The Highs and Lows of Serous Ovarian Cancer

Abstract

Low-grade serous ovarian cancer was initially described as a distinct type of rare epithelial ovarian cancer 20 years ago, however, only recently have we begun to leverage our understanding of the clinical behavior and molecular profile of this disease for its treatment. The use of routine next generation sequencing has allowed for a deeper understanding of the molecular drivers of this disease and shown how molecular alterations in MAPK pathway genes such as KRAS and BRAF can affect overall prognosis and disease behavior. The use of targeted therapies, including MEK inhibitors, BRAF kinase inhibitors and other investigational targeted therapies are changing the way we view and treat this disease. In addition, endocrine therapy can provide prolonged disease stability with generally mild toxicity, as well as promising response rates in recent studies examining combination therapy with CDK 4/6 inhibitors in the upfront and recurrent setting. Once seen merely as a chemo-resistant form of ovarian cancer, recent studies have worked to harness the unique features of low-grade serous ovarian cancer to provide individualized treatment options for patients with this disease.

Condensed Abstract:

The use of routine next generation sequencing has allowed for a deeper understanding of the molecular drivers of low grade serous ovarian cancer and shown how molecular alterations in MAPK pathway genes such as KRAS and BRAF can affect overall prognosis and disease behavior. The use of targeted therapies, including MEK inhibitors, BRAF kinase inhibitors and other investigational targeted therapies, as well as novel endocrine strategies, are changing the way we view and treat this disease.

Introduction:

It has now been almost 20 years since a seminal publication established a two-tiered grading system in serous ovarian cancer with low-grade serous ovarian cancer (LGSOC) and high-grade serous ovarian cancer (HGSOC), based primarily on the assessment of nuclear atypia with mitotic rate used as a secondary feature[1]. This two-tiered system has now largely replaced the 3-tiered systems utilized by the Shimizu/Silverberg and FIGO grading systems for serous ovarian cancer grading [2, 3]. During this time, our understanding of LGSOC as a disease has evolved. Previously considered to be a spectrum of disease ranging from grade 1 to grade 3 serous ovarian cancer, we now understand that LGSOC is a rare entity, with clinical, histologic and molecular characteristics that are distinct from HGSOC. Improved understanding of the behavior and molecular drivers of LGSOC has led to a paradigm shift in the classification and treatment of this disease. LGSOC was classically viewed as a more indolent and less chemo-responsive form of serous ovarian cancer yet was treated similar to HGSOC. Recent insight into the clinical and molecular features of LGSOC, and its precursor, serous borderline tumor (SBT), have led to improved tools for prognostication and prospective clinical trials focused on optimizing both the initial management as well as treatment of recurrent disease.

Histology and Epidemiology of LGSOC:

LGSOC accounts for 2% of all epithelial ovarian cancers and 4.7% of serous ovarian cancer[4]. LGSOC is thought to arise via a stepwise progression, from benign serous cystadenoma, to SBT, which can progress to micropapillary SBT, and ultimately invasive LGSOC (Figure 1A–D). Patients diagnosed with SBT may experience recurrence in the form of SBT or malignant progression to LGSOC, which may occur years, or even decades, after their original diagnosis and surgical resection. A comprehensive population study performed by Vang and colleagues examined all patients who were diagnosed with SBT in Denmark between 1978 and 2002 and found that amongst 942 cases of SBT, a subsequent diagnosis of carcinoma occurred in 4% of patients, of which 93% were LGSOC and 7% were HGSOC. Amongst those patients with no extraovarian disease (implants) at time of diagnosis 17/809 (2.1%) developed subsequent serous carcinoma, while amongst the 114 with noninvasive implants at time of initial diagnosis, 18 (15.8%) developed subsequent serous carcinoma. The median time to development of subsequent invasive cancer was 10 years, with a range up to 25 years[3]. Molecular analyses revealed concordant mutational profiles in matched SBT and subsequent carcinomas in 92% of cases (including those that occurred decades after initial diagnosis) [5]. Therefore, most serous carcinomas occurring in patients with prior ovarian SBT represent true recurrences rather than new independent primary tumors.

Figure 1.

Histologic features of low-grade serous ovarian neoplasms and high-grade serous ovarian cancer. (A-D) The stepwise sequence of progression of low-grade serous ovarian tumors: (A) Serous cystadenoma lined by benign tubal-type epithelium. Occasional papillary projections likely represent the earliest morphologic changes preceding development of serous borderline tumor (SBT). (B) SBT, composed of hierarchically branching papillary structures. (C) Micropapillary SBT is considered a more advanced form of SBT, characterized by long slender papillae emanating directly from fibrous cores. (D) Finally, low-grade serous ovarian cancer (LGSOC) is defined by stromal invasion by tumor cells, arranged in nests and micropapillae. The nuclei are small, round and monotonous. (E) In contrast, high-grade serous ovarian cancer (HGSOC) show marked nuclear pleomorphism, including tumor giant cells, and abundant mitotic figures (inset: higher magnification of boxed area with mitoses indicated by arrowheads). (F) Immunohistochemistry for p53 shows a wildtype (heterogeneous) expression pattern in LGSOC and aberrant (diffuse) overexpression in a representative case of HGSOC.

By immunohistochemistry, ovarian serous tumors (SBT, LGSOC, and HGSOC) express PAX8, WT1, ER and PR (all showing nuclear immunoreactivity). While HGSOC typically exhibits marked nuclear pleomorphism (≥3:1 size variation) and high mitotic rate (≥12 mitoses per 10 high-powered fields; Figure 1E) [1], in cases with equivocal morphologic features, immunohistochemical analysis for p53 can be useful for making the distinction between HGSOC versus LGSOC. HGSOC exhibits aberrant expression of p53 (either strong diffuse staining in tumor cells, or complete absence of expression), corresponding with near-universal TP53 mutations in this tumor type, while LGSOC, which lacks TP53 mutations, shows a wild-type p53 expression pattern (heterogeneous staining of tumor cells; Figure 1F)[6].

Germline testing is recommended for all women with ovarian cancer and will help to better define the prevalence of alterations in rare tumor populations over time but thus far LGSOC does not appear to be associated with presence of BReast CAncer gene 1 or 2 (BRCA) mutations. In a study which employed secondary pathologic review, Manning-Geist et al found no germline BRCA mutations in a population of 79 patients with LGSOC who were treated at a comprehensive cancer center[7]. Similarly, prior studies have found BRCA mutation rates of ≤ 5% in patients with LGSOC[8–10].

Prognostication of LGSOC:

Residual disease following primary surgery, younger age at diagnosis, obesity, and current smoking have all been clinically associated with worse outcomes for patients with LGSOC. Patients with LGSOC generally have a younger median age of diagnosis (43–47 years) than those with HGSOC [11]. Fader and colleagues performed an ancillary analysis of GOG 182, examining 182 enrolled patients with grade 1 serous carcinoma (synonymous with current LGSOC nomenclature). On multivariate analysis, only residual disease status following primary surgery was significantly associated with survival (p=0.006). Patients with microscopic residual disease had a significantly longer median progression-free (33.2 months) and overall survival (96.9 months) compared with those with residual disease of 0.1–1.0 cm (14.7 months and 44.5 months, respectively) and residual disease of 1.0 cm or greater (14.1 months and 42.0 months, respectively; P<.001) [12]. A study examining the outcomes of 350 women with LGSOC treated at MD Anderson Cancer Center (MDACC) found that women diagnosed at age > 35 had a longer median PFS compared with women diagnosed at ≤ 35 years old (32.6 months; 95% CI, 26.8 to 38.5; and 18.8 months; 95% CI, 14.2 to 23.5; P < 0.001, respectively)[13]. In a separate study of 194 patients with LGSOC treated at MDACC, smoking had a negative association with both overall survival (OS) (HR, 1.73; 95% CI, 1.03–2.92; p=0.4) and PFS (HR, 1.72; 95% CI, 1.00–2.96; p=0.5) on multivariable analysis. In this study, the median OS was shorter in current smokers compared to former/never smokers (48 months vs 79.9 months; p=0.002). Further, on univariable analysis BMI ≥ 35 kg/m2 was also associated with an increased risk of death (HR, 2.53; 95% CI, 1.19–5.38; p=0.02)[14].

The presence of a somatic BRAF V600E mutation in patients with SBT may pathologically predict for improved prognosis and decreased likelihood of progression[15] [16] [17]. In a retrospective study by Grisham and colleagues, 75 SBT or LGSOC tumors were examined, 57 of which harbored a BRAF V600E (n=26) or KRAS (n=17) mutation. The presence of BRAF V600E mutation was associated with early disease stage (stage I/II; p< 0.001) and SBT histology (p=0.002)[16]. In a study by Chui and colleagues, 201 SBTs were analyzed for presence of KRAS or BRAF mutations following central pathology review. SBTs with a BRAF V600E mutation were less likely to exhibit micropapillary features (p<0.0001) and were more frequently stage I (p=0.0023). After adjusting for stage and age, the risk of subsequent progression to serous carcinoma was significantly lower among women with BRAF-mutated SBTs [HR 0.27 (0.08–0.93), p=0.038] but not with KRAS- mutated SBTs [HR 1.00 (0.45–2.23), p=0.99], compared to SBTs without a mutation in either gene [17]. While presence of a BRAF mutation has shown the strongest association with improved outcomes for patients with SBT, the broader group of MAPK pathway alterations are associated with improved prognosis in patients with LGSOC (Figure 2). A study by Manning-Geist and colleagues evaluated 119 patients with LGSOC whose tumors had completed central pathology review. Alterations affecting the MAPK pathway were found in 71/119 (60%) of LGSOC’s, most commonly somatic KRAS (33%), NRAS (11%), EIFIAX (10%), and BRAF (11%) alterations. On multivariate analysis, MAPK pathway alterations (p= 0.02) and platinum sensitivity (p= 0.005) were both significantly associated with improved OS[18].

Figure 2.

Simplified schematic of the Mitogen-Activated Protein Kinase (MAPK) pathway. A key signaling pathway that regulates a wide variety of cellular processes including proliferation, apoptosis, and stress responses, the MAPK pathway plays a crucial role in LGSOC and serves as an important target for therapies such as MEK and RAF inhibitors. This figure was made using BioRender.

Initial Treatment of LGSOC:

For patients with stage IC1 and greater disease, following surgical staging by a gynecologic oncologist, treatment has traditionally consisted of six cycles of carboplatin and paclitaxel chemotherapy; however this practice has recently been brought into question with observation considered for certain patients with stage IC disease and endocrine therapy alone considered as a maintenance strategy for some patients with stage IC and greater disease. For patients with extensive preoperative disease precluding optimal resection, or for those who are not surgical candidates, neoadjuvant chemotherapy (NACT) may be offered. In patients with newly diagnosed epithelial ovarian cancer, two large randomized trials have evaluated outcomes in epithelial ovarian cancer patients receiving NACT versus primary debulking surgery followed by adjuvant chemotherapy. The phase III CHORUS study enrolled 550 women across 87 hospitals. This study found OS was noninferior with the use of NACT when compared to primary debulking surgery, with a median OS of 22.6 months in the primary debulking group and 24.1 months in the NACT group [HR for death 0.87 (90% CI 0.72 −1.05)]. However, the vast majority of patients treated on this study presented with HGSOC (73%), with LGSOC representing only 4% of the study population [19]. Similarly, in a separate study of 670 women with newly diagnosed stage IIIC or IV epithelial ovarian cancer who were randomly assigned to primary debulking surgery or NACT; NACT followed by interval debulking was not inferior to primary debulking surgery followed by adjuvant chemotherapy. Similar to the CHORUS study, 3.6% of patients enrolled to this study exhibited well-differentiated histology[20].

Existing literature suggests lower response rates of LGSOC to NACT than are seen in HGSOC. For example, a study by Schmeler and colleagues evaluated 25 women with newly diagnosed LGSOC who were treated with platinum-taxane NACT. Although 50% of patients had a >50% reduction in CA-125 following NACT there was only 1/25 complete (4%) and no partial responses observed [21]. A study by Grabowski and colleagues examined 5114 patients with newly diagnosed ovarian cancer who were treated across 4 randomized phase III trials with platinum-based chemotherapy. Amongst the 145 patients with LGSOC, 39 had suboptimal debulking and were eligible for response evaluation. An objective response to first-line chemotherapy was observed in 90.1% of the patients with HGSOC versus only 23.1% of the patients with LGSOC [22]. while a study by Cobb and colleagues found that 4/36 (11%) of patients with LGSOC treated with NACT achieved a partial response [23]. Overall, given the historically lower response rates to chemotherapy in patients with LGSOC vs HGSOC, primary debulking surgery is generally recommended for patients with newly diagnosed LGSOC.

While adjuvant platinum-based chemotherapy has traditionally been administered following primary debulking surgery for patients with advanced LGSOC, recent studies have questioned whether there is sufficient data to justify this practice. In addition, the broad adoption of letrozole maintenance therapy following completion of adjuvant therapy for patients with advanced LGSOC has led many to wonder if letrozole maintenance by itself, without adjuvant chemotherapy, may provide similar outcomes for patients with newly diagnosed disease [24]. An ongoing international phase III study, NRG-GY019, is studying this question by randomizing patients with advanced LGSOC who have undergone primary debulking surgery to 6 cycles of platinum-taxane chemotherapy followed by letrozole vs letrozole alone (NCT04095364).

Treatment for recurrent LGSOC:

Patients with recurrent LGSOC should be considered for secondary, and in some cases later line, debulking surgery. Similar to the upfront setting, the best outcomes in LGSOC patients are observed when complete gross resection of disease is achieved at the time of surgical debulking. In a retrospective study of 41 women with recurrent LGSOC, those who achieved a complete gross resection at time of secondary surgery had a significantly better PFS (60.3 months vs 10.7 months) and OS (93.6 months vs 45.8 months) from the time of secondary surgery than those patients who had gross residual disease following surgery[25].

In the recurrent setting, systemic treatment decisions for patients with HGSOC are generally determined based on platinum resistant or platinum sensitive status, as determined from time of last platinum based therapy to time of progression. However, it is unclear if this same paradigm should be applied to LGSOC. In a study evaluating patients at a single institution with recurrent LGSOC, 25% of patients responded to cytotoxic chemotherapy in the second-line setting, with a response rate of 22% to platinum-based chemotherapy and 27% to nonplatinum-based chemotherapy[26].

Distinct Opportunities for Treatment of LGSOC in the Recurrent Setting:

Targeting Angiogenesis:

While chemotherapy has historically shown lower response rates in LGSOC than in HGSOC, retrospective studies examining the use of bevacizumab (generally in combination with chemotherapy) have shown promising results. A single-institution study examining 17 patients with SBT or recurrent LGSOC who received bevacizumab as a single agent or in combination with chemotherapy found a response rate of 55% (6/11) amongst the patients with LGSOC; no responses were observed in the 4 patients with SBT[27]. Another single-institution study evaluating 40 patients with recurrent LGSOC who received 45 separate “bevacizumab containing patient-regimens”, generally in combination with chemotherapy, found a response rate of 47.5% and a median PFS of 10.2 months (95% CI 7.9, 12.4)[28].

Targeting the MAPK pathway:

LGSOC has a distinct molecular profile. Unlike HGSOC, in which TP53 mutations are ubiquitous and alterations affecting the MAPK pathway are quite rare, LGSOC tumors seldom display TP53 mutations and commonly harbor MAPK pathway alterations, most frequently KRAS, which is mutated in approximately 33% of cases, and less frequently NRAS (8%), BRAF (6%), NF1 (5%) and RAF1 (2%)[29–32]. This distinct molecular profile affords unique therapeutic options for LGSOC.

Two randomized phase III studies have been performed comparing single agent MEK inhibitor to standard of care therapy in patients with recurrent LGSOC. GOG 281, a phase II/III study of trametinib versus physicians’ choice of chemotherapy or endocrine therapy (PCCe) in patients with recurrent disease and unlimited prior lines of therapy, found a significant difference in median PFS with trametinib (13 months) versus PCC (7.2 months; HR=0.48; P< 0.001). The response rate was 26% in those patients treated with trametinib and 6.2% for those treated with PCCe (p< 0.0001)[33]. A contemporaneous phase III study(MILO/ENGOT-ov11) of binimetinib vs physicians’ choice of chemotherapy (PCC) in patients with recurrent LGSOC and a maximum of 3 prior lines of chemotherapy was discontinued based on an interim analysis showing that the PFS HR crossed the predefined futility boundary. Notably, for patients treated with binimetinib, the response rate and median PFS in the KRAS mutant group (RR= 44%; median PFS 17.7 months) were significantly better than in the KRAS wild-type group (RR= 19%; median PFS=10.8 months; p=0.006), indicating that patients with KRAS mutations may be more likely to respond to MEK inhibitor treatment[34]. Trametinib (2a) and binimetinib (2b) are both now NCCN compendium listed as options for treatment of recurrent LGSOC[35].

An ongoing study examining biomarker driven therapies for rare ovarian cancer subtypes inclusive of LGSOC (GOG-3051/ENGOT-GYN2; NCT04931342) explores different targeted therapy approaches based on a variety of biomarkers including ER/PR positivity and MAPK pathway alterations. The recent Food and Drug Administration (FDA) approval of dabrafenib and trametinib for patients with advanced solid tumors harboring a BRAF V600E mutation has also afforded this combination as an option for the 8% of LGSOC patients with this alteration[36].

A dual RAF/MEK inhibitor VS-6766 in combination with the FAK inhibitor, defactinib, has shown promising response rates in an ongoing phase I study exploring the combination across multiple solid tumors. Of the 24 evaluable patients presented at time of interval analysis, 11 (46%) had KRAS mutations and 10 (42%) had received a MEK inhibitor prior to study entry. The ORR in all patients was 46% (11/24, 95% CI: 28% to 65%), with ORR of 64% (7/11, 95% CI: 35% to 85%) in patients with KRAS mutations and 44% (4/9, 95% CI: 19% to 73%) in patients with KRAS wildtype tumors. Responses were observed in patients with and without prior MEK inhibitor exposure[37]. An international phase II study is currently enrolling patients with recurrent LGSOC to this combination (NCT04625270).

A recent study of 179 specimens of LGSOC performed by El Naggar and colleagues found that amongst those tumor specimens that were hormone receptor negative, nearly 50% contained KRAS or NF1 mutations, suggesting that combination therapy with endocrine and MAPK targeted therapy may be a promising strategy for this population [38].

Targeting endocrine strategies:

Hormone receptor positivity is more prevalent in LGSOC than in HGSOC, with 58–87% of LGSOC tumors being ER+ and 43–58% PR+, versus HGSOC where 27% of tumors are ER+ and 17% PR+ [39, 40]. Prior studies of endocrine treatment in epithelial ovarian cancer, performed primarily in patients with HGSOC, have shown disappointing results[41, 42]. A phase 2 study of letrozole in women with advanced or recurrent estrogen receptor positive (ER+) HGSOC displayed a response rate of 1/33 (3%) and a median duration of clinical benefit of 9 weeks (range, 7–46 weeks)[43]. A phase II study of anastrozole in women with ER+ or progesterone receptor positive (PR+) platinum resistant ovarian cancer displayed a 0% response rate amongst the 49 evaluable patients and a median PFS of 2.7 months (95% CI, 2.0–2.8 months)[44].

However, endocrine treatment studies focused specifically on the LGSOC population have demonstrated prolonged disease stabilization, and in some cases meaningful response rates. A phase II study of anastrozole in patients with ER+ recurrent or metastatic LGSOC or serous borderline ovarian tumors (PARAGON) displayed partial response by RECIST criteria in 5/36 (14%) patients and a median PFS of 11.1 months (95% CI 3.2–11.9)[45]. Similar results were seen with the use of letrozole in the phase II/III GOG281 study of trametinib vs PCCe in patients with recurrent LGSOC. Patients treated in the PCCe arm were allowed to receive treatment with either chemotherapy or endocrine therapy with letrozole or tamoxifen. While 0/27 patients treated with tamoxifen achieved an objective response, the response rate to letrozole was 13.6% (6/44)[33]. In a single-institution study of 64 patients with recurrent LGSOC treated with 89 endocrine regimens over time (most commonly: anastrozole, fulvestrant, letrozole, leuprolide, or megestrol acetate) the objective response rate to endocrine therapy was 9%, while 61.8% of patients experienced stable disease[46]. Some studies have found that IHC levels of ER and PR+ may be associated with response to endocrine therapy[47]. An open-label phase II study of letrozole in patients with recurrent ovarian cancer found that those patients who demonstrated Union International Contre Cancer (UICC) stable disease on letrozole had higher ER (P=0.027) and PR (P=0.0066) histoscores than the progressive disease group, however the majority of patients evaluated on that study displayed poorly differentiated histology[48].

Amongst epithelial ovarian cancers, 60–82% are androgen receptor positive (AR+)[49, 50]. Preclinical data have shown that AR+ ovarian cancer cells display increased cell division when exposed to androgens and that this activity is reversed with the second-generation AR antagonist, enzalutamide [51]. A phase II study of enzalutamide was performed in women with advanced or recurrent AR+ HGSOC or LGSOC with RECIST 1.1 measurable disease and 1–3 prior lines of chemotherapy. The study enrolled 45 patients with HGSOC and 14 patients with LGSOC. The study met its primary endpoint with 13/59 (22%) remaining progression free for ≥ 6 months of treatment, however median PFS was 1.7 months for those patients with HGSOC and 4.6 months for those with LGSOC, and response rate had just one confirmed and one unconfirmed response reported out of the study population of 59 patients[52].

Given that endocrine therapy is generally well tolerated, with moderate single agent activity in LGSOC, there has been recent interest in trying to enhance this activity through combinations with CDK 4/6 inhibitors. Cobb and colleagues described interim results of a 15 patient pilot that enrolled patients with stage III or IV unresectable newly diagnosed LGSOC to treatment with the CDK 4/6 inhibitor abemaciclib in combination with fulvestrant and reported that 47% of patients experienced a partial response to this combination in the front-line setting[35]. Encouraging signals have also been seen with this combination in the recurrent setting. The GOG Foundation recently completed a single arm, phase II study of the CDK 4/6 inhibitor ribociclib, in combination with letrozole for women with advanced or recurrent LGSOC. The response rate was 23% and the clinical benefit rate was 79%. Of interest, the progression free survival was 19.1 months and the duration of response among responders was 19.1 months[35].

Future Directions:

Recent advances in the molecular characterization of LGSOC and novel endocrine combinations have opened the door to a new era of treatment for LGSOC—one that focuses on targeting the unique characteristics of LGSOC and learning from prognostic markers to better tailor sequencing and treatment selection. Recent success in enrolling patients with LGSOC to prospective clinical trials performed by NRG Oncology, ENGOT, and GOG Partners highlight the ability to complete clinically meaningful research for this rare disease. Current clinical trials focused on determining the best front-line strategies in the neoadjuvant and adjuvant settings, as well as recurrence strategies focused on maximizing the effectiveness of endocrine therapy and MAPK pathway targeting drugs will likely shape the next decade of treatment for this rare disease.