Clinical outcome of treatment with serine-threonine kinase inhibitors in recurrent epithelial ovarian cancer: a systematic review of literature

Abstract

Introduction

While serine-threonine kinases (STK) are attractive therapeutic targets in epithelial ovarian cancer, clinical outcomes of STK inhibitors in the management of recurrent disease have not been completely described.

Areas covered

A systematic literature review of published clinical studies on STK inhibitors targeting mTOR, MAPK, and aurora kinase pathways in recurrent epithelial ovarian cancer was conducted, revealing 18 clinical trials (497 patients). Pooled analyses were performed to assess treatment response, survival time, and adverse events. Median progression-free survival was 3.4 months in STK inhibitor-based therapy, and the average response rate and clinical benefit rate were 13% and 67%, respectively. Among regimens comprised of only STK inhibitors (11 trials, 299 patients), median progression-free time was 2.7 months, response rate was 10%, and clinical benefit rate was 64%. Compared to single STK inhibitor monotherapy (52.5%), clinical benefit rates significantly improved when STK inhibitors were combined with a cytotoxic agent (71.4%), other class biological agent (74.2%), or an additional STK inhibitor (95.0%) (all, P ≤ 0.002).

Expert opinion

STK inhibitor-based therapy showed modest activity for recurrent epithelial ovarian cancer with reasonable clinical benefit rates, suggesting its potential utility for maintaining disease stability if supported by future studies. Efficacy appears greatly improved in appropriately selected patient populations, especially those with low-grade serous ovarian carcinoma, platinum-sensitive disease, cancers with somatic RAS or BRAF mutations, and when used in a combination regimen with a cytotoxic or biological agent.

1. Introduction

Ovarian cancer ranks as the fifth most common cause of cancer death among women and the leading cause of gynecologic cancer mortality in the United States.[1] Standard treatment consists of cytoreductive surgery followed by platinum-based combination chemotherapy with carboplatin and paclitaxel. Though the majority of patients respond to initial therapy, cancer recurs in up to 70% of patients, and obtaining a cure is challenging.[2] While platinum compounds are often given to patients with platinum-sensitive recurrences, no standard treatment exists for platinum-resistant ovarian cancer, which often displays concomitant resistance to other cytotoxic agents. A meta-analysis of 40 clinical trials found an average overall response rate of only 16% among patients with platinum-resistant recurrences receiving any type of chemotherapy,[3] highlighting the need to uncover therapies that improve outcomes, especially for patients with relapsed disease.

Because response to currently available cytotoxic agents for recurrent ovarian cancer is limited, attention has shifted to a newer class of agents that target biological mechanisms underlying cancer development and progression. One such category includes drugs targeting serine–threonine kinases (STK), key components in signal transduction pathways controlling metabolism, cell-cycle regulation, cell division, and angiogenesis.[4] STK inhibitors can be grouped into their action on three main signal transduction pathways: (1) the PI3K/AKT/mTOR (phosphatidylinositol 3-kinase/protein kinase B/mammalian target of rapamycin) signaling pathway, (2) the MAPK (mitogen-activated protein kinase) pathway, and (3) the aurora kinase pathway.[5–7] Hyperactivation of these pathways have been linked to cancer development.

The PI3K/AKT/mTOR pathway regulates cell proliferation, survival, mRNA translation, and angiogenesis in response to growth factors, nutrients, hypoxia, and energy input. Other molecules implicated in this pathway include PTEN (phosphate and tensin homolog) and STK11.[5] Indeed, germline mutations in PTEN are responsible for autosomal dominant Cowden Syndrome in humans,[8] while mouse models that combine conditional deletion of PTEN within the ovarian surface epithelium with expression of oncogenic KRAS perfectly recapitulate invasive, widely metastatic endometrioid ovarian cancer.[9] STK11 is a serine threonine kinase tumor suppressor gene that controls the activity of AMP-activated protein kinase (AMPK) family members, and germline mutations in STK11 are responsible for Peutz–Jeghers syndrome,[10,11] although sporadic somatic mutations in STK11 are rare in ovarian tumors.[12] In ovarian cancers, mutations have been observed in several related proto-oncogenes, including PIK3CA, PIK3R1, and AKT1/2.[4]

In 2004, AKT and mTOR were found to be activated in 87% of epithelial ovarian cancers.[13] Since then, several other studies have shown a prevalence of activating mutations in genes encoding regulatory units for proteins in this pathway, including PIK3CA, PIK3R1, AKT, and PTEN.[14] Everolimus (for renal cell carcinoma, breast cancer, pancreatic cancer, and astrocytoma) and temsirolimus (for renal cell carcinoma) are the only two FDA-approved drugs acting on this pathway. However, a multitude of other inhibitors are in development including those acting on AKT, PI3K, and other pathway components. Using a combination of drugs acting at multiple stages of this pathway may help overcome resistance due to feedback inhibition.[4] For example, mTOR forms two complexes to enact downstream effects, mTORC1 and mTORC2. Currently available first-generation mTOR inhibitors selectively act on mTORC1, which may shift downstream effects to mTORC2. Furthermore, it is theorized that diminished output from mTORC1 feeds back to increase activity of AKT and PI3K, causing resistance to mTOR inhibitors. To overcome resistance, some suggest combining mTOR inhibitors with PI3K, AKT, or MAPK inhibitors.[4] While it is difficult to generalize across all drugs that modulate this pathway, the most common side effects include hyperglycemia, rash, stomatitis, myelosuppression, and gastrointestinal effects.[14,15] Metformin is also being studied as an adjunct to cancer therapy due to its indirect inhibition of mTOR and is extremely well tolerated.[14]

The MAPK pathway regulates cell proliferation, metabolism, apoptosis, and immune response. It involves both STK and tyrosine kinases, including BRAF (v-raf murine sarcoma viral oncogene homolog), downstream effectors MEK1/2 (mitogen-activated protein kinase kinase), and ERK1/2 (extracellular signal-regulated kinase).[16,17] Input to the pathway includes metabolic stress, DNA damage, and growth factors. ERK1/2 mediates cross talk with the PI3K/AKT/mTOR pathway.

Alteration of the MAPK pathway has been implicated in cancer survival, dissemination, and resistance to treatment. [17] Mutations in EGFR (epidermal growth factor receptor), RAS (v-ras viral oncogene), and BRAF, which are found frequently in melanoma and in low-grade serous ovarian carcinoma (LGSOC), lead to overactivation of the MAPK pathway. [18] Sorafenib targets the RAF/MEK/ERK pathway and is currently FDA-approved for use in renal cell carcinoma, and trametinib is a MEK inhibitor that is FDA-approved for BRAF-mutated melanoma.[19] Though multiple clinical trials have been conducted, to date, no drugs acting on these pathways have been approved for ovarian cancer.[20] Side effects of drugs interacting with this pathway generally include diarrhea, nausea/vomiting, rash, and fatigue.[21]

The aurora kinase pathway regulates the cell cycle and mitosis through the mitotic spindle, centrosome separation, and cytokinesis.[6] These kinases assist in monitoring the mitotic checkpoint and therefore play a critical role in chromosome distribution to daughter cells.[6] Dysregulation, however, can lead to unrestrained cell proliferation and is linked to poor prognosis and to resistance to chemotherapy, especially to microtubule-based therapy. Aurora kinase A subtype has been implicated in solid malignancies including breast, ovarian, colon, pancreas, and prostate cancers.[22] One study found overexpression of aurora kinase A in up to 91% of ovarian malignancies; higher levels of expression were correlated with worse overall survival (OS).[23] Inhibition of aurora kinase A leads to defective spindle formation and chromosome separation, causing cell cycle arrest and apoptosis.[24] In xenograft models of ovarian cancer, aurora kinase inhibitors induce growth arrest and tumor regression.[25] Though multiple aurora kinase inhibitors have been developed to treat cancer, none have been FDA-approved. Major side effects include neutropenia, mucositis, and alopecia.

The above findings have suggested STK pathway inhibitors as attractive anticancer therapies, though the efficacy of STK inhibitors as a class of medications for ovarian cancer has not been fully assessed. The objective of this study was to conduct a systematic review of clinical outcomes of STK inhibitor-based treatment in recurrent epithelial ovarian cancer.

2. Methods

2.1. Literature search and study eligibility

A systematic literature review was conducted using a public search engine, PubMed/MEDLINE. Forty-one target searches were conducted by entering each of the keywords listed in Table 1 and ‘ovarian cancer.’ The search was conducted without limits on date or language (24 September 2015). Included publications were required to evaluate recurrent ovarian cancer patient outcomes in a clinical study (randomized controlled trials and prospective or retrospective studies) using STK inhibitor-based treatment regimens and to include data on patient demographics, treatment regimen, response to therapy, survival outcomes, and adverse events. Publications were excluded from the analysis if they were duplicated between searches, if reports of individual subjects were duplicated between studies, if they lacked data specific to ovarian cancer treatment outcomes and adverse events, if no STK inhibitor was given, or if they were classified as review articles, case series, commentaries, or preclinical studies. Studies evaluating STK inhibitors for primary prevention or initial treatment of ovarian cancer were also excluded from the study. This literature review involved the use of published data only, was not considered a human subjects study, and was exempt from IRB review. ‘Meta-analysis of Observational Studies in Epidemiology (MOOSE)’ was consulted as a guideline for reporting methods and results.[26]

Table 1.

Terms for literature search.

mTOR pathway	MAPK pathway	Aurora kinase pathway
mTOR	MAPK	Aurora
PI3K	MEK	Danusertib
AKT	Raf	Alisertib
STK11	Refametinib	Hesperadin
Metformin	Pimasertib	Barasertib
Temsirolimus	Binimetinib	Tozasertib
Everolimus	Trametinib
Ridaforolimus	Selumetinib
Pictilisib	Cometinib
Buparlisib	Vemurafenib
Idelalisib	Dabrafenib
Perifosine	Encorafenib
Apitolisib	Sorafenib
Voxtalisib	Tafinlar
Copanilsib
Dactolisib
Miltefosine
Afuresertib
Ipatasertib
Trisiribine

STK: serine–threonine kinase; PI3K: phosphatidylinositol 3-kinase; AKT: protein kinase B; mTOR: mammalian target of rapamycin; MEK: mitogen-activated protein kinase kinase; MAPK: mitogen-activated protein kinase.

2.2. Clinical information abstracted

For each study in the meta-analysis, we recorded drug regimens, year of publication, number of patients enrolled and evaluable for treatment, and type of study. Demographic data were compiled on individual study participants including the number of prior treatment regimens, platinum status, performance status, tumor histology, primary site (epithelial ovarian, primary peritoneal, or fallopian tubal), race, and median age of participants.

Treatment data were also abstracted including rates of overall response, complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Median progression-free survival (PFS) and OS were recorded where available. Adverse event information was collected, including incidences of overall and grade 3/4 events.

2.3. Study definitions

Responses were defined by the RECIST criteria at best response for all patients.[27] Both confirmed and unconfirmed responses were included. Overall response rate included patients with CR and PR, while clinical benefit rate was defined as the percentage of patients with CR, PR, and SD. PFS was defined as the time interval between the initiation of treatment and the date of first progression or last follow-up if the patient was alive. OS was defined as the time interval between the treatment initiation and the date of death due to disease or last follow-up if the patient was alive. Adverse events were graded per the CTCAE guidelines.[28]

2.4. Statistical analysis

The primary objective of the analysis was to examine treatment response and survival outcome of patients on STK inhibitor-based therapy. The secondary objective was to evaluate adverse events. A pooled analysis was conducted to tabulate response rates and clinical benefit rates of all STK inhibitors. Response rates among the various treatment regimens and STK pathways were compared with chi-square test or Fisher’s exact test as appropriate. The magnitude of statistical significance for treatment response was described with odds ratios (OR) and 95% confidence intervals (CI). The survival outcome and response rates were correlated among studies available for both outcome measurements, and treatment regimens above the median of PFS and treatment response (response and clinical benefit rates) were described. Frequency of adverse events per patient across the CTCAE system was compared with Mann–Whitney U test or Kruskal–Wallis test as appropriate. Statistical significance was set at a P value of less than 0.05 (two-tailed).

3. Results

3.1. Search results

Selection schema for the study is shown in Figure 1. The initial search terms returned 2020 publications. Of those, 632 were duplicated and excluded, leaving 1388 unique publications. Abstracts were then reviewed to eliminate 1351 publications not classified as clinical studies or not involving STK inhibitor therapy. This left 37 relevant studies, including 6 retrospective studies and 31 clinical trials.

Figure 1.

Selection schema. PubMed was searched on 9/24/2015. Abbreviation: STKi, serine-threonine kinase inhibitors.

3.2. Clinical trials cohort

3.2.1. Demographic data

Of the 31 qualifying clinical trials, 13 were excluded (Figure 1). Eight studies lacking ovarian cancer-specific outcomes were excluded in addition to three that did not use STK inhibitors as salvage therapy, one that did not specify treatment regimens used, and one in which patients were represented in another included study. The remaining 18 clinical trials represented 497 patients and 20 drug regimens.[29–46] Table 2 outlines demographic data. The median age of women in these trials was typically 50 years or older (77.8%). Only 6 studies (265 subjects representing 53% of the study population) reported data on race. Of these patients, the vast majority were Caucasian (96%). Only 3% were African-American and 1% was Asian. The majority of studies were published after 2010 (83.3%) and were phase II studies (52.6%). No phase III studies were reported. The majority (82.3%) of patients had performance status of 0–1, had serous histology (60%), had platinum resistant or refractory disease (73.2% among known platinum status), and had received at least two prior treatment regimens (55%). Notably, 63.6% of patients were given treatment with a MAPK pathway inhibitor, and 71% of these patients were treated with regimens that included sorafenib. Approximately one-fourth of patients (23.5%) were treated with mTOR-pathway inhibitors, and only 17% were given an aurora kinase inhibitor. Most patients (66%) received STK inhibitor therapy alone, while the remainder received a combination regimen. Nearly a quarter of patients (24.5%) received traditional cytotoxic chemotherapy with STK inhibitor.

Table 2.

Compiled demographics for clinical trial cohort.

	N = 497	%
Year of publication
Before 2000	0	0
2001–2010	81 (3 studies)	16.3
2011 or later	416 (15 studies)	83.7
Study type
Phase I	126 (9 studiesa)	25.4
Phase II	371 (10 studiesa)	74.6
Phase III	0 (0 studies)	0
Performance status
0	243	48.9
1	166	33.4
2	0	0
Unknown	88	17.7
Prior treatment regimens
0	2	0.4
1	115	23.1
2	165	33.2
≥3	108	21.7
Unknown	107	21.5
Platinum status
Sensitive	70	14.1
Resistant	158	31.8
Refractory	41	8.2
Platinum allergic	3	0.6
Unknown	225	45.3
Histology
Serous	298	60.0
Endometrioid	20	4.0
Mucinous	1	0.2
Clear cell	11	2.2
Other or unknown	167	33.6
Primary cancer
Epithelial ovarian	366	73.6
Fallopian tube	1	0.2
Primary peritoneal	47	9.5
Unknown	83	16.7
STK inhibitor pathway type
mTOR/AKT/PI3Kb	117	23.5
MAPK/MEK/Rafb	316	63.6
Aurora	85	17.1
Drug regimens
Single STK inhibition	307	61.8
Multiple STK inhibitionb	21	4.2
STK inhibitor + cytotoxic agent	122	24.5
STK inhibitor + hormonal agent	9	1.8
STK inhibitor + biologic agent	38	7.6

^aOne publication combined phase I and II trials.

^bRegimen of 21 pts taking both MEK 1/2 inhibitor trametinib and PI3K inhibitor buparlisib.[32]

mTOR: Mammalian target of rapamycin; AKT: protein kinase B; PI3K: phosphatidylinositol 3-kinase; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase; STK: serine–threonine kinases.

3.2.2. Treatment and survival outcomes

Of the original 497 patients, 442 were analyzable for treatment outcome by the RECIST criteria. We first performed a focused analysis of the 299 patients (11 trials) receiving only STK inhibitors, revealing an overall response rate of 10% and a clinical benefit rate of 64% due to high rates of SD (Table 3). The overall response and clinical benefit rates were notably higher among PI3K inhibitors at 40% and 95%, respectively. These responses were seen in a single study of 20 patients with RAS or BRAF mutations given a multiple STK inhibitor regimen of buparlisib (PI3K inhibitor) and trametinib (MEK1/2 inhibitor), which also resulted in one of the only two CRs seen without cytotoxic chemotherapy.[32] Similarly, selumetinib therapy induced a single CR without an adjunct cytotoxic agent in a patient with recurrent LGSOC.[34] Data on PFS were available from eight studies and on OS from five studies (Table 3). For STK inhibitor monotherapy, median PFS was 2.7 months and median OS was 14.1 months. PI3K inhibitors resulted in the longest median PFS at 7 months, tested in a regimen involving multiple STK inhibitors in patients with RAS and BRAF mutations.[32] PFS and OS were lowest for patients on aurora kinase inhibitors (2.0 and 9.8 months, respectively).

Table 3.

Treatment outcomes by STK pathways for STK inhibitors only (clinical trial cohort).

		mTOR/AKT/PI3K pathway				Aurora pathway	MAPK/MEK/Raf pathway
	All STK inhibitors	All pathways	mTOR	AKT	PI3K	All pathways	All pathways	Sorafenib
Response rates (RECIST)
No. of subjects (n)	299a	74	54	0	20	82	163	76
Response rate (%)	10	18	9	n/a	40	6	12	5
Clinical benefit rateb (%)	64	95	n/a	n/a	95	69	63	47
Complete response (%)	1	1	0	n/a	5	0	1	0
Partial responseb (%)	9	16	9	n/a	35	6	11	5
Stable diseaseb (%)	53	55	n/a	n/a	55	61	50	42
Progressive disease	35	5	n/a	n/a	5	29	37	53
Response rates (CA125)
No. of subjects (n)	98	0	0	0	0	51	47	0
Response rates (%)	16	n/a	n/a	n/a	n/a	20	13	n/a
Survival time (months)c
PFS	2.7	5.1	3.1	n/a	7.0	2.0	5.6	2.1
OS	14.1	11.6	11.6	n/a	n/a	9.8	16.3	16.3

Response rate includes complete response and partial response. Clinical benefit rate includes complete response, partial response, and stable disease.

^aNot all subjects included in Table 2 were evaluable for treatment response so numbers differ.

^bTwo trials (87 subjects) did not include data on rates of stable disease and progressive disease, so these subjects were removed from analysis of clinical benefit rate, stable disease, and progressive disease (for study with mTOR pathway inhibitor with temsirolims, subgroup of the study showed at least 53% of clinical benefit rate based on 51 cases reporting treatment response).

^cSurvival time is described from study enrollment to final status, numbers are described as median of median reported in each study (not available for all studies).

STK: serine–threonine kinases; mTOR: mammalian target of rapamycin; AKT: protein kinase B; PI3K: phosphatidylinositol 3-kinase; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase; PFS: Progression-free survival; OS: overall survival.

When all drug regimens were examined as a whole, the overall response rate was 13%, and the clinical benefit rate was 67% due to high rates of SD (Table 4). This result was similar among the three STK inhibitor pathways except for the previously mentioned study of buparlisib and trametinib.[32] Another notable exception was AKT inhibitor therapy with the lowest response rate (4%) of all drugs, the lowest clinical benefit rate (28%), and the highest rate of PD (72%). Data on PFS were available from 12 studies and on OS from 9 studies (Table 4). For all STK inhibitors combined, median PFS was 3.4 months and median OS was 13 months. The longest PFS, at 7.0 months, again resulted from the trial of multiple STK inhibitors.[32] PFS and OS were lower for patients on AKT inhibitors (1.9 and 4.5 months, respectively) than any other drug classes. For comparison, treatment response and survival outcome for only regimens combining STK inhibitors and cytotoxic agents are shown in Table 5.

Table 4.

Description of treatment outcomes by STK pathway (clinical trial cohort).

		mTOR/AKT/PI3K pathway				Aurora pathway	MAPK/MEK/Raf pathway
	All STK inhibitors	All pathways	mTOR	AKT	PI3K	All pathways	All pathways	Sorafenib
Response rates (RECIST)
No. of subjects (n)	442a	108	63	25	20	82	272	185
Response rate (%)	13	13	8	4	40	6	17	17
Clinical benefit rateb (%)	67	56	44	28	95	69	71	68
Complete response (%)	1	1	0	0	5	0	2	2
Partial responseb (%)	12	12	8	4	35	6	15	15
Stable disease (%)	52	39	44	24	55	61	53	51
Progressive diseaseb (%)	32	44	56	72	5	29	28	30
Response rates (CA125)
No. of subjects (n)	175	21	0	21	0	51	103	56
Response rates (%)	21	10	0	10	0	20	23	32
Survival time (months)c
PFS	3.4	3.2	3.2	1.9	7.0	2.0	4.6	3.7
OS	13.0	8.0	11.6	4.5	n/a	9.8	15.2	15.2

Response rate includes complete response and partial response. Clinical benefit rate includes complete response, partial response, and stable disease.

^aNot all subjects included in Table 2 were evaluable for treatment response so numbers differ.

^bTwo trials (87 subjects) did not include data on rates of stable disease and progressive disease, so these subjects were removed from analysis of clinical benefit rate, stable disease, and progressive disease (for study with mTOR pathway inhibitor with temsirolims, subgroup of the study showed at least 53% of clinical benefit rate based on 51 cases reporting treatment response).

^cSurvival time described from study enrollment to final status, numbers are described as median of median reported in each study (not available for all studies).

STK: Serine–threonine kinases; mTOR: mammalian target of rapamycin; AKT: protein kinase B; PI3K: phosphatidylinositol 3-kinase; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase; PFS: progression-free survival; OS: overall survival; SD: stable disease; PD: progressive disease.

Table 5.

Treatment outcomes by STK pathway for STK inhibitor with cytotoxic chemotherapy (clinical trial cohort).

		mTOR/AKT/PI3K pathway				Aurora pathway	MAPK/MEK/Raf pathway
	All STK inhibitors	All pathways	mTOR	AKT	PI3K	All pathways	All pathways	Sorafenib
Response rates (RECIST)
No. of subjects (n)	105	21	0	21	0	0	84	84
Response rate (%)	19	5	n/a	5	n/a	n/a	23	23
Clinical benefit rate (%)	71	24	n/a	24	n/a	n/a	83	83
Complete response (%)	3	0	n/a	0	n/a	n/a	4	4
Partial response (%)	16	5	n/a	5	n/a	n/a	19	19
Stable disease (%)	52	19	n/a	19	n/a	n/a	61	61
Progressive disease (%)	26	76	n/a	76	n/a	n/a	13	13
Response rates (CA125)
No. of subjects (n)	77	21	0	21	0	0	56	56
Response rates (%)	26	10	n/a	10	n/a	n/a	32	32
Survival time (months)a
PFS	3.7	1.9	n/a	1.9	n/a	n/a	3.7	3.7
OS	13.5	4.5	n/a	4.5	n/a	n/a	14.0	14.0

Only four trials included regimens with cytotoxic chemotherapy, represented above. All three studies in the MAPK/MEK/Raf pathway involved a combination of Sorafenib with traditional chemotherapy. The one study regimen involving an mTOR/Akt/PI3K pathway inhibitor involved the AKT inhibitor perifosine with docetaxel.

^aSurvival time is described from study enrollment to final status; numbers are described as median of median reported in each study. PFS and OS were not available for all studies.

STK: Serine–threonine kinases; mTOR: mammalian target of rapamycin; AKT: protein kinase B; PI3K: phosphatidylinositol 3-kinase; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase; PFS: progression-free survival; OS: overall survival.

Treatment responses were stratified by drug regimen classification (Figure 2(a)). Response rates differed significantly across the five study regimens (multiple STK inhibitors 40%, STK inhibitors with other class biological agents 25.8%, STK inhibitors with cytotoxic agents 19.0%, single STK inhibitors 7.9%, STK inhibitors with hormonal agents 0%; P < 0.001). STK inhibitor combination therapy resulted in significantly greater response rates than STK inhibitor monotherapy (STK inhibitors with cytotoxic agents vs. STK inhibitors alone, OR for response rate 2.75, 95%CI 1.43–5.28, P = 0.003; and STK inhibitors with other class biological agents vs. single STK inhibitors, OR 4.06, 95%CI 1.63–10.1, P = 0.005). Patients treated with multiple STK inhibitors responded at significantly higher rates than those taking single STK inhibitors (OR 7.79, 95%CI 2.88–21.1, P < 0.001).

Figure 2.

Treatment response based on study regimen and pathway. Response rates and clinical benefit rates broken down by (a) class of drug regimen and (b) STK pathway. mTOR inhibitors represent all STKi falling under the mTOR/AKT/PI3K pathway. Abbreviation: STKi, STK inhibitor.

Figure 2(a) also displays clinical benefit rates stratified by drug regimen class, demonstrating statistical significance across the five study regimens (multiple STK inhibitors 95%, STK inhibitors with other class biological agents 74.2%, STK inhibitors with cytotoxic agents 71.4%, single STK inhibitors alone 52.5%, STK inhibitors with hormonal agents 42.9%; P < 0.001). Clinical benefit rates with combination STK inhibitor therapy again were noteworthy for significantly greater response than single STK inhibitors (STK inhibitors with cytotoxic agents vs. single STK inhibitors, OR 2.26, 95%CI 1.39–3.67, P = 0.001; and STK inhibitors with other class biological agents vs. single STK inhibitors OR 2.60, 95%CI 1.12–6.01, P = 0.023). Multiple STK inhibitors improved the clinical benefit rate compared to STK inhibitor monotherapy (OR 17.2, 95%CI 2.27–130, P < 0.001).

Treatment responses were then compared among the three STK pathway types (Figure 2(b)). Response rates significantly differed across the pathways: inhibition for both mTOR and MAPK 40%, MAPK inhibitors 15.5%, mTOR pathway inhibitors 6.8%, and aurora kinase inhibitors 6.1%, P < 0.001. Patients receiving multiple STK inhibitors for the mTOR and MAPK pathways responded at higher rates than those given regimens based on a single STK pathway (mTOR/MAPK vs. mTOR, OR for response rate 9.11, 95%CI 2.69–30.8, P < 0.001; mTOR/MAPK vs. MAPK, OR 3.64, 95%CI 1.40–9.49, P = 0.011). Clinical benefit rate demonstrated statistical significance across the pathway types (mTOR and MAPK inhibitors 95.0%, MAPK inhibitors 68.7%, mTOR inhibitors 44.7%, and aurora kinase inhibitors 42.7%, P < 0.001). When compared with mTOR inhibitors alone, inhibition of both mTOR and MAPK pathways significantly improved clinical benefit rate (OR 23.5, 95%CI 3.01–184, P < 0.001), as it did when compared with use of MAPK inhibitors (OR 9.01, 95%CI 1.19–68.4, P = 0.01).

Treatment results (response and clinical benefit rates) for each regimen are displayed in Figure 3. The most effective treatment regimens included sorafenib with carboplatin/paclitaxel (response rate 60.9%),[31] sorafenib with bevacizumab response rate 42.1%),[36] and buparlisib and trametinib (response rate 40.0%).[32] Notably, the trial that showed the highest response rate (sorafenib with carboplatin/paclitaxel) used a regimen administered only to platinum-sensitive patients.[18] Similarly, when sorafenib was solely given to platinum-sensitive patients, the response rate was amongst the highest in STK inhibitor monotherapy (15.6%). The trial of buparlisib and trametinib included patients with somatic RAS or BRAF mutations (these upregulate MAPK/mTOR pathways), increasing the likelihood of response.[32] Lastly, the selumetinib regimen, which demonstrated the highest response rate among STK inhibitor monotherapy regimens (16.0%), was given only to patients with LGSOC.[34] These data indicate that the utility of STK inhibitors is maximized when combined with another class biological or cytotoxic agent in select patient populations.

Figure 3.

Treatment response per individual study regimen. Response rates and clinical benefit rates for each of 20 drug regimens represented in 18 clinical trials included in the analysis, listed in order of decreasing clinical benefit rates. Two trials randomized patients to different treatment regimens, which are listed separately in this table. *clinical benefit rate at least 53% based on best estimation among 51 cases. **SD and PD were not reported and clinical benefit rate is determined by the response rate.

Because clinical efficacy includes survival outcome, treatment responses were correlated with PFS (Figure 4). Twelve studies (42%) reported both treatment response and PFS. In this subset, the median response rate was 8.1% and the median clinical benefit rate was 57.6%. Median PFS was 3.4 months. The study regimens above the median for treatment response and PFS include sorafenib with carboplatin and paclitaxel,[31] selumetinib alone,[34] buparlisib and trametenib,[32] sorafenib alone,[31] and sorafenib with gemcitabine.[30]

Figure 4.

Correlation of treatment response with progression-free survival. Correlations of (a) response rate and (b) clinical benefit rate with PFS by individual drug regimen. Twelve studies with published PFS data were examined. Median PFS was 3.4 months demarcated horizontal lines. Median response rate was 8.1%, and median clinical benefit rate was 57.6% demarcated by vertical lines. Data points for drug regimens with both response rate and clinical benefit rate, and PFS above the median values for this analysis are labeled. Abbreviations: TC, paclitaxel and carboplatin; and PFS, progression-free survival.

3.2.3. Adverse events in clinical trials

Adverse events were analyzed in 13 clinical trials meeting the inclusion criteria. The most common grade 1–4 adverse effects per the CTCAE system were blood/lymphatic disorders, followed by gastrointestinal disorders, and skin/subcutaneous tissue disorders (Table 6). This pattern was similar for grade 3–4 adverse events. Adverse event rates between STK inhibitor monotherapy and STK inhibitors with chemotherapy were statistically similar (Table 6). Adverse event rates differed significantly across the three STK pathways with aurora kinase inhibitors having the highest adverse event rates and MAPK inhibitors having the lowest (Table 7).

Table 6.

Adverse event rate compared between STK inhibitor alone and STK inhibitor + cytotoxic agent regimens (clinical trial cohort).

	Grades 1–4				Grades 3 and 4
	All patients	STK alone	STK + cytotoxic agent		All patients	STK alone	STK + cytotoxic agent
Subject no.(n)	426	286	121	P value	426	286	121	P value
Blood/Lymphatic disorders	1.6	1.7	1.8	0.62	0.5	0.4	0.7	0.83
Cardiac disorders	0.1	0.1	<0.1		<0.1	0	0
Constitutional	0.3	0.5	<0.1		<0.1	0	0
Endocrine disorders	<0.1	<0.1	0		0	0	0
Eye disorders	<0.1	<0.1	<0.1		<0.1	0	0
Gastrointestinal disorders	1.4	1.3	1.7		0.2	0.1	0.3
General disorders/administration site conditions	0.5	0.4	0.8		0.1	0.1	0.2
Immune system disorder	<0.1	<0.1	<0.1		<0.1	0	0
Infections /Infestations	0.1	0.1	0.1		<0.1	<0.1	<0.1
Investigations	0.3	0.2	0.4		<0.1	<0.1	<0.1
Metabolism/Nutrition disorders	0.7	0.7	0.8		0.1	92 (0.1)	0.1
Musculoskeletal/Connective tissue disorders	0.1	0.1	0.1		<0.1	<0.1	<0.1
Neoplasms benign, malignant and unspecified	<0.1	<0.1	<0.1		<0.1	0	0
Nervous system disorders	0.3	0.3	0.3		<0.1	<0.1	0.1
Pain	0.3	0.3	0.1		<0.1	<0.1	<0.1
Renal/Urinary disorders	<0.1	<0.1	<0.1		<0.1	<0.1	0
Reproductive system/breast disorders	<0.1	<0.1	0		0	0	0
Respiratory, thoracic, and mediastinal disorders	0.1	0.2	0.1		<0.1	<0.1	0
Skin/Subcutaneous tissue disorders	0.7	0.7	0.8		0.1	0.1	0.2
Syndromes	<0.1	0	0		<0.1	<0.1	<0.1
Vascular disorders	0.1	<0.1	0.2		<0.1	<0.1	<0.1

Mann–Whitney U test for P values for comparison between STK inhibitor monotherapy versus STK inhibitor with chemotherapy. Rates of adverse events per patient were calculated by adding all patients experiencing adverse events in a particular organ system class and dividing by the total number of patients exposed to the drug regimen.

STK: Serine–threonine kinases.

Table 7.

Adverse event rate per patient based on STK pathways (clinical trial cohort).

	Grades 1–4				Grades 3 and 4
	Aurora kinase	mTOR-AKT-PI3K	MAPK-MEK-Raf		Aurora kinase	mTOR-AKT-PI3K	MAPK-MEK-Raf
Subject no. (n)	65	75	286	P value	65	75	286	P value
Blood/Lymphatic disorders	2.8	1.7	1.3	0.024	1.3	<0.1	0.4	0.012
Cardiac disorders	0	<0.1	0.1		0	<0.1	<0.1
Constitutional	0	0.6	0.3		0	<0.1	<0.1
Endocrine disorders	0	0	<0.1		0	0	0
Eye disorders	0	<0.1	<0.1		0	0	<0.1
Gastrointestinal disorders	1.9	1.8	1.1		0.1	0.3	0.1
General disorders/administration site conditions	0.8	0.3	0.5		<0.1	<0.1	<0.1
Immune system disorder	0	<0.1	<0.1		0	0	<0.1
Infections and infestations	0	0.1	<0.1		0	<0.1	<0.1
Investigations	0.7	0	0.2		<0.1	0	<0.1
Metabolism and nutrition disorders	1.1	0.5	0.6		<0.1	0.1	0.1
Musculoskeletal and connective tissue disorders	0	<0.1	<0.1		0	<0.1	<0.1
Neoplasms benign, malignant, and unspecified	0	0	<0.1		0	0	<0.1
Nervous system disorders	0	0.2	0.3		0	0	<0.1
Pain	0	0.5	0.2		0	0.1	<0.1
Renal and urinary disorders	0	<0.1	<0.1		0	<0.1	<0.1
Reproductive system and breast disorders	0	<0.1	<0.1		0	0	0
Respiratory, thoracic, and mediastinal disorders	0	0.4	<0.1		0	<0.1	<0.1
Skin/Subcutaneous tissue disorders	0.4	0.5	0.9		0	<0.1	0.2
Syndromes	0	0	<0.1		0	0	<0.1
Vascular disorders	0	0	0.2		0	0	<0.1

Krsukal–Wallis test for P values (comparison across the three groups). Rates of adverse events per patient were calculated by adding all patients experiencing adverse events in a particular organ system class and dividing by the total number of patients exposed to the drug regimen.

STK: Serine–threonine kinases; mTOR: mammalian target of rapamycin; AKT: protein kinase B; PI3K: phosphatidylinositol 3-kinase; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase.

3.3. Retrospective study cohort

All six retrospective studies involved treatment with metformin. Three described metformin’s impact on ovarian cancer incidence rather than treatment response or survival time; these were excluded from our analysis. Of the remaining three studies, published between 2012 and 2014, two were retrospective cohort studies, and one was a case-control study. While all three studies found that diabetes negatively affected survival, two studies reported that metformin significantly improved survival outcomes, even above those of nondiabetics.[47,48] One study reported no survival difference in diabetic patients based on metformin status [49]; however, authors noted the limitation of a small sample size.

4. Conclusion

In this predominantly platinum-resistant and heavily pretreated ovarian cancer population, the overall response rate of STK inhibitors was 13%. STK inhibitor-based therapy maintained a considerable proportion of SD (clinical benefit rate 67%), suggesting a possible utility for this purpose. When used as monotherapy, STK inhibitors seem to have minimal effect on recurrent ovarian cancer. However, response rates in combination regimens (sorafenib with carboplatin/paclitaxel, gemcitabine, or bevacizumab) were significantly greater. Lastly, certain conditions maximized the effectiveness of STK inhibitor-based therapy that warrant special attention. These include platinum-sensitive disease (sorafenib), LGSOC (selumetinib), and somatic RAS/BRAF mutations (buparlisib with trametnib). To elucidate the true benefit of STK inhibitors in ovarian cancer, future research should focus on the efficacy of STK inhibitors in these subtypes of ovarian cancer and on differentiating response rates in randomized controlled trials comparing chemotherapy with and without the addition of STK inhibitors.

5. Expert opinion

Responses to STK inhibitor-based therapy in this meta-analysis are comparable to published literature on overall response rates for all recurrent ovarian cancer therapies. Previously reported response rates of platinum-resistant disease to single-agent cytotoxic chemotherapy are around 10% with median PFS 3–4 months.[50] Combination cytotoxic regimens have higher antitumor activity than monotherapy with response rates ranging from 21.7 to 33.7%.[3] Our results for STK inhibitor-based therapy for recurrent ovarian cancer were on par with other available treatments currently used for platinum-resistant ovarian cancer. Combination therapy with STK inhibitors and cytotoxic agents resulted in higher response rates when compared with STK inhibitor monotherapy (19.0% vs. 7.9%, Figure 2(a)). Because adding cytotoxic chemotherapeutic agents to STKi does not seem to increase the risk of adverse events (Table 5), it may be a reasonable approach to add cytotoxic agents when considering STK inhibitor-based therapy for recurrent ovarian cancer. However, additional trials directly comparing STK inhibitor-based regimens with chemotherapy alone are needed to determine if STK inhibitors provide true additional benefit.

Notably, several studies not included in this meta-analysis found STK inhibitors ineffective when used as an adjunct to primary adjuvant, neoadjuvant, or maintenance therapy. One randomized control trial compared sorafenib to placebo as maintenance therapy following primary treatment with surgery and adjuvant chemotherapy but did not show significant benefit in PFS.[51] In another trial, patients were randomized to carboplatin and paclitaxel with or without sorafenib as primary therapy after surgical cytoreduction. Authors detected no difference in efficacy but substantially increased toxicity (skin reactions, hand–foot syndrome, mucositis, and hypertension).[52] Sorafenib was similarly studied as an adjunct to neoadjuvant carboplatin and paclitaxel; however, only four patients were enrolled before the study was terminated due to severe postoperative toxicities.[53] Though there does not seem to be a role for STK inhibitors in primary treatment, our study suggests possible efficacy in maintaining SD based on high clinical benefit rates. Additional research on alternate STK pathway inhibitors for maintenance therapy is warranted.

Our data suggest a possible use for STK inhibitors in a combination regimen for salvage therapy and for maintenance of disease stability, particularly in KRAS and BRAF mutant tumors, which are frequently seen in LGSOC. A recent meta-analysis showed rates of KRAS and BRAF mutations in LGSOC as high as 63% and 33%, respectively, which may explain susceptibility to MEK inhibitors.[54] In general, LGSOC and KRAS/BRAF mutations are associated with treatment resistance but longer survival.[54,55] We noted a unique 16% response with selumetinib monotherapy in LGSOC patients. In this trial, 47% of the tumor samples from 34 patients with tissue available for analysis harbored KRAS/BRAF mutations. However, response stratified by mutation status showed no difference between those with and without the mutation.[34] The single patient in this study experiencing CR was noted to have an ongoing response (>5 years), despite previous progression on multiple lines of intravenous and intraperitoneal chemotherapy. Molecular profiling of this patient’s tumor by Grisham et al. revealed a mutation in a negative regulatory region of MAP2K1, which encodes MEK1, a driver mutation deemed to explain her unique response. These authors then sequenced 48 LGSOC and serous borderline tumors and found that patients with exceptional responses to MEK inhibitors had multiple novel MAPK pathway alterations, rather than conventional hotspot KRAS/BRAF mutations.[56] Several clinical trials are currently in progress to further study STK inhibitors in LGSOC (Table 8).

Table 8.

Summary of ongoing trials for STKi for ovarian cancer.

Compound	Target	Study (NCT no.)	Phase	Country	Study start date	Status	Condition	Interventions
mTOR/AKT/PI3K pathway
Metformin	mTOR pathway	NCT01579812	Phase II	USA	10/2011	Recruiting	Advanced (primary treatment)	Met + chemotherapy
Metformin	mTOR pathway	NCT02437812	Phase II	USA	1/2014	Recruiting	Newly diagnosed (primary treatment)	Met + TC
Metformin	mTOR pathway	NCT02122185	Phase II	USA	6/2014	Recruiting	Stage III–IV (primary therapy)	TC (docetaxel) ± met (followed by met vs. placebo maintenance)
Metformin	mTOR pathway	NCT02312661	Phase I	Netherlands	10/2015	Recruiting	Recurrent	Met + TC
Everolimus	mTORC1	NCT01031381	Phase II	USA	9/2010	Completed 5/2013	Recurrent	Everolimus + bevacizumab
Everolimus	mTORC1	NCT01281514	Phase I	USA	12/2010	Recruiting	Platinum-sensitive recurrent	Everolimus + Carboplatin + PLD
Everolimus	mTORC1	NCT02188550	Phase II	USA	6/2014	Recruiting	Platinum-resistant or -refractory recurrent or persistent	Everolimus + letrozole
Everolimus	mTORC1	NCT02283658	Phase II	USA	11/2014	Ongoing Not recruiting	Recurrent hormone receptor positive	Everolimus Letrozole
Ridaforolimus	mTORC1	NCT01256268	Phase I	USA	6/2011	Ongoing Not recruiting	Recurrent or metastatic platinum-sensitive	Ridaforolimus + TC
Sirolimus	mTORC1	NCT01536054	Phase I	USA	8/2012	Ongoing Not recruiting	Recurrent	Sirolimus NY-ESO-1
Temsirolimus	mTORC1	NCT01065662	Phase I	USA	2/2010	Ongoing Not recruiting	Recurrent	Temsirolimus + cediranib
Temsirolimus	mTORC1	NCT01196429	Phase II	USA, Japan, Korea	8/2010	Completed (1/2015)	Stages III and IV OCCC (primary treatment)	Temsirolimus + TC (then temsirolimus consolidation)
Temsirolimus	mTORC1	NCT01460979	Phase II	Germany	10/2011	Completed (11/2015)	Platinum-refractory	Temsirolimus
GSK2141795	AKT	NCT01266954	Phase I	UK	6/2010	Completed (12/2014)a	Recurrent or persistent	GSK2141795
MK2206	AKT	NCT01283035	Phase II	USA	4/2011	Completed (12/2014)	Recurrent platinum-resistant	MK2206
GSK2110183	AKT	NCT01653912	Phase I/II	Australia, Russia, UK	12/2012	Ongoing Not recruiting	Recurrent platinum-resistant	GSK2110183 + TC
Triciribine	AKT	NCT01690468	Phase I/II	USA	9/2014	Ongoing Not recruiting	Recurrent or persistent platinum-resistant	Triciribine + carboplatin
COTI2	AKT	NCT02433626	Phase I	USA	12/2015	Recruiting	Recurrent platinum-resistant	COTI2
MAPK pathway
Sorafenib	MAPK	NCT00436215	Phase II	USA	9/2006	Completed (6/2014)	Recurrent platinum-resistant/refractory	Sorafenib + bevacizumab
Sorafenib	MAPK	NCT01047891	phase II	Germany	1/2010	Completed (2/2015)	Recurrent platinum-resistant	Sorafenib
Trametinib	MAPK	NCT02101788	Phase II/III	USA, UK	2/2014	Recruiting	Recurrent or progressive LGSOC	Trametinib vs. letrozole, paclitaxel, PLD, tamoxifen, or topotecan
LY2228820	MAPK	NCT01663857	Phase I/II	Multi-countriesc	9/2012	Recruiting	Recurrent platinum-sensitive	LY2228820 + gemcitabine + carboplatin
MEK 162	MEK	NCT01649336	Phase I	USA	7/2012	Ongoing Not recruiting	Platinum-resistant or -refractory	MEK 162 + paclitaxel
MEK 162	MEK	NCT01849874	Phase III	Multi-countriesd	6/2013	Recruiting	Recurrent, persistent, or progressive LGSOC	MEK 162 vs. PLD, paclitaxel, or topotecan
Aurora kinase pathway
Alisertib	Aurora kinase	NCT01091428	Phase I/II	USA	5/2010	Ongoing Not recruiting	Recurrent	Alisertib + paclitaxel
Multiple STKi
Pimasertib + SAR245409	MAPK, PI3K	NCT01936363	Phase II	Multi-countriesb	9/2013	Ongoing Not recruiting	Previously treated unresectable LGSOC	Pimasertib + SAR245409 vs. placebo

Table compiled through access to clinicaltrials.gov on 24 December 2015. Ongoing, recruiting, and completed studies involving only mullerian cancers are listed.

^aResults were available after our search on 24 September 2015.

^bUSA, Australia, Belgium, Canada, France, Italy, Poland, and Spain.

^cUSA, Australia, Belgium, Germany.

^dUSA, Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Luxembourg, Netherlands, Norway, Poland, Spain, Sweden, and the United Kingdom.

met: Metformin; TC: paclitaxel/carboplatin; PLD: pegylated liposomal doxorubicin; OCCC: ovarian clear cell carcinoma; LGSOC: low-grade serous ovarian carcinoma; STKi: STK inhibitor; STK: Serine–threonine kinases; mTOR: mammalian target of rapamycin; AKT: protein kinase B; PI3K: phosphatidylinositol 3-kinase; MAPK: mitogen-activated protein kinase; MEK: mitogen-activated protein kinase kinase.

Given that the Bedard et al. trial involving multiple STK inhibitors also focused on patients with somatic KRAS and BRAF mutations, we further examined the breakdown of included subjects’ tumor grades. Eighteen of 21 (85.7%) patients had well or moderately differentiated tumors, while only 2 (9.5%) were poorly differentiated. Of these, 90.5% had KRAS mutations while only one patient had a BRAF mutation. Responses were not stratified by tumor histology, but authors did note possible benefit for patients particularly with KRAS G12V-mutated ovarian cancers. They found no association between response and markers in the PI3K and MAPK pathways.[32] Several additional studies have noted an increased prevalence of mTOR pathway activation and PIK3CA mutations in ovarian clear cell carcinoma (OCCC).[14,57] One study published data on patients with somatic PIK3CA mutations treated on various study protocols involving mTOR pathway drugs. Of nine patients with ovarian cancer, two (22.2%) had PRs.[58] Furthermore, Grisham et al. suggested that MEK inhibitors may only be effective in patients with specific MAPK pathway alterations, similar to findings in the literature on melanoma. [56] These findings highlight the growing importance and complexity of biomarker identification and individualized cancer treatment, as the key to maximizing response is not a one drug fits all approach.

Another explanation for the efficacy of the multiple STK inhibitor regimen may lie in the ability of one pathway inhibitor to overcome resistance to other STK inhibitor therapies due to the existence of significant cross-talk and feedback inhibition between all three STK pathways. mTOR forms two complexes (mTORC1 and mTORC2) enacting downstream effects with feedback inhibition on pathway initiation.[14] First-generation mTOR inhibitors (including temsirolimus, everolimus, and ridaforolimus) only inhibit mTORC1, which may result in resistance via increased PI3K, AKT, and mTORC2 activation. The interaction between the MAPK and mTOR pathways has been extensively studied. Normally, RAS, RAF, MEK1/2, and ERK indirectly stimulate AKT activation of mTOR via inhibition of TSC1/2. As a result, inhibition of the MAPK pathway downregulates the mTOR pathway.[4] Moreover, studies have shown that mTOR inhibition leads to activation of the MAPK pathway via a PI3K-dependent feedback loop, reversible with MAPK inhibitors.[59] Potential mechanisms to overcome resistance include use of multiple mTOR and PI3K or AKT inhibitors. Other suggested combinations include mTOR inhibitors with MAPK inhibitors, VEGF inhibitors, and IGF-1 inhibitors. Examples include sorafenib with bevacizumab and ridaforolimus with dalotuzumab. [36,39,60] Furthermore, second-generation mTOR inhibitors that inhibit both mTORC1/2 are in development and may be of interest for ovarian cancer.[4] Further studies are needed to evaluate the efficacy of multiple mTOR and MAPK pathway inhibitor therapy.

Though less well understood, there is also evidence of interaction between the mTOR, MAPK, and aurora kinase pathways that may provide mechanisms for overcoming resistance. Preclinical studies have suggested that overexpression of aurora kinase A activates the mTOR pathway [61] and that cellular transformation induced by aurora kinase A is determined by co-activation of mTOR pathway.[62] Aurora kinase B phosphorylation of mTOR substrates has similarly been shown to promote cell-cycle progression in T lymphocytes.[63] Demonstrating cross-talk between the MAPK and aurora kinase pathway, inhibition of aurora kinases was indirectly shown to block ERK activity in renal cell carcinoma.[64] Additional research focused on ovarian cancer is needed to determine if combining aurora kinase inhibitors with mTOR or MAPK inhibitors may also assist in overcoming resistance.

Preclinical studies suggesting a potential use for PI3K/AKT/mTOR pathway inhibitors and MAPK pathway inhibitors in combating cancer stem cells may further explain the efficacy of multiple STK inhibitor therapy. Current theory on cancer stem cells proposes that stem cells comprise a small population of tumors that are relatively resistant to conventional chemotherapy. Recurrence following apparent complete remission may originate from these cells. This model seems relevant in ovarian cancer, given that the majority of patients relapse following complete remission.[65] Furthermore, presence of these cancer stem cells is associated with poor prognosis.[66] Studies have demonstrated that agents inhibiting the PI3K/AKT/mTOR pathway, especially newer class inhibitors that act on both mTORC1 and mTORC2 or on both the PI3K and mTOR molecules, preferentially target cancer stem cells. [67,68] A 2011 study by Alvero et al. demonstrated stem cell cytotoxicity using NV-128, which decreases adenosine triphosphate and indirectly inhibits both the mTOR and MAPK pathways.[69] Based on these findings, further preclinical investigation of the impact of second-generation mTOR inhibitors and multiple pathway inhibitors on ovarian cancer stem cells is warranted. Furthermore, this suggests a potential therapeutic role for these agents in patients whose tumors test positive for ovarian cancer stem cell markers such as aldehyde dehydrogenase and CD133.[66]

Metformin is a biguanide antidiabetic agent typically used as an insulin sensitizer. However, epidemiologic data show decreased incidence of ovarian cancer in diabetic women taking metformin.[70,71] Through multiple mechanisms, including AMPK activation and mTORC2 inhibition, metformin also inhibits the mTOR pathway.[72,73] Preclinical studies demonstrate that metformin inhibits proliferation of ovarian cancer cells.[74,75] Two of three studies in our analysis described a survival benefit for metformin.[47,48] The one study that did not detect a difference examined metformin effect as a secondary outcome and noted low numbers of patients on metformin.[49] Although no results are available from clinical trials examining the effects of metformin on ovarian cancer survival, the above studies provide the basis for an ongoing clinical trial evaluating the role of metformin as an adjunct to chemotherapy, even in nondiabetics.

Table 6 shows a list of ongoing, recruiting, and completed clinical trials as per a search of the clinical trial registration site (accessed 24 December 2015).[76] The most commonly examined target was the mTOR pathway. In addition to the trials mentioned previously, there are phase III trials underway for recurrent LGSOC. There is similarly a phase II study examining the efficacy of inhibition of MAPK and PI3K pathways, as well as a trial of temsirolimus specifically for OCCC.

This is the first meta-analysis in the literature examining the performance of STK inhibitors as a whole for ovarian cancer treatment. There were, however, several notable limitations. By necessity, the studies included in the analysis were comprised of a variety of different drug regimens at different doses in varying ovarian cancer histologies, contributing to a complex analysis. This study did not analyze the impact of differential dosing on response rates or survival. Due to the inclusion of multiple combination regimens, none of which compare regimens with and without an STK inhibitor, it is also difficult to estimate the relative contributions of STK inhibitors and chemotherapy to improved disease outcomes. Furthermore, the meta-analysis included studies with vastly different base patient populations, including variations in platinum status and tumor histology, which makes interpretation of our results difficult. That is, while platinum-sensitive disease was associated with high response rates in one particular trial, our study was not able to differentiate treatment outcomes between platinum-sensitive and -resistant cases. Though two studies notable for high response rates involved LGSOC, we were unable to fully stratify results by tumor histology, which would have greatly strengthened the analysis. The data on STK inhibitors in the select populations mentioned above each came from single trials. These data would be strengthened if combined in the future with data from the ongoing clinical trials mentioned in Table 6 and if supported by appropriate randomized control trials comparing the effects of STK inhibitors with chemotherapy.

Though Sorafenib is a dual-action inhibitor affecting both the RAF/MEK/ERK and VEGFR/PDGFR pathways, its antitumor effects are attributed more to VEGF pathway inhibition than MEK pathway inhibition.[77–79] Therefore, it is debatable if sorafenib can be included in this study. However, the clinical benefit rate of 12 studies excluding sorafenib was 53%, still supporting the role of STK inhibitors as potential agents to maintain SD in the management of recurrent ovarian cancer.

We were unable to stratify results by race, which may affect both mutational status and responsiveness to treatment. Though the majority of trials did not provide data on race, we were able to ascertain that at least 51%, but probably a much larger proportion, of the patients were Caucasian, likely a reflection of the patient populations in North America, Australia, and Europe, where the trials were conducted. In addition, data on adverse events were differentially classified between the studies included. For this reason, we were unable to calculate the actual proportion of patients experiencing side effects in each system organ class. Instead, this study calculated the frequency of adverse events per patient, enabling us to compare the relative occurrence of adverse events but not the proportion of patients that can expect to experience a given effect.

Overall, this study showed a modest impact of STK inhibitors in recurrent ovarian cancer. However, STK inhibitors may have potential to maintain SD in patients with ovarian cancer recurrence, especially in select populations with somatic mutations in STK pathway molecules and in patients with LGSOC. Additional research is needed to confirm this benefit and to separate the effects of STK inhibitors from those of cytotoxic chemotherapy. Further studies are expected to identify additional biomarkers for STK inhibitor efficacy. Potential directions for future inquiry in ovarian cancer include combination use of multiple STK inhibitors and second-generation mTOR inhibitors, combination regimens of STK inhibitors with other class biological agents, the effect of metformin on survival, the effect of STK inhibitors on ovarian cancer stem cells and the prevention of recurrent cancer, and the efficacy of these regimens in specific subsets of patients with LGSOC.

Article highlights.

• Clinical outcomes of serine-threonine kinase inhibitors targeting mTOR, MAPK, and aurora kinase pathways for recurrent ovarian cancer were examined in 18 clinical trials.

• Median progression-free survival was 3.4 months for STK inhibitor-based therapy with response rate being 13% in largely platinum-resistant study population.

• The main benefit of STK inhibitor-based therapy seems to be its ability to maintain disease stability in ovarian cancer recurrence as the majority of patients undergoing this treatment (67%) attained clinical benefit; however, additional research is needed to support this idea.

• Treatment responses improved when STK inhibitors were combined with other class biological agents or with cytotoxic agents. This was most notable in a single study of multiple STK inhibitors, where both agents may act together to overcome drug resistance. Randomized controlled trials are needed to directly compare the effects of STK inhibitors to cytotoxic chemotherapy.

• Treatment response was significantly improved in selected patient populations with somatic mutations affecting STK inhibitor pathways, with platinum-sensitive disease, and with low-grade serous ovarian carcinoma, indicating the importance of appropriate patient selection when using these drugs.

This box summarizes key points contained in the article.