M. Rowland

Initiation of a formalized precision medicine program in gynecologic oncology

OBJECTIVE In an effort to better incorporate precision medicine into clinical practice, we initiated a pilot project to screen, discuss, and genetically characterize patients with metastatic or recurrent gynecologic malignancies for whom no curative standard of care exists. METHODS In 7/2014, we initiated a multi-disciplinary Precision Medicine Board (PMB) whose purpose was to apply molecular profiling to select and prioritize early phase clinical trial enrollment for high-risk gynecologic malignancies. Additional objectives were to record outcomes and enable scientific discussions of mutations which may foster local translational research. FoundationOne was the preferred genomic platform; results were reviewed by a team comprised of disease site specialists, phase I trialists, and basic and translational scientists affiliated with the Gynecologic Cancer Program. A detailed database for each patient was created and is followed prospectively for treatment use and resultant outcomes. RESULTS To date, we have presented 62 cases with interpretable FoundationOne testing on 60 tumor samples (31 ovarian, 18 uterine, 9 cervical, and 4 other female genital tract). Significant genomic alterations were commonly found in all tumor types (median: 3); TP53 (45%) and PIK3CA (27%) were the most frequently noted mutations; however, molecular profiling resulted in identification of few actionable mutations (6%). To date, we have matched 4 patients on therapies based on actionable mutations. CONCLUSIONS The predominant function of our PMB is establishment of a forum to enhance research while providing clinical care for refractory malignancies. We have matched patients with specific mutations to ongoing trials and are developing investigator-initiated studies based on trends within genomic profiling results. Longer-term follow up will be required to determine the success of this strategy.

Contribution of age to clinical trial enrollment and tolerance with ovarian cancer

INTRODUCTION Increasing age has been correlated with shorter survival in ovarian cancer patients, a finding attributed to diminished tolerance of standard therapy. Elderly patients, however, are less likely to enroll on clinical trials; thus, limited data exists to evaluate their response to front line treatment. This study describes how elderly patients on trial fared, with respect to toxicity and response, compared to younger women. METHODS A retrospective cohort study was performed of ovarian cancer patients enrolled in front line chemotherapy trials at our institution between 2000 and 2013. Patients were dichotomized by age: <70 and ≥70years. Clinical, pathologic, and treatment characteristics were recorded and analyzed using SAS version 9.3. RESULTS 336 patients were enrolled. Of these, 79 (23.5%) were ≥70yrs. Demographics were similar between the two groups. Compared to patients <70, those ≥70 completed a comparable number of chemotherapy cycles (p=0.16) and had similar numbers of dose modifications (p=0.40) and delays (p=0.26). Both hematologic and non-hematologic toxicities occurred at similar rates as well. Age≥70 (HR 1.8, 95% CI 1.27-2.54, p=0.0009), stage III/IV (HR 3.44, 95% CI 1.08-10.95, p=0.036), and residual disease (HR 2.63, 95% CI 1.82-3.78, p<0.0001) were independently predictive of shorter overall survival. CONCLUSION Our data continues to support reports of shorter survival for older women with ovarian cancer. With physician bias removed and similar chemotherapy tolerance noted, our study suggests that inherent tumor biology may be a significant contributor. Further research is needed to identify the mechanisms which contribute to the inequality that age imposes on outcomes.

Treatment Advances in Locoregionally Advanced and Stage IVB/Recurrent Cervical Cancer: Can We Agree That More Is Not Always Better?

Although relatively rare in the developed world, cervical cancer is the fourth most common cancer among women worldwide with over 527,000 cases in 2012. Primary and secondary prevention efforts will be most impactful, but research aimed at improving outcomes and individualizing therapies is also critical. Developments in individualizing this therapy to the population most likely to benefit may allow evidence-based triage of therapies, even in resource-limited areas, and improve outcomes overall. Accompanying this editorial are two articles helpful for treatment decision making. Rose et al evaluated data from over 2,000 women treated for locoregionally advanced cervix cancer and developed a nomogram that proposes to stratify women by prognostic risk. Kitagawa et al report on a large randomized phase III trial (RCT) comparing standard of care cisplatin and paclitaxel (PT) to carboplatin and paclitaxel (CT) seeking to investigate whether the latter, more tolerable, regimen was noninferior to PT. In 1999, the National Cancer Institute recommended that treatment for locally advanced cervical cancer (stage IB2-IVA) should include the addition of cisplatin-based chemotherapy to radiation as demonstrated by five RCTs showing a decrease in the risk of death of 40%. Despite significant improvements with the addition of cisplatin to radiation (chemoradiotherapy [CRT]), local recurrences still hover around a rate of 20% and 3-year overall survival (OS) ranges from 67% to 74% depending on the study (Table 1). To further decrease the rate of recurrences, which are almost universally lethal, treatment strategies have included combining another cytotoxic therapy with CRT, addition of targeted therapies with CRT, neoadjuvant chemotherapy, and consolidation cytotoxic therapy after CRT. With the exception of the Duenas-Gonzalez et al study evaluating RT with cisplatin and gemcitabine followed by two cycles of cisplatin and gemcitabine as compared with CRT, no other strategy has demonstrated improvement in response rate (RR) and survival over CRT alone. The concept of using adjuvant chemotherapy is currently being evaluated in an international phase III RCT comparing CRT with CRT plus four cycles of CT (NCT01414608). It is important to note that early trials with patients receiving radiation alone demonstrated that approximately 50% of patients never had recurrences and for these patients, additional treatment adds only harm. Furthermore, with the addition of sensitizing cisplatin, the rate of recurrence drops an additional 15%. Maximizing the gain of additional treatment for the 35% who will recur and sparing toxicity for the other 65% is critical to maximizing OS and quality of life. Although prognostic factors for recurrence have been previously evaluated to help guide identification of patients at risk for recurrence, the extent to which these factors may continue to apply with the widespread use of CRT is not defined. Rose et al present an analysis of locally advanced cervical cancer patients treated on phase III RCT of RT alone versus CRT and nomograms were constructed for 2-year progression-free survival, 5-year OS, and pelvic recurrence. The nomograms provide improved risk stratification, by weighting each characteristic rather than a simple additive risk model and offer an opportunity to better estimate an individual patient’s risk and target more aggressive therapies (adjuvant CT for example) to those at highest risk. Although these nomograms are validated for their predictive abilities, we must consider them within the context of individual patients. Patient treatment preferences, goals of care, comorbidities, and many other factors sit outside predictive models, but play an equal role in developing best treatment plans for our patients. Integration of this information is key. We must also consider the context of the studies that were included to create the nomograms. In this case, these six trials span 26 years from first study launch to last study closure. During this time, protocol mandated radiation treatment length decreased from 10 to 8 weeks and all studies mandated 2or 4-field external beam and assigned dose based on anatomic landmarks (80 Gy to Point A, for example, in Table 1). These techniques are still widely used in certain parts of the world and the nomograms may be most applicable where current techniques mirror the included studies. In the developed world, where much of radiation for cervix cancer has transitioned to three-dimensional conformal therapy, intensity-modulated radiation therapy, and volumetric dosing, we wonder if this nomogram is as applicable. In contrast to the management of patients with earlier-stage disease, treatment of stage IVB/recurrent cervical cancer is generally noncurative. For such patients, we have progressed over the past 30 years from a model of best supportive care to doublet cytotoxic therapy with or without bevacizumab, with concordant improvements in JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 33 NUMBER 19 JULY 1 2015

Factors influencing clinical trial enrollment among ovarian cancer patients

OBJECTIVE To characterize patients who did not enroll on a clinical trial and identify barriers that may limit enrollment among patients with advanced epithelial ovarian cancer (EOC) presenting for first-line chemotherapy. METHODS We conducted a retrospective review of patients diagnosed with stage II-IV EOC from 10/2009-4/2013, a time period during which multiple trials were available to all EOC patients, including optimally debulked, suboptimally debulked, or undergoing neoadjuvant chemotherapy. Enrollment status, demographics, tumor characteristics, and treatment details were recorded. SAS version 9.3 was used for all analyses. RESULTS 144 patients met study criteria; 67% were enrolled on a trial. Enrolled patients were significantly younger (median 61 vs 68years, p=0.002). Stage (p=0.30), race (p=0.75), and performance status (p=0.38) were similar between enrolled and non-enrolled patients. Distance did not impact enrollment, as nearly half of patients in both groups lived >50miles from the treatment center (39.0% vs 47.8%, p=0.36). Mode of chemotherapy administration significantly differed based on participation (all p<0.05). Despite similar residual disease status (p=1.00) and number of chemotherapy regimens received (p=0.59), patients treated on trial had a higher 3-year survival rate (70.7% vs 51.7%, p=0.031). The difference in median progression-free survival approached significance (20.2 vs 9.2months, p=0.091). CONCLUSION In an institution where the culture is to offer clinical trials to all eligible patients, 33% of front-line EOC patients did not participate. Increasing age was associated with non-participation. Modifiable barriers must be overcome so that trial enrollment can better reflect true EOC demographics.

Evaluation of the efficacy and toxicity profile associated with intraperitoneal chemotherapy use in older women

OBJECTIVE Intraperitoneal (IP) chemotherapy (CT) for treatment of epithelial ovarian cancer (EOC) has been shown to provide a substantial OS advantage. This study aims to compare the toxicity and benefits of IP CT in patients ≥70 with those <70. METHODS We performed a single institution retrospective review of patients diagnosed with Stage IIA-IIIC EOC from 2000 to 2013 who received IP CT. Clinicopathologic characteristics were extracted, and survival was calculated. RESULTS 133 patients were included with 100 pts. <70years old and 33 pts. ≥70years old. Clinical trial enrollment was similar despite age. In trial enrolled patients, older patients received statistically fewer cycles of therapy (6.4 vs 5.8, p=0.002) but had similar dose delays (0.9 vs 0.7, p=0.72), and modifications (0.9 vs 0.36, p=0.11). Median PFS (27 vs 31months) and OS (71 and 62months) were not statistically different. Grade 3/4 neutropenia was significantly worse in the older patients (82% vs 100%, p=0.04). Neuropathy grade ≥2 and other non-hematologic toxicities were not different between age groups. CONCLUSIONS Despite completing fewer cycles of IP CT, older EOC patients had comparable survival to younger patients. The population of older patients receiving IP CT in this study were on clinical trial and likely to be heartier than the general older population. IP CT appears well tolerated and effective among select older patients and is likely under-utilized outside of clinical trials.

How do palliative care needs vary across the disease trajectory in patients with gynecologic cancer

106 Background: The WHOs integrated model emphasizes that palliative care (PC) should be provided concurrently with curative and life-prolonging care, with disease-directed treatments decreasing and PC increasing over time. This study aimed to understand how accurately this theoretical model matches the PC needs of gynecologic cancer (GC) patients. METHODS GC patients at a comprehensive cancer center completed a symptom and needs assessment questionnaire as part of routine care. Needs were divided into physical, emotional, social, practical, and informational domains. Patients reporting at least one symptom score ≥ 4/10 made the physical need domain positive. For the remainder of the categories, patients endorsing at least one need in a given domain made that domain positive. Time points in the illness trajectory were categorized as diagnosis, primary treatment, remission, and progression/recurrence. Univariate analyses were utilized to assess differences in needs at various time points. RESULTS Questionnaires from 1348 patient visits over 3 months were included. Patients had ovarian (39%), uterine (36%), cervical (16%), and vulvar (5%) cancer. Visits occurred around the time of diagnosis (6%), during primary therapy (28%), during remission (42%), and during disease progression or recurrence (23%). Physical needs were most common at diagnosis and during progression/recurrence (p < 0.01). Emotional needs (p < 0.01), social needs (p < 0.01), and informational needs (p < 0.01) all varied throughout the course of disease and were highest at diagnosis. At each time point, at least 65% of patients had one or more PC needs. Having needs in multiple domains ( ≥ 3) was associated with time point, with patients around the time of diagnosis having the highest rate of need in multiple domains (p < 0.01). CONCLUSIONS GC patients have a broad range of PC needs across the trajectory of their illness. The WHO schema fails to capture the full scope of these needs and under-appreciates their prevalence earlier in the course of the disease. These results underscore the importance of training gynecologic oncologists in primary PC as well as increasing referrals to specialist PC providers.

Abstract A79: Comprehensive genomic profiling (CGP) of adult granulosa cell tumors (aGCT) identifies clinically relevant genomic alterations (CRGA) and targeted therapy options

Background: Adult granulosa cell tumors account for ∼70% of sex cord stromal tumors of the ovary. Most behave as low-grade malignancies and present as stage I disease. However 20% recur, often after a prolonged time (median time to recurrence 6 years). Current therapeutic options (bleomycin/etoposide/cisplatin (BEP), hormone therapy, bevacizumab) have shown modest efficacy in the setting of advanced or recurrent disease. Prior genomic studies of aGCT identified recurrent FOXL2 mutations as characteristic of this tumor type, however this mutation is not currently amenable to targeted therapy. We present the CGP of 70 advanced/recurrent stage aGCT with identification of CRGAs and include a description of a patient response to CGP-matched targeted therapy. Methods: DNA was extracted from 70 FFPE aGCT clinical specimens. Hybridization captured libraries of 236 (FoundationOne, n = 28) or 315 (FoundationOne, n = 42) genes, plus select introns frequently rearranged in cancer, which were sequenced to high (median 780x), uniform coverage. All classes of genomic alterations (base subs, small in/dels, rearrangements, and copy number alterations) were evaluated and reported. CRGA were defined as GA associated with on-label targeted therapies and targeted therapies in mechanism-driven clinical trials. Results: 70 samples, 10% from primary site tissue and 90% from metastatic sites were included. The patients were women aged 30-80 (median 56.5y) with predominantly advanced stage aGCT. 68 cases (97%) had the FOXL2 402C>G mutation and 61% of cases had at least one additional GA (total n = 165) including 37 different genes (avg 2.4 GA per tumor) of which 27 were CRGA (avg 0.44 per tumor). 40% of aGCT cases featured ≥ 1 CRGA, including 14 (20%) cases with CRGA in the PI3K/Akt/mTOR pathway. The most common CRGAs observed were: MLL2 (10%),PIK3CA (8.6%), CDKN2A/B (8.6%), AKT1 (4.3%), KRAS (4.3%); and NRAS (2.8%). To date, we are aware of one patient with an AKT1 missense mutation in the pleckstrin homology domain (AKT1 Q79K) who showed a durable 6-month partial response to AKT-directed targeted treatment. Conclusions: Almost two-thirds of advanced stage aGCT demonstrate GAs in addition to the pathognomonic FOXL2 mutation. 40% of cases demonstrate other targetable mutations, most commonly in the PI3K/Akt/mTOR pathway which present the opportunity for targeted therapy. This first account of a clinical response to CGP-directed targeted therapy in aGCT demonstrates potential efficacy of an AKT inhibitor in a subset of patients suffering from this otherwise treatment refractory tumor. Patients who have exhausted other SOC therapy for metastatic aGCT may similarly achieve clinical benefit from CGP-directed therapeutic decision- making and this provides support for development of further mutation-matched therapeutic trial designs. Citation Format: Michelle Rowland, Scott McMeekin, Kathleen Moore, Mark Bailey, Siraj M. Ali, Rosemary Zuna, Jo-Anne Vergilio, James Suh, Juliann Chmielecki, Garrett M. Frampton, Doron Lipson, Philip J. Stephens, Vincent A. Miller, Jeffrey S. Ross, Julia A. Elvin. Comprehensive genomic profiling (CGP) of adult granulosa cell tumors (aGCT) identifies clinically relevant genomic alterations (CRGA) and targeted therapy options. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr A79.

Predictors of referral to outpatient specialty palliative care (SPC) in gynecologic cancer (GC) patients.

159 Background: Prior studies in GC patients have described predictors of inpatient palliative care (PC) consultation, but predictors of outpatient SPC consultation have not been elucidated. We sought to identify factors predictive of referral and associated care outcomes. METHODS We performed a cross-sectional study of GC patients seen in the gynecologic oncology clinic at a comprehensive cancer center over a three month period. As a part of routine care, patients completed a symptom questionnaire. Patients previously seen at the outpatient PC clinic were compared to those who had not with respect to demographics, disease characteristics, symptom scores, and provider factors using univariate statistics. A multivariate model was created to identify independent predictors of referral. RESULTS 913 patients completed the symptom survey. 76 patients (8%) had been seen in the outpatient PC clinic. Disease factors associated with referral included site (p < 0.01), stage (p < 0.01), evidence of disease (p < 0.01), active treatment (p < 0.01), and time point in the disease trajectory (p < 0.01). Women with moderate to severe pain (p < 0.01), sadness (p = 0.03), distress (p < 0.01), fatigue (p < 0.01), neuropathy (p = 0.03), and sexual dysfunction (p < 0.01) were more likely to have seen PC. Marital status, number of symptoms, and patient provider were also predictive of referral (all p < 0.01). In a multivariate model, site, stage, number of symptoms, moderate to severe sexual dysfunction, and provider were independently associated with referral. Compared to women who had not been referred, patients seen in the PC clinic were more likely to have a health care proxy documented in the electronic medical record (p < 0.01). Among patients with related symptoms, patients referred to PC more often had an opioid prescribed for pain (p < 0.01) and medications prescribed for depression (p < 0.01), anxiety (p = 0.04), insomnia (p < 0.01), and fatigue (p < 0.01). CONCLUSIONS Women with depression, anxiety, insomnia, and fatigue were more likely to receive pharmacologic treatment for these symptoms from a SPC provider. Future research should identify referral triggers for those patients most likely to benefit from outpatient SPC consultation.