Agents targeting the vascular endothelial growth factor pathway

Bevacizumab

Ovarian cancer.

Bevacizumab, a humanized mAb to human VEGF, was the first FDA-approved drug targeting angiogenesis. In ovarian cancer, bevacizumab has been evaluated both as a single agent and in combination therapy for primary and recurrent disease. After encouraging preclinical studies and multiple case studies, the Gynecologic Oncology Group (GOG) instituted a phase II trial of single-agent bevacizumab (15 mg/kg every 3 weeks) in persistent or recurrent refractory epithelial ovarian cancer. Despite a heavily pretreated cohort, the authors reported 13 of 62 patients (21%) experienced a clinical response, including 11 patients with partial response and 2 patients with a complete response. The median number of cycles was seven, and 25 patients (40.3%) had a progression-free survival (PFS) time of at least 6 months. Overall, toxicity was low, and there were no reports of bowel perforation. Another phase II trial of bevacizumab (15 mg/kg every 3 weeks) in 44 patients with ovarian cancer receiving third- or fourth-line chemotherapy demonstrated a partial response rate of 16%. The median PFS time was 4.4 months, and the median overall survival (OS) period was 10.7 months at study closure. This study was terminated early secondary to five patients (11.4%) experiencing spontaneous bowel perforation. The risk of bowel perforation appeared to be higher in patients with a higher median number of prior treatments and in whom impending bowel obstruction was suspected.

In the setting of relapsed disease, bevacizumab was evaluated in combination with both cytotoxic and biologic therapies. A phase II study of bevacizumab (10 mg/kg every 2 weeks) combined with oral cyclophosphamide (50 mg/day) demonstrated a partial response rate of 24% (17 of 70 patients) at a median follow-up period of 23.2 months. The probability of being progression free at 6 months was 56% in this study. Overall toxicity was acceptable in this study. This combination has also been evaluated retrospectively at several institutions with similar encouraging results (objective response rates, 44% to 53.3%). A phase II trial combining bevacizumab (15 mg/kg every 21 days) and oral erlotinib (150 mg/day) in 13 patients with recurrent müllerian cancer showed an objective response rate of 15% and a stable disease (SD) rate of 54%. However, this trial was stopped early secondary to lack of clear benefit of the combination over single-agent bevacizumab and a higher than expected rate of bowel perforation (15%).

Success in the platinum-resistant recurrent setting and in several phase II trials in the upfront setting led to five major phase III trials in ovarian cancer. In the upfront setting, GOG-0218 and International Collaborative Ovarian Neoplasm (ICON) 7 both include bevacizumab in combination with standard cytotoxic drugs followed by maintenance bevacizumab. Key characteristics of these studies are summarized in Table 15.4 . Data from GOG-0218 revealed a PFS benefit of 3 months in the arm that included bevacizumab treatment upfront and continued as single-agent maintenance (14.1 vs. 11.2 months; hazard ratio [HR], 0.72; P < .001). It is interesting that there was no PFS benefit in the patients who received only adjuvant bevacizumab compared with standard therapy alone. No OS benefit was noted in the arms containing bevacizumab, although this was not a primary endpoint of the study. Although different in design, ICON7 reached similar conclusions. In the arm receiving paclitaxel, carboplatin, and bevacizumab followed by bevacizumab maintenance, the PFS time was improved by 1.7 months (24.1 vs. 22.4 months; HR, 0.87; P = .04). OS was similar between the arms. Importantly, an ancillary study of GOG-0218 provided further data to indicate that bevacizumab should be used with caution in the primary setting in patients with history of inflammatory bowel disease or bowel resection.

TABLE 15.4

Comparison of Trial Design of Gynecologic Oncology Group (GOG) 218 and International Collaborative Ovarian Neoplasm (ICON) 7

	GOG 218	ICON 7
Patients (n)	1873	1528
Type	Randomized	Randomized
	Placebo controlled	Open Label
Primary endpoints	Overall survival	Progression-free survival
	Progression-free survival
Secondary endpoints	Toxicity	Overall survival
	Quality of life	Response rate
	Translational research	Biologic progression-free survival
		Toxicity
		Quality of life
		Economics
Strata	Stage (III ≤1 cm vs. >1 cm vs. IV)	Stage (I–III ≤1 cm vs. >1 cm vs. IV)
	PS: (0 vs. 1–2)	Chemotherapy start (≤4 weeks vs. > 4)
		Enrolling center
Sites	490	142
Opened	September 2005	April 2006
Closed	June 2009	February 2009

 

PS, Performance status.

As further targeted therapies and immunotherapies have been developed (discussed later), there has been interest in moving these therapies in combination into the frontline setting. IMAGYN050/GOG-3015 was a randomized control trial to assess the benefit of addition of the PD-1 inhibitor, atezolizumab, to chemotherapy with bevacizumab in newly diagnosed advanced ovarian cancer. There was no improvement in PFS in the group treated with atezolizumab (19.5 vs. 18.4 months, HR 0.92, P = .28). We have yet to determine the appropriate patient population for combination strategies; accordingly, a number of additional trials are planned that combine bevacizumab with poly (ADP-ribose) polymerase (PARP) inhibitors and immune oncology agents in upfront advanced ovarian cancer.

More promising results for bevacizumab have been observed in the phase III trials in the recurrent setting. GOG-0213 is a randomized study evaluating two primary objectives: (1) the effect of the addition of bevacizumab to standard carboplatin and paclitaxel and (2) the role of secondary cytoreduction in patients with platinum-sensitive recurrent ovarian cancer ( Fig. 15.4 ). With respect to the first objective, GOG-0213 demonstrated a significant improvement in PFS as well as a trend toward improved OS in the arm containing bevacizumab. In the OCEANS trial, patients with platinum-sensitive ovarian cancer were treated with gemcitabine and carboplatin with or without bevacizumab. This trial aimed to evaluate PFS and potential GI toxicity of this combination. Similar to GOG-0213, the arm containing bevacizumab demonstrated a significant improvement in PFS of 4 months. However, this did not yield a difference in OS. Finally, the combination of liposomal doxorubicin, carboplatin, and bevacizumab yielded improved PFS when compared with the to the OCEANS regimen of gemcitabine, carboplatin, and bevacizumab (13.3 vs. 11.6 months, HR 0.81, P = .012).

Gynecologic Oncology Group 213 schema.

In the platinum-resistant setting, the AURELIA trial compared bevacizumab with a physician-choice standard agent, including paclitaxel, liposomal doxorubicin, or topotecan, with the standard agent alone. Bevacizumab provided a statistically significant improvement in response rate and PFS when added to standard chemotherapy. Interestingly, when a subanalysis stratified by each chemotherapy cohort was performed, the greatest PFS benefit (6 months) was found in the cohort that combined weekly paclitaxel with bevacizumab. The high crossover to bevacizumab after trial participation likely influenced the lack of OS difference noted in this trial. The results of this trial yielded an FDA approval for bevacizumab in combination with chemotherapy in platinum-resistant ovarian cancer receiving one or two prior regimens.

Bevacizumab has also shown promise in nonepithelial ovarian cancer. A retrospective review of eight patients with recurrent granulosa cell tumors demonstrated a partial response rate of 38% and SD rate of 25%. This study has encouraged the development of a phase II trial by the GOG evaluating bevacizumab for women with recurrent ovarian sex cord–stromal tumors. Among 36 patients treated, 16.7% had a partial response and 77.8% achieved SD. Certainly, these results are promising in this notoriously chemoresistant disease.

Two important recent follow-up studies have addressed other clinical questions with respect to bevacizumab use in the primary treatment setting: first, MITO-16b addressed the hypothesis that prior bevacizumab use would negatively impact subsequent use in platinum-sensitive recurrent disease and, second, AGO-OVAR17/BOOST trial addressed the role that doubling the exposure to bevacizumab in primary management would have on PFS and OS. In the former trial, no detriment was apparent because the HR for subsequent treatment of gemcitabine, carboplatin, and bevacizumab was similar (HR: 0.51, 95% confidence interval [CI]: 0.41 to 0.64) to the similarly designed OCEANS trial. In the latter, doubling bevacizumab from 15 to 30 months of exposure had no effect on PFS or OS.

Uterine cancer.

VEGF expression has been correlated with adverse outcomes in endometrial cancer, and bevacizumab has demonstrated encouraging results in early-phase clinical settings. The GOG initiated a phase II study of single-agent bevacizumab (15 mg/kg every 21 days) for advanced endometrial cancer demonstrating a 13.5% response rate and 40.4% surviving progression free at 6 months. The median OS time was 10.5 months in this trial. A subsequent trial attempted to maximize this benefit by combining bevacizumab with temsirolimus, an mTORC1 inhibitor described later in this chapter. The combination was deemed active based on reasonable response rate (24.5%) and 47% of patients surviving progression free at 6 months; however, its development has been limited by increased toxicity.

Given the activity of bevacizumab for relapsed disease, this agent was added to chemotherapy in two separate randomized phase II trials. In both the GOG-86P study (discussed in detail later) and the MITO-END-2 study, the addition of bevacizumab did not yield an improvement in PFS. Interestingly, the presence of p53 mutation was associated with better PFS and OS in GOG-86P, indicating this may be a potential biomarker to select combination therapy. In the upfront setting, the addition of bevacizumab to intensity-modulated radiation therapy (IMRT) with cisplatin was evaluated in high-risk endometrial cancer. Overall, toxicity was reasonable, and the OS rate for the cohort was 96.7% at 2 years.

Cervical cancer.

Bevacizumab is the first targeted agent that has demonstrated activity in recurrent cervical cancer. A phase II trial of single-agent bevacizumab (15 mg/kg every 21 days) had promising results in a cohort of patients with fewer than three prior regimens, achieving a median OS period of 7.29 months and acceptable toxicity. Five patients (10.9%) had a partial response, and an additional 11 patients were progression free for a minimum of 6 months. Furthermore, an analysis of six patients treated with bevacizumab in combination with 5-fluorouracil or capecitabine demonstrated a clinical benefit rate of 67%. There are several trials with bevacizumab in cervical cancer actively accruing or completed in the upfront and recurrent settings. GOG 240 combined bevacizumab with standard chemotherapy in four regimens, described in Fig. 15.5 . This study found a significant improvement in OS among the arms that received bevacizumab (17.0 vs. 13.3 months, HR, 0.71, 95% CI: P = .004), which led to an FDA approval for bevacizumab in combination with chemotherapy for advanced and recurrent cervical cancer. These successful results have led to exploration of the addition of immune oncology agents (discussed later) to bevacizumab and chemotherapy in this setting.

Gynecologic Oncology Group 240 schema.

Agents targeting vascular endothelial growth factor receptors

AZD2171 (cediranib)

Cediranib is a small molecule inhibitor of VEGFR-2, platelet-derived growth factor receptor (PDGFR), and c-kit; it has shown promise in several phase II trials. In a study of 46 patients with recurrent ovarian cancer, the clinical benefit rate of single-agent cediranib was 30%. Eight patients achieved partial response, and six patients had SD, and median PFS for the group was 5.2 months. Hirte and colleagues reported a response rate of 41% in platinum-sensitive and 29% in platinum-resistant ovarian malignancy. Toxicities in both studies included diarrhea, hypertension, mucositis, fatigue, and anorexia. Cediranib was evaluated in the upfront setting in combination with standard paclitaxel and carboplatin as part of ICON6 ( Fig. 15.6 ). Toxicity of the combination was well tolerated and yielded a PFS benefit of 3 months. Importantly, the use of cediranib in the maintenance setting was not associated with a significant improvement in OS. This combination of paclitaxel and carboplatin with cediranib was studied in cervical cancer as well, yielding 2 months of increased PFS at the expense of increased toxicity. Single-agent cediranib in endometrial cancer had promising activity, with a 12% response rate and 30% of patients progression free at 6 months, leading to the development of ongoing combination trials with multiple targeted agents.

International Collaborative Ovarian Neoplasm 6 schema. Randomization is 2:3:3 in this trial.

Agents targeting multiple vascular endothelial growth factor–related molecules

Pazopanib

Pazopanib inhibits all of the VEGFRs (VEGFR1, VEGFR2, and VEGFR3), PDGFR-α and PDGF-β, and the KIT receptor. This small molecule inhibitor (800 mg/day) has been evaluated in a phase II study of 36 patients with recurrent ovarian cancer by CA 125 and nonbulky disease. This study revealed a 31% response by CA 125 level and a 56% SD rate. Among the 17 patients with measurable disease, 18% had a partial response. Ongoing ovarian cancer studies of pazopanib include combination with liposomal doxorubicin in the recurrent setting and in combination with paclitaxel and carboplatin in the upfront setting. A phase III, placebo-controlled study of pazopanib for consolidation after completion of primary chemotherapy in ovarian cancer yielded a prolonged PFS by 5.6 months when compared with placebo. Adverse events were increased in the pazopanib arm, with 33% of patients terminating this agent early. Unfortunately, pazopanib did not result in benefit in OS in this population. Interestingly, a subset analysis of East Asian women included in this study revealed a negative impact of pazopanib on PFS among this cohort. Further studies are ongoing to understand the mechanism underlying this difference. Pazopanib has also been combined with chemotherapy in the recurrent ovarian cancer setting. One randomized phase II trial comparing weekly paclitaxel with or without pazopanib revealed a 3-month PFS benefit in the setting of moderately increased adverse events, including neutropenia and fatigue. However, another placebo-controlled trial failed to demonstrate a benefit with the addition of pazopanib to weekly paclitaxel in recurrent ovarian cancer. In combination with gemcitabine, pazopanib yielded improved PFS, especially among women with platinum-resistant ovarian cancer. In cervical cancer, pazopanib was explored alone and in combination with lapatinib (a small molecule EGFR inhibitor to be discussed later) in the treatment of advanced and recurrent disease ( Fig. 15.7 ). The combination pazopanib and lapatinib arm was closed early after a futility analysis, leaving the randomized phase II trial to compare pazopanib with single-agent lapatinib. Both PFS (HR, 0.66; 90% CI 0.48 to 0.91) and OS (HR, 0.67; 90% CI 0.49 to 0.99) were superior in the monotherapy pazopanib arm. Median OS was 50.7 weeks versus 39.1 weeks for pazopanib and lapatinib, respectively. Pazopanib has also been evaluated as a treatment for metastatic soft tissue sarcoma. In the PALETTE study, pazopanib yielded improved PFS and OS as compared with placebo. This yielded an FDA approval in this indication.

Schema of trial of pazopanib and lapatinib for advanced cervical cancer.

Vascular disrupting agents

This broad group of antiangiogenic drugs acts to occlude preexisting vasculature in the tumor rather than prevent neovascularization. The disturbance of existing vessels leads to ischemia, hemorrhagic necrosis, and ultimately cellular death. Of note, these agents are able to selectively target tumor blood vessels by taking advantage of the differences between normal and tumor endothelial cells. Two major types of vascular disrupting agents exist: small molecule based and ligand based. The majority of vascular disrupting agents under evaluation in gynecologic cancers target small molecules.

Agents targeting mammalian target of rapamycin

mTOR is a key downstream protein of the PI3K/AKT pathway, consisting of two major complexes (mTORC1 and mTORC2). Initial attempts at targeting this pathway node were directed at mTORC1 (rapalogs). Overall, trials in gynecologic cancer with rapalogs have demonstrated only modest success. Thus there are newer agents in development that compete with ATP on the mTOR catalytic site and inhibit mTORC1 and mTORC2.

Agents targeting AKT

Given the modest activity of targeting downstream regulators of the PI3K/AKT pathway, there is interest in targeting higher level nodes of the pathway. There are several AKT-inhibiting agents in development, and the majority are undergoing evaluation in phase I trials. MK-2206 is a highly selective non-ATP competitive allosteric Akt inhibitor that is equally potent against Akt1 and Akt2 and demonstrated efficacy in vitro and in vivo in several tumor models. A phase I dose escalation study in patients with solid tumors has been completed to identify tolerance of the compound at 60 mg orally every other day. Dose-limiting toxicities were skin rash, mucosal inflammation, and hyperglycemia. Three ovarian cancer patients treated in this study had reducing CA 125 values. A subsequent study of MK-2206 in molecularly selected patients (PTEN protein loss, PIK3CA or AKT mutation) was terminated early due to slow accrual. Best response was SD in four of five patients evaluable for response. Similarly, a single-agent study of MK-2206 in recurrent endometrial cancer stratified by presence of PI3KCA mutation had only modest response rates (6%) and little overall clinical benefit. Importantly, single-agent AKT inhibitors seem to work best in patients with a somatic AKT mutation. In a large phase I study of the AKT inhibitor capivasertib, Hyman and colleagues found a 17% response rate in solid tumors harboring this aberration.

Agents targeting PI3K

There are a number of PI3K inhibitors that are under exploration in gynecologic cancers. Pilaralisib is a highly selective oral inhibitor of PI3K and successfully inhibits tumor growth in vivo. A phase I trial revealed acceptable toxicity and durable clinical benefit. A phase I trial combining pilaralisib with carboplatin and paclitaxel incorporated a dose expansion for ovarian and endometrial cancer patients secondary to favorable responses in those tumor types. Unfortunately, there was no improvement in efficacy with the addition of the PI3K inhibitor to chemotherapy. In recurrent endometrial cancer, pilaralisib as a single agent yielded only a 6% response, and the rate of PFS at 6 months was only 12%.

Copanlisib is a selective pan-class inhibitor of PI3K that demonstrated activity in PI3KCA mutant tumor models. A trial exploring copanlisib as a single agent in endometrial cancer harboring PI3KCA mutations was stopped early due to no observed objective response. This agent is under further exploration in combinations with other relevant targeted agents in ovarian and gynecologic cancers. The PI3K inhibitor LY3023414 demonstrated more promising activity in endometrial cancer with activation of the PI3K pathway. Out of 25 evaluable patients, there was a 16% response rate and a 28% clinical benefit rate. Importantly, the responses were durable with two responses lasting for more than 9 months.

Alpelisib (formerly BYL719) is an alpha-specific PI3K inhibitor that is currently approved in combination with fulvestrant for hormone receptor–positive, HER2-negative, PI3K-mutated breast cancer. There has been renewed interest in studying this agent in combination with PARPi because it has been found that PI3K inhibition can resensitize cells with homologous recombination proficiency, described later. A phase III trial is underway (NCT04729387) versus physician-choice chemotherapy in BRCA -wildtype recurrent platinum-resistant ovarian cancer.

Enzastaurin is an oral multikinase inhibitor that primarily acts to suppress tumor growth through the inhibition of PI3K. A phase I study of enzastaurin combined with bevacizumab revealed promising responses in advanced solid tumors, especially among ovarian cancer (29% partial response or complete response). Furthermore, 51% of 21 ovarian cancer patients remained in the study for greater than 6 months. A phase II study of enzastaurin alone in recurrent ovarian cancer had limited activity, with a response rate of only 7.4%. These results led to the development of several phase II trials of enzastaurin in combination with standard chemotherapy for advanced ovarian cancer. Vergote and colleagues reported that the combination was well tolerated; however, it did not yield a statistically significant improvement in PFS.

Combination agents

In addition to the agents targeting specific aspects of the PI3K/AKT pathway noted previously, there are a large number of combination drugs in development and in early clinical trials. These include combination PI3K/mTOR, PI3K/AKT, and PI3K/MEK inhibitors. Given the extensive cross-talk and feedback loops involved in this pathway, the inhibition of two major nodes in one or more pathways is rational and may lead to improved outcomes. Unfortunately, early reports from single-agent studies of PI3K/mTOR inhibitors have not yielded increased efficacy over the single node agents reported earlier. Furthermore, the majority of published studies combining PI3K/AKT pathway inhibitors with MEK inhibitors have reported limited activity secondary to significant toxicities found with the combinations. In recurrent endometrial cancer, the combination of the MEK inhibitor, trametinib, with the AKT inhibitor, GSK2141795, had only modest activity and significant toxicity. Similarly, a study of the combination of SAR245409 (PI3K/mTOR inhibitor) with pimasertib (MEK inhibitor) in low-grade serous ovarian cancer was unsuccessful in achieving improved clinical benefit over pimasertib alone. Furthermore, the combination required a high proportion of dose interruptions and therapy discontinuation due to toxicity.

Agents targeting MEK

The highest rates of Ras/Raf pathway aberrations are found in low-grade serous and mucinous ovarian cancer and in endometrial cancer. The rarity of mucinous cancer has limited the exploration of targeted agents for this indication. However, MEK inhibitors have been quite successful for the treatment of low-grade ovarian cancer. Selumetinib was explored in a phase II trial by the GOG, yielding a 15% objective response rate and a 65% SD rate in this notoriously chemoinsensitive disease. Based on the results of this trial, two large phase III studies were designed to compare MEK inhibition therapy with physician-choice chemotherapy in recurrent low-grade serous ovarian cancer. In GOG-0281, single-agent trametinib yielded improved PFS and response as compared with physician-choice treatment, including hormonal agents. Conversely, the MILO study explored binimetinib in a similar patient population but did not find an improved PFS in the broader population. However, there was an improvement in PFS among women with a KRAS mutation. In recurrent endometrial cancer, single-agent results for MEK inhibition have been limited. A phase II GOG trial of selumetinib reported a 6% objective response ( n = 3) and an event-free survival rate at 6 months of 12%. As noted earlier, these modest results have led to exploration of MEK inhibitor combinations for endometrial cancer.

Olaparib

Olaparib is an orally active PARP inhibitor that induces synthetic lethality in cells deficient in BRCA . In a study of two doses of olaparib in BRCA -mutant patients with advanced ovarian cancer, both doses were well tolerated and demonstrated modest effect. Olaparib at 100 mg twice a day (bid) achieved a response rate of 12.5% and a clinical benefit rate of 16.7% among 24 patients. The 400-mg bid dose had higher levels of response and clinical benefit among 33 patients, 33% and 57.6%, respectively. Additional phase II studies of olaparib in recurrent high serous ovarian cancer confirmed favorable response in patients with and without BRCA mutations.

Although olaparib has been well tolerated as a single agent with primarily adverse events consisting of fatigue, nausea, and vomiting, studies indicate that there is increased toxicity in combination with cytotoxic chemotherapy, necessitating dose reduction. Oza and colleagues compared the combination of olaparib with paclitaxel and carboplatin versus paclitaxel and carboplatin alone in platinum-sensitive recurrent ovarian cancer. This study included a maintenance arm of olaparib alone. Although there was no significant difference in response rate, there was improvement in PFS in the olaparib arm, which became clear during the maintenance phase of the trial. Additional studies of maintenance are discussed later.

Activity as a single agent in a platinum-resistant setting led to the FDA approval of olaparib in a tight niche, including those treated for ovarian cancer with more than three prior chemotherapies for recurrent disease. The approval of olaparib for BRCA -related ovarian cancer marks the first therapy approved for gynecologic malignancies with an associated biomarker. This approval was confirmed in the subsequent SOLO3 trial which compared olaparib with physician-choice nonplatinum chemotherapy in women with platinum-sensitive recurrent BRCA mutant ovarian cancer. Olaparib monotherapy achieved improved PFS and response in this patient population.

As noted previously, the greatest benefit of olaparib in combination with chemotherapy appeared to be from the maintenance portion. Thus olaparib as a maintenance strategy has been explored in both the upfront and platinum-sensitive recurrent settings after chemotherapy. These studies are discussed in detail in Chapter 14. Study 19 evaluated second-line olaparib maintenance versus placebo in an all-comers’ setting. This yielded a PFS benefit in all comers, but especially in patients with BRCA mutations. This finding was confirmed in SOLO2, which included only patients with BRCA mutations. Furthermore, olaparib maintenance gained an OS benefit for these patients. In the frontline setting, 2 years of olaparib maintenance provided PFS benefit compared with placebo in women with BRCA mutant ovarian cancer. Olaparib was also combined with bevacizumab in the frontline PAOLA-1 trial. This trial revealed PFS benefit for women with homologous recombination deficient tumors, but there was no benefit of the combination over bevacizumab alone in women with homologous recombination proficient tumors. Given these positive results, olaparib now has an indication for second-line maintenance in all comers and as frontline maintenance alone and in combination with bevacizumab for patients with BRCA- mutant and homologous recombination–deficient tumors, respectively.

Olaparib is under evaluation alone and in combination with cediranib in endometrial cancer. A preliminary report of this trial revealed that there was no improvement in PFS from olaparib or olaparib/cediranib compared with cediranib alone in an unselected group of patients with recurrent endometrial cancer. Further combinations are under exploration in recurrent endometrial cancer.

Rucaparib

Preclinical studies of rucaparib have revealed that this agent is a potent inhibitor of PARP-1 and PARP-2 with significant activity in ovarian cancer. The phase I trial of the single agent discovered a phase II dose that is well tolerated, with toxicities primarily including neutropenia, anemia, and elevated liver enzymes. A subset analysis of four patients with ovarian and primary peritoneal cancer reported 80% with objective response leading to the ARIEL2 study in women with platinum-sensitive recurrent ovarian cancer. An objective of this trial was to evaluate a genomic signature to predict response to PARP inhibition in pretreatment tumor tissue. BRCA -mutant tumors had impressive overall response rate of 69% and median PFS of 9.4 months. Conversely, BRCA -wildtype tumors had the lowest evidence of activity with response rate and median PFS of 13% and 3.7 months, respectively. BRCA -like tumors (homologous recombination deficient) included those with evidence of genome-wide loss of heterozygosity in the absence of BRCA mutation. Interestingly, these tumors had evidence of efficacy with response rates of 30% and median PFS of 7.1 months. This led to an FDA approval for rucaparib for the treatment of patients with BRCA mutant ovarian cancer with two or more prior therapies. This approval was subsequently confirmed in the ARIEL4 trial which randomized patients with BRCA- mutant recurrent ovarian cancer to rucaparib versus physician- choice platinum-based chemotherapy. Patients treated with rucaparib achieved improved PFS as compared with the standard therapy.

The final FDA approval for rucaparib is in the platinum-sensitive second-line maintenance space for patients with ovarian cancer regardless of BRCA status based on the results of the ARIEL 3 trial. This trial found improved PFS for women given rucaparib maintenance as compared with placebo. The phase III ATHENA trial (NCT03522246) is evaluating primary maintenance rucaparib alone or in combination nivolumab.

Niraparib

Similar to the aforementioned agents, niraparib inhibits PARP-1 and PARP-2 across a variety of advanced solid tumors. The phase I trial identified a recommended phase II dose of 300 mg/day, with the most common adverse events including myelosuppression, fatigue, and GI toxicity. The trials included BRCA- mutant and proficient tumors, achieving objective response rates of 40% and 16%, respectively. A subsequent phase II trial, QUADRA, explored the role of niraparib in patients with homologous recombination–deficient ovarian cancer with three or more prior lines of therapy. Based on the 28% response rate in a population with a high proportion of platinum resistance, this agent was granted an FDA approval in homologous recombination deficient ovarian cancer in the fourth line or greater.

Similar to the other PARP inhibitors, niraparib has demonstrated efficacy in the second-line and frontline maintenance settings. The NOVA trial found that niraparib improved PFS compared with placebo in both BRCA -mutant and BRCA- wildtype cohorts. This led to an FDA approval in all women with recurrent platinum-sensitive ovarian cancer after response to chemotherapy. Based on this success in the second-line setting, niraparib was assessed as a frontline maintenance in a population of ovarian cancer at high risk of recurrence in the PRIMA study. This study found niraparib maintenance provided improved PFS compared with placebo across both homologous recombination deficient and proficient ovarian cancer subtypes, leading to an FDA approval in this indication.

Veliparib

Veliparib, an oral inhibitor of PARP-1 and PARP-2, has also demonstrated encouraging activity in recurrent ovarian cancer. A phase II study of veliparib as a single agent yielded a response rate of 26% in women with BRCA- mutant ovarian cancer. Adverse events were modest and expected, consisting primarily of hematologic toxicities. Thus far, compared with other PARP inhibitors, this agent has been the most successfully combined with cytotoxic agents. However, it is unclear what benefit the addition of veliparib may add to cytotoxic chemotherapy. A randomized phase II trial of oral cyclophosphamide with veliparib did not find additional benefit in terms of response rate when compared with cyclophosphamide alone. A subsequent large-scale phase I trial identified a tolerable recommended phase II dose of veliparib in combination with paclitaxel, carboplatin, and bevacizumab therapy in upfront advanced ovarian cancer.

Thus the combination of veliparib with chemotherapy followed by veliparib maintenance was explored in the VELIA (GOG 3005) study. When compared with chemotherapy alone, concurrent veliparib followed by veliparib maintenance yielded improved PFS in this study. However, the lack of a veliparib maintenance alone arm hampered interpretation of this study, and it is unclear if it will be incorporated into clinical practice.

Small molecule inhibitors targeting epidermal growth factor receptor

Monoclonal antibodies targeting epidermal growth factor receptor

Combination agents

Sorafenib

Sorafenib is an interesting small molecule inhibitor that effectively blocks several angiogenesis-related receptors (VEGFR, PDGFR, and KIT) and Raf, a key component of the Raf-MAPK kinase (MEK)-MAPK proliferation pathway. Sorafenib is of considerable interest in ovarian cancer that has been shown to harbor alterations in this pathway. Consequently, several trials are ongoing to evaluate the efficacy of sorafenib in the upfront, recurrent, and consolidation setting. Matei and colleagues reported the results of a phase II trial of sorafenib (400 mg bid) in recurrent ovarian cancer treated with fewer than two prior therapies. Of 73 patients, 22 achieved partial response or SD with toxicities, including rash and metabolic abnormalities. Of note, among the 59 patients with measurable disease, 12 were progression free for at least 6 months. The combination of sorafenib (200 mg twice daily) with bevacizumab (5 mg/kg) for the treatment of solid tumors was explored in a phase I study. Although the combination appeared to have significant activity in 13 patients with ovarian cancer (43% response rate), toxicity was common, leading to sorafenib dose reductions in 74% of patients. The most prevalent toxicities were diarrhea, elevated liver transaminases, hypertension, hand and foot syndrome, and fatigue.

In combination with traditional cytotoxic chemotherapy, including topotecan and gemcitabine, sorafenib has shown similar clinical benefit rates in recurrent ovarian cancer with minimal additional toxicity. A randomized phase II trial in newly diagnosed ovarian cancer after cytoreductive surgery sought to determine if the addition of sorafenib (400 mg bid) to standard paclitaxel (175 mg/m ² ) and carboplatin (AUC 6) provided any PFS benefit. Unfortunately, this trial revealed similar clinical outcomes and worse toxicity in the arm that included sorafenib. Similar disappointing results were found in a trial of sorafenib as a maintenance agent in ovarian cancer where frequent dose interruptions and dose reductions yielded no improvement in PFS. Unfortunately, a trial exploring the addition of sorafenib as a chemosensitizer in the upfront treatment of advanced cervical cancer with radiotherapy was stopped early when it was determined that sorafenib was increasing tumor hypoxia, thus potentially worsening disease control in this setting.

Vandetanib

Vandetanib inhibits the tyrosine kinase activity of two parallel pathways through action on VEGFR-2 and EGFR. This leads to suppression of angiogenesis, cell proliferation, migration, and survival mechanisms. A phase II trial of vandetanib alone to treat recurrent ovarian cancer was terminated after the first stage of accrual because no patients achieved response or SD. Of note, molecular testing in this study revealed evidence of blockage of the EGFR pathway but no activity against VEGFR. Although single-agent activity has been disappointing, this agent underwent exploration in combination with docetaxel and liposomal doxorubicin for the treatment of ovarian cancer. A randomized phase I/II study of liposomal doxorubicin alone or in combination with vandetanib was limited by significant toxicity, which led to cessation in a high number of patients treated with the combination. Conversely, in a separate randomized phase II study, vandetanib in combination with docetaxel was tolerable; however, there was no improvement in PFS compared with docetaxel alone.

Imatinib

Imatinib mesylate is an example of a small molecule that inhibits multiple pathways, including c-kit, bcr-abl, and PDGFR. Although imatinib has been quite successful for the treatment of chronic myelogenous leukemia, in which the tyrosine kinase bcr-abl is constitutively activated, the success in gynecologic malignancies as a single agent has been limited. This may result in part from the impact of the PDGF axis on pericytes, which are mesenchymal cells that support endothelial cells and newly developing blood vessels. PDGF activity recruits pericytes, which produce VEGF to stimulate angiogenesis and survival. Research on pericytes has indicated that blocking the PDGF axis in addition to the VEGF axis may be necessary to achieve improved response rates. Thus far, imatinib has been explored only as a single agent or in combination with cytotoxic chemotherapy for gynecologic malignancies.

Several phase II trials for recurrent ovarian cancer, both high and low grade, have been performed using imatinib (400 mg/day) as a single agent. Unfortunately, rates of response have been low, with only one objective responder among all trials. Rates of SD ranged from 9% to 33%. Similar disappointing results have been seen in cervical cancer. In a trial of 12 patients with recurrent cervical cancer treated with imatinib, there were no responses and only one patient with SD. In all trials, expression of drug-related targets has not been associated with response. The use of imatinib in combination with docetaxel (30 mg/m ² weekly) in 23 patients with recurrent ovarian cancer achieved five responses, including one complete response. Three additional patients had SD lasting longer than 4 months. Further evaluations of this drug in combination with other cytotoxic chemotherapy for recurrent ovarian cancer have yielded minimal success.

Dasatinib

Dasatinib is a strong inhibitor of the Src family of kinases, which participate in the regulation of cellular growth, adhesion, invasion, and migration. At higher concentrations, dasatinib has effects on a variety of different targets relevant to tumor progression, including bcr-abl, c-kit, and PDGF. In addition, dasatinib targets the EphA2 receptor, which modulates cell migration, survival, proliferation, and angiogenesis. EphA2 overexpression correlates with poor outcome in ovarian cancer. Dasatinib (150 mg/day) has been explored in multiple solid tumors in early-phase trials with promising response rates. A phase I trial of this agent in combination with paclitaxel (175 mg/m ² ) and carboplatin (AUC 6) was tolerable with reasonable activity. A subsequent phase II trial in chemonaïve advanced or recurrent endometrial cancer demonstrated modest response and acceptable toxicity.

Cabozantinib (XL-184)

Exploration into mechanisms of resistance to anti-VEGF therapies has suggested pathways responsive to the observed upregulation of HIF-1α. One pathway of interest is c-Met, which is also implicated in tumor growth, survival, and metastases. Cabozantinib is a small molecule inhibitor of VEGFR-2 and c-Met, which has been evaluated in a number of solid tumors including, recently, ovarian cancer. In a phase II randomized discontinuation trial, cabozantinib was administered to 68 platinum-sensitive and platinum-resistant patients. Overall response was 24%, with 35% of patients also exhibiting prolonged SD. Response rates were similar between the platinum-sensitive and platinum-resistant cohorts. Because of better-than-expected results, the trial was terminated early. A subsequent GOG study compared this agent with weekly paclitaxel in recurrent ovarian cancer. Cabozantinib failed to meet prespecified endpoints of interest, demonstrating lower levels of response than weekly paclitaxel.

Angiopoietin (Ang)/Tie-2

In addition to the classic VEGF pathways of angiogenesis, other pathways involved in later stages of neovascularization hold promise for targeted therapy. Ang-1 and Ang-2 bind the receptor Tie-2 to activate pathways that stimulate endothelial cell proliferation, motility, and survival. In addition, the angiopoietins recruit pericytes to support vascular maturation. Trebananib (AMG-386) is a peptibody, a peptide fusion protein, that selectively binds Ang-1 and Ang-2, preventing interaction with Tie-2. In phase I studies of solid tumors, this compound was well tolerated, without many of the classic toxicities associated with antiangiogenic agents. A phase II study comparing two different dose levels of trebananib (3 mg/kg, 10 mg/kg) versus placebo in combination with weekly paclitaxel (80 mg/m ² ) demonstrated a trend toward improved PFS in patients who achieved higher steady-state concentrations of trebananib. These results lead to the dose selection for TRINOVA-1, a study comparing trebananib versus placebo plus weekly paclitaxel in recurrent ovarian cancer. The addition of trebananib yielded a significant improvement in PFS period of 2 months. A phase IB study of trebananib in combination with paclitaxel and carboplatin was found to be tolerable in women with advanced ovarian cancer. This led to development of TRINOVA-3, a randomized phase III study in upfront ovarian cancer. The addition of trebananib did not yield improved PFS in this population, although the combination was well tolerated. Trebananib (10 mg/kg) was also explored in combination with liposomal doxorubicin (50 mg/m ² ) or topotecan (4 mg/m ² ) for recurrent ovarian cancer in a phase IB trial. Interim results have demonstrated acceptable toxicity and modest activity. A subsequent phase III trial assessing the addition of trebananib to liposomal doxorubicin yielded improved response but no impact on PFS. In recurrent endometrial cancer, trebananib as a single agent did not have sufficient efficacy to pursue further evaluation, with only one objective response of 32 patients treated.

Aurora kinase

Three members of the aurora kinase family (Aurora kinase A, B, and C) play large roles in the phases of mitosis. These proteins are serine/threonine kinases that act to control cell division through regulation of spindle assembly, chromosome separation, and centrosome maturation. Overexpression of Aurora A and Aurora B is associated with tumorigenesis in multiple solid tumor types. Currently, there are 14 agents targeting Aurora kinase activity, which are under investigation in phase I and II trials. One agent targeting Aurora kinase A, alisertib (MLN8237), demonstrated significant activity in recurrent ovarian cancer and was evaluated as a single agent in a phase II study. Ten percent of patients treated with alisertib achieved a partial response, and 20% had SD for more than 3 months. A subsequent randomized phase II study of weekly paclitaxel alone or in combination with alisertib yielded improved PFS in the combination arm (6.7 vs. 4.7 months; HR, 0.75; 95% CI 0.58 to 0.96). Objective responses by RECIST were higher as well in the combination arm. An interesting new Aurora kinase A inhibitor, ENMD-2076, has additional activity against VEGFR, Flt-3, and FGFR3 kinases and has demonstrated activity in a phase I expansion cohort study of refractory ovarian cancer patients. Based on RECIST (defined later) or GCIG CA 125 response criteria, 2 of 20 patients and 6 of 20 patients, respectively, responded to single-agent therapy. This encouraged pursuit of a phase II study in ovarian cancer which achieved a PFS rate at 6 months of 22%. A subsequent study in clear cell ovarian cancer did not meet predefined criteria to move forward into further study, although ARID1A mutation was associated with benefit from this agent.

Delta-like 4 (DII4)/Notch

The Notch signaling pathway participates in the regulation of angiogenesis, proliferation, and apoptosis, making it an attractive potential target for anticancer therapy. One major avenue of Notch activation is through one of its ligands, Dll4, which creates a negative feedback loop to block cellular proliferation stimulated through VEGFR2. In addition, Notch signaling through Dll4 appears to activate cell cycle arrest through a variety of mechanisms. Notch overexpression has been demonstrated in multiple tumor types, including ovarian carcinoma. Two major therapeutic avenues exist that block signaling through the Notch pathway. The gamma secretase complex acts as a mediator of Notch signaling. Gamma secretase inhibitors are nonspecific agents that are currently in multiple phase I trials for solid tumors. Their lack of specificity has led to significant toxicity in the GI tract. Thus antibodies that target the Dll4 ligand have been developed to reduce toxicity. These compounds have demonstrated significant activity in in vivo models and are undergoing evaluation in early-phase clinical trials. Demcizumab, an anti-Dll4 antibody, achieved reasonable dose levels with adverse events, including hypertension, fatigue, anemia, and headache. In a subsequent phase IB trial the combination of demcizumab with weekly paclitaxel was tolerable and demonstrated encouraging activity in recurrent ovarian cancer. This target has been explored further with the bispecific antibody navicixizumab, which targets both Dll4 and VEGF. In an early phase I trial, this agent was well tolerated and demonstrated preliminary antitumor activity. In combination with weekly paclitaxel, navicixizumab yielded a 41% response rate in recurrent ovarian cancer in a phase IB trial. The most common adverse events were hypertension and fatigue.

Folate receptor α

Folate receptor α binds folic acid with high affinity and is involved in folate transport. This is quite relevant in gynecologic malignancy because this receptor is overexpressed in more than 90% of ovarian cancers. Furthermore, there is minimal expression of this target in normal tissues, making this an attractive target for anticancer therapy. Farletuzumab (MORAb-003), a humanized mAb to folate receptor α, demonstrated inhibition of cellular proliferation and stimulation of cell death in preclinical models. A phase II trial of this agent (100 mg/m ² ) in platinum-sensitive ovarian cancer was designed as a single-agent study with subsequent addition of traditional chemotherapy (carboplatin and paclitaxel) at the time of disease progression. In the single-agent arm, 24 of 28 patients remained on therapy at least 9 weeks. With combination therapy, 80% of patients achieved normalization of their CA 125 and 75% had complete or partial response. Unfortunately, farletuzumab in combination with carboplatin and paclitaxel did not yield superior outcomes compared with chemotherapy alone. Of note, a planned subset analysis found that lower pretreatment CA 125 and higher exposure to farletuzumab yielded improved PFS compared with placebo.

Another agent leveraging the folate receptor α is vintafolide. This novel ligand drug conjugate combines folic acid with desacetylvinblastine hydrazide, a chemotherapeutic agent difficult to deliver systemically. In a randomized phase II trial, vintafolide combined with pegylated liposomal doxorubicin (PLD) was associated with a significantly longer PFS than PLD alone (5 vs. 2.7 months; HR, 0.63; P = .031). Of interest, and in support of the target selectivity, there were few differences in the observed toxicities between the two arms. Furthermore, patients with global tumor expression of folate receptor α as measured by EC-20, a tracer conjugated to folic acid, had a PFS that was similar to that of the overall population but substantially better than the control arm, suggesting that high receptor expression is associated with poor prognosis. Unfortunately, a subsequent phase III trial was stopped early because of futility, although final results have not yet been reported.

Mirvetuximab soravtansine (IMGN853) is an ADC targeting folate receptor α. The folate receptor antibody is attached to maytansinoid DM4, a highly toxic agent that targets microtubules to induce cellular death. This agent is quite potent against cells that are folate receptor positive and even to nearby folate receptor–negative cells. A phase I trial of mirvetuximab soravtansine revealed that the agent was well tolerated, and in a small expansion cohort of ovarian cancer ( n = 17), 53% had an objective response despite platinum-resistant disease. These results led to the FORWARD 1 trial, which compared mirvetuximab soravtansine with standard chemotherapy (paclitaxel, PLD, or topotecan) in women with FRα-positive tumors. Unfortunately, there was no improvement in PFS from the investigational agent; however, this is thought secondary to incorrect identification of FRα positivity. Thus this agent is moving forward with an improved selection marker in further trials. The combination of mirvetuximab soravtansine and bevacizumab has also been found to be tolerable and have intriguing activity with a 39% response rate in platinum-resistant disease.

Antimesothelin antibodies

Mesothelin is an antigen that is highly expressed in several advanced solid tumors, including ovarian cancer. Growing data indicate that this protein may have a role in the peritoneal spread of ovarian cancer. Thus a number of ADCs containing an antimesothelin antibody are in development. DMOT4039A was studied in the phase I setting for unresectable pancreatic and platinum-resistant ovarian cancer. Toxicities included GI and constitutional, and many were cumulative over time. Objective activity was seen, and this agent is under further evaluation. Another ADC of antimesothelin antibody is anetumab ravtansine. Similar to mirvetuximab soravtansine, this antibody is linked to maytansinoid DM4. In phase I trials, this agent was well tolerated and demonstrated modest clinical activity.

Tissue factor

Tissue factor is a transmembrane glycoprotein that initiates the extrinsic coagulation pathway and cell signaling which can promote a malignant phenotype. Tissue factor is overexpressed in a number of solid tumors and is associated with poor prognosis in cervical cancer. Tisotumab vedotin is an ADC directed against tissue factor linked to monomethyl auristatin E, a microtubule-disrupting agent which causes cell cycle arrest and cell death. A phase I/II study of this agent included a cohort of patients with cervical cancer. Despite a heavily pretreated cohort, the objective response rate was 24% and PFS at 6 months was 29%. Thus tisotumab vedotin was explored in a larger study of patients with cervical cancer. Among 101 women, the response rate was confirmed at 24% and duration of response was greater than 8 months.

Histone acetyltransferases and histone deacetylases

The epigenetic modulation of gene expression is a promising target for the treatment of gynecologic malignancies. The posttranslational modification of histones is a key mechanism to regulate the expression of certain genes and has been implicated in tumorigenesis. Histones may be modified through acetylation, methylation, phosphorylation, and ubiquitination. Two enzymes, histone acetyltransferase and histone deacetylases (HDACs), mediate the addition and removal of acetyl groups, respectively. HDAC activity leads to a change in the histone structure, which leads to a change in transcription activity. Through activation of oncogenes or inactivation of tumor suppressors, cell growth, proliferation, and evasion of apoptosis may occur. HDAC inhibitors have been explored in a variety of solid tumor types. Suberoylanilide hydroxamic acid (Vorinostat) is FDA approved for the treatment of refractory cutaneous T-cell lymphoma and has been explored in recurrent ovarian cancer. A phase II study of single-agent vorinostat (400 mg/day) in 27 patients had limited success with only one partial response (4%) and two patients with SD for 6 months (7%). The HDAC inhibitor belinostat was evaluated as a single agent in platinum-resistant ovary and micropapillary low malignant potential tumors. This study of 32 patients demonstrated modest activity with only one partial response and 19 patients with SD. There is evidence to suggest that the combination of HDAC inhibitors with traditional cytotoxic chemotherapy may yield improved results; thus these agents have been explored in combination with traditional chemotherapy regimens. The combination of vorinostat with gemcitabine and carboplatin was quite toxic, although some activity was achieved. The study was terminated early secondary to overall poor tolerance. Conversely, belinostat was successfully combined with carboplatin as well as with carboplatin and paclitaxel. Although both combinations were tolerable, only belinostat in combination with carboplatin and paclitaxel yielded response rates worthy of further study.

Platelet-derived growth factor

In addition to the combination agents noted previously, there are agents that solely target the PDGFR. This was supported by recent research indicating that the v-sis oncogene is derived from the PDGF gene. Furthermore, the PDGF pathway is involved not only in angiogenesis but also in stimulation of cancer growth and regulation of fibroblasts. The use of a mAb to target this pathway promises to improve efficacy and reduce off-target effects. Olaratumab (IMG-3G3) is a mAb to PDGF-α that has demonstrated significant antitumor activity in preclinical studies in ovarian cancer cell lines. Activity in phase I trials in advanced solid malignancy led to a randomized phase II study of this agent (20 mg/kg) in combination with liposomal doxorubicin (40 mg/m ² ) for platinum-resistant recurrent ovarian cancer. The combination did not yield PFS or OS benefit in this patient population. Similarly, in soft tissue sarcoma, the combination of olartulumab and doxorubicin was found to improve PFS and OS in an early phase II study. However, the larger phase III study was unable to confirm this OS benefit among this population.

P53

Given its high prevalence across a number of gynecologic malignancies and advanced solid tumors, there is great interest to leverage targeted agents against p53 mutations. By far, Wee1 inhibitors are the furthest developed against this important target. Tumors with mutations in p53 are dependent on the G2 cell cycle checkpoint, which is regulated by Wee1, to repair damaged DNA. Inhibition of Wee1 creates a situation of synthetic lethality among p53-mutant tumors. Adavosertib is a Wee1 inhibitor that has been combined with carboplatin and paclitaxel in the platinum-sensitive recurrent setting of ovarian cancer. The combination of chemotherapy with a Wee1 inhibitor yielded an improved response rate and PFS compared with chemotherapy alone. These findings were further confirmed in a population of high-grade serous ovarian cancer selected by p53 mutation, although toxicities were increased. This agent is now under evaluation alone and in combination with several targeted agents for ovarian cancer. Furthermore, single-agent adavosertib demonstrated efficacy in recurrent uterine serous cancer, yielding a response rate of 29%. In addition to Wee1 inhibitors, several molecules are in early development that appear to restore the function of p53 across a number of p53 mutant tumors in the preclinical setting.

Immune therapy

As noted earlier, immune escape is an essential contribution to successful tumorigenesis. Over the past 5 years, the number of successful immune therapy agents has skyrocketed. Early studies have focused on notoriously immune-driven cancers, including melanoma and non–small cell lung cancer. Single-agent studies of agents targeting programmed death (PD)-1 and programmed death ligand (PD-L1), including avelumab, nivolumab, and pembrolizumab, have been less successful in ovarian cancer, yielding modest response rates approaching 15% in an unselected platinum-resistant setting, primarily among patients with clear cell histology. Responses among high-grade serous ovarian cancer have been limited. However, growing data indicate that activity of immune therapy may be related to common features of gynecologic malignancies, such as mutations in BRCA , POLE, or the presence of microsatellite instability. This has led to development of immune therapy trials in ovarian, endometrial, and cervical cancer both alone and in combination with chemotherapy or targeted therapy.

In endometrial cancer, the results of a basket trial of patients with mismatch-repair–deficient tumors yielded a tissue-agnostic indication for pembrolizumab in this setting. Importantly, among the 49 women with mismatch-repair–deficient endometrial cancer, the response rate was 57%. These results have been recapitulated in several other trials of PD-1/PD-L1 inhibitors including avelumab, dostarlimab, and durvalumab, where response rates range from 27% to 47%. Of note, dostarlimab has an indication in recurrent MSI-H endometrial cancer as a single agent.

Activity of single-agent checkpoint inhibition has been limited in microsatellite stable endometrial tumors. Results in PD-L1–positive tumors demonstrated little activity in the absence of MSI-H disease. However, the combination of lenvatinib and pembrolizumab has significant activity in patients with MSS tumors. In a trial of 94 patients, response rate with 36% and duration of response was 21 months. This yielded an FDA approval for the combination in this patient population.

In cervical cancer, single-agent checkpoint inhibition with pembrolizumab has an FDA approval based on a trial in PD-L1–positive disease. This agent yielded a 17% response rate, which was adequate to achieve the indication. Subsequent trials are exploring dual checkpoint combinations as well as adding PD-1 inhibitors to standard of care chemotherapy and bevacizumab in advanced and recurrent disease. Furthermore, it is likely, based on reported results of the EMPOWER/GOG-3016/ENGOT-cx9 trial, that cemiplimab will also receive approval after demonstrating superior objective response, PFS, and OS as compared with physician-choice chemotherapy in patients progressing after combination chemotherapy with recurrent cervix cancer. Several other PD-1/L1– and cytotoxic T-lymphocyte associated protein 4 (CTLA-4)–based strategies are being explored in endometrial and cervix cancer, including bispecific antibodies containing both elements.

In addition to the checkpoint inhibitor trials described earlier, the use of tumor-directed T cells (adoptive T-cell therapy) has been explored in cervical malignancies. A small study of this therapy in nine women with metastatic cervical cancer yielded three objective responses, including two durable complete responses. LN-145 is a tumor-infiltrating lymphocyte (TIL)-based therapy that was explored in a small phase II study ( n = 27) of previously treated cervix cancer patients. An objective response rate of 44%, including three patients with complete response enabled a Breakthrough Designation (BTD) in 2019. Certainly, these therapies are promising and results from larger clinical trials are eagerly anticipated. In addition, approaches to enhance cytotoxic T lymphocyte (CTL) entry into tumor islets will be important for improving the efficacy of immune therapies.

Hypertension

Hypertension is one of the most commonly experienced side effects related to targeted agents, with a prevalence that approaches 40%. This adverse event is most commonly related to the use of antiangiogenic agents, including bevacizumab, and combination agents such as sorafenib. The mechanism of hypertension is currently hypothesized to be twofold. First, the inhibition of VEGF results in the reduction of production of nitric oxide by endothelial cells. Resultant vasoconstriction and endothelial cell dysfunction are compounded by the occurrence of microvascular rarefaction. Microvascular rarefaction is a process found in a majority of patients with hypertension, regardless of etiology, and consists of the loss of the arterioles and capillaries that make up the microcirculation. Vasoconstriction from loss of nitric oxide serves to exacerbate nonperfusion and elevate peripheral vascular resistance, which can result in permanent microvessel extinction.

Hypertension appears to be dose dependent for several of the antiangiogenic agents. Bevacizumab at the typical doses given for gynecologic malignancies (10 to 15 mg/kg) has an incidence of hypertension between 18% and 36% compared with an incidence of 3% to 30% in patients treated with low-dose bevacizumab (3 to 7.5 mg/kg). This incidence may be further increased when bevacizumab is combined with cytotoxic agents known to have antiangiogenic activity such as topotecan or paclitaxel. Agents affecting the HIF-1α pathway reduce nitric acid synthesis and metabolism, leading to exacerbation of antiangiogenic drug-induced hypertension. Of note, small molecular inhibitors to VEGF have reduced incidence and severity of hypertension compared with mAbs targeting the receptor.

Hypertension induced by antiangiogenic agents is most commonly found in patients with preexisting hypertension. Thus active management of blood pressure and cardiac risk factors such as elevated lipids is paramount. The use of several antihypertensive agents is often necessary to control hypertension in this setting. Acceptable treatments for angiogenesis inhibitor–induced hypertension include angiotensin-converting enzyme inhibitors, beta-blockers, and dihydropyridine calcium channel blockers (amlodipine, felodipine). Nondihydropyridine calcium channel blockers should be avoided given their inhibition of the P450 CYP3A4, which can affect the metabolism of tyrosine kinase inhibitors. Nifedipine is also not an optimal therapeutic choice secondary to the increase in VEGF levels associated with chronic use.

Reverse protein leukoencephalopathy

Reverse protein leukoencephalopathy (RPLS) is a rare complication of antiangiogenic agents, including bevacizumab, sorafenib, and sunitinib. Occurring in fewer than 1% of all patients treated with these agents, this syndrome is characterized by finding of diffuse, hyperintense T2-weighted signaling secondary to leaking of the capillaries in the brain. Symptoms of RPLS include change in mental status, headache, seizures, and lethargy. Aggressive antihypertensive management and supportive care typically lead to full recovery; however, in rare cases, this syndrome may be fatal. Typically, further use of anti-VEGF therapy in these patients is contraindicated.

Cardiotoxicity

Other effects of targeted agents on the cardiovascular system can include left ventricular dysfunction, congestive heart failure, and myocardial infarction. Uncontrolled hypertension can lead to left ventricular hypertrophy and cardiac musculature that outgrows its vascular supply. This results in traumatic cardiovascular events, which may be seen in any of the anti-VEGF agents. Furthermore, any tyrosine kinase inhibitor may have off-target effects that damage cardiac muscle cells and impair their survival. For example, sorafenib and sunitinib have been associated with decreased left ventricular ejection fraction in up to 30% of patients. Of note, this damage appears to be reversible secondary to mitochondrial dysfunction rather than the permanent free radical–induced cardiac damage found after anthracycline cardiac toxicity.

Thus far, prediction of the development of cardiac toxicity from targeted agents has been difficult. It does appear more common in patients with preexisting cardiac disease and prior treatment with anthracyclines. Again, aggressive medical management of preexisting conditions and careful monitoring of cardiac function are essential to maximize outcomes.

Renal toxicity

Renal adverse effects may be found in response to a variety of targeted agents as a result of the expression of key targets along the nephron. The most common toxicity is proteinuria, especially among the antiangiogenic agents. This is thought to originate secondary to damage of endothelial cells, which depend on VEGF for integrity and regulation of permeability. In addition, thrombotic microangiopathy that leads to further damage of the glomerular capillaries may compound the loss of protein from the vasculature. Proteinuria is typically low grade, is asymptomatic, and may be monitored through assessment of albumin levels by urine dipstick. If protein readings exceed 2+ protein, collection of 24-hour urine protein is indicated for treatment decision making. Urine protein levels exceeding 2 g require treatment interruption and possible dose reduction depending on protocol requirements. Spot urine protein to creatinine ratios may also be used to monitor renal effect of targeted agents. Renal failure caused by targeted agents is rare and appears to be primarily related to glomerulonephritis, interstitial nephritis, and acute tubular necrosis. Prevention and management of renal toxicity is based on thorough assessment of renal risk factors before therapy institution and optimization of any preexisting conditions.

Thromboembolic events

The risk of venous thromboembolism (VTE) is increased at baseline among patients with active malignancy receiving chemotherapy. Antiangiogenic agents also appear to increase the propensity for thrombosis, especially arterial thromboembolism (ATE), including myocardial, cardiovascular, cerebral, and peripheral vascular events. In a large meta-analysis of patients treated as part of a randomized trial for a variety of solid tumors, patients treated with bevacizumab in addition to chemotherapy had two times greater incidence of ATE. Multivariate analysis revealed that bevacizumab remained a significant risk factor for ATE, in addition to age 65 years or older and prior history of ATE. Of note, the rate of thromboembolic complications appears low for VEGF-Trap and VEGF tyrosine kinase inhibitors. This may be secondary to formation of complexes by anti-VEGF antibodies that create a nidus for the formation of thromboses in the vasculature. Further study of this complication in large randomized controlled trials is necessary to determine the true incidence rates. At this point, patients with a history of VTE warrant the careful assessment and aggressive use of anticoagulation therapy. Anti-VEGF therapy should be avoided in patients with a history of prior ATE. Support for the use of routine thromboprophylaxis has not been established.

Gastrointestinal perforation and fistula

GI toxicity is a life-threatening complication that has been primarily associated with anti-VEGF therapy. Risk of intestinal perforation or fistula in any tumor type is associated with prior GI surgery, acute diverticulitis or active abdominal infection, obstruction, carcinomatosis, or history of prior radiotherapy. Rates of bowel perforation are highest among patients with ovarian cancer, ranging as high as 11% in one study of bevacizumab in the recurrent setting. In this study, a higher risk of perforation was associated with a history of more than two prior chemotherapy regimens. Furthermore, the high perforation rate in a trial combining bevacizumab and erlotinib in ovarian cancer led to early closure. Of note, frontline treatment does not appear to lead to increased GI toxicities, as evidenced by equivalent rates of perforation in the control and experimental arms of GOG-0218 (reviewed in Table 15.4 ). Proposed mechanisms for bowel injury include the negative effect on wound healing secondary to damage of normal endothelial cells. Anti-VEGF therapies may induce microangiopathic events in the intestinal lining that cause edema and hemorrhage, making the wall more susceptible to bacterial dissemination and infection. In addition, rapid tumor shrinkage may cause the bowel wall to become weak, leading to perforation.

The management of anti-VEGF therapy-induced GI perforation must be individualized based on patient history, amount of disease, and status. In general, surgical exploration with intestinal diversion is the standard of care for patients with acceptable performance status; however, consideration should be made for risk of further poor wound healing. If a patient cannot undergo surgical exploration, conservative management with decompression, bowel rest, drainage, antibiotics, and nutritional support may be indicated.

Diarrhea and mucositis are common among patients treated with a majority of the small molecule inhibitors, regardless of target. These adverse events may be the dose-limiting toxicity for a patient, requiring admission for IV hydration and drug discontinuation. Prevention with antimotility agents or neutralizing mouthwashes may be helpful to reduce incidence of these off-target effects. Esophagitis or gastritis may be an issue in patients treated with sorafenib or sunitinib; therefore prevention with a proton pump inhibitor is reasonable.

Cutaneous complications

Skin toxicities are quite common in patients treated with anti-EGFR agents (50% to 100%), are dose dependent, and appear to be indicative of drug efficacy. Typically, an acneiform rash consisting of pruritic papules and pustules presents within 3 weeks of treatment on skin areas that have sebaceous glands. A similar rash may be seen in patients treated with mTOR inhibitors. The mechanism is thought to be related to follicular hyperkeratosis that causes glands blockage and inflammatory eruption.

This rash may be managed with topical antiseptics and antiinflammatory agents. Oral antibiotics such as doxycycline or minocycline may be used daily for a short course to improve rash severity and symptoms. In addition to this rash, patients may experience significant xerosis, which may be managed with over-the-counter emollients. Changes to the nail bed may be seen, including fragility, paronychia (inflammation of nail folds), and fissures. The hair may also be affected by the use of these agents. Most commonly, hypertrichosis, including growth of the eyelashes and scalp hair, has been reported.

Hand and foot skin reaction or acral erythema is a classic skin toxicity of sorafenib and sunitinib. Occurring in 20% to 25% of patients, it may consist of erythema, hyperkeratosis, xerosis, and dysesthesias and paresthesia in the bilateral hands, feet, or both. Patients can develop callous-like blisters and desquamation in both pressure- and non–pressure-bearing areas. Treatment for hand and foot skin reaction is primarily supportive, consisting of elevation, compresses, avoidance of trauma, and topical application of corticosteroid cream. Symptoms usually resolve during the course of therapy, but some patients may require dose reduction or complete cessation of therapy. In addition to hand and foot skin reaction, patients treated with sorafenib or sunitinib can develop stomatitis, scalp dysesthesias, perianal rash, and a skin rash that resembles seborrheic dermatitis.

Metabolic abnormalities

The nonspecific nature of some of the targeted agents has led to a variety of unique metabolic side effects, requiring careful monitoring and action. Endocrine effects of tyrosine kinase inhibitors may include altered thyroid function, glucose metabolism, and bone metabolism. For example, sunitinib and sorafenib have been associated with onset of hypothyroidism in previous euthyroid patients. mTOR inhibitors, including temsirolimus and everolimus, have been associated with incidence of any hyperglycemia as high as 50%. Of note, alterations in glucose metabolism can occur in either direction in patients treated with sunitinib or imatinib. Glucose levels must be monitored closely because these changes may necessitate change in dosing of antihyperglycemic agents among patients with diabetes. Glycemic control should be maximized before treatment with these agents. Lipid abnormalities, including elevated triglycerides and cholesterol, are common toxicities experienced by patients treated with mTOR inhibitors. These levels should be followed during treatment, and, depending on patient cardiac risk factors, lipid-lowering agents may be instituted. Electrolytes should be monitored during treatment with targeted agents because many tyrosine kinase inhibitors have caused profound hypophosphatemia manifested as muscle weakness and fatigue.