The current study evaluated three biomarkers [homologous recombination deficiency (HRD), tumor BRCA1/2 (tBRCA) mutations, and CCNE1 copy-number variation (CNV)] in ovarian tumors from patients enrolled on the SCOTROC4 clinical trial for associations with outcome following carboplatin monotherapy. Ovarian tumors (n = 250), with high-grade serous (HGSOC) subgroup analysis (n = 179) were classified as HRD positive (HRD score ≥42 or tBRCA mutation) and as CCNE1 amplification positive (CCNE1 CNV score >2.4). Seventy-four (30%) tumors were HRD positive, including 34 (14%) with tBRCA mutations. Forty-seven (19%) were CCNE1 amplification positive, all of which were tBRCA wild-type. HRD and tBRCA, but not CCNE1 amplification, were significantly associated with CA125 complete response in the entire cohort (HRD, P = 0.00015; tBRCA P = 0.0096), and the HGSOC subgroup (HRD, P = 0.0016; tBRCA P = 0.032). HRD and lack of CCNE1 amplification were associated with improved progression-free survival (PFS) and overall survival (OS) in the full cohort and HGSOC subgroup (HRD, P = 0.00021; CCNE1 status P = 0.038). HRD remained significant for OS and PFS after adjusting for clinical factors, while CCNE1 status only remained significant for PFS. Patients with HRD-positive tumors had greater PFS and OS benefit from platinum dose intensification than HRD-negative tumors (P = 0.049 and P = 0.035, respectively). An alternative exploratory HRD score threshold (≥33 or tBRCA mutation) was also significantly associated with both PFS and OS in the HGSOC subset.

Implications: HRD, tumor BRCA1/2 mutations, and absence of CCNE1 amplification are associated with improved survival of ovarian cancer patients treated with platinum monotherapy and HRD-positive patients may benefit from platinum dose intensification. Mol Cancer Res; 16(7); 1103–11. ©2018 AACR.

Defects in the homologous recombination (HR) pathway are associated with increased sensitivity to DNA-damaging agents and targeted agents, such as PARP inhibitors, across many cancer types. The most well-studied markers of HR pathway defects are mutations in BRCA1 or BRCA2 (BRCA1/2). For example, previous studies have shown that triple-negative breast cancer (TNBC) tumors and ovarian cancer tumors with BRCA1/2 mutations show improved sensitivity to platinum-based chemotherapy relative to BRCA1/2 wild-type tumors (1, 2). Similarly, ovarian cancer tumors with mutations in BRCA1/2 have shown improved sensitivity to PARP inhibitors (3–5). However, defects in the HR pathway are not confined to mutations in BRCA1/2 in ovarian cancer. Studies report HR pathway defects in as many as 50% of epithelial ovarian cancers, a third of which may be caused by something other than a mutation in BRCA1 or BRCA2 (6).

In order to improve the identification of tumors with HR pathway defects that are likely to respond to DNA-damaging agents, a three-biomarker measure of homologous recombination deficiency (HRD) has been developed. The HRD assay quantitates genomic instability in a tumor genome (7) based on three independent measures of genomic instability: loss of heterozygosity (LOH; ref. 8), telomeric allelic imbalance (TAI; ref. 9), and large-scale state transition (LST; ref. 10). Each individual measure has been shown to be associated with response to platinum-based therapy in either TNBC or ovarian cancer (9–11), and the combined score has been shown to be a better predictor of HRD than any of the individual scores (12).

An HRD score threshold of 42 was recently developed in a cohort of breast and ovarian chemotherapy-naïve tumor samples with known BRCA1/2 deficiency status (13). This threshold is used in combination with tumor BRCA1/2 mutation status to differentiate tumors with HR deficiency (HRD positive; HRD score ≥42 or a tumor BRCA1/2 mutation) from HR nondeficient tumors (HRD negative; HRD score < 42 and wild-type BRCA1/2). In an independent cohort, HRD positive was significantly associated with response to platinum-based treatment in TNBC (13).

Copy-number amplification of the cell-cycle regulator cyclin E1 (CCNE1) is observed only in tumors with wild-type BRCA1/2 and has been associated with early primary treatment failure and reduced patient survival in ovarian cancer (14, 15). In a recent study, Etemadmoghadam and colleagues demonstrated that CCNE1-amplified ovarian tumors require the presence of functional BRCA1 protein and may be responsive to the proteasome inhibitor bortezomib (16). In addition, CCNE1-amplified ovarian xenograft models were observed to be sensitive to a combination of a CDK2 inhibitor and an AKT1 inhibitor in a high-throughput screen (17).

Here, we evaluated using a predefined analysis plan the association of three molecular biomarkers (HRD status using an HRD score of ≥42 or tBRCA mutation, BRCA1/2 mutations, and CCNE1 copy-number amplification) with clinical outcomes following monotherapy with the DNA-damaging agent carboplatin at primary presentation. This was done in a cohort of tumors from patients enrolled in the SCOTROC4 phase III trial of stage IC to IV epithelial ovarian carcinoma, primary fallopian tube carcinoma, or ovarian-type peritoneal carcinomatosis treated with platinum monotherapy, with or without dose intensification (18). Available clinical endpoints in this study included CA125 response, progression-free survival (PFS), and overall survival (OS). All three biomarkers were assessed for their ability to predict response to platinum monotherapy and for their association with patient survival outcomes.

Recently, the predictive power of the HRD threshold of ≥42 (5th percentile of HRD scores observed in BRCA1/2-mutant tumors) was evaluated for the prediction of PFS benefit due to the PARP inhibitor niraparib in second-line platinum-sensitive germline BRCA1/2-negative HGSOC (4). While the HRD ≥42 threshold was associated with significant niraparib PFS benefit, the patient group falling below this threshold also received significant, albeit reduced, benefit. These data suggest that a revision of the threshold might better define the responding patient group. To explore this concept in this study, we tested an HRD threshold of ≥33 (1st percentile of HRD scores observed in BRCA1/2-mutant tumors) against CA125 response, PFS, and OS in the HGSOC patient set.

The SCOTROC4 trial was a randomized trial of flat dosing versus intrapatient dose escalation of single-agent carboplatin as first-line chemotherapy for advanced ovarian cancer (18). Although the trial showed that intrapatient dose escalation of carboplatin based on nadir blood counts is feasible and safe, it provided no improvement in PFS or OS compared with flat dosing. However, we hypothesized that HRD-positive tumors might gain additional benefit from dose intensification and have explored potential differences depending on HRD status between patients in the dose escalation and flat dosing arms of the SCOTROC4 trial.

Patients and treatment

SCOTROC4 was a phase III randomized trial that enrolled patients with stage IC to IV epithelial ovarian carcinoma, primary fallopian tube carcinoma, or ovarian-type peritoneal carcinomatosis (18). Patients were randomized into treatment arms and received 6 cycles of 3 weekly carboplatin either at a flat dose or with an intrapatient dose escalation. The flow of patients and samples through the study is described in Supplementary Fig. S1. Tumor collection for this study was approved by local Ethics Committee and informed written consent was obtained from patient. Among patients from SCOTROC4 with epithelial ovarian carcinoma, 250 were included in this study based on patient consent and tumor sample availability. This includes 120 patients in the arm without dose intensification and 130 patients in the dose intensification arm. Based on pathologic review of tumor slides from all samples and TP53 mutation status, 179 samples were classified as HGSOC. Of 179 patients with HGSOC tumors, 115 were in the flat dose arm and 64 were in the dose escalation arm.

Clinical assessments and endpoints

Response to therapy was monitored by CA125 response (19). CA125 measurements were carried out at baseline, before each cycle of treatment, and then twice monthly. Patients were followed up for 2 years every 2 months and then every 3 months. PFS was determined according to RECIST version 1.0 (20). CT scans were carried out at baseline and after 6 cycles of treatment and also carried out if CA125 rose or clinical progression was suspected. PFS was the time from randomization until PD or death from any cause (whichever occurred first).

Molecular analysis

DNA from patient samples was extracted from three to five 10-μm formalin-fixed paraffin-embedded (FFPE) tissue sections from each available tumor sample after scraping areas with the highest tumor cell density (Promega Maxwell 16 LEV FFPE Plus kit AS1290, Promega). FFPE tissue was incubated overnight in 20 μL Proteinase K and 180 μL incubation buffer at 70°C in a shaking heat block. An additional 20 μL Proteinase K was then added, followed by 3 hours digestion at 70°C. RNase A (A1973, 10 μL; Promega) was added followed by RNA digestion at 37°C for 20 minutes. Lysis buffer (420 μL) was then added, and the samples were loaded into Maxwell cartridges. gDNA was eluted in 110 μL of water.

The DNA analysis approach used here has been previously described (13). Genome-wide SNP data were generated using a custom hybridization enrichment panel, which targets 54,091 SNPs distributed across the human genome. TP53, BRCA1, and BRCA2 mutation data were also evaluated in the context of this study. Details of the methods used for identification of BRCA1- and BRCA2-deficient tumors are provided in Timms and colleagues (7). Deleterious and suspected deleterious mutations were included in the analysis (21, 22).

Allelic imbalance profiles were generated to determine the scores for each individual biomarker component (TAI, LST, and LOH), and the combined HRD score is the sum of the individual biomarker scores (7, 13). An HRD score threshold of 42 (5th percentile of HRD scores observed in BRCA1/2-deficient tumors) has been previously developed to identify HR-deficient tumors (13). Tumors are considered HR deficient (HRD positive) if they have a high HRD score (≥42) or a tumor BRCA1 or BRCA2 (tBRCA) mutation and HR nondeficient (HRD negative) if they have a low HRD score (<42) and wild-type BRCA1/2 (13). In this study, we explored whether lowering the threshold from the 5th percentile level of HRD scores observed in BRCA1/2-deficient tumors (HRD score ≥42) to the 1st percentile (HRD score ≥33) might better define the responding patient group. In these analyses, HRD-positive status was defined as an HRD score either greater than or equal to the exploratory threshold of 33 or a BRCA1/2 mutant with any HRD score. This exploratory threshold was evaluated in the HGSOC subgroup only.

To identify tumors with CCNE1 copy-number amplification, the copy number was averaged for the 3 SNPs on the HRD SNP assay which surround the CCNE1 locus. The average copy number was then adjusted by the average copy number across all SNPs of the sample to produce a relative amplification score (Supplementary Fig. S2). CCNE1 amplification values of between 0.5 and 2 were considered to be within the accepted range for tumor sample variability and therefore did not represent CCNE1 amplification. Assuming these nonamplified samples to be log-normally distributed, the derived mean and standard deviation yielded at 99th percentile gave an amplification value of 2.4. Samples that exceed a CCNE1 amplification score of 2.4 were designated as CCNE1 amplification positive.

Statistical analysis

Clinical and molecular variables were evaluated as predictors of CA125 response in terms of odds ratios (OR) and Wald confidence intervals (CI) from logistic regression models. Associations with PFS and OS were assessed with hazard ratios (HR) from Cox proportional hazards (PH) models; categorical variables were also evaluated with Kaplan–Meier (KM) curves and Mantel–Cox log-rank tests. P values from logistic regression and Cox PH models were based on likelihood ratio tests. P values are reported as two-sided unless otherwise noted.

Study cohort

Patient demographic and clinical data are shown in Supplementary Table S1. CA125 response was available for 139 patients, while PFS and OS were available for all patients (N = 250). Overall, 74 (30%) of tumors were HRD positive (≥42), including 34 (14%) with tBRCA mutations, and 47 (19%) were identified as having amplification of CCNE1 (Supplementary Table S1). CCNE1 amplification was observed only in tumors without BRCA1/2 mutations, which is consistent with previous reports (14, 15). CCNE1 amplification was observed more frequently in HRD-negative tumors in this cohort (logistic P = 1.6 × 10−4; OR, 5.50; 95% CI, 1.89–16.0) compared with HRD positive (≥42). The HGSOC subset included 64 (36%) HRD positive (≥42) tumors, 29 (16%) of which had tBRCA mutations, and 39 (22%) tumors with CCNE1 amplification.

Association with response to platinum monotherapy

CA125 response and molecular results were available for 139 tumors from the entire cohort and 113 HGSOC tumors. The distribution of HRD scores stratified by CA125 response category is shown in Fig. 1. HRD (≥42) and tBRCA mutation status were both significantly associated with CA125 complete response (CR) in the entire cohort (P = 0.00015 and P = 0.0096, respectively), and in the subgroup of HGSOC patients (P = 0.0016 and P = 0.032, respectively; Supplementary Table S2). In the HGSOC subgroup, the HRD-positive rate increases from 37% to 52% when the HRD threshold is reduced from ≥42 to ≥33. HRD status defined as ≥33 or BRCA1/2 mutant remains statistically significantly associated with CA125 complete response (P = 5.0 × 10−4) (Supplementary Table S2). A receiver-operating curve (ROC) was used to compare sensitivity and specificity of different thresholds as predictors of CA125 response (Supplementary Fig. S2). CCNE1 amplification was not significantly associated with CA125 response in either the overall cohort or the HGSOC subgroup (Supplementary Table S2).

Figure 1.

Biomarker status and CA125 response. HRD status, tBRCA mutation status, and CCNE1 amplification as predictors of CA125 response in (A) the overall cohort (n = 137) and (B) the HGSOC subgroup. One BRCA1 mutation carrier is not shown due to failed HRD score. CR, complete response; PR, partial response; none, no response.

Figure 1.

Biomarker status and CA125 response. HRD status, tBRCA mutation status, and CCNE1 amplification as predictors of CA125 response in (A) the overall cohort (n = 137) and (B) the HGSOC subgroup. One BRCA1 mutation carrier is not shown due to failed HRD score. CR, complete response; PR, partial response; none, no response.

Close modal

In a multivariate logistic regression analysis of CA125 complete response adjusted for clinical variables (age at surgery, histology, grade, stage, bulk of residual disease after surgery, and performance status), HRD status remained significantly associated with response in the overall cohort (P = 3.6 × 10−4; Supplementary Table S3). Similarly, HRD status (≥42) retained statistical significance in the HGSOC subset after adjusting for clinical variables (P = 0.0050). In these multivariable analyses of the overall cohort and HGSOC subset (Supplementary Table S3), HRD status was the only variable that was significantly associated with CA125 response. HRD status as defined using the exploratory threshold of ≥33 also retained statistical significance after adjusting for clinical factors (P = 9.4 × 10−4; Supplementary Table S4). tBRCA was significantly associated with CA125 response in the full cohort (P = 0.049), but not the HGSOC subset after adjusting for clinical factors (Supplementary Table S5).

Association of HRD, tBRCA, and CCNE1 with PFS or OS

HRD status (≥42) was significantly associated with both improved PFS and OS in the overall cohort (P = 0.014 and P = 0.016, respectively) and in the HGSOC subgroup (P = 2.1 × 10−4 and P = 0.0011, respectively; Table 1). The HRD-positive rate in the HGSOC subgroup increases from 35.8% to 48.6% for PFS and OS when the threshold is reduced from ≥42 to ≥33. HRD status remains significantly associated with both improved PFS and OS in the HGSOC subgroup when the threshold is reduced to ≥33 in both univariate (P = 1.4 × 10−4 and P = 3.3 × 10−4, respectively; Table 1) and multivariate (P = 3.0 × 10−6and P = 3.1 × 10−4, respectively) Cox proportional hazards models (Supplementary Table S6). Improvements in median PFS and OS were similar to those observed for the prespecified threshold (Supplementary Fig. S3).

Table 1.

Univariate Cox PH analysis of PFS and OS for HRD and tBRCA

Overall cohortHGSOC subset
VariableLevelsHR (95% CI)PHR (95% CI)P
PFS 
HRD status (≥42) HRD positive 0.65 (0.46–0.93) 0.014 0.50 (0.34–0.73) 0.00021 
 HRD negative Ref  Ref  
HRD status (≥33) HRD positive ND ND 0.51 (0.36–0.72) 0.00014 
 HRD negative ND  Ref  
tBRCA mutation status Mutant 0.61 (0.38–0.99) 0.034 0.48 (0.29–0.79) 0.0017 
 Wild-type Ref  Ref  
CCNE1 amplification statusa Amplified 1.91 (1.32–2.75) 0.0011 1.56 (1.04–2.34) 0.038 
 Not amplified Ref  Ref  
OS 
HRD status (≥42) HRD positive 0.57 (0.36–0.92) 0.016 0.45 (0.27–0.74) 0.0011 
 HRD negative Ref  Ref  
HRD status (≥33) HRD positive ND ND 0.43 (0.27–0.69) 0.00033 
 HRD negative ND  Ref  
tBRCA mutation status Mutant 0.64 (0.35–1.17) 0.12 0.50 (0.26–0.95) 0.022 
 Wild-type Ref  Ref  
CCNE1 amplification statusa Amplified 1.82 (1.15–2.88) 0.015 1.72 (1.04–2.85) 0.043 
 Not amplified Ref  Ref  
Overall cohortHGSOC subset
VariableLevelsHR (95% CI)PHR (95% CI)P
PFS 
HRD status (≥42) HRD positive 0.65 (0.46–0.93) 0.014 0.50 (0.34–0.73) 0.00021 
 HRD negative Ref  Ref  
HRD status (≥33) HRD positive ND ND 0.51 (0.36–0.72) 0.00014 
 HRD negative ND  Ref  
tBRCA mutation status Mutant 0.61 (0.38–0.99) 0.034 0.48 (0.29–0.79) 0.0017 
 Wild-type Ref  Ref  
CCNE1 amplification statusa Amplified 1.91 (1.32–2.75) 0.0011 1.56 (1.04–2.34) 0.038 
 Not amplified Ref  Ref  
OS 
HRD status (≥42) HRD positive 0.57 (0.36–0.92) 0.016 0.45 (0.27–0.74) 0.0011 
 HRD negative Ref  Ref  
HRD status (≥33) HRD positive ND ND 0.43 (0.27–0.69) 0.00033 
 HRD negative ND  Ref  
tBRCA mutation status Mutant 0.64 (0.35–1.17) 0.12 0.50 (0.26–0.95) 0.022 
 Wild-type Ref  Ref  
CCNE1 amplification statusa Amplified 1.82 (1.15–2.88) 0.015 1.72 (1.04–2.85) 0.043 
 Not amplified Ref  Ref  

aCCNE1 amplification status was determined for 248 of 250 patients in the full cohort, and 178 of 179 patients in the HGSOC subcohort.

tBRCA mutation status was significantly associated with only PFS in the entire cohort (P = 0.034), and with both PFS and OS in the HGSOC subgroup (P = 0.0017 and P = 0.022, respectively; Table 1). CCNE1 amplification was significantly associated with both PFS and OS in the overall cohort (0.0011 and 0.015, respectively) and in the HGSOC subgroup (P = 0.038 and 0.043, respectively; Table 1).

In the overall cohort, significant improvements in median survival were observed for all three biomarkers (Fig. 2). HRD status was associated with a 7-month improvement in PFS (18.9 months for HR deficient vs. 11.6 months for nondeficient) and a 20-month improvement in OS (48.5 months for HR deficient vs. 28.1 months for nondeficient; Supplementary Table S7). Similarly, tBRCA mutations were associated with an 8-month improvement in PFS and 18-month improvement in OS. CCNE1 amplification was associated with a 6-month reduction in PFS and a 27-month reduction in OS. Similar associations were observed in the HGSOC subset (Fig. 3; Supplementary Table S8).

Figure 2.

Biomarker status and survival in the overall SCOTROC4 cohort. Kaplan–Meier survival curves for the overall cohort (N = 250) according to (A) HRD status, (B) tBRCA mutation status, and (C) CCNE1 amplification. Details of numbers of events and median survival with 95% CI are shown in Supplementary Table S7.

Figure 2.

Biomarker status and survival in the overall SCOTROC4 cohort. Kaplan–Meier survival curves for the overall cohort (N = 250) according to (A) HRD status, (B) tBRCA mutation status, and (C) CCNE1 amplification. Details of numbers of events and median survival with 95% CI are shown in Supplementary Table S7.

Close modal
Figure 3.

Biomarker status and survival in the HGSOC SCOTROC4 cohort. Kaplan–Meier survival curves for the HGSOC subgroup (N = 179) according to (A) HRD status, (B) tBRCA mutation status, and (C) CCNE1 amplification. Details of numbers of events and median survival with 95% CI are shown in Supplementary Table S8.

Figure 3.

Biomarker status and survival in the HGSOC SCOTROC4 cohort. Kaplan–Meier survival curves for the HGSOC subgroup (N = 179) according to (A) HRD status, (B) tBRCA mutation status, and (C) CCNE1 amplification. Details of numbers of events and median survival with 95% CI are shown in Supplementary Table S8.

Close modal

In multivariate Cox PH analyses including all patients, HRD status remained significantly associated with both PFS (P = 2.1 × 10−5) and OS (P = 0.0012; Table 2). Clinical variables, which were also significantly associated with outcome, were grade (P = 0.013 and 0.0064), stage (PFS only, P = 0.00014), and bulk of residual disease after surgery (PFS only, P = 0.0049; Table 2). Age at surgery, histology, and performance status were not significantly associated with either PFS or OS in this analysis. When multivariate analysis was restricted to HGSOC, HRD status remained significant for PFS and OS (P = 2.2 × 10−4 and P = 0.0048, respectively). Stage and bulk of residual disease also remained significant in the HGSOC subset for only PFS (P = 0.019 and P = 0.0055, respectively; Table 2). Age at surgery and performance status were not significantly associated with outcome in this analysis. In multivariable models restricted to the subset of tBRCA nonmutant patients, HRD status was significantly associated with PFS (P = 0.0023, HR, 0.50; 95% CI, 0.31–0.80) and OS (P = 0.015; HR, 0.47; 95% CI, 0.25–0.91) in the entire cohort (N = 216), and in HGSOC patients (N = 150; PFS P = 0.017, HR, 0.55; 95% CI, 0.33–0.92; OS P = 0.037, HR, 0.49; 95% CI, 0.24–0.99).

Table 2.

Multivariate Cox PH analysis of HRD as a predictor of PFS and OS

PFSOS
VariableLevelsPatients N (%)HR (95% CI)PHR (95% CI)P
All patients 
HRD status HRD positive 71 (31) 0.44 (0.30–0.65) 2.1 × 10−5 0.45 (0.27–0.74) 0.0012 
 HRD negative 155 (69) Ref  Ref  
Age at surgery Years 226 (100) 1.01 (0.99–1.02) 0.55 1.00 (0.98–1.03) 0.68 
Histology Serousa/clear cell 189 (84) 1.34 (0.72–2.49) 0.34 1.18 (0.53–2.63) 0.68 
 Other 37 (16) Ref  Ref  
Grade Low 20 (9) Ref 0.013 Ref 0.0064 
 High 206 (91) 2.59 (1.11–6.05)  4.70 (1.13–19.51)  
Stage IC–II 56 (25) Ref 0.00014 Ref 0.12 
 III 144 (64) 3.33 (1.80–6.16)  1.84 (0.84–4.05)  
 IV 26 (12) 2.37 (1.12–4.98)  1.13 (0.42–3.05)  
Bulk of residual disease None/microscopic 85 (38) Ref 0.0049 Ref 0.091 
 Macroscopic <2 cm 54 (24) 1.35 (0.80–2.30)  1.41 (0.69–2.86)  
 Macroscopic >2 cm 87 (38) 2.04 (1.28–3.24)  1.92 (1.03–3.61)  
Performance status 69 (31) Ref 0.19 Ref 0.17 
 122 (54) 1.17 (0.75–1.84)  1.02 (0.57–1.83)  
 35 (15) 1.66 (0.94–2.92)  1.73 (0.85–3.56)  
HGSOC 
HRD status HRD positive 63 (36) 0.46 (0.30–0.70) 2.2 × 10−4 0.47 (0.28–0.81) 0.0048 
 HRD negative 110 (64) Ref  Ref  
Age at surgery Years 173 (100) 1.01 (0.99–1.03) 0.39 1.02 (0.99–1.04) 0.19 
Stage IC–II 31 (18) Ref 0.019 Ref 0.12 
 III 120 (69) 2.12 (1.07–4.20)  1.59 (0.63–4.00)  
 IV 22 (13) 1.28 (0.56–2.90)  0.78 (0.25–2.49)  
Bulk of residual disease None/microscopic 49 (28) Ref 0.0055 Ref 0.32 
 Macroscopic <2 cm 48 (28) 1.37 (0.77–2.44)  1.11 (0.52–2.36)  
 Macroscopic >2 cm 76 (44) 2.15 (1.28–3.60)  1.54 (0.78–3.03)  
Performance status 41 (24) Ref 0.083 Ref 0.18 
 100 (58) 1.27 (0.75–2.15)  1.13 (0.56–2.30)  
 32 (18) 1.98 (1.05–3.75)  1.91 (0.83–4.38)  
PFSOS
VariableLevelsPatients N (%)HR (95% CI)PHR (95% CI)P
All patients 
HRD status HRD positive 71 (31) 0.44 (0.30–0.65) 2.1 × 10−5 0.45 (0.27–0.74) 0.0012 
 HRD negative 155 (69) Ref  Ref  
Age at surgery Years 226 (100) 1.01 (0.99–1.02) 0.55 1.00 (0.98–1.03) 0.68 
Histology Serousa/clear cell 189 (84) 1.34 (0.72–2.49) 0.34 1.18 (0.53–2.63) 0.68 
 Other 37 (16) Ref  Ref  
Grade Low 20 (9) Ref 0.013 Ref 0.0064 
 High 206 (91) 2.59 (1.11–6.05)  4.70 (1.13–19.51)  
Stage IC–II 56 (25) Ref 0.00014 Ref 0.12 
 III 144 (64) 3.33 (1.80–6.16)  1.84 (0.84–4.05)  
 IV 26 (12) 2.37 (1.12–4.98)  1.13 (0.42–3.05)  
Bulk of residual disease None/microscopic 85 (38) Ref 0.0049 Ref 0.091 
 Macroscopic <2 cm 54 (24) 1.35 (0.80–2.30)  1.41 (0.69–2.86)  
 Macroscopic >2 cm 87 (38) 2.04 (1.28–3.24)  1.92 (1.03–3.61)  
Performance status 69 (31) Ref 0.19 Ref 0.17 
 122 (54) 1.17 (0.75–1.84)  1.02 (0.57–1.83)  
 35 (15) 1.66 (0.94–2.92)  1.73 (0.85–3.56)  
HGSOC 
HRD status HRD positive 63 (36) 0.46 (0.30–0.70) 2.2 × 10−4 0.47 (0.28–0.81) 0.0048 
 HRD negative 110 (64) Ref  Ref  
Age at surgery Years 173 (100) 1.01 (0.99–1.03) 0.39 1.02 (0.99–1.04) 0.19 
Stage IC–II 31 (18) Ref 0.019 Ref 0.12 
 III 120 (69) 2.12 (1.07–4.20)  1.59 (0.63–4.00)  
 IV 22 (13) 1.28 (0.56–2.90)  0.78 (0.25–2.49)  
Bulk of residual disease None/microscopic 49 (28) Ref 0.0055 Ref 0.32 
 Macroscopic <2 cm 48 (28) 1.37 (0.77–2.44)  1.11 (0.52–2.36)  
 Macroscopic >2 cm 76 (44) 2.15 (1.28–3.60)  1.54 (0.78–3.03)  
Performance status 41 (24) Ref 0.083 Ref 0.18 
 100 (58) 1.27 (0.75–2.15)  1.13 (0.56–2.30)  
 32 (18) 1.98 (1.05–3.75)  1.91 (0.83–4.38)  

aOne patient with serous or endometrioid histology was categorized as serous for this analysis.

CCNE1 amplification was associated with PFS (P = 1.8 × 10−4) in the overall cohort after adjusting for clinical factors (Table 3). When multivariate analysis was restricted to HGSOC, CCNE1 amplification remained significant for PFS (P = 0.0033; Table 3). tBRCA was associated with PFS in the overall cohort (P = 0.0015) and the HGSOC subcohort (0.0019) after adjusting for clinical factors (Supplementary Table S9).

Table 3.

Multivariate Cox PH analysis of CCNE1 as a predictor of PFS and OS

PFSOS
VariableLevelsHR (95% CI)PHR (95% CI)P
Overall cohort (N = 225) 
CCNE1 status Amplified 2.19 (1.49–3.22) 1.8 × 10−4 1.63 (1.01–2.63) 0.052 
 Not amplified Ref  Ref  
Age at surgery Years 1.02 (1.00–1.03) 0.051 1.02 (0.99–1.04) 0.15 
Histology Serousa/clear cell 1.31 (0.71–2.43) 0.37 1.18 (0.53–2.62) 0.68 
 Other Ref  Ref  
Grade Low Ref 0.072 Ref 0.020 
 High 2.03 (0.87–4.73)  3.89 (0.94–16.1)  
Stage IC–II Ref 7.9 × 10−5 Ref 0.17 
 III 3.35 (1.86–6.03)  1.77 (0.83–3.80)  
 IV 2.57 (1.25–5.29)  1.17 (0.44–3.08)  
Bulk of residual disease None/microscopic Ref 0.0036 Ref 0.099 
 Macroscopic ≤2 cm 1.10 (0.66–1.83)  1.19 (0.59–2.38)  
 Macroscopic >2 cm 1.91 (1.21–3.04)  1.81 (0.96–3.38)  
Performance status Ref 0.48 Ref 0.28 
 1.06 (0.68–1.65)  0.92 (0.51–1.65)  
 1.37 (0.78–2.42)  1.45 (0.71–2.99)  
HGSOC subset (N = 172) 
CCNE1 status Amplified 1.95 (1.28–2.99) 0.0033 1.69 (1.01–2.84) 0.056 
 Not amplified Ref  Ref  
Age at surgery Years 1.02 (1.00–1.04) 0.019 1.03 (1.00–1.06) 0.018 
Stage IC–II Ref 0.010 Ref 0.13 
 III 2.45 (1.26–4.75)  1.71 (0.70–4.20)  
 IV 1.61 (0.73–3.57)  0.89 (0.29–2.75)  
Bulk of residual disease None/microscopic Ref 0.0031 Ref 0.25 
 Macroscopic ≤2 cm 1.07 (0.61–1.87)  0.94 (0.45–1.97)  
 Macroscopic >2 cm 1.99 (1.19–3.30)  1.45 (0.74–2.85)  
Performance status Ref 0.29 Ref 0.36 
 1.12 (0.67–1.88)  0.99 (0.49–2.01)  
 1.58 (0.84–2.97)  1.52 (0.66–3.51)  
PFSOS
VariableLevelsHR (95% CI)PHR (95% CI)P
Overall cohort (N = 225) 
CCNE1 status Amplified 2.19 (1.49–3.22) 1.8 × 10−4 1.63 (1.01–2.63) 0.052 
 Not amplified Ref  Ref  
Age at surgery Years 1.02 (1.00–1.03) 0.051 1.02 (0.99–1.04) 0.15 
Histology Serousa/clear cell 1.31 (0.71–2.43) 0.37 1.18 (0.53–2.62) 0.68 
 Other Ref  Ref  
Grade Low Ref 0.072 Ref 0.020 
 High 2.03 (0.87–4.73)  3.89 (0.94–16.1)  
Stage IC–II Ref 7.9 × 10−5 Ref 0.17 
 III 3.35 (1.86–6.03)  1.77 (0.83–3.80)  
 IV 2.57 (1.25–5.29)  1.17 (0.44–3.08)  
Bulk of residual disease None/microscopic Ref 0.0036 Ref 0.099 
 Macroscopic ≤2 cm 1.10 (0.66–1.83)  1.19 (0.59–2.38)  
 Macroscopic >2 cm 1.91 (1.21–3.04)  1.81 (0.96–3.38)  
Performance status Ref 0.48 Ref 0.28 
 1.06 (0.68–1.65)  0.92 (0.51–1.65)  
 1.37 (0.78–2.42)  1.45 (0.71–2.99)  
HGSOC subset (N = 172) 
CCNE1 status Amplified 1.95 (1.28–2.99) 0.0033 1.69 (1.01–2.84) 0.056 
 Not amplified Ref  Ref  
Age at surgery Years 1.02 (1.00–1.04) 0.019 1.03 (1.00–1.06) 0.018 
Stage IC–II Ref 0.010 Ref 0.13 
 III 2.45 (1.26–4.75)  1.71 (0.70–4.20)  
 IV 1.61 (0.73–3.57)  0.89 (0.29–2.75)  
Bulk of residual disease None/microscopic Ref 0.0031 Ref 0.25 
 Macroscopic ≤2 cm 1.07 (0.61–1.87)  0.94 (0.45–1.97)  
 Macroscopic >2 cm 1.99 (1.19–3.30)  1.45 (0.74–2.85)  
Performance status Ref 0.29 Ref 0.36 
 1.12 (0.67–1.88)  0.99 (0.49–2.01)  
 1.58 (0.84–2.97)  1.52 (0.66–3.51)  

In Cox PH analyses of the full cohort adjusted for clinical factors, HRD and CCNE1 amplification, HRD was associated with both PFS and OS (P = 7.3 × 10−4 and P = 0.0052 respectively) while CCNE1 amplification was associated with PFS only (P = 0.0087). When the same models were examined in the HGSOC subset, HRD maintained significant associations with both PFS and OS (P = 0.0027 and P = 0.019, respectively; Supplementary Table S10).

Association of HRD with dose intensification

We hypothesized that the improved outcomes observed for HRD-positive tumors were due to increased platinum sensitivity, and that these tumors might gain additional benefit from dose intensification. One hundred thirty patients were in the dose-intensified arm (42 HRD positive) and 120 patients (32 HRD positive) were in the arm without dose intensification. In subset analyses of both arms combined, there were no significant differences in PFS rates due to dose intensification in either the HRD-negative (HR, 1.13; 95% CI, 0.79–1.62) or HRD-positive (HR, 0.62; 95% CI, 0.33–1.14) groups. However, Cox PH analysis of the full cohort stratified by treatment arm suggested that the effect on PFS of platinum dose intensification was greater in the HRD-positive group (one-sided interaction P = 0.049). Similarly, for OS there were no significant differences in OS rates in the HRD-negative (HR 1.54; 95% CI, 0.96–2.45) or HRD-positive (HR, 0.61; 95% CI, 0.25–1.48) groups, but the effect of dose intensification on OS was significantly greater for HRD-positive tumors (one-sided interaction P = 0.035). These data support the hypothesis that patients with HR-deficient tumors may benefit from dose intensification by intrapatient carboplatin dose escalation.

The HR deficiency score based on measures of genomic instability and BRCA1/2 mutations are markers of HR pathway defects, and previous studies have demonstrated that these molecular markers predict response to DNA-damaging agents in some cancer types (1–5, 13–15, 23). In addition, CCNE1 amplification has been associated with chemotherapy resistance and poor prognosis in HGSOC (14, 15). Standard of care for first-line treatment of advanced ovarian cancer is carboplatin/paclitaxel combination therapy. However, the SCOTROC4 study provided an opportunity to investigate the ability of these three molecular markers to predict treatment response and outcomes following platinum monotherapy in a cohort of women with ovarian cancer and in the subset with HGSOC, thus avoiding potential confounding effects of paclitaxel. HGSOC histotype was based on pathological review of tumor slides by two gynecologic pathologists and TP53 mutation status. While we recognize the importance of defining histotype in this heterogeneous disease, some non–high-grade serous tumors (endometrioid and mucinous) can have defective homologous repair as determined by the HRD score (see Supplementary Table S1). Because the SCOTROC4 trial was all epithelial ovarian cancer, we had a predetermined analysis plan that would analyze all available tumors and then a high-grade serous subgroup analysis.

A positive relationship was observed between the HRD score and BRCA1/2 mutation status, which is consistent with previously published data (7, 13, 22). In addition, CCNE1 amplification was observed only in tumors without BRCA1/2 mutations, as previously reported (14, 15). A similar relationship was observed between low HRD score (<42) and CCNE1 amplification here, suggesting that CCNE1-amplified tumors may require functional homologous recombination repair or represent alternative tumor development pathways.

CA125 response data showed significant association with both HRD status and BRCA1/2 mutation status, but not with CCNE1 amplification. In multivariate analysis, only HRD status retained statistical significance. This result is consistent with previously published observations in both TNBC and ovarian cancer (3–5, 13, 23) and supports the hypothesis that HRD status (as defined by HRD score in combination with BRCA1/2 mutation screening) predicts sensitivity to DNA-damaging agents.

An exploratory analysis of an alternate HRD score threshold at the first percentile (≥33) of HRD scores in BRCA1/2-deficient tumors showed that HRD status remained significantly associated with CA125 response, while the fraction of biomarker positive to biomarker negative patients increased with the reduction in the HRD threshold. In a companion diagnostic context, such a threshold adjustment would enable more patients to receive drug benefit, although it will also increase the number of patients receiving treatment with limited benefit.

HRD and BRCA1/2 mutation status was also significantly associated with improved patient survival in this study, in both the overall cohort and in the HGSOC subgroup. CCNE1 amplification was also significantly associated with reduced survival in the overall study cohort, consistent with previous reports (14, 15). Both HRD status and CCNE1 amplification remained significantly associated with outcome in multivariate analysis.

Based on the positive association between HRD status and both response and outcome in this cohort, it was hypothesized that HRD-positive tumors would show more benefit from platinum dose intensification than HRD-negative tumors. The effect of dose intensification on PFS and OS was significantly greater in the HRD-positive group, suggesting that patients whose tumors are defective in HR may benefit from dose escalation based on intrapatient measures of toxicity as in the dose escalation arm of SCOTROC4 (18).

HRD status as defined by a three-biomarker HRD score in combination with BRCA1/2 mutation screening provided significant improvement over clinical variables in identifying patients with ovarian cancer who had improved response to platinum monotherapy, and was prognostic in this setting. HRD-positive tumors were observed predominantly in HGSOC tumors. In the clinical setting, the HRD test could be used to identify patients with increased likelihood of response to DNA-damaging agents, or other agents that target the DNA-damage repair pathways. CCNE1 amplification is also prognostic with patients whose tumors have amplification of this locus having significantly worse outcomes. Therapies that target this defect may provide an opportunity to improve outcomes for patients with CCNE1-amplified ovarian tumors.

K.M. Timms is employed with Myriad Genetics, Inc. and has ownership interest (including patents) in the same. E. Hughes and K. Brown are employed with Myriad Genetics, Inc. A. Gutin is SVP of Bioinformatics at Myriad Genetics, Inc. and is manager of automation engineering for the same. H. Gabra is vice president at AstraZeneca. J.S. Lanchbury is CSO at Myriad Genetics Inc. and has ownership interest (including patents) in the same. R. Brown reports receiving a commercial research grant from Myriad Genetics Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: E.A. Stronach, K.M. Timms, M. Perry, H. Gabra, J.S. Lanchbury, R. Brown

Development of methodology: E.A. Stronach, K.M. Timms, M. Perry, J.S. Lanchbury

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.A. Stronach, K.M. Timms, C. Neff, M. El-Bahrawy, J.H. Steel, X. Liu, N. Siddiqui, R. Brown

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.A. Stronach, J. Paul, K.M. Timms, E. Hughes, C. Neff, M. Perry, A. Gutin, M. El-Bahrawy, X. Liu, H. Gabra, R. Brown

Writing, review, and/or revision of the manuscript: E.A. Stronach, J. Paul, K.M. Timms, E. Hughes, K. Brown, C. Neff, M. Perry, M. El-Bahrawy, X. Liu, N. Siddiqui, H. Gabra, J.S. Lanchbury, R. Brown

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E.A. Stronach, J. Paul, K.M. Timms, K. Brown, J.H. Steel, L.-A. Lewsley

Study supervision: E.A. Stronach, K.M. Timms, N. Siddiqui, J.S. Lanchbury, R. Brown

Support for this study was provided by Cancer Research UK (A13086), Ovarian Cancer Action, Imperial Biomedical Research Centre, and CRUK/NIHR Imperial Experimental Cancer Medicine Centre.

The authors acknowledge support from Naina Patel, Yuepeng Wang, Nona Rama, and Charlotte Wilhelm-Benartzi for sample coordination. We also thank all the patients enrolled in SCOTROC4 who consented for their tumor samples to be used in this study.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Dann
RB
,
DeLoia
JA
,
Timms
KM
,
Zorn
KK
,
Potter
J
,
Flake
DD
 2nd
, et al
BRCA1/2 mutations and expression: response to platinum chemotherapy in patients with advanced stage epithelial ovarian cancer
.
Gynecol Oncol
2012
;
125
:
677
82
.
2.
Byrski
T
,
Huzarski
T
,
Dent
R
,
Gronwald
J
,
Zuziak
D
,
Cybulski
C
, et al
Response to neoadjuvant therapy with cisplatin in BRCA1-positive breast cancer patients
.
Breast Cancer Res Treat
2009
;
115
:
359
63
.
3.
Domchek
SM
,
Aghajanian
C
,
Shapira-Frommer
R
,
Schmutzler
RK
,
Audeh
MW
,
Friedlander
M
, et al
Efficacy and safety of olaparib monotherapy in germline BRCA1/2 mutation carriers with advanced ovarian cancer and three or more lines of prior therapy
.
Gynecol Oncol
2016
;
140
:
199
203
.
4.
Mirza
MR
,
Monk
BJ
,
Herrstedt
J
,
Oza
AM
,
Mahner
S
,
Redondo
A
, et al
Niraparib maintenance therapy in platinum-sensitive, recurrent ovarian cancer
.
N Engl J Med
2016
;
375
:
2154
64
.
5.
Swisher
EM
,
Lin
KK
,
Oza
AM
,
Scott
CL
,
Giordano
H
,
Sun
J
, et al
Rucaparib in relapsed, platinum-sensitive high-grade ovarian carcinoma (ARIEL2 Part 1): an international, multicentre, open-label, phase 2 trial
.
Lancet Oncol
2017
;
18
:
75
87
.
6.
Cancer Genome Atlas Research Network
. 
Integrated genomic analyses of ovarian carcinoma
.
Nature
2011
;
474
:
609
15
.
7.
Timms
KM
,
Abkevich
V
,
Hughes
E
,
Neff
C
,
Reid
J
,
Morris
B
, et al
Association of BRCA1/2 defects with genomic scores predictive of DNA damage repair deficiency among breast cancer subtypes
.
Breast Cancer Res
2014
;
16
:
475
.
8.
Abkevich
V
,
Timms
KM
,
Hennessy
BT
,
Potter
J
,
Carey
MS
,
Meyer
LA
, et al
Patterns of genomic loss of heterozygosity predict homologous recombination repair defects in epithelial ovarian cancer
.
Br J Cancer
2012
;
107
:
1776
82
.
9.
Birkbak
NJ
,
Wang
ZC
,
Kim
JY
,
Eklund
AC
,
Li
Q
,
Tian
R
, et al
Telomeric allelic imbalance indicates defective DNA repair and sensitivity to DNA-damaging agents
.
Cancer Discov
2012
;
2
:
366
75
.
10.
Popova
T
,
Manie
E
,
Rieunier
G
,
Caux-Moncoutier
V
,
Tirapo
C
,
Dubois
T
, et al
Ploidy and large-scale genomic instability consistently identify basal-like breast carcinomas with BRCA1/2 inactivation
.
Cancer Res
2012
;
72
:
5454
62
.
11.
Telli
ML
,
Jensen
KC
,
Vinayak
S
,
Kurian
AW
,
Lipson
JA
,
Flaherty
PJ
, et al
Phase II study of gemcitabine, carboplatin, and iniparib as neoadjuvant therapy for triple-negative and BRCA1/2 mutation-associated breast cancer with assessment of a tumor-based measure of genomic instability: PrECOG 0105
.
J Clin Oncol
2015
;
33
:
1895
901
.
12.
Mills
GB
,
Timms
KM
,
Reid
JE
,
Gutin
AS
,
Krivak
TC
,
Hennessy
B
, et al
Homologous recombination deficiency score shows superior association with outcome compared with its individual score components in platinum-treated serous ovarian cancer
.
Gynecol Oncol
2016
;
141
:
2
3
.
13.
Telli
ML
,
Timms
KM
,
Reid
JE
,
Hennessy
B
,
Mills
GB
,
Jensen
KC
, et al
Homologous recombination deficiency (HRD) score predicts response to platinum-containing neoadjuvant chemotherapy in patients with triple negative breast cancer
.
Clin Cancer Res
2016
;
22
:
3764
73
.
14.
Mayr
D
,
Kanitz
V
,
Anderegg
B
,
Luthardt
B
,
Engel
J
,
Lohrs
U
, et al
Analysis of gene amplification and prognostic markers in ovarian cancer using comparative genomic hybridization for microarrays and immunohistochemical analysis for tissue microarrays
.
Am J Clin Pathol
2006
;
126
:
101
9
.
15.
Etemadmoghadam
D
,
deFazio
A
,
Beroukhim
R
,
Mermel
C
,
George
J
,
Getz
G
, et al
Integrated genome-wide DNA copy number and expression analysis identifies distinct mechanisms of primary chemoresistance in ovarian carcinomas
.
Clin Cancer Res
2009
;
15
:
1417
27
.
16.
Etemadmoghadam
D
,
Weir
BA
,
Au-Yeung
G
,
Alsop
K
,
Mitchell
G
,
George
J
, et al
Synthetic lethality between CCNE1 amplification and loss of BRCA1
.
Proc Natl Acad Sci U S A
2013
;
110
:
19489
94
.
17.
Au-Yeung
G
,
Lang
F
,
Azar
WJ
,
Mitchell
C
,
Jarman
KE
,
Lackovic
K
, et al
Selective targeting of cyclin E1-amplified high-grade serous ovarian cancer by cyclin-dependent kinase 2 and AKT inhibition
.
Clin Cancer Res
2017
;
23
:
1862
74
.
18.
Banerjee
S
,
Rustin
G
,
Paul
J
,
Williams
C
,
Pledge
S
,
Gabra
H
, et al
A multicenter, randomized trial of flat dosing versus intrapatient dose escalation of single-agent carboplatin as first-line chemotherapy for advanced ovarian cancer: an SGCTG (SCOTROC 4) and ANZGOG study on behalf of GCIG
.
Ann Oncol
2013
;
24
:
679
87
.
19.
Rustin
GJ
,
Vergote
I
,
Eisenhauer
E
,
Pujade-Lauraine
E
,
Quinn
M
,
Thigpen
T
, et al
Definitions for response and progression in ovarian cancer clinical trials incorporating RECIST 1.1 and CA 125 agreed by the gynecological cancer intergroup (GCIG)
.
Int J Gynecol Cancer
2011
;
21
:
419
23
.
20.
Therasse
P
,
Arbuck
SG
,
Eisenhauer
EA
,
Wanders
J
,
Kaplan
RS
,
Rubinstein
L
, et al
New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada
.
J Natl Cancer Inst
2000
;
92
:
205
16
.
21.
Eggington
JM
,
Bowles
KR
,
Moyes
K
,
Manley
S
,
Esterling
L
,
Sizemore
S
, et al
A comprehensive laboratory-based program for classification of variants of uncertain significance in hereditary cancer genes
.
Clin Genet
2014
;
86
:
229
37
.
22.
Richards
S
,
Aziz
N
,
Bale
S
,
Bick
D
,
Das
S
,
Gastier-Foster
J
, et al
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American college of medical genetics and genomics and the association for molecular pathology
.
Genet Med
2015
;
17
:
405
24
.
23.
Kaklamani
VG
,
Jeruss
JS
,
Hughes
E
,
Siziopikou
K
,
Timms
KM
,
Gutin
A
, et al
Phase II neoadjuvant clinical trial of carboplatin and eribulin in women with triple negative early-stage breast cancer (NCT01372579)
.
Breast Cancer Res Treat
2015
;
151
:
629
38
.

Supplementary data