Multidisciplinary considerations in the maintenance treatment of poly(ADP-ribose) polymerase inhibitors for homologous recombination-proficient, advanced-stage epithelial ovarian cancer

A 73-year-old, para five, postmenopausal woman with a history of type 2 diabetes mellitus, hypertension, hyperlipidemia, sarcoidosis, and osteoarthritis presented to the Emergency Department with shortness of breath, abdominal distention, and early satiety for 1 month. She had a remote history of an abdominal hysterectomy for uterine fibroids. Chest x-ray and subsequent computed tomography (CT) of the chest revealed a moderate, right-sided pleural effusion with compressive atelectasis. A CT of the abdomen and pelvis revealed a large, complex, cystic, and solid left adnexal mass measuring 8 × 13 cm with a smaller mass in the right lower pelvis measuring 4.3 cm. Omental thickening and ascites also were noted. The patient was admitted to the hospital and underwent an ultrasound-guided thoracentesis. One liter of fluid was drained and sent for cytology, which returned positive for malignancy. Her cancer antigen 125 level was elevated at 424 U/ml. The patient was discharged home with a plan for outpatient gynecologic oncology follow-up given the concern for ovarian cancer.

The patient was seen in consultation and an extensive history was taken. She denied a family history of cancer. Management options were discussed, including either primary cytoreductive surgery followed by chemotherapy or neoadjuvant chemotherapy with possible interval cytoreduction. Given the extent of her disease on imaging and her presentation, with symptomatic pleural effusions limiting her mobility and functional status, it was recommended she undergo neoadjuvant chemotherapy. While awaiting chemotherapy, the patient was readmitted to an outside hospital with recurrent shortness of breath caused by the re-accumulation of pleural fluid. She underwent repeat thoracentesis, and a PleurX catheter (Becton, Dickinson and Company) was placed. An omental biopsy was also performed and revealed metastatic adenocarcinoma of Mullerian origin.

The patient subsequently completed four cycles of neoadjuvant chemotherapy with paclitaxel, carboplatin, and bevacizumab, with normalization of her cancer antigen 125 level to 18 U/ml after three cycles. She developed worsening peripheral neuropathy grade 2 between cycles three and four despite the use of B6, glutamine, and alpha lipoic acid. Preoperative CT demonstrated an interval decrease in size of her bilateral adnexal masses and resolution of her omental caking, ascites, and pleural effusion. Her PleurX catheter was removed before surgery. At the time of exploratory laparotomy, she had a palpably thickened omentum and normal adnexa. She underwent bilateral salpingo-oophorectomy, infracolic omentectomy, biopsies, and external iliac lymph node sampling, with no gross residual cancer palpated or visualized at the end of the case (R0 resection). Pathology revealed microscopic, high-grade, serous epithelial ovarian adenocarcinoma involving the ovaries and omentum.

The patient's postoperative course was complicated by readmission for recurrent, noninfectious diarrhea grade 2–3, resulting in dehydration and hypomagnesemia. Her case was presented at a multidisciplinary team meeting with the consensus plan to complete an additional three cycles of chemotherapy, exchanging paclitaxel for docetaxel given her progressive neuropathy. The patient tolerated the remaining three cycles of paclitaxel and carboplatin; however, the decision was made to hold bevacizumab in light of her ongoing diarrhea.

The patient was referred to genetic counseling, and a detailed family history revealed no family history of breast, gynecologic, or colon cancer on either side of her family. The patient has two daughters, and she wanted to pursue genetic testing to help them determine whether they were at increased risk of ovarian cancer. Current practice guidelines endorse that all women with epithelial ovarian cancer (EOC) should be offered germline genetic testing at primary diagnosis for cancer susceptibility genes, regardless of their cancer family history or clinical features of their cancer.^{1, 2} In 2022, it is projected that there will be 22,000 new cases of ovarian cancer diagnosed in the United States. Despite advances in diagnostics, surgery, and treatment, an estimated 14,000 women will die of disease. Ovarian cancer currently ranks fifth in the causes of death attributed to cancer among women, accounting for more deaths than those attributed to all other gynecologic malignancies.³

The strongest risk factor for ovarian cancer is a family history of breast or ovarian cancer. Meta-analyses have shown that the relative risk of ovarian cancer is 3.1 (95% CI, 2.6–3.7) for first-degree relatives of a woman with ovarian cancer, although those analyses did not factor in the impact of inherited susceptibility because of germline mutation carrier status.⁴ Approximately 18%–25% of all EOCs are caused by an inherited susceptibility, of which BRCA1 and BRCA2 germline pathogenic mutations comprise almost 75%.^{5, 6} The remaining proportion of heritable ovarian cancers are caused by germline pathogenic mutations in the DNA mismatch-repair genes, as in Lynch syndrome, or by genes involved in homologous recombination, such as those in in the BRCA–Fanconi anemia pathway.^6-8

The discovery that a woman with ovarian cancer has a germline mutation is significant for both the patient and her family. Germline heritable mutations follow an autosomal dominant pattern of inheritance, and each child has a 50% probability of inheriting the pathogenic mutation.⁹ This information could allow for more personalized treatment for the woman with ovarian cancer and tailored screening for other cancers for which she may have a genetic predisposition. First-degree and second-degree relatives of a woman who has ovarian cancer and a known germline mutation should be offered genetic risk evaluation and testing (cascade testing) to help identify germline mutation carriers.¹⁰

The patient had a MyRisk hereditary cancer test (Myriad Genetics Laboratories), a multigene panel consisting of genes that have been associated with inherited susceptibility to ovarian cancer, including: BRCA1, BRCA2, MLH1, MSH2, MSH6, PMS2, EPCAM, STK11, PALB2, RAD51C, RAD51D, and BRIP1. A recent American Society of Clinical Oncology expert consensus panel recommended that a multigene panel should be offered to all women with EOC to include germline sequencing of the genes associated with an inherited risk of ovarian cancer.¹⁰ BRCA germline mutation carriers have a significantly elevated lifetime risk of developing ovarian cancer (BRCA1, 40%–60%; BRCA2, 11%–27%). BRCA mutation-associated ovarian cancers have distinct clinical features, including a younger age of diagnosis and enhanced sensitivity to platinum-based chemotherapy, and have improved survival compared with non-BRCA–associated ovarian cancers.¹¹ MLH1, MSH2, MSH6, PMS2, and EPCAM are mismatch-repair genes associated with Lynch syndrome, also known as hereditary nonpolyposis colorectal cancer.⁸ Mutations in these mismatch-repair genes confer an increased risk of gynecologic cancers, such as ovarian and endometrial cancers, as well as colorectal, stomach, small intestine, hepatobiliary, and upper urinary tract cancers. Defective mismatch-repair gene–associated cancers have a phenotype characterized by frequent microsatellite instability, which may confer an enhanced response to certain types of chemotherapy and novel agents that target DNA repair pathways.¹²

The selection of additional genes to include in a patient’s multigene panel testing will then be guided by the patient’s cancer family history and her preference regarding the number of genes to be tested. A three-generation pedigree to include first-degree and second-degree relatives should focus on the type and age at diagnosis of cancers from both sides of the family. A notable cancer family history will include multiple family members diagnosed with cancer, especially at a younger age, multiple primary cancers within one individual, or certain pathognomonic cancers. Commercially available multigene panels typically include 30–40 genes known to be associated with hereditary cancers, whereas the larger expanded panels will include in excess of 80–90 genes, some of which may have less clearly established associations with ovarian cancer. The cost and convenience of commercial multigene panel testing are comparable to those of germline testing for BRCA only; consequently, the majority of patients tend to choose the broader panels. In one study, out-of-pocket cost was more important to women who chose single-gene testing. Women who chose multigene testing preferred the enhanced probability of detecting both deleterious mutations and variants of uncertain significance. Less than 9% of patients declined testing after genetic counseling.¹³ Variants of uncertain significance are detected in up to 40% of women who are tested with larger gene panels and are considered nonactionable, with no definitive pathologic associations or clinical management recommendations.¹⁴

Additional considerations for the choice of a commercial laboratory multigene gene panel will depend on the provider preference, the in-network preference of the patient's insurance, and the potential for out-of-pocket costs. The majority of women diagnosed with ovarian cancer have been referred for genetic testing after extensive diagnostic evaluation and often surgery; therefore, the issue of a patient not having met her deductible is exceedingly uncommon. One discriminating factor between the different commercial laboratories has been the cost and timeframe to test first-degree or second-degree blood relatives of a patient found to have a germline pathogenic mutation (cascade testing). Some laboratories have offered this free of charge or free of charge within a specified time frame, whereas others have no such policy.

For the highly penetrant, more common susceptibility genes like BRCA and Lynch syndrome-associated mismatch-repair genes, the reproducibility of the test results from one laboratory platform to another is exceedingly high. However, for some of the newer genes with a less well established relation to cancer, the certainty of identifying a definitive pathogenic mutation that results in a nonfunctional protein varies from one laboratory to another. Pretest counseling must include discussion with the patient of some of the current limitations of the testing platforms and how to interpret ambiguous results in the context of cancer risk for the patient and her family.

Women with ovarian cancer who do not have germline mutations detected should be offered somatic tissue testing for BRCA mutations within their tumor.¹⁰ Somatic or epigenetic alterations in BRCA genes can also result in loss of BRCA function within the tumor; however, by definition, they are not hereditary and confer no increased risk of ovarian cancer to family members. An additional 5%–7% of women will be found to have BRCA somatic mutations within their tumor tissue.^{15, 16} Acquired somatic BRCA mutations can only be detected in tumor tissue, whereas germline mutations should be detected in both tumor and DNA sequencing. Although the optimal strategy to test for germline and tumor somatic BRCA mutations is debated, current American Society of Clinical Oncology guidelines advise that germline DNA sequencing should precede somatic testing. Tumor tissue testing may fail to detect the large gene rearrangements and dosage variants that account for up to 5% of germline BRCA mutations.^{10, 15}

The patient opted for the Myriad myChoice companion diagnostic (CDx) assay, a companion diagnostic performed on tissue that identifies tumors with deleterious BRCA variants and homologous recombination deficiency (HRD). The patient's tumor was classified as homologous recombination-proficient, without any identifiable somatic alterations in genetic susceptibility genes. HRD is a phenotype characterized by the inability of a cell to effectively repair double-stranded DNA breaks using the homologous recombination repair gene pathways, resulting in the cell's reliance on alternative, error-prone DNA repair pathways. Approximately one half of women with high-grade serous EOC have HRD⁸ (Figure 1). Homologous recombination status is an important prognostic marker that predicts response to chemotherapy and poly(ADP-ribose) polymerase (PARP) inhibitors, allowing tailored treatment choices for ovarian cancer.

Commercially available HRD assays vary in the molecular measures they rely on to classify tumors as HRD or homologous recombination-proficient. The companion diagnostic that was used in three complete prospective clinical trials to identify patients who would benefit from PARP inhibitor treatment was the myChoice CDx assay from Myriad Genetics Laboratories. This assay measures both the causes and consequences of impaired homologous recombination repair, although the cutoff aggregate score to define HRD status varied from 33 in the VELIA/GOG-3005 study (ClinicalTrials.gov identifier NCT02470585) to 42 in the PRIMA/ENGOT-ov26 study (ClinicalTrials.gov identifier NCT02655016) and the PAOLA-1/ENGOT-ov25 study (ClinicalTrials.gov identifier NCT02477644).¹⁷ The American Society of Clinical Oncology states that no recommendation can be made to support routine HRD testing using HRD assays for patients with EOC given the lack of consensus regarding a uniform method to identify HRD tumors.¹⁰ Further study and standardization of the definition of homologous repair deficiency will be of paramount importance in the selection of companion diagnostic testing in the next generation of clinical studies of PARP inhibitors.

The patient has stage IV EOC with an excellent clinical response to neoadjuvant combination chemotherapy using paclitaxel, carboplatin, and bevacizumab and interval surgical cytoreduction to no gross residual disease. Treatment options for women with advanced-stage EOC who have responded to first-line combination, platinum-based chemotherapy include bevacizumab and/or PARP inhibitor treatment, depending on their primary treatment.¹⁸ Despite an initial response to primary treatment, 70% of women with advanced-stage EOC will have a relapse within 3 years of diagnosis.¹⁹ Recurrent ovarian cancer is generally incurable, with a pressing need to develop better first-line treatment options that can increase the likelihood of long-term remission. The addition of bevacizumab to chemotherapy after primary cytoreductive surgery improved progression-free survival (PFS) in two prospective randomized trials (GOG-218 [ClinicalTrials.gov identifier NCT00262847] and ICON7 [ClinicalTrials.gov identifier NCT00483782]) in women with advanced EOC.^{20, 21}

The patient had no germline or somatic mutations identified in her tumor tissue (homologous-proficient). After discussing the data with the patient, she and her husband chose maintenance treatment with niraparib based on the PRIMA/ENGOT-ov26 trial, which was most applicable to her because the study included patients who had a higher risk of recurrence (she presented with stage IV disease) and was enriched for patients who received neoadjuvant chemotherapy followed by surgery. The patient started on 300 mg daily but had a dose interruption because of grade 2 neutropenia, with an absolute neutrophil count of 1000/mm³, followed by a dose reduction to 200 mg after 2 weeks because of persistent grade 2 anemia (hemoglobin 8.2 g/dl). Over the following 3 months, she complained of persistent grade 2 nausea and grade 1 diarrhea despite antiemetics and dietary alterations and required twice weekly intravenous hydration and treatment for hypomagnesemia. The patient chose to discontinue niraparib after 4 months of treatment because of the significant side effects she experienced, which she felt compromised her quality of life. The patient and her husband weighed the anticipated benefit of PARP inhibitor treatment in view of her homologous-proficient status with her health care preferences and priorities, and the patient chose surveillance. The patient remains without evidence of disease by examination and imaging, with a normal cancer antigen 125 level 12 months after discontinuing niraparib.

Some of the challenges of managing the side effects of PARP inhibitors result from the recommended daily treatment for 2 years (olaparib) or 3 years (niraparib). The oral formulation is highly acceptable to patients; however, the daily continuous dosing for 2 consecutive years means that small side effects can cumulatively compromise quality of life. Dose interruptions and the use of supportive medications to minimize some of the side effects can be very helpful. More severe toxicities require dose reductions, which are quite common, especially at the beginning of treatment. Dose reductions improve tolerability and enable women to remain on PARP inhibitor treatment for the prescribed 2 years of treatment.^29-31 It is imperative to counsel patients who are starting PARP inhibitors to expect that they will have to make some lifestyle changes during their dose-finding journey to balance the clinical benefit of maintenance treatment with quality of life.

PARP inhibitors share many side effects, with some noteworthy effects specific to each drug. Nephrotoxicity is more commonly seen with olaparib than with niraparib and more frequently occurs at the initiation of treatment. PARP inhibitors can cause nasopharyngitis, especially olaparib (Table 2). Some of the important side effects common to all PARP inhibitors include fatigue, nausea, and myelosuppression. Fatigue is a very common side effect of PARP inhibitor treatment, with severe symptoms reported to affect 2%–6% of women. Other contributing factors, such as insomnia, depression, or anemia, must be considered. Brief interruptions in treatment or dose reductions in severe cases can help patients with long-term compliance.³⁰

Severe nausea was reported in 1%–6% of women who received PARP inhibitors as maintenance treatment, and the highest frequency was among the women who received niraparib. Interventions that can be useful to manage nausea includes eating smaller, more frequent snacks rather than large meals; taking the medication at bedtime; and identifying possible trigger foods that may exacerbate symptoms. Prophylactic antiemetic treatment can be used, although coordinating daily antiemetic medication, especially with a twice-a-day medication like olaparib, can be a very onerous for patients. Once-daily dosing with niraparib offers better patient convenience and may enhance patient compliance. Patients are encouraged to implement systems to track taking their pills and to review the use of all dietary supplements and other medications with their provider.

Myelosuppression, especially anemia and thrombocytopenia, are serious side effects associated with PARP inhibitor treatment. Notably, severe thrombocytopenia can be sudden and prolonged in some patients, especially with niraparib. Patients should be completely recovered from any prior chemotherapy-related hematologic toxicity before commencing PARP inhibitors. Complete blood counts are monitored weekly in women who are taking niraparib, every 2 weeks in women who are taking olaparib for the first 4–6 weeks of treatment, or after any dose reductions. A reduced starting dose of niraparib at 200 mg daily is appropriate for women <77 kg or for those who have pre-existing thrombocytopenia (platelets <150K). The interval between complete blood counts can be lengthened thereafter.^{30, 31}

Based on the available studies of PARP inhibitors used as maintenance treatment, dose interruptions were necessary in the majority of patients, and dose reductions were necessary in from one third to one half of patients. Despite these frequent alterations in dose, the majority of women were able to continue their PARP inhibitors. It is imperative that providers allocate the necessary resources to patient education and to clear expectation setting with patients and their families as women embark upon PARP inhibitors. Similarly, processes to regularly monitor side effects, especially within the first 2–3 months of initiating treatment, are advised to optimize patient outcomes.

The widespread use of PARP inhibitors has enhanced our understanding and ability to manage the attendant side effects and optimize patient compliance and quality of life during therapy. However, one of the most consequential risks of PARP inhibitor treatment is the development of therapy-related myeloid neoplasms, which occur in 1%–3% of patients who have ovarian cancer.^{31, 32} The spectrum of therapy-related myeloid neoplasms includes myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML), which are characterized by complex karyotypes and a poor prognosis. A meta-analysis of PARP inhibitor therapy in women with largely recurrent ovarian cancer could not identify clinical risk factors that predicted the development of this uniformly fatal condition. Surprisingly, no relation between germline BRCA mutation status, prolonged prior chemotherapy exposure, or recurrent disease was noted. Recent data suggest that there may be early signals or detectable markers in the peripheral blood that herald an elevated risk of therapy-related myeloid neoplasms. One study found that early pancytopenia may be associated with MDS/AML, whereas another study found that somatically acquired oncogenic mutations present in women who are treated for ovarian cancer before starting PARP inhibitor treatment may increase the risk for developing MDS/AML.^{33, 34} More investigation is necessary to elucidate the link between PARP inhibitors and hematologic toxicities and to identify reliable early indicators associated with significant hematologic toxicity. The increasingly widespread use of PARP inhibitors for women with ovarian cancer, both as maintenance treatment and for recurrent disease, will advance our knowledge and understanding of this relation to optimize patient tolerability and minimize drug-associated toxicity.

The authors made no disclosures.