PARP inhibitors and immunotherapy in ovarian and endometrial cancers

Abstract

Advanced ovarian and endometrial cancers have historically been associated with poor prognosis and few treatment options, limited to single or doublet chemotherapy regimens. The introduction of novel target therapies has transformed the management of these cancers. In contrast to chemotherapy, which inhibits DNA replication and mitosis, targeted therapies target cancer signalling pathways, stroma, immune-microenvironment and vasculature in tumour tissues. The most notable advances in gynaecological cancers have come from the introduction of PARP inhibitors and immune checkpoint inhibitors for ovarian and endometrial cancer, respectively. Several PARP inhibitors, which target defective DNA repair, have been approved as maintenance therapy for advanced ovarian cancer in both the first line and platinum-sensitive relapsed settings. Immune checkpoint inhibitors such as anti-PD-1/PD-L1 antibodies have proven successful in advanced mismatch repair deficient endometrial cancers with use now being investigated beyond this population. This review will explore the biological rationale and clinical evidence behind the use of PARP inhibitors and immunotherapy in ovarian and endometrial cancers.

Introduction

Despite advances in cancer treatment and the introduction of new therapies with novel mechanisms of action, recurrent or metastatic ovarian and endometrial cancers are almost universally fatal with more than 6300 deaths in the UK per annum.¹ Unlike chemotherapy, which non-specifically attacks proliferating cells, targeted therapy is designed to exploit a specific vulnerability within the cancer cell. Targeted therapy works by interfering with a single molecule to manipulate signalling networks. These signalling networks control critical homeostatic regulatory processes which are key drivers of oncogenesis.² Generally speaking, targeted therapies are divided into two categories: small molecules and monoclonal antibodies. Small molecules enter the cell cytoplasm and act on targets such as tyrosine kinases (e.g. PI3K/AKT/mTOR pathways) and DNA repair mechanisms, with poly(ADP-ribose) polymerase (PARP) inhibitors the most established of these in gynaecological cancers. The second are monoclonal antibodies, which bind to ligands and receptors on either the tumour or host cell surface and do not penetrate the cell. An example of this are immune checkpoint inhibitors (ICIs), such as anti-programmed cell death-1 (PD-1) or anti-programmed cell death-ligand 1 (PD-L1) antibodies, which bind to PD-1 or PD-L1 on host T-cells or tumour cells, respectively, and facilitate T-cell-mediated tumour killing. In recent years, the use of immune checkpoint inhibitors has become increasingly common in cancer treatment and has greatly improved outcomes in settings such as lung cancer and melanoma. In principal, these drugs restore or stimulate the body’s own immune system to seek out and destroy cancer cells by means of altering or blocking the mechanisms by which the tumour cells ordinarily evade such destruction. Immune checkpoint inhibitors have proven efficacy in mismatch repair deficient (MMRd) solid tumours³ but there appears to be a variable response to immune checkpoint inhibitors across different molecular subtypes.⁴

This review will explore the development and future directions of PARP and immune checkpoint inhibitors use in ovarian and endometrial cancers.

PARP inhibitors

The PARP family of proteins play critical roles in DNA repair through multiple DNA damage response pathways, with homologous recombination repair deficient (HRD) cells showing a greater reliance on PARP activity to maintain cell survival.⁵ The finding that single-agent PARP inhibition selectively killed BRCA1/2 deficient cells was a key discovery in exploiting synthetic lethal approaches in oncology.^6,7 PARP inhibitors trap PARP protein onto DNA at sites of single-strand DNA breaks. When this trapped PARP is encountered by the DNA replication machinery, it leads to stalling of the replication fork, collapse and the generation of a double-strand break, which cannot be repaired in cells with HRD such as BRCA1/2 mutated cells.⁸

Mutations in BRCA1 and BRCA2 genes are responsible for the majority of inherited ovarian cancers. Inherited (germline) mutations in BRCA1/2 (gBRCA1/2) account for approximately 13–15% of all high-grade serous and endometroid ovarian cancers (HGOC), with a further 5–7% of patients harbouring somatic BRCA1/2 mutations that have arisen in the tumour during cancer development or progression.^9,10 In addition to BRCA1/2 deficient HGOC, approximately 50% of sporadic HGOC are also defective in HRD, via a variety of mechanisms and share the phenotype associated with BRCA1/2 loss.^9,11,12

PARP inhibitors in ovarian cancer

Recurrent disease

The introduction of PARP inhibitors has transformed the management of HGOC. Initially, PARP inhibitors were only licensed for BRCA1/2 mutated cancers.^13,14 Subsequent data quickly revealed that PARP inhibitor benefit extended to a much broader group of HGOCs and three PARP inhibitors; niraparib, rucaparib and olaparib are now approved by European Medicines Agency (EMA) and US Food and Drug Agency (FDA) regulatory agencies for use as maintenance therapy for all patients with platinum-sensitive relapsed ovarian cancer, regardless of BRCA1/2 status.^13–18

Study 19 was the first randomised phase 2 trial to explore olaparib maintenance therapy in patients with recurrent HGOC who had responded to platinum chemotherapy. This trial explored the concept of using maintenance therapy to improve clinical benefit, determined by prolongation of progression-free survival (PFS). 265 patients were randomised to receive olaparib or placebo until disease progression, BRCA1/2 mutation status was not required for trial inclusion but was determined retrospectively.¹⁹ Overall, median PFS was increased from 4.8 to 8.4 months (HR 0.35 95% CI 0.25–0.49; p < 0.001).¹⁹ Subsequent analysis by BRCA1/2 status revealed a greater benefit in the BRCA1/2 mutant group (HR 0.18; 95% CI 0·10–0·31), although there was still a significant benefit for the BRCA1/2 wild-type group (HR 0·54; 95% CI 0·34–0·85).²⁰ These results were confirmed in the phase 3 trial, SOLO2, where patients with gBRCA1/2 mutant HGOC were randomised following a response to platinum-based chemotherapy to receive olaparib or placebo. There was a significant improvement in PFS observed with olaparib (19.1 vs 5.5 months, HR: 0.30; 0.22–0.41, Table 1).²² A similar benefit for maintenance PARP inhibition in platinum-sensitive, relapsed BRCA1/2 mutated HGOC has been observed with both niraparib (NOVA) and rucaparib (ARIEL3) (Table 1).^21,23

Table 1.

Key trials of PARP inhibitor maintenance therapy and combination therapy in HGOC

Trial (NCT number)	PARP inhibitor	N	Primary outcomes	PFS PARP inhibitor vs placebo (months)	HR (95% CI)
Maintenance therapy in platinum-sensitive recurrent HGOC
Study19 (NCT00753545) Ledermann et al20	Olaparib	265	PFS in ITT and BRCA1/2 status	All patients: 10.8 vs 5.4	0.35 (0.25–0.49)
				BRCA1/2mut: 11.2 vs 4.3	0.18 (0.34–0.85)
				BRCA1/2 wt 7.4 vs 5.5	0.54 (0.34–0.85)
NOVA (NCT01847274) Mirza et al21	Niraparib	553	PFS according to BRCA1/2 and HRD status	gBRCA1/2mut: 21 vs 5.5	0.27 (0.17–0.41)
				gBRCA1/2 wt: 9.3 vs 3.9	0.45 (0.34–0.61)
				HRD & BRCA1/2 wt: 12.9 vs 3.8	0.38 (0.24–0.59)
SOLO2 (NCT01874353) Pujade-Lauraine et al22	Olaparib	295	PFS	BRCA1/2mut: 19.1 vs 5.5	0.33 (0.24–0.44)
ARIEL3 (NCT01968213) Coleman et al23	Rucaparib	564	PFS in ITT, HRD and BRCA1/2 mutant group	All patients: 10.8 vs 5.4	0.36 (0.3–0.45)
				BRCA1/2mut: 16.6 vs 5.4	0.23 (0.16–0.34)
				HRD: 13.6 vs 5.4	0.32 (0.24–0.42)
First-line maintenance therapy
SOLO1 (NCT01844986) Banerjee et al24	Olaparib	391	PFS in ITT population	BRCA1/2mut: 56.0 vs 13.8	0.33 (0.25–0.48)
PAOLA-1 (NCT02477644) Ray-Coquard et al25	Olaparib plus bevacizumab	806	PFS in ITT population	All patients: 22.1 vs 16.6	0.59 (0.49–0.72)
PRIMA (NCT02655016) Gonzalez-Martin et al26	Niraparib	733	PFS in ITT and HRD	All patients: 13.8 vs 8.2	0.62 (0.50–0.76)
				HRD: 21.9 vs 10.4	0.43 (0.31–0.59)
VELIA (NCT0247058) Coleman et al27	Veliparib in combination with chemotherapy and as maintenance therapy	1140	PFS in veliparib throughout group v control group in ITT, BRCA1/2mut and HRD	All patients: 23.5 vs 17.3	0.68 (0.56–0.83)
				BRCA1/2mut: 34.7 vs 22	0.44 (0.28–0.68)
				HRD: 31.9 vs 20.5	0.57 (0.43–0.76)

BRCAmut, mutation in BRCA1 or BRCA2 gene; BRCAwt, BRCA1/2 wild-type; HRD, homologous recombination deficiency; ITT, intention to treat; NR, not reached; g, germline.

Benefit from PARP inhibitor is displayed as progression-free survival (PFS) with corresponding hazard ratios (HR) and 95% confidence intervals (CI).

The NOVA and ARIEL3 trials included cohorts of BRCA1/2 wild-type patients. These patients were further subdivided by HRD status using either the Myriad (NOVA) or Foundation Medicine (ARIEL3) HRD assays.^21,23 Both trials found a benefit for PARP inhibition in all patents, regardless of BRCA1/2 or HRD status but there was an incremental reduction in benefit from BRCA1/2 mutated to HRD to non-HRD (or homologous recombination proficient, HRP) (Table 1).^21,23 Neither HRD assay was able to identify a sub-group of patients who did not benefit (e.g. HRP). On the basis of these results, olaparib, niraparib and rucaparib are now all licensed as maintenance treatment in recurrent HGOC that has responded to platinum-based therapy, regardless of BRCA1/2 status. These drugs have now been adopted as standard of care in recurrent HGOC.

First-line maintenance therapy

Despite impressive prolongation in PFS, the majority of patients with recurrent HGOC will eventually relapse and die from their disease. First-line therapy for advanced HGOC is the only opportunity for cure and recent efforts have focused on whether a greater benefit can be derived from the early introduction of PARP inhibitors following cytoreductive surgery and platinum-based chemotherapy. Initial data were restricted to BRCA1/2 mutant HGOCs. In the SOLO1 trial, 391 patients with newly diagnosed HGOC and a BRCA1/2 mutation who had a complete or partial response (PR) after platinum-based chemotherapy were randomised to receive olaparib or placebo.^24,28 The addition of olaparib maintenance led to an unprecedented improvement with a 70% reduction in the risk of disease progression or death compared to placebo (60 vs 27% h 0.30, 95% CI 0.23–0.41) with a median PFS of 56.0 months with olaparib versus 13.8 months on placebo (HR 0.33, 95% CI 0.25–0.48)²⁸ leading to FDA (2018) and EMA²⁷ approval in this indication.^16,29

Three subsequent first-line randomised phase 3 trials have explored the role of maintenance PARP inhibitors in BRCA1/2 wild-type patients, in addition to BRCA1/2 mutation populations: PAOLA-1, PRIMA and VELIA (Table 1).^25–27 In each trial, the presence of a BRCA1/2 mutation consistently predicted a benefit to PARP inhibitors with a similar degree of benefit to that seen in the relapsed setting (HR range 0.30–0.44) although the duration of benefit observed was longer.^25–27 BRCA1/2 wild-type, HRD tumours also appear to benefit although to a lesser degree than the BRCA1/2 mutant population.^25–27 For example; the PRIMA study compared niraparib and placebo with patients stratified by HRD-score (Myriad). BRCA1/2 wild-type/HRD patients benefited from niraparib with a median PFS increase from 8.2 to 19.6 months (HR; 0.5, 95% CI 0.31–0.83). The trial was not powered to detect benefit in the HRP subgroup although exploratory analyses indicate some benefit, albeit of a lesser magnitude (HR 0.68; 95% CI 0.49–0.94).²⁶ Similarly, in the PAOLA1 study, where olaparib (or placebo) was added to maintenance therapy with the anti-angiogenic drug, bevacizumab, a benefit from olaparib was observed with the BRCA1/2 wild-type/HRD tumours (PFS increase from 16.6 to 28.1 months, HR 0.43; 95% CI 0.28–0.66) but not the BRCA1/2 wild-type/HRP tumours (16.0 to 16.9 months, HR 0.92; 95% CI 0.72–1.17).²⁵ Based on these results, the FDA and EMA recently approved maintenance therapy with olaparib and bevacizumab for HRD HGOC and niraparib for all HGOCs following first-line platinum-based chemotherapy.^15,16,30,31 A different approach was adopted in the VELIA study where the PARP inhibitor veliparib was added either concurrently with chemotherapy, as maintenance therapy or both.²⁷ No benefit was observed with concurrent veliparib use only. As with the other trials, a benefit from maintenance therapy was observed within the BRCA1/2 mutant population. Less benefit in the BRCA1/2 wild-type tumours was observed from the addition of veliparib given with chemotherapy and as maintenance therapy, whether HRD (HR 0.80; 95% CI 0.64–0.997) or HRP (HR; 0.81; 95% CI 0.6–1.09). The discrepancy between VELIA and the other two first-line studies may be explained by an unvalidated HRD cut-off score being used.²⁷

In the first-line setting, maintenance PARP inhibition following chemotherapy leads to an unprecedented improvement in outcome as determined by PFS, especially within the BRCA1/2 mutant population. Overall survival data are immature, but it is hoped the impressive PFS will translate into a long-term improvement in survival and may even translate into cure in some patients. It is clear that this benefit is extended to the BRCA1/2 wild-type/HRD patients, although to a lesser magnitude. The role of PARP inhibitors for HRP patients in the first-line setting is less clear cut with conflicting results observed between PRIMA and PAOLA1/VELIA. Whilst BRCA1/2 testing (somatic and/or germline) is routine for all patients with a new diagnosis of HGOC, HRD testing is not. To date, a perfect HRD assay does not exist and the commercial ones available consistently fail to identify a group of patients who do not benefit. Although an in-depth review on different means of HRD testing is outside the scope of this review (reviewed extensively in³²), needless to say urgent efforts are required to optimise this test in order to maximise patient benefit from PARP inhibitors.

Immune checkpoint inhibitors in endometrial cancer

Most endometrial cancer presents at an early stage and is successfully treated with surgery alone. Identifying patients at risk of relapse has traditionally used a combination of stage, histopathological features (histology, depth of myometrial invasion, the presence of lymphovascular space invasion) and age. The Cancer Genome Atlas published a putative reclassification in 2013 based on genomic and molecular analysis of endometrial cancer.³³ Four groups were defined according to mutational load, somatic copy number alterations and specific driver mutations. These were polymerase-Ɛ (POLE)-mutated, microsatellite unstable cancers driven by MMR deficiency (MMRd), copy number low and copy number high somatic mutations (with frequent TP53 mutations). The first three groups were predominately endometrioid and the latter mainly serous histology.³³ Each group was associated with a different prognosis. POLE-mutated cancers, even when high grade, are associated with an excellent prognosis. MMRd cancers have an intermediate prognosis and the somatic copy number high group, which includes serous and high-grade endometrioid cancers, have the poorest prognosis. The group with no specific molecular profile was the largest, mostly comprising endometrioid cancers with high progesterone receptor expression and also with an intermediate prognosis.³³ This new classification not only better defines prognosis according to the molecular and genetic profile but also provides the potential for appropriately targeted therapy.

POLE-mutated and MMRd tumours are associated with high levels of tumour infiltrating lymphocytes and have a high tumour mutational burden and increased expression of immune checkpoint genes. This makes them a potential target for immune checkpoint inhibitors.^3,34

Immune checkpoint inhibitors in advanced or recurrent/relapsed endometrial cancer

In a small preliminary study looking at the efficacy and safety of Pembrolizumab, a humanised monoclonal antibody targeting PD-1, in MMRd tumours, 8 out of 15 patients (53%) with endometrial cancer showed some response, with three patients having a complete response (CR) and five having a PR.³⁵ An additional three patients had stable disease. The KEYNOTE-028 study looked at the efficacy of pembrolizumab in PD-L1-positive patients with advanced or metastatic endometrial cancer. Three patients (3/24, 13%) had a confirmed PR (95% CI, 2.8 to 33.6%), and a further three (3/24, 13%) stable disease.³⁶ On the basis of these results and other studies examining pembrolizumab in MMRd tumours, the FDA granted a tumour agnostic approval for the use of pembrozulimab in any MMRd tumours that has progressed following standard treatment.³⁷

Atezolizumab, an anti-PD-L1 antibody is widely in use in other solid tumours such as lung and bladder cancer. It has been investigated in a phase I study of recurrent endometrial cancer, with an overall response rate of 13% (2/15). Both patients had a PR. Interestingly, one patient was MMRd, whereas the other was MMR proficient.³⁸ This suggests that MMR status is potentially not critical in conveying response to immunotherapy. Response, however, did seem to increase with higher rates of PD-L1 expression.³⁸ The international phase 3, placebo-controlled, ATTEND study investigating the efficacy of atezolizumab versus placebo in combination with carboplatin and paclitaxel chemotherapy is currently ongoing (NCT03603184).

Durvalumab, an anti-PD-L1 antibody, has been shown to be efficacious in patients with advanced endometrial cancer. In the MMRd group, the overall response rate was 40% (14/35, 95% CI 26–56), with four CR and ten PRs. Only one of the observed responses (PR) was in the MMR proficient group (3%, 95% CI 1–14).³⁹

Immune checkpoint inhibitors as combination therapy

Enhancing the effect of immunotherapy by combining immune checkpoint inhibitors with other systemic agents such as chemotherapy, antiangiogenic drugs or other immunotherapy agents is an area of great interest. Currently under investigation is the combination of pembrolizumab with the oral antiangiogenic tyrosine kinase inhibitor, lenvatinib. The underlying hypothesis is that lenvatinib may prevent VEGF-mediated immune suppression thus improving the efficacy of pembrolizumab. In the initial early-phase trial, a PR was seen in 11 out of 23 patients (48%), (95% CI 27–69).⁴⁰ Following on from this, a single-arm phase 2 study has shown good objective response rates in the same setting. Patients with metastatic endometrial cancer (regardless of MMR status) and progression after no more than two previous lines of treatment were given lenvatinib daily in combination with 3 weekly pembrolizumab. 41 out of 108 (38%) were found to have an objective response at 24 weeks, with toxicity profiles being similar to those found in previous studies, save for an increased rate of hypothyroidism.^41,42 This has led to recent FDA approval for the use of pembrolizumab plus lenvatinib for the treatment of patients with non-MSI-H or MMRd advanced endometrial cancer with disease progression after prior systemic therapy but no available curative treatment options.⁴³ This treatment combination is now under further investigation in a phase 3 trial (NCT03517449). A further study exploring the combination of immune checkpoint inhibition with anti-VEGF therapy (Atezolizumab with the anti-VEGF antibody bevacizumab) is also currently under investigation (NCT03526432).

A potential way for tumours to evade recognition by the immune system is thought to be the upregulation of the indoleamine 2,3-deoxygenase 1 (IDO1) enzyme. Inhibitors of this enzyme, such as epacadostat, have been combined with pembrolizumab in the phase 1 setting to investigate a possible additive effect. Analysis of the results showed that two patients (2/7, 29%) with endometrial cancer had a response, with one having a CR and one PR.⁴⁴ Further studies are ongoing to investigate the efficacy of immune checkpoint inhibitors with other novel therapies.

Immune Checkpoint inhibitors with radiotherapy

There is great interest in combining radiotherapy with immune checkpoint inhibitors. The abscopal effect is a little understood, but well-described phenomenon where distant metastases show response to radiotherapy at another site. Furthermore, through immune priming radiotherapy may enhance local tumour control at site of radiation. It is postulated that radiotherapy releases tumour neoantigens, which activate cytotoxic T-cells, that in turn migrate to both irradiated and non-irradiated tumours cells to produce immune-mediated tumour cell death.⁴⁵ To date, these responses are anecdotal, for example this has been observed in a patient receiving nivolumab who underwent multiple courses of palliative radiotherapy, with disease responses both within and distant to radiation field.⁴⁶ Further evaluation of the combination of radiotherapy with ICI is required, with trials already underway in the metastatic (NCT03192059) and adjuvant setting (NCT03932409).

Immune checkpoint inhibitors in ovarian cancer

The role of immune checkpoint inhibitors in ovarian cancer is less well established and benefit observed to date has been modest.⁴⁷ For example, in heavily pre-treated patients, single-agent pembrolizumab yielded an overall response rate of only 8.5%⁴⁸ and in the first-line setting, there was no PFS benefit with addition of avelumab to standard of care chemotherapy.⁴⁹ There have been several reports suggesting tumour infiltrating lymphocytes (TILs) have a prognostic role in ovarian cancer and correlate with survival.⁵⁰ However, TILs in ovarian cancer are often functionally exhausted with high-levels of PD-1 expression contributing to the immune-suppressive tumour microenvironment^51–53 Ovarian cancer is classified as a ‘cold tumour’ with low TMB and lack of T-cell infiltration,⁵⁴ which may in part explain the low response rates to monotherapy. As a consequence, it is necessary to identify optimal combination therapies to improve clinical responses.

Immune checkpoint inhibitors and PARP inhibitor combination therapy

Although HRD is most commonly associated with BRCA1/2 mutations, it is clear that numerous other genes play an important role, including ARID1A, which is frequently mutated in endometrial cancer.⁵⁵ HRD is also linked with P53 mutation in serous-like endometrial cancer⁵⁶ and a number of trials evaluating PARP inhibitors in endometrial cancer are underway (e.g. NCT04080284, NCT03570437).

There has been significant interest in the combination of a PARP inhibitor with immune checkpoint inhibitors. The rationale for this combination is based on two hypotheses; firstly, tumours with defective DNA repair, such as HRD tumours have increased tumour mutational burden resulting in higher neo-antigen loads, which stimulates an increased anti-tumour immune response.^57,58 Secondly, PARP inhibition upregulates PD-L1 expression⁵⁹ and in the absence of a functional BRCA pathway, there is activation of the innate immune response via the STING pathway,⁶⁰ which may enhance the PARP inhibitor immunotherapy combination. Several trials evaluating this combination in the adjuvant and relapsed setting are ongoing in both ovarian (NCT03522246, NCT02734004, NCT03737643) and endometrial cancer (e.g. NCT03016338, NCT03694262, NCT03951415) with the results awaited with interest.

Conclusions

The introduction of PARP inhibitor therapy for advanced ovarian cancer and immune checkpoint inhibitors for MMRd advanced endometrial cancers has led to significant improvement in outcomes with meaningful clinical benefit observed. The ability to personalise treatment by targeting specific molecular and genomic profiles within tumours is increasingly been incorporated into standard of care, allowing us to accurately select the most appropriate treatment for each individual. With this approach, we are moving away from a one-size fits all treatment strategy to a more personalised approach, which should improve outcome. Despite significant advances in the management of both ovarian cancer endometrial cancer, there remains subgroups of patients for which there is an unmet need – for example, HRP ovarian cancer or those who have progressed on PARP inhibitor therapy. Similarly, whilst some clinical activity from immune checkpoint inhibitors has been observed in the copy number high group and group with no specific molecular profile endometrial cancer patients this is limited, and efforts are required to identify better targeted therapies for these cohorts.

Table 2.

Key trials of immunotherapy in endometrial cancer

Trial (NCT number)	Immunotherapy	Phase	N	Primary outcomes	ORR 95% CI (where available)	Type of response and subtype
Relapsed/recurrent endometrial cancer
NCT01876511 Le et al35	Pembrolizumab	II	15	Immune related PFS	53%	3/15 (20%) CR (MMRd)
				Objective response rate		5/15 (33%) PR (MMRd)
KEYNOTE-028 (NCT02054806) Ott et al36	Pembrolizumab	I	24	Best overall response	13% (2.8–33.6)	No CR
						3/24 (13%) PR (2MMRp, 1 UK)
NCT01375842 Flemming et al38	Atezolizumab	I	15	Dose limiting toxicities	13%	2/15 (13%) PR (1MMRd, 1 MMRp)
				Maximum tolerated dose
				Recommended Phase two dose
				Adverse event frequency
PHAEDRA (ACTRN12617000106336) Antill et al39	Durvalumab	II	71	Objective response rate	40% (26–56) MMRd	4/35 (11%) CR
						10/35 (29%) PR
					3% (1–14) MMRp
						1/36 (3%) CR
Combination therapy
KEYNOTE-037 (NCT02178722) Mitchell et al44	Pembrolizumab and Epacadostat	I/II	7	Adverse events	29%	1/7 (14%) CR
				Objective response rate		1/7 (14%) PR
						(MMR unknown)
KEYNOTE-146 (NCT02501096) Makker et al40	Pembrolizumab and Lenvatinib	Ib	23	Maximum tolerated dose Objective response rate	48%	11/23 PR (48%)
KEYNOTE-146 (NCT02501096) Makker et al42	Pembrolizumab and Lenvatinib	II	108	Objective response rate	38% (28.8–47.8) All	41/108 (38%)
					63.6% (30.8–89.1) MMRd	1/11 (9%) CR, 6/11 (54.5%) PR
					36.2% (26.6–46.7) MMRp	2/94 (2.1%) CR, 32/94 (34%) PR

CR, Complete response; MMRd, Mismatch repair deficient; MMRp, Mismatch repair proficient; PR, Partial response.

Benefit from immunotherapy is displayed as overall response rate (ORR) with 95% confidence intervals (CI) where available.