Overcoming the challenges of drug development in platinum-resistant ovarian cancer

The definition of “platinum-resistant ovarian cancer” has evolved; it now also reflects cancers for which platinum treatment is no longer an option. Standard of care for platinum-resistant ovarian cancer is single-agent, non-platinum chemotherapy with or without bevacizumab, which produces modest response rates, with the greatest benefits achieved using weekly paclitaxel. Several recent phase 3 trials of pretreated patients with prior bevacizumab exposure failed to meet their primary efficacy endpoints, highlighting the challenge in improving clinical outcomes among these patients. Combination treatment with antiangiogenics has improved outcomes, whereas combination strategies with immune checkpoint inhibitors have yielded modest results. Despite extensive translational research, there has been a lack of reliable and established biomarkers that predict treatment response in platinum-resistant ovarian cancer. Additionally, in the platinum-resistant setting, implications for the time between the penultimate dose of platinum therapy and platinum retreatment remain an area of debate. Addressing the unmet need for an effective treatment in the platinum-resistant setting requires thoughtful clinical trial design based on a growing understanding of the disease. Recent cancer drug approvals highlight the value of incorporating molecular phenotypes to better define patients who are more likely to respond to novel therapies. Clinical trials designed per the Gynecologic Cancer InterGroup recommendations—which advocate against relying solely upon the platinum-free interval—will help advance our understanding of recurrent ovarian cancer response where platinum rechallenge in the platinum-resistant setting may be considered. The inclusion of biomarkers in clinical trials will improve patient stratification and potentially demonstrate correlations with biomarker expression and duration of response. With the efficacy of antibody-drug conjugates shown for the treatment of some solid and hematologic cancers, current trials are evaluating the use of various novel conjugates in the setting of platinum-resistant ovarian cancer. Emerging novel treatments coupled with combination trials and biomarker explorations offer encouraging results for potential strategies to improve response rates and prolong progression-free survival in this population with high unmet need. This review outlines existing data from contemporary clinical trials of patients with platinum-resistant ovarian cancer and suggests historical synthetic benchmarks for non-randomized trials.

The definition of "platinum-resistant ovarian cancer" has evolved; it now also reflects cancers for which platinum treatment is no longer an option.Standard of care for platinum-resistant ovarian cancer is single-agent, non-platinum chemotherapy with or without bevacizumab, which produces modest response rates, with the greatest benefits achieved using weekly paclitaxel.Several recent phase 3 trials of pretreated patients with prior bevacizumab exposure failed to meet their primary efficacy endpoints, highlighting the challenge in improving clinical outcomes among these patients.Combination treatment with antiangiogenics has improved outcomes, whereas combination strategies with immune checkpoint inhibitors have yielded modest results.Despite extensive translational research, there has been a lack of reliable and established biomarkers that predict treatment response in platinum-resistant ovarian cancer.Additionally, in the platinum-resistant setting, implications for the time between the penultimate dose of platinum therapy and platinum retreatment remain an area of debate.Addressing the unmet need for an effective treatment in the platinum-resistant setting requires thoughtful clinical trial design based on a growing understanding of the disease.Recent cancer drug approvals highlight the value of incorporating molecular phenotypes to better define patients who are more likely to respond to novel therapies.Clinical trials designed per the Gynecologic Cancer InterGroup recommendations-which advocate against relying solely upon the platinum-free interval-will help advance our understanding of recurrent ovarian cancer response where platinum rechallenge in the platinum-resistant setting may be considered.The inclusion of biomarkers in clinical trials will improve patient stratification and potentially demonstrate correlations with biomarker expression and duration of response.With the efficacy of antibody-drug conjugates shown for the treatment of some solid and hematologic cancers, current trials are evaluating the use of various novel conjugates in the setting of platinum-resistant ovarian cancer.Emerging novel treatments coupled with combination trials and 1 Introduction: Current landscape in platinum-resistant ovarian cancer Although surgery and platinum-based chemotherapy are effective treatment strategies for advanced-stage epithelial ovarian cancer, most tumors will relapse within several years and develop resistance to platinum-based therapy (1).This state, termed platinum-resistant ovarian cancer (PROC), is defined as progression within 6 months of the last platinum-based regimen (2, 3).Despite advances in ovarian cancer treatment over the past decade, PROC remains a lethal disease with limited therapeutic options (2).
The current standard treatment for PROC is single-agent nonplatinum chemotherapy-the most common of which are pegylated liposomal doxorubicin (PLD), paclitaxel, gemcitabine, and topotecan-with or without bevacizumab (2).As monotherapies, the clinical benefit of non-platinum single agents is modest in pretreated patients, with low objective response rates (ORR, 6%-13%), median progression-free survival (mPFS) of <6 months, and overall survival (OS) of <1 year (4)(5)(6).However, adding bevacizumab markedly increases clinical benefit (ORR, 31%; mPFS, 6.7 months; median OS [mOS], 16.6 months), particularly when combined with weekly paclitaxel (7,8).Among tumors with BRCA mutations or homologous recombination deficiency (HRD), poly[adenosine diphosphate-ribose] polymerase inhibitors (PARPi) have changed the front-line treatment landscape.The role of PARPi in PROC is limited.For tumors that are PARPi naïve and BRCA mutated, the response rate in a single-arm study approached 30% (9).For tumors that are BRCA wild-type, PARPi are not considered a treatment option, as the response rate is <5% (2, 9).Most recently, mirvetuximab soravtansine, an antibody-drug conjugate (ADC) targeting folate receptor alpha (FRa), was approved by the US Food and Drug Administration (FDA) for patients with PROC and high FRa expression (10).
Bevacizumab, a monoclonal antibody that inhibits vascular endothelial growth factor (VEGF), received approval in the US and Europe in 2014 for use in combination with chemotherapy in PROC based on the results of the phase 3, randomized AURELIA trial (7,11,12).In AURELIA, treatment with bevacizumab combined with PLD, paclitaxel, or topotecan resulted in a longer mPFS compared to chemotherapy alone (6.7 vs 3.4 months [hazard ratio (HR), 0.48; 95% CI, 0.38-0.60;P<.001]) (7).Furthermore, the ORR of chemotherapy plus bevacizumab was also higher than that of chemotherapy alone (30.9% vs 12.6% [95% CI, 9.6-27.0;P<.001]) (7).In a correspondence published after the primary AURELIA manuscript, the paclitaxel plus bevacizumab cohort exhibited the greatest benefit, with an ORR of 53.3% (8).Interestingly, the paclitaxel monotherapy cohort performed notably well, with an ORR of 30.2% (8).Importantly, the AURELIA study population was limited to patients with ≤2 prior anticancer regimens, excluded tumors that progressed on platinum-based therapy, and included few patients with prior bevacizumab exposure (7).Single-agent bevacizumab first showed efficacy in the platinum-resistant, laterline treatment setting (13).
Bevacizumab is approved for first-line treatment of stage III/IV ovarian cancer in platinum-sensitive ovarian cancer (PSOC) and in PROC (<2 prior lines of therapy), impacting the relevance of the AURELIA data in contemporary patient cohorts (11).The AURELIA patient population may no longer accurately represent the PROC population, making response comparisons between historical and contemporary cohorts challenging.
When PROC progresses after treatment with bevacizumabcontaining regimens, single-agent non-platinum chemotherapy is usually implemented (2).The efficacy of single-agent non-platinum chemotherapy with (n=52) or without (n=51) bevacizumab in Japanese patients with PROC that recurred after a bevacizumabcontaining chemotherapy regimen was investigated in the openlabel, randomized, phase 2 JGOG3023 trial (14).Although patients receiving chemotherapy plus bevacizumab had a numerically greater mOS than those receiving chemotherapy alone (15.3 vs 11.3 months [HR, 0.67; 95% CI, 0.38-1.17;P=.1556]) and a higher ORR (25.0%vs 13.7% [P=.0599]), the results were not significant, likely due to a small sample size (14).

Drug development in platinumresistant ovarian cancer
Since the breakthroughs of paclitaxel, topotecan, and PLD in the 1990s, ovarian cancer treatments stagnated until 2014.Bevacizumab and PARPi transformed the front-line treatment of high-grade epithelial ovarian cancer, resulting in multiple new FDA approvals granted between 2014-2020 (Figure 1) (12).This shifted the ovarian cancer treatment paradigm to include combination and maintenance regimens after initial diagnosis and surgical resection.Conversely, 1 PROC treatment has been approved since the approval of bevacizumab plus chemotherapy in 2014 (10,12).

Opportunities and challenges
Numerous unsuccessful clinical trials in PROC underscore the need to identify effective treatments.Novel combinations using existing therapies may improve outcomes.Recent phase 2 and 3 clinical trials of combination therapies are summarized in Table 3.
Antiangiogenic strategies have been pivotal in treating ovarian cancer, and combination therapies show promising results (3).TRIAS, a phase 2 clinical trial that evaluated a sequential combination of sorafenib-a non-selective oral multi-kinase inhibitor of VEGF-with topotecan, was the first to significantly improve OS in patients with PROC or platinum-refractory ovarian cancer (ie, cancer that progressed during platinum therapy) (30).PFS also significantly improved (30).In the open-label, phase 2b CONCERTO trial, the combination of olaparib and cediranib-an antiangiogenic targeting VEGF-demonstrated modest efficacy (ORR, 15.3%; mPFS, 5.1 months; mOS, 13.2 months) in heavily treated patients (all ≥3 prior lines of therapy) with non-germline-BRCA1/2-mutated PROC (38).However, this combination failed to demonstrate superior PFS vs standard-of-care paclitaxel in the open-label, phase 2 BAROCCO trial in a less heavily treated (60%, ≥3 prior lines of therapy) PROC population, of whom 12% had a BRCA mutation (39).
Immunotherapeutic approaches may hold promise in ovarian cancer; however, combination strategies using immune checkpoint inhibitors have yielded modest results to date (3).In the open-label, phase 2 NRG GY003 trial, patients with recurrent or persistent FIGURE 1 FDA-Approved Agents in Ovarian Cancer by Initial Approval Date (10,20,21).Figure 1 shows all drugs approved (and currently approved, as of November 2022) for the treatment of ovarian cancer.Agents in black boxes are indicated for the treatment of platinum-resistant ovarian cancer.FDA, US Food and Drug Administration; PLD, pegylated liposomal doxorubicin.
Of note, baseline clinical characteristics such as number of prior lines of chemotherapy, prior treatment regimens (eg, bevacizumab), histologic subtype, and mutational status are important to consider when interpreting efficacy outcomes in PROC.Therefore, much of this information has been described in Tables 1-3, which report outcomes in relevant clinical trials in PROC.

New approaches to drug development
Addressing the unmet need in PROC requires thoughtful clinical trial designs and novel effective agents.Recent cancer drug approvals emphasize incorporating molecular phenotypes into clinical trial designs to better define patient and tumor characteristics that may benefit from the study drug.The era of precision medicine has witnessed tumor-agnostic approvals, based on expression of a common biomarker rather than simply the tumor type (defined by primary site of origin) (45).In 2017, the FDA approved pembrolizumab for unresectable/metastatic solid tumors progressing after prior treatment in adult and pediatric patients based on microsatellite instability-high status or mismatch repair deficiency (dMMR), irrespective of tumor type (46).In 2020, pembrolizumab received approval to treat adult and pediatric patients with unresectable/metastatic tumor mutational burdenhigh solid tumors (excluding central nervous system cancers) and progression after prior therapy (46).Similarly, larotrectinib and entrectinib, 2 selective inhibitors of tropomyosin receptor kinases, received approval for treating unresectable/metastatic solid tumors with a neurotrophic tyrosine receptor kinase gene fusion in adult and pediatric patients, regardless of tumor origin (47,48).Additional recent tumor-agnostic approvals include dabrafenib plus trametinib (based on BRAF V600E or V600K mutations) and dostarlimab (based on dMMR) (49)(50)(51).
Unfortunately, high-grade serous ovarian cancers are not commonly associated with specific driver mutations, but rather by chromosomal instability and copy number alterations (52).Considering histology in this cancer type, while recognizing that some patients (eg, those with high FRa expression) may benefit from targeted treatment, is necessary (45).Tumor-agnostic indications may be most appropriate for drugs with high response rates and for rare tumors (45).As our understanding of ovarian cancer evolves, we must include biomarkers, companion diagnostics, and molecular profiling in clinical trial designs.
Further, umbrella and basket protocol designs can compare different novel regimens or incorporate different patient populations within 1 trial.Notably, in the investigator-initiated, phase 2 AMBITION trial in PROC, patients with HRD test positive tumors were randomized to combination olaparib/cediranib or olaparib/durvalumab, while those with HRD test negative tumors were randomized by PD-L1 expression to durvalumab/ chemotherapy, durvalumab/tremelimumab at 75 mg/ chemotherapy, or durvalumab/tremelimumab at 300 mg/ chemotherapy (53).Additionally, the open-label, phase 2 BOUQUET trial is an ongoing biomarker-driven trial evaluating multiple biomarker-based therapies in patients with persistent or recurrent rare epithelial ovarian tumors, including but not limited to low-grade serous ovarian carcinoma, clear cell carcinoma, mucinous carcinoma, and carcinosarcoma (54).Data from this trial may inform biomarker-driven patient stratification and treatment selection.
Finally, clinical trials in the PROC setting may have limited generalizability to real-world patient populations due to eligibility criteria that restrict enrollment to moderately pretreated individuals with potentially fewer comorbidities.Additionally, some trials lack an appropriately diverse patient population, limiting our ability to understand clinical benefit across certain racial and ethnic groups.There are active initiatives to address these shortcomings by both ensuring diversity in enrollment while simultaneously mandating more practical inclusion and exclusion criteria in PROC trials.

Accelerated approvals, benefits, and pitfalls
The FDA's accelerated approval program (and analogous non-US programs) can expedite approvals of drugs that address unmet needs for serious or life-threatening conditions (55).Accelerated approval can be based on a surrogate or intermediate clinical endpoint "reasonably likely to predict clinical benefit," providing earlier evaluation of clinical benefit based on well-controlled phase 2 studies (55).Although OS is the most objective benchmark for demonstrating clinical benefit in oncology clinical trials, it requires a larger sample size and longer follow-up compared with more rapidly assessed endpoints (eg, time to tumor progression and PFS) (56).Since ORR can be assessed using a single-arm trial, it is the most common surrogate endpoint for accelerated approvals (57).Duration of response (DOR) is also occasionally used to support ORR (57).The gold standard for endpoints in PROC randomized trials remains OS, but when OS is confounded by long postprogression survival (>18 months) and crossover (common in trial participants), PFS is the preferred endpoint.Open-label studies where PFS is a key endpoint should utilize placebo controls and blinded independent central review to objectify clinical activity.Several treatments for ovarian cancer have benefitted from accelerated approval (58).PLD's accelerated approval occurred in 1999 based on 3 phase 2 studies, with confirmation from a randomized phase 3 trial (58,59).However, some accelerated approvals in oncology do not demonstrate clinical benefit in confirmatory studies (57).Notably, bevacizumab received accelerated approval for metastatic breast cancer in 2008 based on improved PFS in an open-label, phase 3 clinical trial (60,61).Subsequent placebo-controlled, double-blind confirmatory trials failed to confirm that the magnitude of the effect on PFS constituted a clinical benefit and showed more TRAEs compared to chemotherapy, which led to a revoked approval for this indication (62).
In ovarian cancer, the voluntary withdrawals of rucaparib, olaparib, and niraparib for recurrent, late-line treatment demonstrate the vulnerabilities of accelerated approval (63)(64)(65).These withdrawals stemmed from non-hypothesis-tested subset analyses, which suggested a detrimental effect on OS with PARPi exposure when compared to chemotherapy.
Although post-approval studies typically take years to complete (median [range], 3.4 years [0.5-12.6]),as of November 2022, only 21 indications (for 16 different drugs) with accelerated approvals have been withdrawn, while 88 were verified in confirmatory studies (57,66,67).While accelerated approvals usually endure, choosing appropriate surrogate endpoints and designing proper single-arm trials that can predict meaningful treatment effects remain challenging.
6 Moving away from "platinumresistant ovarian cancer" Over the past decade, the definition of "platinum resistant" has shifted from utilizing only the historical definition of disease recurrence <6 months after the completion of last platinum-based chemotherapy (2, 68).In clinical trials and practice, PFI-based classification has been accepted for predicting chemotherapy response, disease prognosis, and patient selection and stratification (68).However, platinum response is neither binary, nor accurately represented by an arbitrary cutoff based solely on the time of diagnosis of recurrence (68,69).Further, evidence suggests that PFI may not be optimal for predicting clinical response in all cases.Previous trials such as AURELIA have shown clinical benefit independent of PFI (7).Thus, a PFI of <6 months may not equate with resistance to platinum agents.A recent meta-analysis reported a 36% response rate to platinum-based regimens vs a 16% response rate to non-platinum-based regimens in a PROC population (70).Therefore, platinum rechallenge remains a viable option for some historically defined patients with PROC.Although the term "PROC" has evolved in clinical practice, it is still useful for regulatory approval due to previously discussed uncertainties, where platinum retreatment may not be the best option.
Several factors can affect time to relapse and platinum sensitivity.Variable timing and application of tools to detect recurrence (eg, cancer antigen-125, computed tomography scan, positron emission tomography scan) can skew the time to relapse and impact historical "platinum sensitive" vs "platinum resistant" designations (68,69).Further, the tumor's molecular profile can influence response.Patients with BRCA1/2-mutated recurrent ovarian cancer may respond to platinum-based and other chemotherapy agents that induce direct DNA damage (69,71,72).In ovarian cancer, the type, number, and outcome of prior therapies are important considerations when assessing potential response to further treatments (68).However, it is unclear what impact maintenance therapy with targeted and biological agents may have on subsequent treatment response (68).Other influential factors include histological subtype, prior surgical interventions, symptoms at recurrence, and the molecular profile beyond BRCA (eg, HRD status) (68,69).
These factors have been incorporated in a new proposed classification system for patients with recurrent ovarian cancer and in the Gynecologic Cancer InterGroup (GCIG) consensus recommendations for recurrent ovarian cancer clinical trials (68,69).The GCIG recommends TFI replaces PFI, with specific reporting of the TFI from last platinum dose (TFIp) and the specific method used to diagnose recurrence (68).Additionally, the GCIG recommends recording the TFI from last non-platinum therapy (TFInp) and last biological agent (TFIb), where applicable (68).The GCIG also notes that biomarkers will likely play a more predictive role than TFIp regarding treatment response (68).
Historically, developing new PROC regimens without reliable and established biomarkers for treatment response has been challenging.The recent validation of FRa as a predictive biomarker for the ADC mirvetuximab soravtansine heralds a new strategy for drug development where tumor surface antigens may be identified, quantified, and targeted.For trials targeting immune mechanisms, biomarkers such as tumor-infiltrating lymphocytes and tumor mutational burden (a tumor's total number of somatic coding mutations) may guide which patients may respond to immunotherapy (3); however, these biomarkers have not been validated.
Including biomarkers in future clinical trials will help establish their relevance in ovarian cancer and their relationship to efficacy endpoints, including DOR.This will help identify measures to inform patient response beyond the 6-month PFI cutoff.

Future perspectives
Developing effective therapies for patients with PROC is complex and necessitates consideration of tumor and patient heterogeneity.Ongoing and recently completed clinical trials feature novel agents, which utilize combination treatments and targeted, biomarker-based therapies (Table 4).
Since ADCs have succeeded in treating other cancers, the use of this novel approach of targeting cytotoxic payloads to tumor cells is also being evaluated in PROC (73).
Given the heterogeneity of PROC, combination therapies may improve clinical outcomes.Chemotherapy is being explored in combination with antiangiogenics, immune checkpoint inhibitors, PARPi, agents targeting DNA damage repair, agents targeting AXL and AKT/ERK, and a glucocorticoid receptor modulator (Table 4).Non-chemotherapeutic immunotherapy regimens being investigated include existing and novel immune checkpoint inhibitors combined with antiangiogenics, an engineered IL-2, an anti-TIGIT monoclonal antibody, and a p53 vaccine.PARPi are being studied in non-chemotherapeutic combinations with antiangiogenics and inhibitors of PI3K and WEE1.Triplet combinations utilizing various agents, including PARPi, anti-PD-1/PD-L1, and antiangiogenics, are underway.Other ongoing and recently completed phase 2 and 3 trials are summarized in Table 4.
These novel PROC treatments may improve response rates and prolong PFS for patients with few options.However, improving the treatment paradigm will require overcoming several obstacles.Disease heterogeneity is a major challenge that may have influenced prior trial failure (41).Identifying a well-defined homogeneous population will be critical to ensure impactful trial outcomes.Increased testing for biomarkers such as tumor mutational burden and dMMR, particularly in mucinous, clear cell, and endometrioid histologies, is encouraged, as is participation in appropriate clinical trials (81).Another challenge in PROC is the frailty of patients-due to age, comorbidities, and/or toxicity from multiple rounds of prior therapy-who are frequently recruited for trials and often have rapidly progressive disease (68).In 1 study, 30%-50% of patients discontinued due to progressive disease before receiving sufficient doses of therapy, and therefore had no chance of clinical benefit (15).Focusing on more specific populations may improve outcomes by excluding patients unlikely to be positively impacted.

Conclusion
Addressing the unmet need for effective therapies in heavily pretreated PROC is an ongoing challenge.As research elucidates the molecular mechanisms of ovarian cancer progression, new insights will provide guidance on developing novel targeted therapies.

TABLE 1
Recent phase 3 trials in patients with platinum-resistant recurrent ovarian cancer comparing single-agent chemotherapy to an experimental agent a .

TABLE 2
Select monotherapy clinical trials with results in heavily pretreated patients with platinum-resistant ovarian cancer.

TABLE 3
Recent clinical trials using combination therapy approaches in pretreated patients with platinum-resistant ovarian cancer.