUBJECTS/METHODS

In this two-part expert opinion-based guide, we provide an expert consensus opinion on the utilization of germline and somatic testing to detect HRR alterations in patients with mCRPC. This guide was developed by a multidisciplinary expert panel that convened in 2023-2024, including representatives from medical oncology, urology, radiation oncology, pathology, medical genomics, and basic science.

RESULTS/CONCLUSION

We argue for the widespread adoption of germline testing in all patients with prostate cancer and for somatic mutations testing in patients at the time of recurrent/metastatic disease. In this first part, we review how genomic testing is performed. We also review how to overcome certain barriers to integrate genetic and biomarker testing into clinical practice.

BRCA genetic testing and counseling in breast cancer: how do we meet our patients' needs?

BRCA1 and BRCA2 are tumor suppressor genes that have been linked to inherited susceptibility of breast cancer. Germline BRCA1/2 pathogenic or likely pathogenic variants (gBRCAm) are clinically relevant for treatment selection in breast cancer because they confer sensitivity to poly(ADP-ribose) polymerase (PARP) inhibitors. BRCA1/2 mutation status may also impact decisions on other systemic therapies, risk-reducing measures, and choice of surgery. Consequently, demand for gBRCAm testing has increased. Several barriers to genetic testing exist, including limited access to testing facilities, trained counselors, and psychosocial support, as well as the financial burden of testing. Here, we describe current implications of gBRCAm testing for patients with breast cancer, summarize current approaches to gBRCAm testing, provide potential solutions to support wider adoption of mainstreaming testing practices, and consider future directions of testing.

Trends in estimated PARP inhibitor eligibility and benefit among US epithelial ovarian cancer patients

OBJECTIVE

To estimate the annual percentage of patients with epithelial ovarian cancer (EOC) who could be eligible for and benefit from PARP inhibitor therapy amidst changing US Food and Drug Administration (FDA)-approved indications.

METHODS

This is a simulated retrospective observational study using publicly available data on patients with advanced-stage EOC. PARPi eligibility is based on FDA approvals and withdrawals from 2014 through 2023, along with published demographic and genomic data. Clinical trial data is used to estimate treatment benefit. PARPi including olaparib, niraparib, and rucaparib are analyzed in aggregate with sub-analyses by molecular classification and treatment timing. Results are reported as the percentage of EOC patients appropriate for any cancer-directed therapy.

RESULTS

PARPi were approved for 9 different indications in EOC between 2014 and 2021; reduced to 6 indications by 2023. Eligibility increased from 2.0% (95% CI,1.3%-1.6%) in 2014 to a maximum of 93.4% (95% CI,90.1%-94.6%) in 2021. The maximum percentage of patients with 2-year PFS benefit was 22.0% (95% CI, 17.2%-26.8%) in 2021, projected to decrease to 13.0% (95% CI, 9.9%-15.9%) in 2024. Most of this decrease was seen in the homologous recombination deficient, BRCA wild-type population (8.4% to 4.0%).

CONCLUSIONS

PARPi eligibility increased at a greater rate than benefit resulting in a low population-level benefit-to-eligibility ratio until 2021. Recent FDA withdrawals improved this ratio with an accompanied decrease in the absolute number of patients benefiting. To further optimize population-level benefit-to-eligibility ratio of targeted therapies in ovarian cancer, we need to identify better biomarkers, treatment combinations, and novel therapeutic targets.

Prevalence of homologous recombination biomarkers in multiple tumor types: an observational study

Aim: To determine the prevalence of deleterious mutations in BRCA1 and BRCA2 and in 13 genes involved in homologous recombination repair (HRR), the prevalence of genomic loss of heterozygosity and the allelic and hereditary status of BRCA1, BRCA2 and other HRR gene mutations in multiple solid tumor types. Patients & methods: This was a retrospective observational study of patients with an advanced/metastatic diagnosis in one of 15 solid tumor types, who were identified in a real-world clinico-genomic database. Results: Tumor tissue samples from 9457 patients were analyzed, among which 4.7% had known or suspected deleterious BRCA1/2 mutations. The prevalence (range) of mutations in HRR genes was 13.6% (2.4%-26.0%) and genomic loss of heterozygosity ≥16% was 20.6% (2.6-34.4%) across all tumor types. Conclusion: The prevalence of mutations varied significantly depending on the type of tumor.

Durvalumab in Combination With Olaparib Versus Durvalumab Alone as Maintenance Therapy in Metastatic NSCLC: The Phase 2 ORION Study

INTRODUCTION

Increased DNA damage triggered through poly (ADP-ribose) polymerase inhibition may modify tumor immunogenicity, sensitizing tumors to immunotherapy. ORION (NCT03775486) evaluated the combination of olaparib with durvalumab as maintenance therapy in patients with metastatic NSCLC.

METHODS

ORION is a phase 2, randomized, multicenter, double-blind, international study. Patients with metastatic NSCLC (without activating EGFR or ALK aberrations) and Eastern Cooperative Oncology Group performance status of 0 or 1 were enrolled to receive initial therapy with durvalumab (1500 mg intravenously; every 3 wk) plus platinum-based chemotherapy for four cycles. Patients without disease progression were then randomized (1:1) to maintenance durvalumab (1500 mg; every 4 wk) plus either olaparib (300 mg orally) or placebo (both twice daily); randomization was stratified by objective response during initial therapy and tumor histologic type. The primary end point was investigator-assessed progression-free survival (PFS) (Response Evaluation Criteria in Solid Tumors version 1.1).

RESULTS

Between January 2019 and February 2020, 269 of 401 patients who received initial therapy were randomized. As of January 11, 2021 (median follow-up: 9.6 mo), median PFS was 7.2 months (95% confidence interval: 5.3-7.9) with durvalumab plus olaparib versus 5.3 months (3.7-5.8) with durvalumab plus placebo (hazard ratio = 0.76, 95% confidence interval: 0.57-1.02, p = 0.074). Safety findings were consistent with the known profiles of durvalumab and olaparib. Anemia was the most common adverse event (AE) with durvalumab plus olaparib (26.1% versus 8.2% with durvalumab plus placebo). The incidence of grade 3 or 4 AEs (34.3% versus 17.9%) and AEs leading to treatment discontinuation (10.4% versus 4.5%) was numerically higher with durvalumab plus olaparib versus durvalumab plus placebo.

CONCLUSIONS

Maintenance therapy with durvalumab in combination with olaparib was not associated with a statistically significant improvement in PFS versus durvalumab alone, although numerical improvement was observed.

Differential Sensitivity of Germline and Somatic BRCA Variants to PARP Inhibitor in High-Grade Serous Ovarian Cancer

PURPOSE

The introduction of PARP inhibitors (PARPis) as a treatment option for patients with high-grade serous ovarian cancer (HGSOC) modified the approach of BRCA testing worldwide. In this study, we aim to evaluate the impact of BRCA1 and BRCA2 variants on treatment response and survival outcomes in patients diagnosed in our institution.

METHODS

A total of 805 HGSOC samples underwent BRCA1 and BRCA2 variant detection by using next-generation sequencing (NGS). Among them, a pathogenic alteration was detected in 104 specimens. Clinicopathological features and germline status were recovered, and alteration types were further characterized. The clinical significance of variant type in terms of response to chemotherapy and to PARPis as well as overall survival were evaluated using univariate analysis.

RESULTS

In our cohort, 13.2% of the HGSOC samples harbored a pathogenic BRCA1 or BRCA2 variant, among which 58.7% were inherited. No difference was observed between germline and somatic variants in terms of the gene altered. Interestingly, patients with somatic variants only (no germline) demonstrated better outcomes under PARPi treatment compared to those with germline ones.

CONCLUSION

The determination of the inheritance or acquisition of BRCA1 and BRCA2 alterations could provide valuable information for improving management strategies and predicting the outcome of patients with HGSOC.

Impact of DNA damage repair alterations on prostate cancer progression and metastasis

Prostate cancer is among the most common diseases worldwide. Despite recent progress with treatments, patients with advanced prostate cancer have poor outcomes and there is a high unmet need in this population. Understanding molecular determinants underlying prostate cancer and the aggressive phenotype of disease can help with design of better clinical trials and improve treatments for these patients. One of the pathways often altered in advanced prostate cancer is DNA damage response (DDR), including alterations in BRCA1/2 and other homologous recombination repair (HRR) genes. Alterations in the DDR pathway are particularly prevalent in metastatic prostate cancer. In this review, we summarise the prevalence of DDR alterations in primary and advanced prostate cancer and discuss the impact of alterations in the DDR pathway on aggressive disease phenotype, prognosis and the association of germline pathogenic alterations in DDR genes with risk of developing prostate cancer.

BRCA2 reversion mutation confers resistance to olaparib in breast cancer

Key Clinical Message

A rare missense mutation was identified as a reversion mutation using cancer genomic profiling and a suspected mechanism underlying resistance to olaparib in breast cancer.

Abstract

A 34-year-old woman with breast cancer and BRCA2: p.Gln3047Ter was treated with olaparib. After tumor progression, cancer genomic profiling testing using liquid biopsy revealed BRCA2 p.Gln3047Ter and p.Gln3047Tyr, with 48.9% and 0.37% allele frequency, respectively. These findings shed light on reversion mutation as a mechanism of resistance to olaparib in breast cancer.

Oncogenic SIRT7 inhibits GATA4 transcriptional activity and activates the Wnt signaling pathway in ovarian cancer

OBJECTIVE

Sirtuin-7 (SIRT7) is a class III histone deacetylase that plays an important role in cancer development and frequently overexpressed in carcinomas. In this study, the tumor-supporting role and underlying mechanisms of SIRT7 were characterized in ovarian cancer (OC) aggressiveness.

METHODS

SIRT7 expression was examined in OC tissues and cells. Interactions among SIRT7, GATA4, Wnt signaling pathway were explored by bioinformatics tools and experimental validations. The effect of SIRT7 and GATA4 on malignant phenotypes of OC cells were examined with gain- and loss-of-function experiments. A nude mouse model of OC was developed to verify the in vitro findings.

RESULTS

It was noted that SIRT7 was highly expressed in OC tissues and cells. Cell lines with higher SIRT7 expression (OVCAR-3 and OVCAR-8) were used for subsequent in vitro experiments. The experimental data indicated that silencing of SIRT7 suppressed the OC cell proliferation, colony formation, migration, and invasion, and promoted cell senescence, which could be abolished by GATA4 knockdown. Mechanistically, SIRT7 promoted deacetylation of GATA4 and consequently inhibited the transcriptional activity of GATA4. In addition, GATA4 induced OC cell senescence by inhibiting Wnt signaling pathway. Further in vivo experiments substantiated that SIRT7 knockdown or overexpressed GATA4 could effectively inhibit tumor growth of nude mice.

CONCLUSION

Taken together, our findings indicated that SIRT7 enhanced development of OC by suppressing GATA4 and activating Wnt signaling pathway, suggesting the potential of SIRT7/GATA4/Wnt axis as a therapeutic target for OC.

Therapeutic targeting of DNA damage repair pathways guided by homologous recombination deficiency scoring in ovarian cancers

The susceptibility of cells to DNA damage and their DNA repair ability are crucial for cancer therapy. Homologous recombination is one of the major repairing mechanisms for DNA double-strand breaks. Approximately half of ovarian cancer (OvCa) cells harbor homologous recombination deficiency (HRD). Considering that HRD is a major hallmark of OvCas, scholars proposed HRD scoring to evaluate the HRD degree and guide the choice of therapeutic strategies for OvCas. In the last decade, synthetic lethal strategy by targeting poly (ADP-ribose) polymerase (PARP) in HR-deficient OvCas has attracted considerable attention in view of its favorable clinical effort. We therefore suggested that the uses of other DNA damage/repair-targeted drugs in HR-deficient OvCas might also offer better clinical outcome. Here, we reviewed the current small molecule compounds that targeted DNA damage/repair pathways and discussed the HRD scoring system to guide their clinical uses.

Direct prediction of Homologous Recombination Deficiency from routine histology in ten different tumor types with attention-based Multiple Instance Learning: a development and validation study

Background

Homologous Recombination Deficiency (HRD) is a pan-cancer predictive biomarker that identifies patients who benefit from therapy with PARP inhibitors (PARPi). However, testing for HRD is highly complex. Here, we investigated whether Deep Learning can predict HRD status solely based on routine Hematoxylin & Eosin (H&E) histology images in ten cancer types.

Methods

We developed a fully automated deep learning pipeline with attention-weighted multiple instance learning (attMIL) to predict HRD status from histology images. A combined genomic scar HRD score, which integrated loss of heterozygosity (LOH), telomeric allelic imbalance (TAI) and large-scale state transitions (LST) was calculated from whole genome sequencing data for n=4,565 patients from two independent cohorts. The primary statistical endpoint was the Area Under the Receiver Operating Characteristic curve (AUROC) for the prediction of genomic scar HRD with a clinically used cutoff value.

Results

We found that HRD status is predictable in tumors of the endometrium, pancreas and lung, reaching cross-validated AUROCs of 0.79, 0.58 and 0.66. Predictions generalized well to an external cohort with AUROCs of 0.93, 0.81 and 0.73 respectively. Additionally, an HRD classifier trained on breast cancer yielded an AUROC of 0.78 in internal validation and was able to predict HRD in endometrial, prostate and pancreatic cancer with AUROCs of 0.87, 0.84 and 0.67 indicating a shared HRD-like phenotype is across tumor entities.

Conclusion

In this study, we show that HRD is directly predictable from H&E slides using attMIL within and across ten different tumor types.

Durvalumab Plus Olaparib in Previously Untreated, Platinum-Ineligible Patients With Metastatic Urothelial Carcinoma: A Multicenter, Randomized, Phase II Trial (BAYOU)

PURPOSE

Homologous recombination repair gene mutations (HRRm) are common in urothelial carcinoma (UC), rendering tumor cells sensitive to poly (ADP-ribose) polymerase (PARP) inhibition. We assessed efficacy and safety of durvalumab (anti-programmed cell death ligand-1) plus olaparib (PARP inhibitor) in patients with metastatic UC (mUC).

METHODS

This randomized, multicenter, double-blind, phase II trial enrolled untreated, platinum-ineligible patients with mUC. Patients (N = 154) were randomly assigned 1:1 to receive durvalumab (1,500 mg intravenously once every 4 weeks) plus olaparib (300 mg orally, twice daily) or durvalumab plus placebo. The primary end point was progression-free survival (PFS) assessed by investigators per RECIST version 1.1. Secondary end points included overall survival in all patients and PFS in patients with HRRm.

RESULTS

Overall, median PFS was 4.2 months (95% CI, 3.6 to 5.6) for durvalumab plus olaparib and 3.5 months (95% CI, 1.9 to 5.1) for durvalumab plus placebo (hazard ratio [HR], 0.94; 95% CI, 0.64 to 1.39; log-rank P value, .789). Median overall survival was 10.2 months (95% CI, 7.0 to 13.9) and 10.7 months (95% CI, 7.2 to 17.3), respectively (HR, 1.07; 95% CI, 0.72 to 1.61). In the 20% of patients with HRRm, median PFS was 5.6 months (95% CI, 1.9 to 8.1) and 1.8 months (95% CI, 1.7 to 2.2), respectively (HR, 0.18; 95% CI, 0.06 to 0.47). Treatment-related grade 3 or 4 adverse events occurred in 18% and 9% of patients, respectively.

CONCLUSION

Adding olaparib to durvalumab did not improve survival outcomes in an unselected mUC population. Efficacy outcomes with durvalumab were similar to those reported for other anti-programmed cell death-1/programmed cell death ligand-1 agents. However, the results of secondary analyses suggest a potential role for PARP inhibition in patients with UC harboring HRRm.

From Samples to Germline and Somatic Sequence Variation: A Focus on Next-Generation Sequencing in Melanoma Research

Next-generation sequencing (NGS) applications have flourished in the last decade, permitting the identification of cancer driver genes and profoundly expanding the possibilities of genomic studies of cancer, including melanoma. Here we aimed to present a technical review across many of the methodological approaches brought by the use of NGS applications with a focus on assessing germline and somatic sequence variation. We provide cautionary notes and discuss key technical details involved in library preparation, the most common problems with the samples, and guidance to circumvent them. We also provide an overview of the sequence-based methods for cancer genomics, exposing the pros and cons of targeted sequencing vs. exome or whole-genome sequencing (WGS), the fundamentals of the most common commercial platforms, and a comparison of throughputs and key applications. Details of the steps and the main software involved in the bioinformatics processing of the sequencing results, from preprocessing to variant prioritization and filtering, are also provided in the context of the full spectrum of genetic variation (SNVs, indels, CNVs, structural variation, and gene fusions). Finally, we put the emphasis on selected bioinformatic pipelines behind (a) short-read WGS identification of small germline and somatic variants, (b) detection of gene fusions from transcriptomes, and (c) de novo assembly of genomes from long-read WGS data. Overall, we provide comprehensive guidance across the main methodological procedures involved in obtaining sequencing results for the most common short- and long-read NGS platforms, highlighting key applications in melanoma research.

Opinion: PARP inhibitors in cancer-what do we still need to know?

PARP inhibitors (PARPi) have been demonstrated to exhibit profound anti-tumour activity in individuals whose cancers have a defect in the homologous recombination DNA repair pathway. Here, we describe the current consensus as to how PARPi work and how drug resistance to these agents emerges. We discuss the need to refine the current repertoire of clinical-grade companion biomarkers to be used with PARPi, so that patient stratification can be improved, the early emergence of drug resistance can be detected and dose-limiting toxicity can be predicted. We also highlight current thoughts about how PARPi resistance might be treated.

Volumetric Measurement of Relative CBV Using T1-Perfusion-Weighted MRI with High Temporal Resolution Compared with Traditional T2*-Perfusion-Weighted MRI in Postoperative Patients with High-Grade Gliomas

BACKGROUND AND PURPOSE

T1-PWI with high temporal resolution may provide a reliable relative CBV value as a valid alternative to T2*-PWI under increased susceptibility. The purpose of this study was to assess the technical and clinical performance of T1-relative CBV in patients with postoperative high-grade gliomas.

MATERIALS AND METHODS

Forty-five MRIs of 34 patients with proved high-grade gliomas were included. In all MRIs, T1- and T2*-PWIs were both acquired and processed semiautomatically to generate relative CBV maps using a released commercial software. Lesion masks were overlaid on the relative CBV maps, followed by a histogram of the whole VOI. The intraclass correlation coefficient and Bland-Altman plots were used for quantitative and qualitative comparisons. Signal loss from both methods was compared using the Wilcoxon signed-rank test of zero voxel percentage. The MRIs were divided into a progression group (n = 20) and a nonprogression group (n = 14) for receiver operating characteristic curve analysis.

RESULTS

Fair intertechnique consistency was observed between the 90th percentiles of the T1- and T2*-relative CBV values (intraclass correlation coefficient = 0.558, P < .001). T2*-PWI revealed a significantly higher percentage of near-zero voxels than T1-PWI (17.7% versus 3.1%, P < .001). There was no statistically significant difference between the area under the curve of T1- and T2*-relative CBV (0.811 versus 0.793, P = .835). T1-relative CBV showed 100% sensitivity and 57.1% specificity for the detection of progressive lesions.

CONCLUSIONS

T1-relative CBV demonstrated exquisite diagnostic performance for detecting progressive lesions in postoperative patients with high-grade gliomas, suggesting the potential role of T1-PWI as a valid alternative to the traditional T2*-PWI.

Homologous recombination deficiency in diverse cancer types and its correlation with platinum chemotherapy efficiency in ovarian cancer

BACKGROUND

Homologous recombination deficiency (HRD) is a molecular biomarker for administrating PARP inhibitor (PARPi) or platinum-based (Pt) chemotherapy. The most well-studied mechanism of causing HRD is pathogenic BRCA1/2 mutations, while HRD phenotype is also present in patients without BRCA1/2 alterations, suggesting other unknown factors.

METHODS

The targeted next-generation sequencing (GeneseeqPrime® HRD) was used to evaluate the HRD scores of 199 patients (Cohort I). In Cohort II, a total of 85 Pt-chemotherapy-treated high-grade serous ovarian cancer (HGSOC) patients were included for investigating the role of HRD score in predicting treatment efficacy. The concurrent genomic features analyzed along HRD score evaluation were studied in a third cohort with 416 solid tumor patients (Cohort III).

RESULTS

An HRD score ≥ 38 was predefined as HRD-positive by analyzing Cohort I (range: 0-107). Over 95% of the BRCA1/2-deficient cases of Cohort I were HRD-positive under this threshold. In Cohort II, Pt-sensitive patients have significantly higher HRD scores than Pt-resistant patients (median: 54 vs. 34, p = 0.031) and a significantly longer PFS was observed in HRD-positive patients (median: 548 vs. 343 days, p = 0.003). Furthermore, TP53, NCOR1, and PTK2 alterations were enriched in HRD-positive patients. In Cohort III, impaired homologous recombination repair pathway was more frequently observed in HRD-positive patients without BRCA1/2 pathogenic mutations. The alteration enrichment of TP53, NCOR1, and PTK2 observed in Cohort II was also validated by the ovarian subgroup in Cohort III.

CONCLUSIONS

Using an in-house HRD evaluation method, our findings show that overall HRR gene mutations account for a significant part of HRD in the absence of BRCA1/2 aberrations, and suggest that HRD positive status might be a predictive biomarker of Pt-chemotherapy.