Adavosertib plus gemcitabine for platinum-resistant or platinum-refractory recurrent ovarian cancer: a double- blind, randomised, placebo-controlled, phase 2 trial
Stephanie Lheureux, Mihaela C Cristea, Jeffrey P Bruce, Swati Garg, Michael Cabanero, Gina Mantia-Smaldone, Alexander B Olawaiye, Susan L Ellard, Johanne I Weberpals, Andrea E Wahner Hendrickson, Gini F Fleming, Stephen Welch, Neesha C Dhani, Tracy Stockley, Prisni Rath, Katherine Karakasis, Gemma N Jones, Suzanne Jenkins, Jaime Rodriguez-Canales, Michael Tracy, Qian Tan, Valerie Bowering, Smitha Udagani, Lisa Wang, Charles A Kunos, Eric Chen, Trevor J Pugh, Amit M Oza
Summary
Background
The Wee1 (WEE1hu) inhibitor adavosertib and gemcitabine have shown preclinical synergy and promising activity in early phase clinical trials. We aimed to determine the efficacy of this combination in patients with ovarian cancer.
Methods
In this double-blind, randomised, placebo-controlled, phase 2 trial, women with measurable recurrent platinum- resistant or platinum-refractory high-grade serous ovarian cancer were recruited from 11 academic centres in the USA and Canada. Women were eligible if they were aged 18 years or older, had an Eastern Cooperative Oncology Group performance status of 0–2, a life expectancy of more than 3 months, and normal organ and marrow function. Women with ovarian cancer of non-high-grade serous histology were eligible for enrolment in a non-randomised exploratory cohort. Eligible participants with high-grade serous ovarian cancer were randomly assigned (2:1), using block randomisation (block size of three and six) and no stratification, to receive intravenous gemcitabine (1000 mg/m² on days 1, 8, and 15) with either oral adavosertib (175 mg) or identical placebo once daily on days 1, 2, 8, 9, 15, and 16, in 28-day cycles until disease progression or unacceptable toxicity. Patients and the team caring for each patient were masked to treatment assignment. The primary endpoint was progression-free survival. The safety and efficacy analysis population comprised all patients who received at least one dose of treatment. The trial is registered with ClinicalTrials.gov, NCT02151292, and is closed to accrual.
Findings
Between Sept 22, 2014, and May 30, 2018, 124 women were enrolled, of whom 99 had high-grade serous ovarian cancer and were randomly assigned to adavosertib plus gemcitabine (65 [66%]) or placebo plus gemcitabine (34 [34%]). 25 women with non-high-grade serous ovarian cancer were enrolled in the exploratory cohort. After randomisation, five patients with high-grade serous ovarian cancer were found to be ineligible (four in the experimental group and one in the control group) and did not receive treatment. Median age for all treated patients (n=119) was 62 years (IQR 54–67). Progression-free survival was longer with adavosertib plus gemcitabine (median 4·6 months [95% CI 3·6–6·4] with adavosertib plus gemcitabine vs 3·0 months [1·8–3·8] with placebo plus gemcitabine; hazard ratio 0·55 [95% CI 0·35–0·90]; log-rank p=0·015). The most frequent grade 3 or worse adverse events were haematological (neutropenia in 38 [62%] of 61 participants in the adavosertib plus gemcitabine group vs ten [30%] of 33 in the placebo plus gemcitabine group; thrombocytopenia in 19 [31%] of 61 in the adavosertib plus gemcitabine group vs two [6%] of 33 in the placebo plus gemcitabine group). There were no treatment-related deaths; two patients (one in each group in the high-grade serous ovarian cancer cohort) died while on study medication (from sepsis in the experimental group and from disease progression in the control group).
Interpretation
The observed clinical efficacy of a Wee1 inhibitor combined with gemcitabine supports ongoing assessment of DNA damage response drugs in high-grade serous ovarian cancer, a TP53-mutated tumour type with high replication stress. This therapeutic approach might be applicable to other tumour types with high replication stress; larger confirmatory studies are required.
Introduction
In 2018, an estimated 295 414 new cases and 184 799 deaths related to ovarian cancer occurred, making it the second leading cause of death from gynaecological malignancy worldwide.1 The most common subtype is high-gradeserous ovarian cancer.2 For newly diagnosed disease, debulking surgery and platinum-based chemotherapy are the mainstay of treatment,2 with or without antiangiogenic therapy and polyADP ribose polymerase (PARP) inhibitors. PARP inhibitors seem to be mostactive in patients whose tumours have BRCA1 or BRCA2 mutations, or both, or homologous recombination defi- ciency.3,4 However, disease typically recurs, and treatment becomes increasingly challenging with diminishing response to subsequent therapies because of emerging drug resistance. Potential therapeutic opportunities exploit the molecular landscape of high-grade serous ovarian cancer, which is characterised by high genomic instability and alterations in cell cycle (eg, TP53, CCNE1, RB1, PTEN, and CDK12) and DNA repair (BRCA1, BRCA2, and other homologous recombination defi- ciency) pathway genes.5,6 In platinum-sensitive recurrent disease, PARP inhibitors are considered standard-of-care maintenance therapy for patients previously untreated with a PARP inhibitor.2 However, platinum-resistant recurrent ovarian cancer remains a clinical challenge with few effective therapeutic options.
The DNA and RNA helicase Schlafen family member 11 (SLFN11) has been studied for its potential role in trans- lation of DNA damage response proteins.7 SLFN11 binds to replication protein A (RPA), and high SLFN11 levels have been postulated to inhibit homologous recom- bination repair by promoting the destabilisation of the interaction between RPA and single-strand DNA.8 Low levels or an absence of SLFN11 have been proposed as potential mechanisms of resistance to DNA-damaging agents (including platinum and gemcitabine) and PARP inhibitors.9–11
Interest is increasing in targeting DNA damage response to overcome platinum resistance. DNA damage response comprises a network of molecules involved in cell cycle regulation and DNA repair, including cell cycle checkpoint coordination of DNA damage repair withcell cycle progression.12,13 Cell cycle genes guard cellular integrity by halting proliferation at various checkpoints (G1/S, G2/M), allowing repair of damaged DNA.14 If the G1/S checkpoint is impaired, as in p53-deficient cancers, cancer cells rely on the G2/M checkpoint for DNA repair.15,16
Wee1 (WEE1hu) kinase is a crucial regulator of the G2/M checkpoint. The G2/M checkpoint prevents entry of the damaged DNA into mitosis and is altered in several cancers.17–19 TP53 mutations, which are ubiquitous in high-grade serous ovarian cancer, lead to increased dependency on S-phase and G2-phase checkpoints. Wee1 inhibition with adavosertib (AZD1775; MK1775) induces G2 checkpoint escape.20 Gemcitabine is an antimetabolite therapy and blocks the progression of cells through the G1/S phase. Combining gemcitabine with Wee1 inhibition can lead to mitotic catastrophe by compromising the G2/M checkpoint.15 Combinations of gemcitabine and adavosertib have shown synergistic effects in preclinical studies and promising activity in early phase clinical trials.21,22 Notably, in one of these preclinical studies, tumour regression of more than 50% was observed with the combination of gemcitabine and adavosertib in 25 (51%) of 49 TP53-mutated tumours compared with none of 23 TP53-wild-type tumours, suggesting that TP53 mutations might serve as a potential predictive factor for benefit from the combination.22 Because high-grade serous ovarian cancer is characterised by TP53 alterations,2 we aimed to assess the adavosertib plus gemcitabine regimen identified in a previous phase 1 study21 as treatment for recurrent platinum-resistant or platinum-refractory epithelial ovarian, fallopian tube, or primary peritoneal cancer (hereafter referred to collectively as ovarian cancer).
Methods
Study design and participants
This double-blind, randomised, placebo-controlled, two- arm, phase 2 trial, with a third non-randomised cohort, was done in 11 recruiting academic cancer centres or departments in the USA (six sites) and Canada (five sites; appendix p 1), all participating as members of the US National Cancer Institute Cancer Therapy Evaluation Program, Princess Margaret, California, Chicago and Mayo Consortia.
Eligible patients were aged 18 years or older, had histologically confirmed platinum-resistant or platinum- refractory high-grade serous ovarian cancer (unlimited number of previous treatment lines) with disease progression and measurable lesions according to Response Evaluation Criteria in Solid Tumors (RECIST; version 1.1), an Eastern Cooperative Oncology Group (ECOG) performance status of 0–2, a life expectancy of more than 3 months, normal organ and marrowfunction (as defined in the protocol), and, for women of childbearing potential, willing to use adequate contra- ception before study entry and for the duration of the study. Patients previously treated with gemcitabine or with active bowel obstruction were ineligible. All patients provided written informed consent and consented to mandatory collection of archival tumour tissue and paired (pre-treatment and on-treatment) tumour biopsy samples. Patients with ovarian cancer of non-high-grade serous histology were eligible for enrolment in a non- randomised exploratory cohort; all other eligibility criteria were identical to those in the randomised cohort. A full list of eligibility criteria is in the protocol (appendix pp 15–18).
The trial was approved by each participating site’s institutional review board or ethics committee and done in accordance with the Declaration of Helsinki and Good Clinical Practice standards. A copy of the protocol is available in the appendix (pp 2–68).
Randomisation and masking
Eligible patients were enrolled by investigators and randomly assigned (2:1) to receive gemcitabine in combination with either oral adavosertib or identical placebo. Random assignment was done centrally using block randomisation (block size of three and six) and computer-generated randomised codes provided by an independent unmasked statistician. Patients and theteam caring for each patient were masked to treatment assignment. There were no stratification factors.
Procedures
Patients with high-grade serous ovarian cancer in the randomisation cohort were given gemcitabine 1000 mg/m² intravenously on days 1, 8, and 15 witheither oral adavosertib (175 mg once daily on days 1, 2, 8,9, 15, and 16, on an empty stomach 1 h before or 2 h after a meal) or matched placebo on the same schedule, in 28-day cycles until disease progression or unac- ceptable toxicity. No crossover was permitted. All eligible patients with ovarian cancer with non-high- grade serous histology were enrolled in an exploratory single-arm cohort and were given adavosertib plus gemcitabine combination therapy as administered in the experimental group of the randomised cohort. Gemcitabine dose modifications were implemented according to the prescribing information (appendix pp 23–27). Primary prophylactic granulocyte colony- stimulating factor was not recommended but secondary prophylaxis was allowed.
Tumours were assessed every 8 weeks by CT scan according to RECIST (version 1.1). Adverse events were assessed continuously by the clinical team as reported by patients throughout the study and graded according to Common Terminology Criteria for Adverse Events (version 4.03).
Venous blood samples were collected before treatment, and at 1, 2, 4, and 8 h after adavosertib or placebo administration on days 1 and 2 of cycle 1. Plasma adavosertib concentrations were profiled on day 1 (0, 1, 2, 4,8, and 24 h) and day 2 (2, 4, and 8 h) using a validated liquid chromatography and tandem-mass spectrometry method (API 5000, Sciex, Framingham, MA, USA).23 Pharmacokinetic parameters were assessed in all evalu- able patients and calculated using non-compartmental methods (PKPlus 2.5, SimulationsPlus, Sunnyvale, CA, USA).
Formalin-fixed paraffin-embedded (FFPE) archival tissue from initial diagnosis was requested for central expert gynaecological pathology review and TP53 analysis. Sanger sequencing was done in a Clinical Laboratory Improvement Amendments-registered laboratory using the BigDye terminator cycle sequencing kit (version 3.1; ThermoFisher, Waltham, MA, USA) on a 3500xL genetic analyser (ThermoFisher), with sequences compared with the National Center for Biotechnology Information reference sequence (TP53: NM_000546.5). DNA was amplified using primers specific for exons 5–8 of the TP53 gene (Advanced Molecular Diagnostics Laboratory, Princess Margaret Cancer Centre, Toronto, ON, Canada; appendix p 69). Immunohistochemistry analysis was done according to previously described methods to act as a surrogate for underlying TP53 mutation status; immunohistochemistry scores of 0 (underexpression) or 2 or 3 (overexpression) were defined as abnormalp53 expression.24 Full details of immunohistochemistry methods are provided in the appendix (pp 70–71).
For exploratory genomic analyses, tumour biopsy was done before treatment and while on treatment (cycle 1,day 2 or 9, within 24 h after adavosertib administration).
Outcomes
The primary outcome was progression-free survival. Secondary outcomes included response rate (according to RECIST version 1.1 and Gynecologic Cancer InterGroup CA-125 criteria), overall survival after 1 year of follow-up, safety, and tolerability. An additional secondary objective was assessment of TP53 mutations and p53 protein expression as potential predictive markers for response and progression-free survival benefit. We used TP53 analysis of archival tissue (immunohistochemistry and Sanger sequencing) as an integrated biomarker, according to the National Cancer Institute definition.25
Exploratory outcomes, to be reported elsewhere, were assessment of patient-reported outcomes using Patient Reported Outcomes-Common Terminology Criteria for Adverse Events (PRO-CTCAE), assessment of concord- ance between TP53 mutations in the tumour specimen and in circulating tumour DNA; assessment of the correlation between response and levels of circulating DNA TP53 mutations; validation of pCDC2 (PCDH- alpha-C2) and gamma-H2AX (phosphorylated H2A.X) in skin and tumour tissue as a pharmacodynamic marker of therapy; and assessment of survival outcomes and response rate with respect to changes in pCDC2 and gamma-H2AX.
In post-hoc exploratory analyses, we assessed potential associations between whole-exome sequencing (including CCNE1) and RNA sequencing with clinical outcomes (progression-free survival, overall survival, and response rate), and associations between SLFN11 levels and response by RECIST. All patients were included in a pharmacokinetics substudy and all available samples were analysed.
Statistical analysis
In the randomised cohort, we planned to enrol75 evaluable patients with high-grade serous ovarian cancer (50 to the experimental group, 25 to the control group) over approximately 15 months. Progression-free survival events in 65 patients (anticipated after approxi- mately 8 months of follow-up) would provide 90% powerto detect a doubling of median progression-free survival from 3·5 to 7·0 months at a one-sided α significance level of 0·10.
In the open-label, non-randomised, exploratory, non- high-grade serous ovarian cancer cohort, an objective response in four of ten patients was considered to repre- sent a clinically meaningful signal of activity for adavosertib plus gemcitabine combination therapy.
The analysis population comprised all patients who received at least one dose of treatment. In the randomised cohort, we compared response rates (by RECIST and by CA-125 criteria) between treatment groups using Fisher’s exact test. We analysed progression-free survival and overall survival using the Kaplan-Meier method and we compared treatment groups using the log-rank test. For progression-free survival, patients who discontinued treatment without progression were censored at the date of treatment discontinuation. Progression-free survival without such censoring was an exploratory analysis. We used Kruskal-Wallis tests in post-hoc exploratory analyses of the association between SLFN11 levels and response.p values of 0·05 or less were considered to be significant for clinical data analyses. We used SAS (version 9.4) for all clinical data analyses and R (version 4.0.2) for genomic statistical analyses. This trial is registered with ClinicalTrials.gov, NCT02151292.
Role of the funding source
The funders Princess Margaret Cancer Foundation, US Department of Defense, Ontario Institute for Cancer Research, and US National Cancer Institute Cancer Therapy Evaluation Program supported trial conduct, patient enrolment, drug supply, genomic analysis, and analysis of samples and AstraZeneca provided study drug and some funds for genomic analysis through the US National Cancer Institute Cancer Therapy Evaluation Program, and did SLFN11 testing. However, the funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Results
Between Sept 22, 2014, and May 30, 2018, 124 women were enrolled, of whom 99 had high-grade serousovarian cancer and were randomly assigned to the adavosertib plus gemcitabine group (65 from ten sites) or placebo plus gemcitabine group (34 from ten sites). The remaining 25 patients with non-high-grade serous ovarian cancer from seven sites were enrolled in the exploratory cohort and given adavosertib plus gemcitabine (figure 1). Five patients with high-grade serous ovarian cancer were found to be ineligible after randomisation (three with bowel obstruction; one with increased creatinine levels; one admitted to hospital before enrolment) and thus received no treatment and were excluded from the analysis population. Median age for all treated patients (n=119) was 62 years (IQR 54–67). Baseline characteristics were generally well balanced between randomised treatment groups, except for a slight numerical imbalance in ECOG performance status and BRCA mutation status between treatment groups (table 1). In the exploratory cohort, histological types included low-grade serous, endometrioid, and clear-cell ovarian cancer.
As of March 1, 2019, in the high-grade serous ovarian cancer cohort, the median duration of therapy was four cycles (IQR 2–6) in the adavosertib plus gemcitabine group and three cycles (2–6) in the placebo plus gemcitabine group. The median gemcitabine dose intensity (delivered divided by planned dose based on the dose at cycle 1 day 1) was 67% (IQR 57–88) in the experimental group versus 75% (67–90) in the placebo group, and the median adavosertib dose intensity was 67% (58–89) in the experimental group.
Median progression-free survival was 4·6 months (95% CI 3·6–6·4) in the adavosertib plus gemcitabine group versus 3·0 months (1·8–3·8) in the placebo plus gemcitabine group (hazard ratio [HR] 0·55 [95% CI 0·35–0·90]; log-rank p=0·015; figure 2). The exploratory analysis of progression-free survival without censoring patients at the date of treatment discontinuation showed similar results (p=0·007). Median overall survival at the time of data cutoff for the final analysis was 11·4 months (95% CI 8·2–16·5) in the adavosertib plus gemcitabine group versus 7·2 months (5·2–13·2) in the placebo plus gemcitabine group (HR 0·56 [95% CI 0·35–0·91]; log-rank p=0·017; figure 2). The proportion of patients with a confirmed partial response (no complete responses) according to RECIST (version 1.1) was14 (23%) of 61 participants in the adavosertib plus gemcitabine group versus two (6%) of 33 in the placebo plus gemcitabine group (p=0·038; figure 3). The median duration of response in the 14 patients given adavosertib plus gemcitabine was 3·7 months (IQR 3·5–9·6). In the two patients who were given placebo plus gemcitabine, the duration of response was 5·8 months in one patient and 12·9 months in the other. A response according to CA-125 criteria was observed in 14 (23%) of 61 patients in the adavosertib plus gemcitabine group versus three (9%) of 33 in the placebo plus gemcitabine group (p=0·16). Additionally, 35 (57%) of 61 participants in theadavosertib plus gemcitabine group versus 24 (73%) of33 in the placebo plus gemcitabine group had stable disease (figure 3), which was durable (≥6 months) in four patients in each group.
Patients with non-high-grade serous ovarian cancer in the exploratory cohort received a median of two cycles (IQR 2–6) of adavosertib plus gemcitabine. Four (16%) of 25 participants had a partial response (one with serous and endometrioid histology, one with low-grade endo- metrioid cancer, one with carcinosarcoma, and one with clear-cell carcinoma) and nine (36%) patients had stable disease (appendix p 76). Samples were available for whole-exome sequencing from two patients with partial response; both had activating KRAS mutations and one also had a TP53 mutation (appendix p 77).
Among all participants, the most common grade 3 or worse adverse events were haematological and fatigue (table 2). Grade 3–4 neutropenia was observed in 38 (62%) of 61 patients in the adavosertib plus gemcitabine group versus ten (30%) of 33 in the placebo plus gemcitabine group. Grade 3–4 thrombocytopenia was also more frequent with adavosertib plus gemcitabine than with placebo plus gemcitabine (19 [31%] of 61 patients vs two [6%] of 33). The adverse events most commonly leading to dose interruption or dose reduction were haematological. There were no treatment-related deaths; two patients (one in each group in the high-grade serous ovarian cancer cohort) died on study (one from sepsis 12 days after the last doses of adavosertib and gemcitabine in the experimental group and one from disease progression 19 days after the last dose of placebo and 20 days after the last dose of gemcitabine in the control group). Adverse events led to treatment discontinuation in13 (21%) patients in the adavosertib plus gemcitabine group, no patients in the placebo plus gemcitabine group, and eight (32%) in the exploratory cohort.
In the high-grade serous ovarian cancer cohort, in the adavosertib plus gemcitabine group, day 1 adavosertib administration was delayed in 13 (21%) of 61 patients after cycle 1 and 52 (85%) patients had at least one dose modification (reduction or delay). Day 1 gemcitabine administration was delayed in 11 (18%) of 61 patients in the adavosertib plus gemcitabine group and seven (21%) of 33 in the placebo plus gemcitabine group, and the dose was modified in 56 (92%) patients in the adavosertib plus gemcitabine group and 23 (70%) in the placebo plus gemcitabine group. The reasons most frequently recorded for dose reduction or delay were haematological (neutropenia and thrombocyto- penia) in both treatment groups of the randomised cohort and in the non-high-grade serous cohort. All patients in the non-high-grade serous ovarian cancer exploratory cohort had adavosertib and gemcitabine dose modifications.
68 (79%) of 86 patients who were given combination therapy (for high-grade serous ovarian cancer or non- high-grade serous ovarian cancer) were evaluable forpharmacokinetic analysis. Adavosertib was rapidly absorbed with a median time to peak concentration of 2 h (IQR 2–2) in the high-grade serous ovarian cancer cohort and 2 h (2–4) in the non-high-grade serous ovarian cancer cohort. The mean elimination half-life was approximately 7·8 h (SD 2·2) in the high-grade serous ovarian cancer cohort and 7·8 h (SD 1·7) in the non-high- grade serous ovarian cancer cohort, similar to in previous reports (appendix pp 72, 78).23 We found no evidence that gemcitabine changed the pharmacokinetic profile of adavosertib compared with the expected profile.
TP53 status was assessed with Sanger sequencing in 108 patients (eight had insufficient or unavailable samples and three were not tested; appendix pp 73–75). 78 (72%) of108 patients were TP53 mutation positive by Sanger sequencing: 48 (86%) of 56 patients in the adavosertib plus gemcitabine group and 22 (67%) of 33 in the placebo plus gemcitabine group in the randomised cohort, andeight (42%) of 19 patients in the exploratory adavosertib plus gemcitabine group. 53 (68%) of 78 patients carried missense TP53 mutations, 19 (24%) carried frameshift or truncation mutations, four (5%) had splice sites, and two (3%) had deletions. In immunohistochemistry analysis, 82 (93%) of 88 evaluable patients with high- grade serous ovarian cancer had abnormal p53 expres- sion. Of 70 patients with high-grade serous ovarian cancer and TP53 mutations, 68 (97%) had abnormal p53 expres- sion by immunohistochemistry analysis, representing 97% concordance between mutation data and protein expression levels in this population (appendix pp 73–75).
Tumour tissue was available immediately before or 1–2 weeks after starting treatment in 59 (63%) of 94 patients who were treated in the high-grade serous ovarian cancer cohort. Whole-exome sequencing showed TP53 mutation in samples from 57 (97%) of 59 patients and BRCA1 or BRCA2 mutation, or both, in samples from 14 (24%) patients. The patients with TP53 and BRCA1 or BRCA2 mutations, or both, were evenly distributed between the two randomised treatment groups (figure 4). In two patients (one in each of the high-grade serous ovarian cancer treatment groups), we observed BRCA1 reversion mutants. A BRCA2 reversion mutant was observed in a patient with non-high-grade serous ovarian cancer (appendix p 77). Mutation profiles in genes related to replication stress, entry into S phase, and the homologousrecombination DNA damage response pathway were similar in the two high-grade serous ovarian cancer treatment groups (figure 4).
In post-hoc analyses, although restricted by small sample sizes and the exploratory nature of the genomic analyses, the addition of adavosertib to gemcitabine appeared to benefit patients with homologous recombination gene mutation (appendix p 79). The same pattern was observed for those with BRCA1 or BRCA2 mutations, or both (appendix p 79). Patients with CCNE1-amplified tumours were significantly more likely to respond to the com- bination (as defined by RECIST version 1.1) than those without (Fisher’s exact test p=0·013; appendix p 79). This association corresponded with a non-significant improvement in progression-free survival (log-rank p=0·13) and overall survival (log-rank p=0·15) in patients with CCNE1-amplified tumours and high-grade serous ovarian cancer in the adavosertib plus gemcitabine group (appendix p 79). This association was not observed in those in the placebo plus gemcitabine group; however, only one responder in this group had copy-number data available.
In the post-hoc exploratory analysis of SLFN11 in 73 patients with high-grade serous ovarian cancer and evaluable samples (46 in the adavosertib plus gemcitabine group, 27 in the placebo plus gemcitabine group), we were unable to detect a correlation between SLFN11 expression and progression-free survival or overallsurvival in either treatment group (appendix p 80). Similarly, best objective response showed no association with SLFN11 levels in the adavosertib plus gemcitabine group (median SLFN11 level: 50% [IQR 7–80] tumour staining by immunohistochemistry analysis in patients with partial response, 40% [8–70] in those with stable disease, and 35% [15–65] in those with progressive disease Kruskal-Wallis p=0·98; appendix p 80). In the placebo plus gemcitabine group, only one of two patients with a partial response had an evaluable tumour biopsy sample (SLFN11 level of 60% vs a median SLFN11 level of 30% [IQR 3–35] in patients with stable disease and 6% [1–33] in those with progressive disease; Kruskal-Wallis p=0·16; appendix p 80). We found no change in SLFN11 levels between pre-treatment and on-treatment samples in either treatment group in the high-grade serous ovarian cancer cohort (Kruskal-Wallis p=0·82 for the adavosertib plus gemcitabine group and p=0·32 for the placebo plus gemcitabine group; appendix p 80).
Discussion
In advanced-stage or heavily pretreated high-grade serous ovarian cancer, few options remain after conventional therapy. This is the first trial to show a significant benefit (progression-free survival, overall survival, and response rate according to RECIST) from the addition of adavosertib to gemcitabine in heavily pretreated platinum-resistant or platinum-refractory high-grade serous ovarian cancer, a setting of clinical need that has yet to be met with standard-of-care therapies. In ovarian cancer trials, improved overall survival is important but often challenging to show.26 However, our study population with heavily pretreated advanced stage disease, including refractory disease with fewer therapy options, allowed the demonstration of overall survival benefit. The adavosertib plus gemcitabine combination was associated with more haematological adverse effects (consistent with previous reports of adavosertib21), but these were generally manageable with intermittent dose modification and did not lead to severe complications. Adavosertib plus gemcitabine also showed signs of activity in rare histological subtypes of ovarian cancer (serous and endometrioid, low-grade endometrioid, carcinosarcoma, and clear cell). Interestingly, both of the two patients in the non-high- grade serous ovarian cancer cohort with available genomic profiling were KRAS mutation positive, suggesting involvement of replication stress in directing treatment response. The pharmacokinetic parameters of adavosertib combined with gemcitabine were similar to those from single-agent studies,23 indicating no drug– drug interaction between these two agents. Genomic analyses revealed that most patients with high-grade serous ovarian cancer had TP53 mutations, as expected. The mutation rate was lower when assessed with Sanger sequencing (72%), as estimated by target sequencing of selected exons, than when profiled with whole-exomesequencing, when 97% of evaluable patients were estimated to have TP53 mutation-positive tumours.
In platinum-sensitive high-grade serous ovarian cancer, the clinical benefit of adavosertib plus carboplatin and paclitaxel has also been seen for patients with different TP53 mutation subtypes, and so these subtypes have been identified as a possible biomarker for response.27 In this study, all assessed subsets of TP53 mutations derived clinical benefit from the addition of adavosertib to gemcitabine; no strong conclusions could be drawn from our study based on the small sample size of individual mutations.
In our post-hoc exploratory analyses, addition of adavosertib to gemcitabine improved progression-free survival in subgroups with homologous recombination gene mutations or pathogenic BRCA mutations in baseline biopsy samples. Overall survival showed a similar trend. We also observed an enhanced rate of partial response, as defined by RECIST (version 1.1), to the combination in patients with CCNE1-amplified tumours versus non- amplified tumours. CCNE1 amplification is a potential predictive marker of cytotoxic chemotherapy resistance. Overexpression of CCNE1 results in dysregulation of the G1/S checkpoint promoting replication stress and genomic instability. Targeted inhibition of Wee1 kinase, in conjunc- tion with CCNE1 amplification, might cause dysregulation of both the G1/S and G2/M phase checkpoints and thus cell death in TP53-mutated ovarian cancer. Explor- atory analyses suggest that patients with homologous recombination or BRCA mutation, or both, or CCNE1 amplification (putatively discordant genomic alterations28) could derive enhanced benefit from the combination. Given that our study population was heavily pretreated and previous platinum-based therapy is known to have an effect on homologous recombination competence,29 we cannot definitively conclude whether these correlations are directly caused by defective homologous recombination or other confounding phenotypic properties in tumours with homologous recombination gene mutation or CCNE1 amplifiation, or both. These hypothesis-generating findings require further investigation in a trial controlled for these molecular variables.
The DNA and RNA helicase SLFN11 is being considered as a potential biomarker to determine sensitivity towards various chemotherapeutic agents because studies have indicated a high correlation between SLFN11 expression and response to DNA damaging agents.30,31 We believe this is the first clinical study exploring a potential association between SLFN11 levels and sensitivity to gemcitabine, with or without DNA damage response inhibition, albeit this was a post-hoc exploratory analysis and was not powered to assess this association. We did not observe any correlation between baseline SLFN11 expression and outcome in either treatment group of the randomised cohort, although our ability to make conclusions is restricted by the relatively small sample size, particularly in the placebo plus gemcitabine group, and post-hoc nature of theanalysis. For the first time in clinical tissues, we found that SLFN11 levels do not change during gemcitabine treatment, with or without adavosertib, showing the relative stability of this biomarker. Preclinical data suggest that combining DNA-damaging agents with ATR inhib- ition can overcome resistance to DNA-damaging agents in low SLFN11-expressing models through the independent role of SLFN11 at the intra-S-phase checkpoint following replication stress.32–34 A randomised phase 2 trial showed the benefit of adding berzosertib, an ATR inhibitor therapy, to gemcitabine in platinum-resistant high-grade serous ovarian cancer,35 yet no molecular subtypes were described. The results from our study provide valuable insight and further assessment in larger clinical trials is recommended. Our study had several limitations, including in our randomisation approach. First, randomisation in a 2:1 ratio is associated with reduced statistical power, a factor that was taken into consideration when the trial was designed by increasing the sample size by 12% compared with a 1:1 randomisation design to detect the same magnitude of effect with equivalent power. Second, we did not have any stratification factors because of the small sample size, potentially resulting in imbalances in BRCA mutation status, which is a known prognostic factor. Moderate numerical differences in BRCA mutation status by treatment group were identified by molecular profiling at the time of study entry, which could lead to bias. However, in this relatively small trial, adjustment for all possible confounding factors is difficult. A related weakness is the absence of information on BRCA status in about a third of participants in each group of the randomised cohort, potentially masking further imbal- ances. Another limitation of our study is the absence of quality-of-life assessment; however, assessment of patient- reported quality-of-life measures using PRO-CTCAE isunderway and will be reported in the future.
At first glance, the gemcitabine control group in our study might be perceived as having poor outcomes. In comparison, response rates without bevacizumab in the AURELIA trial36 were 30% with weekly paclitaxel, 8% with pegylated liposomal doxorubicin, and 0% with topotecan, compared with the 6% response rate (RECIST version 1.1) we observed in the gemcitabine control group of our trial. However, unlike AURELIA, we enrolled a heavily pretreated population with platinum-resistant disease (with no restriction on the number of previous anticancer regimens) and included platinum-refractory high-grade serous ovarian cancer. Most patients in our trial had previously received weekly paclitaxel, topotecan, or pegylated liposomal doxorubicin, representing a setting that could be described as a post-AURELIA population. Results from our study can be generalised to the wider population of patients with platinum-resistant or platinum-refractory high-grade serous ovarian cancer. However, further assessment is required to understand the role of this regimen in non-high-grade serous ovarian cancer and TP53-mutated cancer.
Finally, the relatively high proportion of patients requiring dose reduction might suggest that the regimen could be optimised for further assessment. The combi- nation schedule assessed in this study was based on the recommended regimen from a previous phase 1 study,21 and followed appropriate dose modification guidelines. Although such an approach is reasonable for future phase 3 trials, a lower gemcitabine dose could be considered. An alternative approach would be to assess the combination earlier in the treatment course and in a less heavily pretreated population in which haematological toxicity might be less problematic.
In summary, our results show significantly extended progression-free survival and overall survival with the addition of adavosertib to gemcitabine in platinum- resistant or platinum-refractory advanced high-grade serous ovarian cancer. Activity was also observed in less common histological subtypes of epithelial ovarian cancer. Most adverse events were haematological and managed medically without substantial complication.
References
1 Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.CA Cancer J Clin 2018; 68: 394–424.
2 Lheureux S, Gourley C, Vergote I, Oza AM. Epithelial ovarian cancer. Lancet 2019; 393: 1240–53.
3 González-Martín A, Pothuri B, Vergote I, et al. Niraparib in patients with newly diagnosed advanced ovarian cancer. N Engl J Med 2019; 381: 2391–402.
4 Ray-Coquard I, Pautier P, Pignata S, et al. Olaparib plus bevacizumab as first-line maintenance in ovarian cancer. N Engl J Med 2019;381: 2416–28.
5 Cancer Genome Atlas Research Network. Integrated genomic analyses of ovarian carcinoma. Nature 2011; 474: 609–15.
6 Patch AM, Christie EL, Etemadmoghadam D, et al. Whole-genome characterization of chemoresistant ovarian cancer. Nature 2015; 521: 489–94.
7 Li M, Kao E, Malone D, Gao X, Wang JYJ, David M. DNA damage- induced cell death relies on SLFN11-dependent cleavage of distinct type II tRNAs. Nat Struct Mol Biol 2018; 25: 1047–58.
8 Mu Y, Lou J, Srivastava M, et al. SLFN11 inhibits checkpoint maintenance and homologous recombination repair. EMBO Rep 2016; 17: 94–109.
9 Zoppoli G, Regairaz M, Leo E, et al. Putative DNA/RNA helicase Schlafen-11 (SLFN11) sensitizes cancer cells to DNA-damaging agents. Proc Natl Acad Sci USA 2012; 109: 15030–35.
10 Shee K, Wells JD, Jiang A, Miller TW. Integrated pan-cancer gene expression and drug sensitivity analysis reveals SLFN11 mRNA as a solid tumor biomarker predictive of sensitivity to DNA-damaging chemotherapy. PLoS One 2019; 14: e0224267.
11 Conteduca V, Ku SY, Puca L, et al. SLFN11 expression in advanced prostate cancer and response to platinum-based chemotherapy. Mol Cancer Ther 2020; 19: 1157–64.
12 Hustedt N, Durocher D. The control of DNA repair by the cell cycle.Nat Cell Biol 2016; 19: 1–9.
13 Branzei D, Foiani M. Regulation of DNA repair throughout the cell cycle. Nat Rev Mol Cell Biol 2008; 9: 297–308.
14 Funk JO. Cell cycle checkpoint genes and cancer. In: Encyclopaedia of life sciences. John Wiley & Sons, Jan 27, 2006. https://onlinelibrary.wiley.com/doi/full/10.1038/npg.els.0006046 (accessed Dec 1, 2020).
15 Matheson CJ, Backos DS, Reigan P. Targeting WEE1 kinase in cancer. Trends Pharmacol Sci 2016; 37: 872–81.
16 Dixon H, Norbury CJ. Therapeutic exploitation of checkpoint defects in cancer cells lacking p53 function. Cell Cycle 2002; 1: 362–68.
17 Murrow LM, Garimella SV, Jones TL, Caplen NJ, Lipkowitz S. Identification of WEE1 as a potential molecular target in cancer cells by RNAi screening of the human tyrosine kinome.
Breast Cancer Res Treat 2010; 122: 347–57.
18 Kreahling JM, Foroutan P, Reed D, et al. Wee1 inhibition by MK-1775 leads to tumor inhibition and enhances efficacy of gemcitabine in human sarcomas. PLoS One 2013; 8: e57523.
19 Slipicevic A, Holth A, Hellesylt E, Tropé CG, Davidson B, Flørenes VA. Wee1 is a novel independent prognostic marker of poor survival in post-chemotherapy ovarian carcinoma effusions. Gynecol Oncol 2014; 135: 118–24.
20 Guertin AD, Li J, Liu Y, et al. Preclinical evaluation of the WEE1 inhibitor MK-1775 as single-agent anticancer therapy.Mol Cancer Ther 2013; 12: 1442–52.
21 Leijen S, van Geel RM, Pavlick AC, et al. Phase I study evaluating WEE1 inhibitor AZD1775 as monotherapy and in combination with gemcitabine, cisplatin, or carboplatin in patients with advanced solid tumors. J Clin Oncol 2016; 34: 4371–80.
22 Rajeshkumar NV, De Oliveira E, Ottenhof N, et al. MK-1775,a potent Wee1 inhibitor, synergizes with gemcitabine to achieve tumor regressions, selectively in p53-deficient pancreatic cancer xenografts. Clin Cancer Res 2011; 17: 2799–806.
23 Do K, Wilsker D, Ji J, et al. Phase I study of single-agent AZD1775 (MK-1775), a Wee1 kinase inhibitor, in patients with refractory solid tumors. J Clin Oncol 2015; 33: 3409–15.
24 Mandilaras V, Garg S, Cabanero M, et al. TP53 mutations in high grade serous ovarian cancer and impact on clinical outcomes:a comparison of next generation sequencing and bioinformatics analyses. Int J Gynecol Cancer 2019; 29: 346–52.
25 Dancey JE, Dobbin KK, Groshen S, et al. Guidelines for the development and incorporation of biomarker studies in early clinical trials of novel agents. Clin Cancer Res 2010; 16: 1745–55.
26 Wilson MK, Pujade-Lauraine E, Aoki D, et al. Fifth Ovarian Cancer Consensus Conference of the Gynecologic Cancer InterGroup: recurrent disease. Ann Oncol 2017; 28: 727–32.
27 Oza AM, Estevez-Diz M, Grischke EM, et al. A biomarker-enriched, randomized phase II trial of adavosertib (AZD1775) plus paclitaxel and carboplatin for women with platinum-sensitive TP53-mutant ovarian cancer. Clin Cancer Res 2020; 26: 4767–76.
28 Etemadmoghadam D, Weir BA, Au-Yeung G, et al. Synthetic lethality between CCNE1 amplification and loss of BRCA1. Proc Natl Acad Sci USA 2013; 110: 19489–94.
29 Konstantinopoulos PA, Ceccaldi R, Shapiro GI, D’Andrea AD. Homologous recombination deficiency: exploiting the fundamental vulnerability of ovarian cancer. Cancer Discov 2015; 5: 1137–54.
30 Luan J, Gao X, Hu F, Zhang Y, Gou X. SLFN11 is a general target for enhancing the sensitivity of cancer to chemotherapy(DNA-damaging agents). J Drug Target 2020; 28: 33–40.
31 Winkler C, Armenia J, Jones GN, et al. SLFN11 informs on standard of care and novel treatments in a wide range of cancer models.Br J Cancer 2020; published online Dec 18. https://doi.org/10.1038/ s41416-020-01199-4.
32 Coussy F, El-Botty R, Château-Joubert S, et al. BRCAness, SLFN11, and RB1 loss predict response to topoisomerase I inhibitors in triple-negative breast cancers. Sci Transl Med 2020; 12: eaax2625.
33 Murai J, Feng Y, Yu GK, et al. Resistance to PARP inhibitors by SLFN11 inactivation can be overcome by ATR inhibition. Oncotarget 2016; 7: 76534–50.
34 Murai J, Tang SW, Leo E, et al. SLFN11 blocks stressed replication forks independently of ATR. Mol Cell 2018; 69: 371–84.e6.
35 Konstantinopoulos PA, Cheng SC, Wahner Hendrickson AE, et al. Berzosertib plus gemcitabine versus gemcitabine alone in platinum-resistant high-grade serous ovarian cancer: a multicentre,open-label, randomised, phase 2 trial. Lancet Oncol 2020; 21: 957–68.
36 Poveda AM, Selle F, Hilpert F, et al. Bevacizumab combined with weekly paclitaxel, pegylated liposomal doxorubicin, or topotecan in platinum-resistant recurrent ovarian cancer: analysis by chemotherapy cohort of the randomized phase III AURELIA trial of Adavosertib. J Clin Oncol 2015; 33: 3836–38.