Transforming Ovarian Cancer Care by Targeting Minimal Residual Disease

Summary

Frontline treatment and resultant cure rates in patients with advanced ovarian cancer have changed little over the past several decades. Here we outline a multidisciplinary approach aimed at gaining novel therapeutic insights by focusing on the poorly understood minimal residual disease phase of ovarian cancer that leads to eventual incurable recurrences.

An Opportunity to Transform the Current Treatment Paradigm

High-grade serous ovarian cancer (HGSOC), the most common histologic subtype of the disease, is often diagnosed at advanced stage. Based on patient and disease factors, women with suspected advanced ovarian cancer either undergo upfront cytoreductive surgery followed by adjuvant chemotherapy, or neoadjuvant chemotherapy, interval cytoreductive surgery, and additional adjuvant therapy. Despite high rates of initial response to surgery and platinum/taxane based chemotherapy, the vast majority of patients recur with a median progression free survival (PFS) of approximately 18 months. Recently, poly(ADP-ribose) polymerase inhibitors (PARPi) have helped to prolong PFS in the first-line setting, particularly in patients harboring a germline or somatic BRCA1/2 deleterious mutation. However, cure rates for HGSOC remain low at 15–20% or less,¹ and have not changed significantly over the last four decades. This status quo is a direct result of chemoresistant ovarian cancer cells that survive frontline therapy, and this “minimal residual disease” (MRD) then leads to tumor repopulation and eventual clinical recurrence and incurable disease.²

To date, the main approach for addressing MRD and the resultant high recurrence rate in ovarian cancer has been to add drugs to the carboplatin-paclitaxel backbone and to continue patients on “maintenance” therapies. These additional drugs have been selected based on their anti-tumor activity in the recurrent setting, and a tolerable side effect profile that allows months to years of additional therapy. Unfortunately, maintenance strategies with current drugs have not demonstrated increased cure rates in HGSOC. The limited success of this approach can be traced to knowledge gaps that include: 1) unknown molecular and immunologic characteristics of MRD and consequent inability to apply such knowledge for the development of rational frontline and maintenance therapies, 2) limited predictors of benefit from current maintenance therapies beyond homologous recombination deficiency (HRD) and BRCA status for PARPi, 3) lack of effective maintenance strategies for 50% of patients with homologous recombination proficient (HRP) ovarian cancers, and 4) the long duration and significant expense of frontline trials that are currently designed based on a PFS endpoint, stifling innovation and improvement in cure rates. Recent scientific advances including single cell genomics, spatial TME profiling, circulating tumor DNA (ctDNA), and nanoparticle drug delivery make this an opportune time to formulate a new strategy for increasing cure rates in ovarian cancer by understanding, targeting, and eradicating MRD.

The Current Frontline Maintenance Strategy

Maintenance treatments tested to date have included single agent paclitaxel, bevacizumab, anti-PD-1/PD-L1 immune checkpoint inhibitors (ICI), PARPi, and combinations of these drugs. Thus far, the only FDA approved maintenance approaches are PARPi (niraparib and olaparib) and bevacizumab, but neither have significantly increased cure rates.

The greatest clinical benefit associated with maintenance therapy has been achieved with the use of PARPi in patients with BRCA1/2 mutations, and to a lesser degree, for those with homologous recombination deficient (HRD) tumor phenotype lacking BRCA1/2 mutations. While it is notable that SOLO1, a randomized phase 3 trial comparing olaparib maintenance to placebo in BRCA1/2 mutation carriers, demonstrated that at seven years of follow-up, 45% of patients treated with olaparib maintenance had not received any subsequent therapy,³ it is clear that even in this most favorable prognostic patient population, the extent of benefit is variable.

Potential Impact of a More Personalized and MRD-Guided Approach

Aside from HRD/BRCA status, the most important predictor of benefit from PARPi is the chemo-sensitivity/resistance of the tumor.⁴ Determination of relative chemoresistance during frontline treatment is limited because most patients achieve “clinical remission” (complete radiographic response and normalized CA-125 tumor marker). However, in this setting, the presence of MRD after frontline treatment can be determined through “second look surgery” and may have utility as a biomarker of chemotherapy-resistant residual disease that has persisted after 6 or more cycles of chemotherapy.

Second look surgery has an established track record in ovarian cancer and in the past was performed routinely via laparotomy to determine the duration of adjuvant chemotherapy. Today, this surgical evaluation can be performed via minimally invasive second look laparoscopy (SLL), a low risk, outpatient procedure with minimal recovery time. The SLL procedure includes peritoneal washings and multiple peritoneal biopsies of any visibly abnormal peritoneal areas and other sites of frequent ovarian cancer metastases including the right diaphragm peritoneum and omental remnants. The early experience at MD Anderson Cancer Center reveals that approximately 50% of patients who have had an apparent complete response by imaging and CA-125 after frontline therapy have surgically detectable MRD, and these patients have a much shorter PFS compared to those negative for MRD (Jazaeri et al., manuscript in preparation). Thus, SLL can provide sensitive detection of MRD, the sine qua non of chemotherapy-resistant disease.

Knowledge of MRD may provide a more personalized approach in patients with both BRCA mutated (BRCAm)/HRD and HRP tumors. Understanding the biological underpinnings of MRD is particularly crucial to guide the development of novel therapies for patients with HRP tumors, as these patients who comprise 50% of all ovarian cancer cases have a worse prognosis and lack of proven survival benefit from any of the current maintenance strategies.⁵ In the BRCAm/HRD population, not all patients have the same extent of benefit from PARPi. It is well established that patients with platinum resistant disease are less likely to receive benefit from PARPi.⁴ Therefore, knowing which patients have MRD, an indicator of relatively more platinum resistant disease, may further refine benefit from PARPi and open the door for PARPi combinations or other more effective maintenance strategies. Additionally, there are important translational opportunities in this patient population, as mechanisms of chemo-resistance in BRCAm/HRD positive vs. HRD negative patients may be different and understanding these differences is critical to rationale drug development that may be targeted at specific biomarker groups.

While detection is a necessary first step, we believe that striving to gain a comprehensive understanding of ovarian cancer MRD, at this time of rapidly expanding targeted and immune therapies, represents an important strategic opportunity to address this large unmet need and transform ovarian cancer care.

Unique Aspects of Investigating MRD and Clonal Evolution in Ovarian Cancer

MRD has been explored as a biomarker in other cancer types, frequently defined by the presence of circulating tumor DNA (ctDNA).⁶ Blood-based MRD assays are now being used for prognostic and management purposes in colorectal cancer and can reduce the number of patients treated with adjuvant chemotherapy.⁷ However, there are important limitations of ctDNA-based MRD detection, including lack of information about the TME and its immunologic characteristics.

Currently, the utility of ctDNA for detection of MRD in ovarian cancer is largely unknown. As noted above, surgical detection of MRD in ovarian cancer is an established procedure with significant prognosticating capability which also affords access to low volume/microscopic areas of residual disease. This provides a unique opportunity to simultaneously sample multiple potential metastatic sites and, in parallel, through serial analysis of tumor biopsies and ctDNA, identify resistant clones and subpopulations as they evolve, providing an unprecedented opportunity to understand the emergence of chemo-resistant disease (Figure 1). Given the relative initial chemosensitivity of ovarian cancer, such resistant clones are likely to represent a small fraction of the original (pre-treatment) tumor, and as such may be undetectable in biopsies obtained at the time of diagnosis. Additionally, there are key compositional, topological, and functional phenotypic differences in primary adnexal tumors compared to distal peritoneal foci,⁸ highlighting the need to sample multiple areas to develop a comprehensive understanding of ovarian MRD.

Figure 1. Opportunities for tracking genomic and immune evolution in ovarian cancer to define mechanisms of resistance and develop novel therapies.

Tissues collected at surgical time points will include freshly dissociated single cell suspensions, snap frozen tissue for bulk nucleic acid extractions, FFPE material for spatial analysis, and plasma for circulating tumor DNA sequencing. Serial tissue collections will allow for investigation of longitudinal dynamics of cell-intrinsic and TME-associated determinants of MRD and therapeutic response, as well as creation of organoids and xenografts for further molecular studies aimed at identifying mechanisms of therapy resistance. Abbreviations: MRD=minimal residual disease. HGSOC= high grade serous ovarian cancer.

Clonal analysis and assessment of the TME in surgically detected MRD provides a unique opportunity to better understand the immunogenomic characteristics of MRD, with the goal of novel targeted and immune therapeutic development. A recently published study involving single-cell RNA-seq analysis of metastatic ovarian cancer before and after neoadjuvant chemotherapy demonstrated subclonal enrichment of a stress-associated transcriptional profile after chemotherapy and that stress-associated cancer cells strongly associate with presence of inflammatory cancer-associated fibroblasts within the microenvironment.⁹

Recent studies have demonstrated the feasibility of clonal tracking by leveraging whole genome sequencing (WGS) approaches (both single cell and bulk) to provide insights into clonal structure, structural variation, mutational processes, and copy number alterations driving cell-intrinsic biology in HGSOC. From single cell WGS, phylogenetic trees are inferred, and clone-specified relative abundance measures modeled. High fitness clones are then examined for gene-based mutations and patterns of mutational signatures as putative determinants of response.^8,10 Spatial transcriptomics and multi-color immunofluorescence using the CODEX platform from matched tissue specimens can then be utilized, complemented by multimodal single-cell omics, to profile MRD at unprecedented resolution.

Accelerating Frontline Trials by using MRD as an Early Surrogate Indicator for Efficacy

Currently, most frontline trials for ovarian cancer use PFS as a primary endpoint. This approach typically requires up to 10 years for design, implementation, follow-up, and maturation of survival data prior to informing subsequent clinical care. Such trials are expensive (typically costing hundreds of millions of dollars) and therefore cost-prohibitive for most biotech companies, stifling innovation. Unfortunately, even with significant investment of patients, time, and money, there are ample examples of recent negative randomized, phase 3 frontline trials in ovarian cancer, including GOG 3015 and JAVELIN Ovarian 100. Notably, both the cost and timelines of future frontline trials are expected to further increase, given the approved use of PARPi maintenance therapies incorporated into the current standard of care.

Given data demonstrating that the presence of MRD is a predictive biomarker of PFS in ovarian cancer¹¹ as well as in other solid tumors^6,7, there is an opportunity to use MRD as an early surrogate marker of PFS. This is similar to the evolution of pathologic complete response (pCR) as an early surrogate for overall survival in frontline trials of patients with locally advanced breast cancer.¹² The use of ovarian MRD as a surrogate efficacy endpoint could enable go/no-go decisions in smaller upfront trials, thus lowering cost and de-risking the pursuit of more innovative trial concepts, including the use of immunotherapies, nanoparticles and novel combinations. Such trials can serve as initial signal-generating investigations prior to committing to large, resource-intensive phase 3 studies.

Team Science to Generate a Comprehensive Understanding of MRD

Clinical and basic science investigators from the University of Texas MD Anderson Cancer Center, Memorial Sloan Kettering Cancer Center, Dana-Farber Cancer Institute, Massachusetts Institute of Technology, and the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins are coming together in a collaborative effort, with the goal of increasing cure rates for ovarian cancer by understanding and eliminating MRD.

The inaugural clinical trial as part of this ovarian MRD initiative is set to open in 2023 and will involve treatment of patients with neoadjuvant carboplatin, paclitaxel, and bevacizumab, with or without GEN-1, an IL-12 gene expression plasmid delivered intraperitoneally (NCT05739981; schema depicted in Figure 2). IL-12 is an active immunocytokine that induces anticancer immunity; intraperitoneal nanoparticle formulated IL-12 delivery (GEN-1, Imunon) has the advantage of providing increased local efficacy while minimizing systemic toxicity. GEN-1 in ovarian cancer is biologically active and promotes a pro-immune T-cell population dynamic and conversion of tumor naïve T-cells into cytotoxic effector T-cells in the TME.¹³ We hypothesize that adding GEN-1 to the current standard of care chemotherapy including bevacizumab can overcome the immunosuppressive features of the peritoneal TME. The primary endpoint in this study will be the rate of MRD at the time of SLL, which will be correlated with PFS for validation. Guidelines for standardization of SLL are included in Supplemental Figure 1. To our knowledge, this will be the first ovarian cancer trial utilizing MRD via SLL as a primary clinical endpoint. Working with the FDA and patient advocacy groups, we plan to use this trial as a proof of principle for the utility of MRD as an endpoint to enable more rapid evaluation of novel therapeutics in the frontline setting. Innovative translational aspects embedded in this trial will allow for serial investigation of ovarian cancer clonal evolution and peritoneal tumor involvement at single cell resolution, comprehensive characterization of cancer cell phenotypes and the TME at sites of residual disease, and detailed analysis of associated peritoneal and gut microbiomes. In this regard, it is worth mentioning that recent advances in FFPE-based multi-omic and digital spatial transcriptomic technologies have significantly enhanced the ability to interrogate even small tissue samples, such as those from SLL. Additionally, serial analyses of ctDNA will allow us to both investigate genomic evolution of the tumor over time and correlate this biomarker with surgical MRD. This will afford us the opportunity to validate ctDNA in ovarian cancer MRD, setting the stage for its potential use as a prognostic marker in the future in place of a surgical procedure.

Figure 2. Study schema for recently activated neoadjuvant trial with MRD rate as primary endpoint.

Patients with suspected advanced stage ovarian cancer will undergo diagnostic laparoscopy and be randomized 1:1 to either the investigational arm (carboplatin, paclitaxel, bevacizumab, and i.p. GEN-1) or the control arm (carboplatin, paclitaxel, and bevacizumab), which represents current standard of care. All subjects will undergo interval cytoreductive surgery after 4–6 cycles of therapy and SLL for detection of MRD. Maintenance regimens in each arm are based on HRD status. Abbreviations: OVCA= ovarian cancer. NACT= neoadjuvant chemotherapy. BEV= bevacizumab. I.p.= intraperitoneal. HRD+= homologous recombination deficient. HRP= homologous recombination proficient. cfDNA= cell free DNA. MRD= minimal residual disease. SLL= second look laparoscopy. PFS= progression-free survival.

The ultimate goal of these translational investigations is the eradication of MRD through gaining a more complete understanding of its immune-genomic underpinnings. We envision achieving this goal through development of novel therapeutics for patients with MRD and, in addition, more effective rational frontline combination therapies aimed at reducing the portion of patients with MRD. As noted above, our first trial addressing the latter objective tests the benefit of adding IL-12 to chemotherapy. While many patients across all cancer types have benefited from successful cancer immunotherapy, it is notable that success of immune checkpoint inhibitors in ovarian cancer (both alone and in combination with frontline chemotherapy) has been modest at best. This suggests that there remain un-discovered mechanisms of immune resistance/evasion that are likely to be involved in ovarian cancer, and we believe the investigation of MRD allows unprecedented opportunity to discover such mechanisms and exploit them for therapeutic intervention.

While gaining a better understanding of the biological underpinnings of MRD is the primary goal, the ultimate clinical objective is to develop therapeutic interventions aimed at eradicating MRD. The initial trial in the MRD setting, which is ongoing, evaluates the use of maintenance bevacizumab until clinical progression in patients with HRP residual disease (NCT02884648). It is worth noting that the current maintenance use of bevacizumab is for up to 17 cycles and not until progression, and possible differential benefit based on HRD status remains unknown. The next trial in this queue is a single-arm study of maintenance pembrolizumab, bevacizumab, and oral metronomic cyclophosphamide for the same cohort (residual HRP tumors). The rationale for this combination is based on its demonstrated efficacy in the recurrent setting, especially among patients who received three or fewer lines of previous therapy.¹⁴ The PFS data from this trial will be compared to that of the bevacizumab alone trial to better understand if the combination regimen results in improved outcomes in the MRD setting. These trials can serve as a roadmap for developmental therapeutic interventions aimed at MRD in patients with HRP tumors. For patients with BRCA mutated/HRD tumors, prospective observational studies are required to determine if the presence of residual disease may further differentiate extent of benefit from currently approved PARPi maintenance therapies and whether combination studies and targeted therapies in molecular subtypes may provide additional benefit.

It is important to note that our understanding of the clinical significance of MRD in different ovarian cancer subtypes is still emerging and is not yet at the stage that validated recommendations can be made for general clinical management based on MRD status. Further research is needed to understand which clinical and biological factors (e.g. HRD status, neoadjuvant chemotherapy vs. primary cytoreductive surgery, residual disease status at cytoreductive surgery, histology, cyclin E amplification, etc.) are associated with the presence of MRD. Additionally, second look laparoscopy is currently not considered a part of standard ovarian cancer management and ctDNA assays have not yet been validated in this disease. We hope that insights from in depth, multi-omic investigations of MRD, as well as highlighting the rationale for the broader scientific community to focus on MRD, will help generate data to drive clinical practice guidelines in the future.

The final aim of the ovarian cancer MRD project involves modeling MRD in immune-competent mouse models and engaging in preclinical testing of the most promising novel therapies and delivery modalities (Figure 1). Specifically, we will evaluate therapeutic agents that sensitize tumor cells to immune attack, engage dendritic cells required for induction of tumor-specific T-cell responses, and enhance the activity of tumor-infiltrating T-cells. These novel therapeutics may include proteins, nucleic acids, and/or small molecules to be delivered using nanoparticles engineered for these cargoes that can target ovarian MRD. Specifically, these nanoparticles will be generated using an electrostatic assembly platform which allows for optimized loading of promising therapeutics and development of cancer-targeting nanoparticle surface chemistries.¹⁵ Nanoparticles will incorporate cancer targeting moieties informed by clinical trial patient data. The overall goal of these studies will be to develop novel therapeutic concepts that can be evaluated in future clinical trials, including immune-stimulatory agents potently delivered directly to MRD lesions using rationally designed nanoparticle delivery systems.

Given the large unexplored clinical and translational opportunities in this space, we anticipate future investigations may also include other immunotherapeutic strategies with potential to target MRD. These include but are not limited to targeted therapies for molecular subtypes of MRD, adoptive cell therapies, and oncolytic viruses with possible intraperitoneal route of administration.

Concluding Remarks

When the initial forays into the use of second look surgery in ovarian cancer in the late 20^th century did not result in improved overall survival, the Gynecologic Oncology Group (GOG) investigators concluded, “[Second look laparotomy] remains the best means available for determining the posttreatment status of ovarian cancer and, as such, its correlation with prognosis should be discussed with patients in the interest of full disclosure. However, its current use should be limited to research until additional adjuvant therapy, initiated on the basis of SLL [second look laparotomy] findings, has been shown to improve survival. Its future use may change depending upon the development of new chemotherapies, biological agents, or diagnostics.”¹¹ We feel that this time has arrived and that we owe it to the women diagnosed with this deadly disease to reconsider the decades long approach in favor of understanding ovarian cancer MRD.