Results from a single arm, single stage phase II trial of trametinib and GSK2141795 in persistent or recurrent cervical cancer
Abstract
Background
The urgent need for improved treatment strategies for advanced cervical cancer remains a significant challenge in oncology. Currently, available treatment options for these patients are limited, primarily consisting of either combined chemotherapy regimens with bevacizumab (an anti-angiogenic agent) or pembrolizumab monotherapy, specifically for patients whose disease expresses programmed death-ligand 1 (PD-L1). Genomic profiling of cervical cancers has consistently revealed the presence of recurrent mutations in crucial signaling genes, notably PIK3CA and KRAS. Given the prevalence and oncogenic drive of these mutations, this study was specifically designed to investigate a novel therapeutic approach: dual inhibition of the critical PI3K and RAS signaling pathways. This dual blockade was achieved by combining the MEK inhibitor trametinib, which targets the RAS-MAPK pathway, and the AKT inhibitor GSK2141795 (Uprosertib), which targets the PI3K-AKT pathway, in patients with recurrent cervical cancer.
Methods
This investigation was an investigator-initiated phase II clinical study, meticulously designed to evaluate the combination of trametinib and GSK2141795 in a cohort of patients suffering from recurrent cervical cancer. The primary endpoint established for the study was the best tumor response, assessed according to RECIST criteria. Secondary endpoints included crucial clinical outcomes such as progression-free survival (PFS), overall survival (OS), and a comprehensive assessment of the safety and tolerability profile of the combination regimen. Beyond clinical endpoints, translational objectives were integrated to characterize the molecular alterations present in genes within the PI3K and RAS signaling pathways in enrolled patients, aiming to correlate molecular profiles with treatment response.
Results
The study initially planned to accrue 35 patients to ensure adequate statistical power. However, the study was prematurely terminated after only 14 patients were enrolled and received at least one dose of the study drug. The termination was necessitated by the discontinuation of the development program for GSK2141795 by its manufacturer, preventing further investigation of this specific combination. Despite the early termination, the outcomes in the enrolled patients were evaluated. There were no confirmed objective responses (neither complete response nor partial response). One patient did experience an unconfirmed partial response. Eight patients achieved stable disease, indicating some level of disease control. However, three patients experienced disease progression as their best response, and two patients were unevaluable for response. Regarding safety, toxicities were predominantly mild to moderate, categorized as Grade 1 and 2. Nonetheless, a notable proportion of patients, 57%, experienced Grade 3 or 4 adverse events, indicating severe or life-threatening toxicities. Consequently, 50% of patients required at least one dose reduction due to these adverse events, highlighting challenges with the tolerability of the combination at the initially prescribed doses.
Conclusions
The study demonstrated that the combination of trametinib and GSK2141795 was generally feasible to administer, though it frequently necessitated dose holds and modifications due to the occurrence of adverse events. However, despite the targeted approach, the overall anti-cancer activity observed was minimal, even in the subset of patients who harbored specific molecular alterations within the PI3K or RAS pathway, suggesting that this combination may not be sufficiently potent or that resistance mechanisms were at play. Although the study was terminated prematurely due to the discontinuation of GSK2141795 development, the findings derived from these 14 enrolled patients collectively do not provide sufficient support for the further clinical development of this specific combination regimen in the treatment of recurrent cervical cancer. These results underscore the complexity of targeting these pathways and the ongoing need for more effective strategies.
Introduction
Cervical cancer continues to be a significant global health concern, ranking as the fourth most commonly diagnosed cancer among women worldwide, and tragically, it also holds the fourth position in terms of cancer-related mortality. Annually, there are approximately 569,847 new diagnoses of cervical cancer and a devastating 311,365 women succumb to the disease across the globe. Recent advancements have brought about some improvements in the treatment landscape for advanced and/or recurrent cervical cancer. Notable among these are the findings from the Gynecologic Oncology Group (GOG) 240 study, which demonstrated a significant improvement in overall survival (OS) to 17.5 months and 15.0 months, respectively, when bevacizumab (an anti-angiogenic agent) was incorporated into platinum-based or non-platinum-based chemotherapy regimens. Additionally, in June 2018, the United States Food and Drug Administration (FDA) granted approval to the PD-1 inhibitor pembrolizumab for the treatment of PD-L1-positive cervical cancer, specifically defined by a combined positive score of 1 or higher. Two key trials underscored the single-agent activity of pembrolizumab in this context: KEYNOTE 028, a 24-patient cohort of PD-L1 positive cervical cancer, reported a 17% response rate. Subsequently, the KEYNOTE 158 study, which included 98 women (77 of whom had PD-L1 positive cervical cancer), reported an overall response rate of 14.3%. It is critical to note that no objective responses were observed in women whose cancers were characterized as PD-L1–negative. Despite these important advances, metastatic cervical cancer regrettably remains an incurable disease, underscoring the persistent and urgent need for the identification of novel treatment options and therapeutic strategies.
Numerous examples exist in oncology where significant treatment advances for advanced cancers have been achieved through the successful targeting of specific molecular alterations identified within the tumor. Several comprehensive studies have likewise identified recurring genomic alterations in cervical cancer that could potentially serve as actionable targets for biologic therapies. Specifically, alterations within the PI3 kinase pathway have been consistently identified in both squamous cell carcinomas and adenocarcinomas of the cervix. One trial, for instance, reported a PIK3CA mutation rate of 36% in squamous cervical cancer. Furthermore, mutations in KRAS and HRAS genes have also been reported at frequencies of 10% and 22% respectively. In our own institutional study at the Dana-Farber Cancer Institute (DFCI), which analyzed a group of 80 patients with either squamous cell carcinoma or adenocarcinoma of the cervix, validated mutations were detected in a substantial 60.0% (48 out of 80) of the tumors examined. Within this cohort, the highest mutation rates were observed for PIK3CA (31.3%) and KRAS (8.8%). Interestingly, PIK3CA mutations were found at similar frequencies in both adenocarcinoma and squamous cell carcinomas (25.0% vs. 37.5%, p = 0.33). In contrast, KRAS mutations were identified exclusively in adenocarcinoma cases (17.5% vs. 0%, p = 0.01). These compelling genomic data served as the initial and crucial rationale for the design and execution of the present study.
Preclinical and translational studies have provided strong support for the rationale behind combining PI3K/AKT inhibitors with MEK inhibitors to effectively target these interconnected signaling pathways. This approach is justified given the inherent redundancy and extensive crosstalk that exists between the PI3K-AKT and RAS-ERK pathways, which often lead to compensatory feedback loops that limit the efficacy of single-agent inhibition. For instance, mouse models of KRAS-driven tumors, which are notoriously aggressive, typically do not respond to either PI3K or MEK inhibition alone but demonstrate a robust response when these two inhibitors are combined. Specifically, PI3K pathway inhibition can paradoxically result in the activation of several receptor tyrosine kinases that are capable of signaling through the RAS-ERK pathway, thus providing a clear rationale for combining inhibitors that target both arms of this critical oncogenic network. In the phase II study reported here, we rigorously tested the combination of the MEK inhibitor trametinib and the pan-AKT inhibitor GSK2141795. This specific combination had been previously investigated in a phase I clinical trial, which successfully determined the recommended phase II dosing for these agents.
Methods
Patients for this study were recruited from the Dana-Farber Cancer Institute and Beth Israel Deaconess Medical Center, both located in Boston, MA. All participating patients were required to provide written informed consent, a document that had been meticulously reviewed and approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board, ensuring ethical conduct of the research. This study was initiated by the investigators, and the University of Alabama at Birmingham (UAM) held the FDA Investigational New Drug (IND) application, overseeing the regulatory aspects. Funding for the study was provided by the National Comprehensive Cancer Network, and the study drugs (trametinib and GSK2141795) were generously supplied by GlaxoSmithKline.
Eligibility
Strict eligibility criteria were established for patient enrollment. Patients were required to have histologically or cytologically confirmed cervical cancer that was deemed recurrent or metastatic and refractory to established standard treatments. Participants had to be 18 years of age or older, possess an ECOG (Eastern Cooperative Oncology Group) performance status of 0–2, indicating good functional capacity, and exhibit normal organ and bone marrow function. Additionally, patients needed to have controlled blood pressure and the ability to tolerate oral medications, as both study drugs were administered orally. The presence of measurable cancer, as per Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria, was also mandatory. Importantly, patients were excluded if they had received any prior treatment with drugs specifically targeting the PI3K pathway (including PI3K inhibitors or mTOR inhibitors) or the RAS-ERK pathway. All histologic types of cervical origin were permitted. Patients were required to have received at least one prior chemotherapeutic regimen for the management of cervical carcinoma. Chemotherapy administered concurrently with primary radiation as a radio-sensitizer was explicitly not counted as a systemic chemotherapy regimen for eligibility purposes. Patients were allowed, but not required, to have received one additional prior treatment regimen (which could include a single chemotherapeutic agent, a combination of chemotherapeutics, or biologic drugs such as bevacizumab) for the management of recurrent cancer, allowing for some flexibility in prior treatment history.
Study Design
This investigation was designed as an open-label, single-arm study, meaning all enrolled patients received the same experimental treatment and there was no blinding to the assigned therapy. Each treatment cycle was defined as 28 days. The starting doses for the therapeutic agents were precisely set: trametinib at 1.5 mg orally daily and GSK2141795 at 50 mg orally daily. Patients were instructed to self-administer both medications simultaneously each day. Dose reductions were permitted for trametinib, down to 1.0 mg, and for GSK2141795, down to 25 mg per day, specifically for certain pre-specified toxicities, such as Grade 3 rash and ophthalmologic toxicities, to ensure patient safety and tolerability. Treatment was continued until documented disease progression or until adverse effects prohibited further administration of the drugs. Patients underwent radiographic assessments by RECIST 1.1 criteria, utilizing either computed tomography (CT) or magnetic resonance imaging (MRI) scans, performed every 8 weeks (equivalent to every 2 cycles) until disease progression was definitively confirmed. Toxicities were meticulously measured and graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0.
The primary endpoint of this study was the best objective tumor response, defined as the proportion of patients achieving either a complete response (CR) or a partial response (PR) as per RECIST version 1.1 criteria. Secondary endpoints encompassed several crucial clinical and molecular outcomes: progression-free survival (PFS) and overall survival (OS) following the initiation of therapy with trametinib and GSK2141795; a comprehensive assessment of all toxicities encountered; the determination of mutation and co-mutation rates of genes within the PI3K and RAS-ERK signaling pathways in recurrent cervical cancer, utilizing high-throughput targeted mutational analysis performed on participant tumor samples; and an exploratory investigation into the association between the detected mutational status and the observed clinical benefit derived from trametinib and GSK2141795.
Mutational Analysis
To identify specific genomic alterations, mutational and copy number assessment was rigorously performed using OncoPanel on formalin-fixed paraffin-embedded (FFPE) archival tumor samples. The OncoPanel assay is a validated next-generation sequencing platform that surveys exonic DNA sequences of 447 cancer-related genes and probes 191 regions across 60 genes for potential gene rearrangements. DNA was meticulously isolated from tissue samples confirmed to contain at least 20% tumor nuclei, ensuring sufficient tumor cellularity for accurate analysis. The extracted DNA was then analyzed by massively parallel sequencing, utilizing a solution-phase Agilent SureSelect hybrid capture kit for target enrichment and an Illumina HiSeq 2500 sequencer for high-throughput sequencing. The OncoPanel assay has been previously described and extensively validated for its robust and reliable detection of genomic alterations and somatic mutations in various cancers.
Statistical Considerations
With a planned accrual of 35 participants, the study was statistically powered at 91% to detect a significant improvement in the response rate from a historical baseline of 0.07 to an anticipated 0.22. This power calculation utilized a one-sided, 0.09-level exact binomial test. Consequently, the null hypothesis, which posited no significant improvement, would be rejected if 5 or more participants demonstrated an objective response to the combination therapy with trametinib and GSK2141795. The historical null hypothesis response rate of 0.07 was established based on a weighted average of observed response rates from 11 cohorts previously enrolled in the GOG 127 and 227 series, providing a relevant benchmark.
The statistical analyses encompassed all participants who initiated study treatments, adhering to an intent-to-treat principle. The objective response rate (ORR), defined as the sum of complete response (CR) and partial response (PR), was estimated with a 90% exact binomial confidence interval. Progression-free survival (PFS) was defined as the time interval from study registration until documented disease progression or death from any cause without prior progression, whichever occurred first. Participants who did not experience a PFS event were censored at the last date of documented disease evaluation. Overall survival (OS) was defined as the time from study registration to the time of death from any cause, with follow-up censored at the last date the patient was known to be alive. The PFS and OS analyses were summarized using the Kaplan-Meier product-limit estimator, with reported event rates and estimates of median survival calculated using Greenwood’s formula, accompanied by 95% confidence intervals (CI). Safety data were comprehensively described by enumerating the number and proportion of patients who experienced treatment-related adverse events, classified and graded according to CTCAE version 4.0.
Results
The clinical trial for this combination therapy was initiated in October 2013 and subsequently closed prematurely in June 2015. The early termination of the study was necessitated by two primary factors: firstly, the discontinuation of the development program for the AKT inhibitor GSK2141795 by its manufacturer; and secondly, GlaxoSmithKline’s determination that the drug combination exhibited an unacceptable toxicity profile in other patient populations, including those with breast cancer and melanoma. Despite an initial planned accrual of 35 patients, 16 patients were consented for the study. However, 2 patients never commenced therapy due to ineligibility, ultimately leading to 14 patients who were formally enrolled and received at least one dose of the study treatment.
A detailed summary of the patient and tumor characteristics is provided. Of the enrolled patients, 93% were identified as white. The median age of the cohort was 52 years, with an age range spanning from 46 to 61 years. A significant majority, 71% of patients, had an ECOG performance status of 0, indicating good functional capacity. Histologically, 50% of the patients presented with squamous cell carcinoma. The median number of prior line therapies received by patients was 2, with a range from 1 to 3, reflecting the heavily pre-treated nature of this recurrent cancer population. Ten patients (71%) had received prior radiation therapy (RT), and 7 patients (50%) had undergone prior chemoradiation. The estimated median follow-up time for the study was 6.3 months (interquartile range 2.4 to 12.9 months), based on known follow-up data for all patients. The median number of treatment cycles completed by patients on study was 3, with a wide range from 0 to 22 cycles. All 14 treated patients eventually discontinued treatment. Among these, 8 patients had documented progressive disease, while 6 patients went off treatment without a formally documented progression. Of those who discontinued without progression, 2 did so by treating investigator decision (one in the setting of a 16% increase in tumor burden, and another due to clinical decline related to disease progression and entry into hospice care), 2 patients withdrew consent, and 2 discontinued due to toxicity (one for colonic perforation and another for Grade 3 diarrhea who declined dose reduction).
Regarding objective response, one patient achieved an unconfirmed partial response, resulting in a response rate of 7.1%. Eight patients (57.1%) experienced stable disease; among these, stable disease was confirmed in 5 patients, while it remained unconfirmed in 3. The median duration of stable disease was relatively short at 1.9 months. Three patients (21.4%) exhibited progressive cancer as their best response. Two patients (14.3%) were ultimately unevaluable for response. The waterfall plot provides a visual representation of the maximum tumor shrinkage observed in the 12 patients for whom response data were available.
Kaplan-Meier curves for progression-free survival (PFS) and overall survival (OS) are presented. A total of 11 PFS events were reported, comprising 6 by RECIST 1.1 criteria, 3 by clinical progression, and 2 deaths without prior progression. A total of 8 deaths were reported during the study follow-up period, with 5 patients having previous documentation of progression (4 by RECIST 1.1, 1 by clinical progression). At a median follow-up of 6.3 months (with a maximum of 24.5 months), the median PFS (based on RECIST 1.1) was 3.7 months (95% CI: 1.9, Not Available); median PFS (combining RECIST 1.1 and clinical progression) was 3.6 months (95% CI: 1.6, Not Available); and the median OS was 14.8 months (95% CI: 6.7, Not Available).
Table 2 provides a detailed description of toxicities deemed related to the study treatment that occurred in at least 10% of patients. All 14 enrolled patients received at least one treatment dose, and adverse event (AE) forms were submitted for all. The overall rate of Grade 3/4 AEs that were possibly, probably, or definitely related to treatment was 57.1% (8 out of 14 patients), with 7 patients experiencing Grade 3 events and 1 patient a Grade 4 event. Grade 3 AEs related to the study drugs included 3 cases of rash (2 acneiform, 1 maculopapular), 2 cases of diarrhea, 1 case of mucositis, 1 case of hypertension, 1 case of hyponatremia, 1 case of INR increase, and 1 thromboembolic event. There was one Grade 4 treatment-related AE, which was a colonic perforation. The most commonly occurring AEs (any grade) included acneiform rash (71.4%), diarrhea (64.3%), nausea (36%), and ALT (alanine aminotransferase) elevation (29%), with most of these being Grade 1 or 2 in severity. Dose reductions were required in 50% of patients due to adverse events, indicating significant manageability challenges at the initial dosing.
Finally, the observed mutations and amplifications in genes related to PI3K or RAS signaling are presented for the 13 patients for whom tissue was available for testing; one patient did not have archival tissue available. Among the 7 squamous cell carcinoma cases, 2 had a PIK3CA mutation, and among the 5 adenocarcinoma cases, 2 also had a PIK3CA mutation. The only cancer with a PIK3CA amplification was of adenocarcinoma histology. Three cancers exhibited KRAS aberrations: the one mucinous cancer had a KRAS mutation, and the other 2 were adenocarcinomas (one with a KRAS mutation and one with a KRAS amplification). Of note, while the one patient who achieved prolonged stable disease (lasting over 20 months) did harbor a PIK3CA mutation, there was no clear overall pattern of association between the time spent on the study drug and any specific detected genomic alteration across the entire cohort.
Discussion
This study meticulously investigated the therapeutic potential of combining a MEK inhibitor with a PI3K pathway inhibitor, specifically a pan-AKT inhibitor (GSK2141795), for the treatment of relapsed and advanced cervical cancer. The trial, however, was prematurely terminated due to the discontinuation of GSK2141795′s clinical development by its manufacturer, GlaxoSmithKline (GSK), primarily because GSK identified an unacceptable toxicity profile for this combination in other cancer populations. Among the 14 patients who were enrolled in our study, the combination regimen was deemed tolerable, albeit with a need for dose adjustments and interruptions. Crucially, despite the treatment, no confirmed objective responses were observed within this cohort, indicating limited clinical efficacy.
The preclinical literature has consistently shown promising anti-cancer activity for the dual inhibition of MEK and PI3K pathways, providing a strong biological rationale for combining these agents. Consequently, numerous phase I and phase II clinical studies, including our own, have explored this therapeutic strategy. A comprehensive analysis of data from 236 patients enrolled in various PI3K pathway inhibitor studies (with or without a MAPK pathway inhibitor) appeared to support this dual inhibition approach, suggesting that simultaneous blockade of these pathways might indeed offer improved efficacy compared to inhibiting either pathway alone.
The specific combination of trametinib and GSK2141795 had been previously evaluated in earlier phase I and phase II studies. In the initial phase I investigations, the maximally tolerated doses for this combination were established at varying regimens, including trametinib/GSK2141795 at 1.5 mg/50 mg, 2 mg/25 mg, and 0.5 mg/75 mg daily. For our study, the trametinib 1.5 mg daily and GSK2141795 50 mg daily regimen was selected based on these prior findings. The adverse events observed in our study were generally consistent with those reported in the phase I study, encompassing gastrointestinal issues, fatigue, and rash. Notably, the rate of Grade 3 adverse events in the phase I study (13% of 62 patients) was lower than in our study and included diarrhea, stomatitis, and an increase in AST (aspartate aminotransferase). Partial responses were observed at 8 weeks in three patients in the phase I study: one with triple-negative breast cancer (TNBC), one with endometrial cancer, and one with ocular melanoma, suggesting some initial signals of activity in other tumor types.
In addition to our trial in advanced cervical cancer, phase II testing of the trametinib/GSK2141795 combination has been conducted in other challenging cancer types, including endometrial cancer, triple-negative breast cancer, and melanoma. In each of these studies, the observed efficacy was generally deemed limited, with only one response reported among 26 endometrial cancer patients, one unconfirmed partial response in 16 TNBC patients, and no objective responses in 20 melanoma patients (comprising both NRAS-mutated and NRAS-wildtype cases). The most common adverse events observed in our study, namely rash and diarrhea, occurred at a rate comparable to that seen with the same combined dose of trametinib (1.5 mg daily) and GSK2141795 (50 mg daily) in the melanoma study. However, the overall rate of Grade 3 or higher adverse events was higher in our study. This increased toxicity might reflect the eligibility criteria in our trial, which permitted patients with a performance status of 2, or potentially an increased sensitivity to this drug combination within our specific study population. In other studies, increased toxicity has also been observed in endometrial cancer patients. For instance, in the phase II endometrial cancer study, 8 dose-limiting toxicities (including 2 cases of hypertension, 2 mucositis, 2 rash, 1 dehydration, and 1 acute kidney injury) were reported at the initial dose level of trametinib 1.5 mg daily and GSK2141795 50 mg daily, which was also utilized in our trial and other phase II studies. A lower dose of trametinib 1.5 mg daily and GSK2141795 25 mg daily was subsequently tested in 12 patients with endometrial cancer, proving to be better tolerated with no dose-limiting toxicities. Nevertheless, given the consistent lack of substantial efficacy observed across multiple patient populations and the notable toxicities associated with the combination, our trial was prematurely closed before achieving its accrual goal, and the clinical development of the AKT inhibitor GSK2141795 was ultimately terminated.
Other MEK inhibitors, such as trametinib, have also been explored in combination with alternative inhibitors of the PI3K pathway. A phase I trial involving 20 patients treated with trametinib and another AKT inhibitor, afuresertib (GSK2110183), reported toxicities including elevated liver enzymes, stomatitis, rash, and diarrhea. Interestingly, one partial response was observed in a BRAF wild-type melanoma patient. Continuous dosing in this combination was poorly tolerated, whereas intermittent dosing showed better tolerability. Similarly, another trial reported poor tolerability of continuously dosed trametinib when combined with everolimus, an mTOR inhibitor. Combinations of MEK inhibitors with PI3K inhibitors (as opposed to AKT inhibitors) have also been investigated. A phase I/II trial combining BKM120, a pan-PI3K inhibitor, with trametinib reported dose-limiting toxicities including stomatitis, diarrhea, dysphagia, and creatine kinase elevation, with Grade 3/4 toxicities primarily consisting of creatine kinase elevation, stomatitis, liver enzyme elevation, and rash. Encouragingly, clinical activity was observed in the 21 patients with ovarian cancer enrolled in this trial (4 in the dose escalation and 17 in the dose expansion phases), achieving an overall response rate of 29% and a median PFS of 7 months. Of these 21 patients, 71% had low-grade serous histology, and a remarkable 91% harbored a KRAS mutation. In a separate phase I trial, BYL719, an alpha-specific PI3K inhibitor, was combined with the MEK inhibitor binimetinib (MEK162). This study, enrolling 58 patients, reported Grade 3/4 toxicities such as diarrhea, increased creatine kinase, rash, nausea, and hyperglycemia. Again, clinical activity was observed in ovarian cancer patients, with 3 out of 4 enrolled ovarian cancer patients achieving a partial response. Two of these three responders had a known KRAS mutation, while the third patient’s somatic sequencing results were unavailable. The fourth patient with ovarian cancer also had a somatic KRAS mutation and exhibited stable disease for over 41 weeks.
A few combinations of MEK inhibitors and PI3K pathway inhibitors have advanced to both phase I and randomized phase II settings. These include the combination of the AKT inhibitor MK2206 with the MEK inhibitor selumetinib, and the combination of SAR24509, an oral PI3K and mTORC1/TORC2 inhibitor, with the MEK inhibitor pimasertib. In the phase I trial of MK2206 and selumetinib, RECIST 1.0-confirmed partial responses were observed in patients with non-small cell lung cancer and low-grade ovarian cancer. However, subsequent single-arm phase II testing of this combination in colorectal cancer unfortunately showed no objective responses. Furthermore, this combination was compared to modified FOLFOX chemotherapy in an open-label phase II randomized trial, which found no statistical difference in overall survival between the two groups, and a shorter median PFS with the MEK/AKT inhibitor combination. Toxicities were also higher in the MK2206/selumetinib arm, with an increased incidence of Grade 3 or higher toxicities. Phase I testing of SAR245409 and pimasertib also demonstrated significant toxicity: at the recommended phase II dosing, 74 patients (73%) required dose interruption, 20% required dose reduction, and 26% of patients discontinued treatment due to adverse events. Moreover, limited anti-cancer activity was observed in patients with advanced solid tumors. In a randomized phase II study in recurrent low-grade serous ovarian cancer comparing pimasertib and SAR245409 versus pimasertib alone, the trial (NCT01936363) was closed for futility after an interim analysis in May 2016, having enrolled only 65 patients.
The totality of experience gathered from these various clinical trials investigating the dual inhibition of the PI3K and RAS signaling pathways reveals a consistent pattern of low overall response rates in unselected patient populations. This strongly emphasizes the critical importance of identifying precise biomarkers to select those patients who are most likely to derive clinical benefit from such interventions, particularly given that these drug combinations frequently impart significant toxicities. Considering the documented presence of alterations in both PI3K and RAS signaling pathways in cervical cancers, there was a robust preclinical rationale to explore dual inhibition in this specific disease. Molecular characterization of tumors from the patients enrolled in our study demonstrated a pattern of alterations consistent with existing literature, showing no clear difference in the frequency of PIK3CA mutations between squamous cell and adenocarcinoma histologies. However, alterations in KRAS, ERBB2, and ERBB3 were observed exclusively in adenocarcinoma patients. While alterations in KRAS, ERBB2, and ERBB3 do occur in squamous cell carcinomas as per The Cancer Genome Atlas data, their frequency is lower, and given the small number of patients in our study, it is not unexpected that we did not observe these alterations in our squamous cell carcinoma cohort. Nevertheless, despite the molecular characteristics of our study population appearing consistent with published data, the presence or absence of alterations in PI3K or RAS signaling did not appear to have any significant correlation with the observed clinical activity of the combination therapy. While the single patient who experienced long-term stable disease did harbor a PIK3CA mutation, other patients with PIK3CA alterations had short durations of treatment or experienced rapid disease progression, preventing a clear correlative pattern.
Overall, while a strong preclinical rationale continues to suggest that dual inhibition of PI3K and RAS signaling pathways holds significant promise as a therapeutic strategy, the clinical implementation of this strategy has been consistently challenging. Although occasional signals of activity have been observed in specific cases or patient subsets, identifying a reliable biomarker of response has proven elusive across broad populations, and the overall toxicities associated with these combinations have been significant. Our trial in cervical cancer patients regrettably failed to demonstrate significant clinical activity for the MEK/AKT inhibitor combination of trametinib and GSK2141795. Furthermore, while expected alterations in the PI3K and RAS signaling pathways were indeed identified in our cohort, these did not appear to serve as reliable biomarkers for predicting clinical activity. It is important to note that 93% of the patients in this study were white, and therefore, the generalizability of these specific results across all racial and ethnic backgrounds may be limited. Nonetheless, despite the premature closure of this trial due to the termination of GSK2141795′s clinical development, the collective findings from these 14 patients strongly suggest that further exploration of combination MEK and PI3K inhibition in an unselected advanced cervical cancer population is not currently supported.
Author Contributions
All authors actively contributed to the writing, critical review, and/or revision of the manuscript, ensuring intellectual integrity and scientific accuracy. JFL, AAW, and UAM played crucial roles in the conception and design of the study, laying its foundational framework. Data collection was meticulously performed by JFL, AAW, SC, PAK, AP, KM, SM, CW, DD, and UAM. The complex tasks of data analysis and interpretation were primarily carried out by JFL, KPG, AAW, PAK, DD, and UAM, leading to the conclusions drawn in this paper.