Phase II, 2-Stage, 2-Arm, PIK3CA Mutation Stratified Trial of MK-2206 in Recurrent Endometrial Cancer

Abstract

Endometrial cancers have high rates of phosphoinositide 3-kinase (PI3K) pathway alterations. MK-2206 is an allosteric inhibitor of AKT, an effector kinase of PI3K signals. We hypothesized patients with tumors harboring PIK3CA mutations would be more likely to benefit from MK-2206 than those without PIK3CA mutation. A phase II study was performed in patients with recurrent endometrial cancer; all histologies except carcinosarcoma were eligible. Up to 2 prior chemotherapy lines were permitted, excluding prior treatment with PI3K pathway inhibitors. The first 18 patients were treated with MK-2206 200mg weekly. Due to unacceptable toxicity, dose was reduced to 135mg. Co-primary endpoints were objective response rate (ORR) and progression free survival at 6 months (6moPFS). Thirty-seven patients were enrolled (1 ineligible). By somatic PIK3CA mutation analysis, 9 patients were mutant (MT) (1 with partial response [PR]/6moPFS, 2 with 6moPFS). Twenty-seven patients were wildtype (WT) (1 PR and 4 6moPFS). Most common toxicities were rash (44%), fatigue (41%), nausea (42%) and hyperglycemia (31%). Grade 3 and 4 toxicities occurred in 25% and 17% of patients, respectively. Exploratory analysis found serous histology had greater 6moPFS as compared to all other histologies (5/8 vs. 2/28, p=.003). PTEN expression was associated with median time to progression (p= 0.04). No other significant associations with PI3K pathway alterations were identified. There is limited single agent activity of MK-2206 in PIK3CA MT and PIK3CA WT endometrial cancer populations. Activity was detected in patients with serous histology and due to their poor outcomes warrants further study. (NCT01307631)

Introduction

Endometrial cancer is the most common gynecologic malignancy in the developed world; it affects greater than 60,000 women and is the cause of death for more than 10,000 women in the U.S. annually¹. Contrary to trends in other solid tumors, endometrial cancer incidence and mortality continue to increase. Although early stage disease is often curable, there are limited effective therapies for advanced and recurrent disease². Thus, development of novel therapies for endometrial cancer is critical.

The phosphoinositide 3-kinase (PI3K) signaling pathway regulates key aspects of cancer biology. Briefly, upon activation of receptor tyrosine kinases, PI3K phosphorylates and alters the cellular lipid membrane, resulting in localization of the protein kinase AKT to the membrane. AKT is then activated, setting off a cascade of protein modifications that result in the activation of mTORC1, a key regulator of cellular energy use, growth, and survival³.

Endometrial cancers have the highest reported rate of PI3K pathway alterations among any solid tumor. Mutations or loss of protein expression of PTEN occur in more than 80% of endometrioid endometrial cancers. In endometrioid and serous tumors, PI3KCA mutations are found in 32–40% and PI3KR1 mutations in 20%^4–6. In addition, mutations in AKT occur, albeit rarely (2–3%)⁴. As these abnormalities activate the PI3K pathway, this pathway is an attractive target in endometrial cancer^4–6.

Clinical trials of rapamycin analogs (rapalogs), allosteric inhibitors of mTORC1, demonstrate modest clinical activity in endometrial cancer^7–14. Rapalogs are incomplete inhibitors of mTORC1 targets and, due to feedback loops, can result in activation of upstream PI3K signals¹⁵. MK-2206 (Merck Pharmaceuticals, New Jersey) is an allosteric inhibitor of all three isoforms of AKT (AKT1, 2, and 3), with greater activity against AKT1 and 2¹⁶. The present study was based on the hypotheses that; first, the inhibition of AKT would result in greater clinical efficacy than rapalogs and, second, presence of a PIK3CA activating mutation would yield more sensitivity to MK-2206 in women with endometrial cancer.

Patients and Methods

This was an open label, phase II, 2-stage, 2-arm, PI3KCA mutation stratified study of MK-2206 in patients with recurrent endometrial cancer (Figure 1). The study was conducted at 6 sites in the United States as part of the Stand Up to Cancer “PI3K in Women’s Cancer” Dream Team. All trial funding was provided by AACR. MK-2206 supply was provided by Merck. The trial was registered at ((clinicaltrials.gov NCT03043001.

Figure 1:

Study schema

Patient Population

Eligible patients had histologically confirmed recurrent endometrial cancer with measurable disease by RECIST 1.1 ¹⁷. All histologies except carcinosarcoma were eligible. An ECOG performance status of ≤ 2 was required. Patients were required to have one and permitted to have no more than 2 prior systemic regimens, including chemotherapy given as a radio-sensitizer. Prior treatment with PI3K pathway inhibitors, inclusive of rapalogs, was not permitted. All patients were required to have adequate bone marrow (absolute neutrophil count (ANC) ≥ 1,500/mcL, hemoglobin ≥ 9.0 g/dL, platelets ≥ 100,000/mcL ), liver (total bilirubin within institutional limits and AST/ALT ≤ 2.5 x institutional upper limit of normal) (ULN)), and renal function (creatinine within institutional normal limits or creatinine clearance ≥ 60 mL/min/1.72 m²). Exclusion criteria included brain metastasis, anti-cancer treatment within 3 weeks of enrollment, or residual toxicities not resolved to ≤ grade 1 (excluding alopecia). Patients with diabetes or hyperglycemia on screening labs (HgA1C > 7.5% or fasting blood glucose ≥ 130mg/dL) were excluded. All subjects provided written informed consent; the study was Institutional Review Board approved at all centers.

Treatment Plan

MK-2206 was initially administered at a starting dose of 200mg orally once a week (QW), the previously identified recommended Phase II dose^{18, 19}. Cycle length was 28 days. Dose level −1 was 135mg QW and dose level −2 was 90mg QW.

Each cycle of MK-2206 was continued based on these criteria, ANC ≥ 1,000/mcL, platelets ≥ 75,000/mcL, AST/ALT ≤ 2.5 institutional ULN, creatinine within normal institutional limits or creatinine clearance > 60 mL/min/1.73 m2, and fasting glucose < 150 mg/dL). Delay in therapy was permitted for a maximum of 14 days. Patients with nonhematologic grade 4 toxicities were discontinued from the study. Patients received MK-2206 until disease progression, unacceptable adverse events, intercurrent illness preventing further MK-2206 administration or patient choice to withdraw.

Assessments

Toxicities were monitored through the duration of the study and 30 days after cessation of study treatment. Common Terminology Criteria for Adverse Events (CTCAE), Version 4.03 was utilized to grade adverse events. All patients underwent disease radiographic assessment with CT or MRI based on RECIST 1.1 every 2 cycles.

Sample Collection/Molecular Testing

Sample collection:

Fixed primary tissue and whole blood were collected prospectively from patients enrolled. DNA was isolated from macro-dissected tumor and matched whole blood specimens, if available.

Molecular Testing:

Immunohistochemistry (IHC) analysis for Stathmin and PTEN was performed and categorized as previously published^{20, 21}. Next–generation sequence analysis was performed on a panel of 504 genes (Oncopanel_v2) with relevance in cancer by the Center for Cancer Genome Discovery at the Dana Farber Cancer Institute (as previously described)^{22, 23} including PTEN, PIK3CA, KRAS, and AKT. For the purposes of stratification, PIK3CA mutations included R88Q, K111N, E110K, E418K, C420R, E453K, E542K/V,E545K, Q546R, H701P, M1043V, H1047R/L/Y change. For retrospective analyses, KRAS mutations included G12D, G12V, and G13D. AKT1 mutations were limited to E17K.” All non-synonymous changes were included for PTEN and PIK3CA genes.

Statistical Methods

This study evaluated efficacy of MK-2206 using co-primary endpoints of objective response rate (ORR, defined as complete response (CR) or partial response (PR) per RECIST version 1.1) and frequency of patients who survived progression free for at least 6 months (6moPFS) in two distinct stratum defined by PI3KCA mutation status (mutant (MT) versus wildtype (WT)). Secondary endpoints included safety, time on treatment, progression free and overall survival. Exploratory endpoints included association between molecular aberrations and efficacy.

Independent Simon 2-stage tests were planned within PIK3CA MT and WT strata. For the PI3KCA MT stratum, a total of 25 patients were planned, 15 in the first stage and 10 in the second stage, to discriminate ORR < 5% and 6moPFS < 10% versus ORR > 25% or 6moPFS > 35%. A maximum of 55 PI3KCA WT patients were planned, 31 in the first stage and 24 in the second stage, to discriminate ORR< 5% and 6moPFS <10% versus ORR > 20% or 6moPFS > 25%. The study was closed to accrual in July of 2012 due to feasibility; analysis of PIK3CA MT and WT strata were performed as a single-stage.

Event rates in each stratum were reported with 90% confidence intervals (90%CI) from an exact Binomial distribution. For secondary endpoints, dichotomous and categorical variables were summarized as proportions with 90%CI, and time-to-event variables evaluated using the Kaplan-Meier product-limit estimator. Contrasts between stratum and patient subgroups were based on Fisher’s Exact test and Score tests from a Cox-proportional hazard model. Time on treatment was defined as duration of time from treatment start to off-treatment date. Patient time on treatment was grouped by molecular phenotypes and analyzed. Due to limited sample size, these analyses are underpowered and provided for illustration. Association between molecular and clinical variables utilized Wilcoxon rank sum, Kruskal-Wallis and Spearman correlation tests where indicated.

All statistical analyses were performed using SAS 9.4 (Cary, NC) and figures plotted using SAS 9.4 and R (https://CRAN.R-project.org/doc/FAQ/R-FAQ.html).

Results

Thirty-seven patients were enrolled on the study; one patient withdrew before dosing. Among the eligible patients, 9 had PI3KCA MT and 27 had PI3KCA WT tumors. Patient characteristics are described in Table 1. Median age was 62 (range, 39 – 76), and the majority of patients had endometrioid histology (44%). Median time to receipt of result of PI3KCA mutation testing was 119 days (range 50 – 369). Due to rate of ≥ grade 3 rash (39%) among the first 18 patients treated at the 200mg dose, the protocol was amended to a starting dose of MK-2206 135mg QW. Eighteen patients (50%) were treated at each dose level, 200mg and 135mg QW. Enrollment was held after 36 patients due to feasibility of timely molecular analysis as well as apparent low activity of the compound.

Table 1.

Patient demographics and clinical characteristics (n=36)

	PIK3CA				Overall
	WT		MT
	N	%	N	%	N	%
N	27	75	9	25	36	100
Age: Median (range)	62.4 (39.3 – 75.7)		61.1 (51.3 – 72.8)		62.1 (39.3 – 75.6)
Race
White	20	74.1	9	100	29	80.5
Black	2	7.4	0	0	2	5.6
Asian	1	3.7	0	0	1	2.8
Other	4	14.8	0	0	4	11.1
Stage
I	8	29.7	6	66.7	14	38.9
II	3	11.1	0	0	3	8.3
III	10	37.0	2	22.2	12	33.3
IV	6	22.2	0	0	6	16.7
Unknown	0	0	1	11.1	1	2.8
Pathology
Adenocarcinoma	2	7.4	1	11.1	3	8.3
Clear Cell	2	7.4	0	0	2	5.6
Grade 1 Endometrioid Adenocarcinoma	6	22.2	1	11.1	7	19.4
Grade 2 Endometrioid	3	11.1	4	44.4	7	19.4
Grade 3 Endometrioid	2	7.4	0	0	2	7.4
Mixed Epithelial	5	18.6	2	22.2	7	19.4
Mixed Serous/Endometrioid	1	3.7	0	0	1	2.8
Serous Adenocarcinoma	6	22.2	1	11.1	7	19.4
Dose
200mg	14	51.9	4	44.4	18	50.0
135mg	13	48.1	5	55.6	18	50.0

Ninety two percent of patients experienced at least one drug-related adverse event. Most commonly observed toxicities were rash (44%), fatigue (42%), nausea (42%), hyperglycemia (31%), diarrhea (25%), fever (19%), and vomiting (17%) (Table 2). Overall, rates of grade 3 and 4 toxicities were 25% and 17%, respectively. Rates of grade 3 rash were 39% (90% CI 20%−61%) and 0% (90% CI 0%−15%) in patients on the 200mg and 135mg dose, respectively, p=0.008). Among the 18 patients enrolled at the 200mg QW dose level, 5 required hospitalization for adverse events of rash. Patients typically presented after the second dose of the MK-2206 with an erythematous, pruritic rash involving the trunk, extremities and face (Figure S1) and systemic symptoms, including fevers, tachycardia and hypotension. All patients recovered to pre-study baseline with supportive care including fluid, steroids, anti-pyretics, and anti-histamines. All were attributed to MK-2206 treatment and all patients were discontinued from the study. Based on these events, the protocol was amended to the reduced starting dose. Eight (22%) patients required a dose reduction and 5 (14%) required dose delay due to toxicity at the 200mg QW level. Of these 13 patients, 5 (39%) had a prior history of whole pelvic radiotherapy.

Table 2-.

Toxicities experienced by patients during treatment with MK-2206 at both dose levels

		200 mg (N=18)				135 mg (N=18)				All patients (N=36)
		All grades		Grade ¾		All grades		Grade ¾		All grades		Grade ¾
		N	%	N	%	N	%	N	%	N	%	N	%
Category of Toxicity	Toxicity	2	11.1	-	-	3	16.7	-	-	5	13.9	-	-
Blood and lymphatic system disorders	Anemia
	Leukocytosis	-	-	-	-	2	11.1	1	5.6	2	5.6	1	2.8
Cardiac disorders	Atrial flutter	-	-	-	-	1	5.6	1	5.6	1	2.8	1	2.8
	Atrioventricular block first degree	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Supraventricular tachycardia	-	-	-	-	1	5.6	1	5.6	1	2.8	1	2.8
Eye disorders	Blurred vision	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Conjunctivitis	1	5.6	-	-	-	-	-	-	1	2.8	-	-
Gastrointestinal disorders	Bloating	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Colonic perforation	-	-	-	-	1	5.6	1	5.6	1	2.8	1	2.8
	Constipation	1	5.6	-	-	3	16.7	-	-	4	11.1	-	-
	Diarrhea	5	27.8	-	-	4	22.2	-	-	9	25.0	-	-
	Dry mouth	1	5.6	-	-	1	5.6	-	-	2	5.6	-	-
	Enterocolitis	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Gastroesophageal reflux disease	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Mucositis oral	2	11.1	-	-	-	-	-	-	2	5.6	-	-
	Nausea	9	50.0	-	-	6	33.3	-	-	15	41.7	-	-
	Rectal pain	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Vomiting	2	11.1	-	-	4	22.2	-	-	6	16.7	-	-
General disorders and administration site conditions	Chills	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Edema face	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Edema limbs	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Edema trunk	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Fatigue	9	50.0	1	5.6	6	33.3	-	-	15	41.7	1	2.8
	Fever	5	27.8	1	5.6	2	11.1	-	-	7	19.4	1	2.8
	Non-cardiac chest pain	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Immune system disorders	Allergic reaction	2	11.1	2	11.1	-	-	-	-	2	5.6	2	5.6
Infections and infestations	Kidney infection	-	-	-	-	1	5.6	1	5.6	1	2.8	1	2.8
	Lung infection	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Nail infection	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Skin infection	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Urinary tract infection	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Injury, poisoning and procedural complications	Kidney anastomotic leak	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Investigations	Alanine aminotransferase increased	2	11.1	-	-	-	-	-	-	2	5.6	-	-
	Alkaline phosphatase increased	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Aspartate aminotransferase increased	3	16.7	1	5.6	-	-	-	-	3	8.3	1	2.8
	Creatinine increased	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	INR increased	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Lymphocyte count decreased	2	11.1	2	11.1	-	-	-	-	2	5.6	2	5.6
	Neutrophil count decreased	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Platelet count decreased	1	5.6	-	-	1	5.6	-	-	2	5.6	-	-
	Weight loss	2	11.1	-	-	-	-	-	-	2	5.6	-	-
	White blood cell decreased	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Metabolism and nutrition disorders	Anorexia	3	16.7	-	-	1	5.6	-	-	4	11.1	-	-
	Hypercalcemia	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Hyperglycemia	6	33.3	2	11.1	5	27.8	-	-	11	30.6	2	5.6
	Hypoalbuminemia	-	-	-	-	2	11.1	-	-	2	5.6	-	-
	Hypomagnesemia	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Hyponatremia	-	-	-	-	2	11.1	-	-	2	5.6	-	-
Musculoskeletal and connective tissue disorders	Bone pain	-	-	-	-	2	11.1	-	-	2	5.6	-	-
	Generalized muscle weakness	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Musculoskeletal and connective tissue disorder - Other, specify	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Pain in extremity	-	-	-	-	2	11.1	-	-	2	5.6	-	-
Nervous system disorders	Headache	1	5.6	-	-	1	5.6	-	-	2	5.6	-	-
Psychiatric disorders	Agitation	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Insomnia	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Renal and urinary disorders	Acute kidney injury	1	5.6	-	-	1	5.6	1	5.6	2	5.6	1	2.8
	Proteinuria	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Renal and urinary disorders - Other, specify	2	11.1	-	-	-	-	-	-	2	5.6	-	-
	Urinary tract pain	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Reproductive system and breast disorders	Vaginal perforation	-	-	-	-	1	5.6	1	5.6	1	2.8	1	2.8
Respiratory, thoracic and mediastinal disorders	Cough	1	5.6	-	-	-	-	-	-	1	2.8	-	-
	Dyspnea	-	-	-	-	2	11.1	-	-	2	5.6	-	-
	Pneumonitis	1	5.6	-	-	-	-	-	-	1	2.8	-	-
Skin and subcutaneous tissue disorders	Alopecia	2	11.1	-	-	1	5.6	-	-	3	8.3	-	-
	Dry skin	2	11.1	-	-	1	5.6	-	-	3	8.3	-	-
	Erythema multiforme	2	11.1	2	11.1	-	-	-	-	2	5.6	2	5.6
	Periorbital edema	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Pruritus	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Rash acneiform	-	-	-	-	1	5.6	-	-	1	2.8	-	-
	Rash maculo-papular	12	66.7	7	38.9	4	22.2	-	-	16	44.4	7	19.4
	Skin and subcutaneous tissue disorders - Other, specify	-	-	-	-	1	5.6	-	-	1	2.8	-	-
Vascular disorders	Hypotension	1	5.6	1	5.6	-	-	-	-	1	2.8	1	2.8
	Thromboembolic event	-	-	-	-	4	22.2	2	11.1	4	11.1	2	5.6

Median cycle length was 2 (range 0,11). Within the entire cohort, there were 2 patients with partial response (PR, 6%, 90%CI 1–16%) and 7 who survived progression free for at least 6 months (19%, 90%CI 9–33%). Among the patients with PI3KCA MT endometrial tumors, there was one patient with PR (11%) and three who survived 6 months progression free (33%, 90%CI 10–66%). Of the patients with PI3KCA WT tumors, there was 1 PR (4%) and four who survived 6 months progression free (15%, 90%CI 5–31%). Of note, five of seven patients progression free at 6 months were of serous histology (Table S1).

The median PFS for all patients was 2.0 months. Median PFS for patients with and without PIK3CA mutation was 1.7 months and 2.5 months, respectively (Figure 2A). Median OS for all patients was 8.4 months. The median OS for patients with and without PIK3CA mutation was 8.4 months and 11.1 months, respectively (Figure 2B).

Figure 2:

Progression free and overall survival by PIK3CA mutation status

Molecular characteristics are presented in Table S2. Overall, the proportion of endometrial tumors with a molecular aberration in the PI3K pathway was consistent with the literature^{4, 6}. Figure 3A demonstrates time on MK-2206 treatment based on PI3KCA mutation status, there was no difference between the two groups (PI3KCA WT 1.8 versus PI3KCA MT 1.6 months, p=0.73). Assessment of time on treatment based on additional molecular aberrations, including PI3KCA mutation by NGS, PTEN mutation, KRAS mutation, PTEN expression, and Stathmin expression was performed (Figures 3B – 3F, Figure S2). Indeterminate PTEN protein expression was associated with significantly shorter time on treatment with MK-2206 (Table 3).

Figure 3:

Time on treatment based on molecular aberrations. Off treatment reasons are provided at the end of each plot (P, progressive disease); T, unacceptable toxicity; W, patient withdrawal; C, patient continue on treatment)

Table 3.

Time to treatment failure by molecular phenotypes

Molecular	Phenotype	Median Time-to-Event (Months)	95% LCL	95% UCL	Log-Rank P-value
PIK3CA per protocol	WT	1.8	1.3	4.4	0.7347
	MT	1.6	0.3	7.4
PIK3CA NGS	WT	1.7	0.9	2.7	0.2105
	MT	1.9	0.3	7.4
PTEN IHC expression	Positive	2.2	0.3	7.4	0.0403
	Heterogeneous	1.5	0.3	1.8
	Negative	2.0	0.3	4.8
PTEN	WT	1.6	0.3	2.7	0.2939
	MT	1.8	1.2	4.8
STMN expression	Yes	1.8	0.3	7.4	0.1182
	No	1.7	0.6	3.8
	Other	2.1	1.6	2.7
AKT mutation	WT	1.8	1.5	3.8	*
	MT	0.3	.	.
KRAS mutation	Yes	1.8	1.5	4.4	0.2807
	No	1.7	0.3	3.8

^*Sample sizes are too small for inference (n ≤ 3); logrank tests would require pooling across phenotypes. LCL, lower confidence limit; UCL upper confidence limit

Discussion

This phase II, biomarker stratified study of MK-2206 demonstrated limited efficacy in recurrent endometrial cancer, regardless of PI3KCA mutation status. Interestingly, modest clinical benefit was found in patients with serous histology, warranting further study. These findings are consistent with studies of single agent MK-2206 in solid tumors and lymphoma, where response rates range from 0 – 14% ^{18, 19, 24, 25}. In the current study, 13 (36%) of patients were unable to tolerate the original dose level of 200mg, requiring dose delay, dose reduction or removal from study. A similar result was seen in the single agent study of apitolisib, a dual PI3K/mTor inhibitor, in recurrent endometrial cancer reported by Makker and colleagues. In that study, 34% of patients discontinued the study for toxicity before the first imaging evaluation and an additional 40% required a dose reduction ²⁶.

Single agents targeting the PI3K/AKT pathway have, in general, had only modest success in advanced and recurrent endometrial cancer. The response rate seen in the current study is consistent with reports of single agent mTORC1 and PI3K inhibitors in endometrial cancer ^{7–14, 26, 27}. Only a recent report of temsirolimus achieved a more promising response rate across an unselected group of patients with recurrent endometrial cancer⁸. Low levels of objective response have led to exploration of dual-target agents as well as combination trials of relevant pathway targets.

The high number of toxicities experienced by patients treated with MK-2206 on this trial was surprising. Specifically, the occurrence of fever in the setting of grade 3 rash appears to be unique to the endometrial cancer population. In the phase I study of MK-2206, DLTs were experienced in 4 of 17 patients treated at 200mg/week (G3 rash in 3 patients, G3 dermatitis acneiform in 1 patient). None of these patients were described to have fever or require hospital admission ¹⁹. Overall, there appears to be a higher level of toxicity among patients with endometrial cancer compared to other solid tumors. This finding has been observed in other studies in recurrent endometrial cancer, where previously identified recommended phase II doses had to be reduced due to patient tolerance^{26, 27}. To date, the etiology of the increased toxicity has not been elucidated. Certainly, it could be due to the overall functional status of the trial participants, although patients on this trial were required to have a performance status of 0 or 1. Further, many patients with recurrent endometrial cancer have been treated with whole pelvic radiotherapy, which may predispose them to increased toxicity with future treatments. The high number of molecular aberrations in the PI3K pathway among endometrial tumors could lead to increased sensitivity to agents targeting this pathway, yielding more adverse events. These theories are only speculation and would be worthwhile to study in depth across the clinical trials employing PI3K-targeting agents in endometrial cancer.

When evaluating other potential factors associated with response, the majority of patients with evidence of clinical benefit were of serous histology. Uterine serous tumors have abnormalities in the PI3K/AKT pathway, although less frequently than endometrioid histology. Serous cancers have previously achieved modest benefit when treated with mTORC1 inhibitors ¹². Given the promising finding of 62.5% clinical benefit among patients with recurrent uterine serous tumors, single agent MK-2206 was thought worthy of further exploration in this disease with limited therapeutic alternatives. An expansion of the trial in uterine serous tumors will be reported in the future.

This trial was initially designed with the hypothesis that the presence of a PI3KCA mutation in an endometrial tumor would be associated with improved response to MK-2206. PI3KCA mutations can lead to constitutive activation of the PI3K/AKT pathway, thus, inhibition of AKT is reasonable option for targeted therapy³. In addition, endometrial cancer has the largest number of aberrations in the PI3K/AKT pathway of any solid tumor including frequent mutations in more than one pathway member⁴. Disappointingly, there was little activity in this patient population, regardless of PI3KCA status. Further, none of the molecular abnormalities assessed correlated to the time patients remained on treatment with MK-2206. Certainly, the molecular aberrations studied in this trial are not exhaustive. Additional markers of interest, such as phosphorylated Akt or Ki67, would be interesting to explore in future studies of agents targeting the PI3K/AKT pathway.

To date, attempts to identify molecular abnormalities associated with response to PI3K pathway-directed single agents in endometrial cancer have been unsuccessful. Translational components of phase II studies of mTORC1 inhibitors have yielded little in terms of predictive markers for response. Mackay and colleagues utilized mutation analysis and immunohistochemistry for PTEN and Stathmin to predict response in a series of phase II trials investigating ridaforolimus and temsirolimus in endometrial cancer. None of these markers was associated with response to therapy, although the authors note this may have been limited by the use of primary tumor samples for study²⁸. In a study of everolimus, loss of PTEN expression or KRAS mutation was not correlated to clinical benefit. However, a KRAS mutation in the presence of phosphorylation of s6 protein was able to predict nonresponse to everolimus with 100% specificity²⁹. Similarly, Tredan and colleagues found that PI3K-pathway related proteins were not associated with response to everolimus. However, no patient with a KRAS mutation responded to single agent mTOR therapy³⁰. In the current study, only one patient of 8 with a KRAS mutation garnered clinical benefit (PFS > 6 months) from MK-2206. These results suggest that further exploration of KRAS mutation as a predictive marker is warranted.

Interestingly, in a study of gynecologic and breast cancers in a phase I clinic, Janku and colleagues demonstrated that patients with PI3KCA mutation treated with a PI3K-directed therapy had improved response rates compared to those patients with wildtype PI3KCA. It is important to note that many of these patients were not treated on a single agent trial but rather trials of a PI3K pathway agent in combination with chemotherapy or other targeted therapies³¹. There was a trend toward improved response among patients treated with combination therapy. It may be that combinations are required to achieve true benefit among patients with molecular aberrations in the PI3K pathway.

In summary, the orally bioavailable AKT inhibitor, MK-2206, has limited activity in patients with recurrent endometrial cancer. There was a significant issue with tolerability at the recommended phase II dose, which may have impacted overall efficacy. A promising proportion of patients with uterine serous cancer gained benefit from MK-2206, prompting a histology-specific expansion that will be reported separately.

Novelty and Impact:

Endometrial cancer has a number of actionable molecular aberrations yielding great interest in targeted therapy given the lack of FDA approved agents. To date, success of single agents targeting PI3K pathway members has been modest. In this biomarker-stratified phase II study of the AKT inhibitor, MK-2206, toxicity was higher than expected at the recommended phase II dose and clinical benefit was limited. However, there was encouraging clinical benefit among serous histology which is being explored.

Abbreviations:

PI3K

phosphoinositide 3-kinase

ORR

objective response rate

6moPFS

6 month progression free survival

MT

mutant

WT

wildtype

PR

partial response

AACR

American Association of Cancer Research

RECIST

Response Evaluation Criteria in Solid Tumors

ECOG

Eastern Cooperative Oncology Group

ANC

absolute neutrophil count

AST

aspartate aminotransferase

ALT

alanine aminotransferase

ULN

upper limit of normal

QW

every week

CTCAE

Common Terminology Criteria for Adverse Events

IHC

immunohistochemistry

CR

complete response

CI

confidence interval