Evaluating the role of aromatase inhibitors (AIs) in the treatment of endometrial stromal sarcomas (ESS)

5575

Background: Endometrial stromal sarcomas (ESS) account for < 20% of uterine sarcomas. They usually express estrogen and progesterone receptors (ER/PR) and are considered hormone sensitive. Due to the rarity of these tumors, large clinical trials studying optimal treatment have not been possible. This study represents the largest retrospective study of ESS treated with AI. Methods: The clinicopathological variables and outcomes of patients (pts) with pathologically confirmed low grade ESS treated with AI at our institution between 1998-2020 were recorded. Results: 48 pts with ESS treated with AI were identified. They had a median age of 54 years (range 23-84) and BMI of 27 (range 20-50). 79% were white. 6 (12%), 9 (19%), 14 (29%) and 19 pts (40%) had stage 1,2,3,4 ESS, respectively. 37 (77%) were ER+/PR+; 2 (4%) ER+/PR- and 9 pts (19%) had unknown ER/PR status. All pts were postmenopausal at AI initiation. 12 pts (25%) had a synchronous cancer (5 of these had breast cancer {3 of the 5 presented post tamoxifen}). 23 pts (48%) received megestrol acetate and 25 (52%) an AI as first line hormonal manipulation. During their disease course, 35 pts (73%) received letrozole, 21 (44%) anastrozole and 19 (39.6%) exemestane. 22 pts (46%) were treated with more than one AI. 28 pts (58%) reported side-effects; arthralgia (33%) being the most common. 10 pts (21%) discontinued AI due to toxicity; 12 pts (25%) switched AI for toxicity (with improved tolerance in 67% of these pts). Among the 24 pts (50%) with measurable disease there were 2 partial responses (objective response rate of 8.3%). 1-year disease control rate (DCR) was (79%) for all pts and 58% in stage 4 disease. Median PFS for 1st line AI was 161.6 months (95% CI 48.5 to 274.7). Conclusions: This study represents the largest study of AI use in ESS to date. We found the ORR to be more modest than previously reported. The majority of pts had prolonged stable disease with a DCR of 58% even in stage 4 disease. Pts who progress on one AI may benefit from trial of a 2nd AI. A phase 2 study of interruption versus maintenance AI in locally advanced/metastatic ESS is currently underway (NCT03624244).

Early age of onset and broad cancer spectrum persist in MSH6 and PMS2-associated Lynch syndrome

10516

Background: Recently updated NCCN guidelines for management of germline MSH6 and PMS2 pathogenic/likely pathogenic (P/LP) variant carriers suggest a more modest phenotype with later onset colorectal cancer (CRC) and limited extra-colonic cancers compared to other Lynch Syndrome (LS) genes. However, data are limited, and a comprehensive understanding of the risk spectrum and age of cancer onset is critical for cancer screening and risk-reduction. We sought to characterize MSH6 and PMS2-associated cancers and age of diagnosis in those with mismatch repair deficient (MMRD)/ microsatellite instability-high (MSI-H) tumors, a hallmark of LS pan-cancer. Methods: Pts consented to IRB-approved protocols of tumor/germline sequencing or to a prospective registry of LS pts at a single institution from 2/2005-01/2021 were reviewed to identify those with germline heterozygous MSH6/ PMS2 P/LP variants; pts with constitutional MMRD (CMMRD) were excluded. In cancer-affected pts, tumors were evaluated for MSI and/or MMR protein expression via immunohistochemistry. Tumor types were tabulated, and clinical variables were correlated with MMR/MSI status using non-parametric tests. Results: We identified 243 pts (159 female, 94 male) with P/LP germline MSH6/PMS2 variants, and 186 (77%) pts had a confirmed cancer [ MSH6 111/148 (75%); PMS2 75/95 (79%)]. Overall, 51 (21%) pts had multiple primary cancers, 35 (24%) in MSH6 and 16 (17%) in PMS2 (p = 0.20), resulting in 261 total tumors, 160 in MSH6 and 101 in PMS2. Of the 191 tumors with molecular assessment, 118 (62%) were MMRD/MSI-H, including CRC (n = 54), endometrial (EC, n = 34), small bowel (SBA, n = 6), ovarian (OC, n = 5), urothelial (n = 5), pancreato-biliary (n = 4), gastroesophageal (n = 3), non-melanoma skin (n = 3), prostate (n = 2), breast (n = 1), and brain (n = 1). While CRC and EC were more likely to be MMRD/MSI-H (79% each) compared to other cancers (37%) (p < 0.001 overall, p = 0.001 for MSH6, and p < 0.001 for PMS2), 25% of all MMRD/MSI-H tumors in both genes were comprised of non-CRC/EC cancers. Notably, there were 6 SBAs (5 in PMS2, 1 in MSH6), and all were MMRD/MSI-H. There were 17 OCs (12 in PMS2, 5 in MSH6), and of the 12 that underwent molecular assessment, 5 (41.7%) were MMRD/MSI-H (3 PMS2, 2 MSH6). Among MMRD/MSI-H CRC and EC, median age of diagnosis was 51.5 (range 27-80) and 55 (range 39-74) respectively, with 9/54 (17%) of CRC (4 in MSH6, 5 in PMS2) diagnosed < age 35, the suggested upper threshold for initiation of colonoscopy per NCCN. Conclusions: Despite being lower penetrance LS-associated genes, pts with MSH6/PMS2 P/LP variants remain at risk for a broad-spectrum of cancers and very early-onset CRC, with 17% of MMRD/MSI-H CRC presenting prior to upper threshold of initiation of colonoscopic screening per NCCN.

Olaparib treatment (Tx) in patients (pts) with platinum-sensitive relapsed ovarian cancer (PSR OC) by BRCA mutation (BRCAm) and homologous recombination deficiency (HRD) status: Overall survival (OS) results from the phase II LIGHT study

5515

Background: LIGHT (NCT02983799) evaluated olaparib Tx in pts with PSR OC in cohorts with known BRCAm and HRD status. We report the final OS analyses. Methods: We conducted anopen-label, non-randomized, multicenter study of pts with PSR OC and ≥1 prior line of platinum-based chemotherapy (CTx). Pts were assigned to one of 4 cohorts: germline (g) BRCAm; somatic (s) BRCAm; HRD +ve (non-BRCAm); and HRD -ve. Genomic instability score (GIS) and gBRCAm status were determined by Myriad myChoice and BRACAnalysis CDx tests, respectively. HRD +ve tumors were defined by a GIS ≥42. Pts received olaparib Tx (starting dose 300 mg bid) until disease progression or unacceptable toxicity. OS was a secondary endpoint and was analyzed at 12 months (mo) after the primary analysis and 18 mo after the last pt was enrolled. Safety was assessed in pts who received ≥1 dose. Results: Data cut-off (DCO) was Aug 27, 2020. Of 272 enrolled pts, 271 received olaparib; of these, 270 had measurable disease at baseline and were included in efficacy analyses (Table). At DCO, 40% of pts had died (maturity) with a median follow-up in censored pts of 26.3 mo. Kaplan–Meier 18-mo OS rates were 86%, 88%, 79%, and 60% in the gBRCAm, sBRCAm, HRD +ve (non-BRCAm), and HRD -ve cohorts, respectively. Platinum-based CTx was the most common first subsequent Tx and was received by 39% pts after olaparib discontinuation. At DCO, the median duration of Tx was 7.4 mo and 244 pts had discontinued treatment, mainly due to disease progression (72%); 5% discontinued due to treatment-emergent adverse events (TEAEs). The only TEAE leading to discontinuation in >1 pt was nausea (2 pts). Serious TEAEs were reported in 25% of pts. The most common serious TEAE was small intestinal obstruction (6%). Three adverse events of special interest occurred, each in 1 pt (<1%): acute myeloid leukemia (post discontinuation), pneumonitis, and pulmonary fibrosis. Conclusions: In the final OS analysis, 18-mo OS ranged from 60–88% across the 4 cohorts. Consistent with the primary analysis, the 18-mo OS rate was highest in the BRCAm cohorts (similar OS in g and sBRCAm); among pts without a BRCAm, 18-mo OS was highest in the HRD +ve cohort. No new safety signals were observed compared with the primary analysis and with prior olaparib studies. Clinical trial information: NCT02983799. [Table: see text]

Prevalence and clinical characterization of MMR-D/MSI extra-colonic cancers among germline PMS2 mutation carriers

1527

Background: PMS2-associated Lynch syndrome (LS) may have a more modest phenotype than that associated with other mismatch repair (MMR) genes ( MLH1, MSH2, MSH6, EPCAM). Recent studies suggest limited extra-colonic cancers, and modified risk-reducing measures can be provided. Understanding the spectrum of risk is of critical importance as some LS-associated cancers do not have effective screening, requiring risk-reducing surgery (endometrial, ovarian). As MMR-deficiency (MMRD)/ microsatellite instability (MSI) is associated with LS pan-cancer, we sought to characterize PMS2-associated malignancies according to MMR/MSI status. Methods: Review of cancer patients (pts) consented to an IRB-approved protocol of tumor/germline next-generation sequencing (NGS) identified 43 germline heterozygous PMS2 mutation carriers. Tumors were evaluated for MSI via MSIsensor and/or corresponding MMR protein expression via immunohistochemical staining (IHC). Clinical variables were correlated with MMR/MSI status, comparing via Chi-square or standard T-test. Results: There were > 10 tumor types; 69.8% (30/43) were extra-colonic cancers (endometrial (n = 4), ovarian (n = 6), small bowel (n = 3), urothelial (n = 2), pancreas (n = 3), prostate (n = 3), breast (n = 3), brain (n = 3), biliary (n = 1), spindle cell sarcoma (n = 1), and hepatoblastoma (n = 1)). 46.5% (20/43) of tumors were MMRD/MSI. 61.5% (8/13) of colorectal cancers (CRC) were MMRD/MSI, compared to 40% (12/30) of extra-colonic tumors. All endometrial and small bowel cancers were MMRD/MSI. Of 6 ovarian cancers, 3 were clear-cell, 1 endometrioid, and 2 high-grade serous (HGS). The only MMRD/MSI ovary tumor was HGS. 73.9% (17/23) of pts with MMRP/MSS tumors had recurrent/metastatic disease vs 30% (6/20) of pts with MMRD/MSI tumors ( p= 0.004). Mean age at diagnosis did not differ significantly between MMRP/MSS and MMRD/MSI groups (49 vs. 57, respectively, p= 0.146). 11.6% (5/43) of pts had a prior cancer, with only one patient having prior CRC. Pts with extra-colonic tumors were less likely to meet clinical pt and family history LS testing criteria than those with CRC (63.3% (19/30) vs. 7.7% (1/13); p< 0.001). Conclusions: While PMS2-related LS may have a more modest clinical phenotype, in this single-institution study, 60% (12/20) of patients with MMRD/MSI tumors presented with extra-colonic cancers. We caution counseling pts with PMS2-associated LS about reduced extra-colonic risk until more complete information about penetrance, spectrum, and age distribution of cancer is available.

Prospective agnostic germline testing in pediatric cancer patients

1589

Background: We report our large cohort of pediatric cancer patients undergoing prospective agnostic germline sequencing. Our dataset is a significant addition to the 1,573 children reported to date who have undergone agnostic germline sequencing in previous large sequencing studies, each with ascertainment bias. Methods: 676 patients with pediatric solid tumors underwent matched tumor-normal targeted DNA sequencing from July 2015 to February 2020. At least 76 genes associated with cancer predisposition were analyzed in the germline, and variants were classified per American College of Medical Genetics guidelines. Pathogenic and likely pathogenic (P/LP) variants were reported to patients/families, who were offered genetic counseling and cascade testing with screening recommendations and referral to a surveillance clinic as appropriate. Results: One or more P/LP variants were found in 17% (115/676) of individuals when including low, moderate and high penetrance mutations in recessive and dominant genes, or 12% (81/676) when including moderate and high penetrance mutations in dominant genes. P/LP variants were detected in 40% (21/53) of patients with retinoblastomas, 8% (13/161) with neuroblastomas/ganglioneuroblastomas, 13% (14/112) with brain/spinal tumors, 8% (20/245) with sarcomas, and 12% (13/105) with other solid tumors. The most frequent mutations were in RB1 (n = 28) and TP53 (n = 8) in patients with associated tumors. Of patients with moderate/high penetrance mutations, 30% (24/81) had unexpected tumor types, with potential therapeutic relevance in 58% (14/24) including BRCA1 n = 2, BRCA2 n = 3, RAD51D n = 1, ATM n = 1 MLH1 n = 1, MSH2 n = 1, MSH6 n = 1, PMS2 n = 3, and SUFU n = 1. Two patients received immunotherapy based on their germline finding. Conclusions: P/LP germline variants are frequently present in patients with pediatric cancer. We are contributing significantly to the cohort size of agnostic sequencing in pediatric cancers. Our experience is similar to other studies with a ~12% detection rate of moderate and high penetrance mutations. Moderate/high penetrance mutations were concordant with the patient’s cancer history in 70% of cases, higher than previously reported, likely due to an enrichment of retinoblastoma. While many mutations are identified in patients with associated tumor types, a large proportion of mutations are unexpected based on the patient’s history. Clinical actionability of these findings may include screening, risk reduction, family planning, and increasingly targeted therapies.

The impact of poly ADP ribose polymerase (PARP) inhibitors on clonal hematopoiesis

1513

Background: Poly (ADP-ribose) polymerase (PARP) inhibitors are an important new class of anti-cancer therapies. Therapy-related myeloid neoplasia (tMN) has been reported following PARPi therapy, and is associated with adverse outcomes. Further insight is required into the risk of tMN conferred by PARPi therapy, independent of germline genetic background and prior therapy. We have shown that oncologic therapy selects for acquired mutations in the blood (clonal hematopoiesis; CH) particularly those in the DNA damage response pathway (DDR) including PPM1D, TP53 and CHEK2 and that CH confers an increased risk of tMN. We hypothesized that characterization of the relationship between CH and PARPi therapy provides insight into its potential for leukemogenesis and may offer opportunities for tMN prevention. Methods: We assessed for CH in the blood of 10,156 cancer patients, including 54 who received PARPi therapy, 5942 who received another systematic therapy or radiation therapy and 4160 untreated prior to blood draw. Results: Patients exposed to PARPi therapy were more likely to have CH (33%) compared to those exposed to other systemic therapies or radiation (18%) or untreated patients (16%). This was particularly pronounced for DDR CH; 25% of PARPi treated patients had DDR CH compared to 2% of untreated patients. In a multivariable model accounting for demographics, exposure to chemotherapeutic agents, radiation therapy and germline BRCA mutation status, exposure to PARPi conferred an increased risk of DDR CH (OR = 3.6, 95% CI 1.5-8.5, p = 0.004). This effect was attenuated after accounting for cumulative exposure to therapy (OR = 2.8, 95% CI 0.97-8.2, p = 0.06) suggesting a multifactorial contribution to the enrichment of CH following PARPi therapy. To characterize this further we performed a prospective collection of patients with CH over a median follow-up time of 58 months. During the follow-up period, 17 patients received PARPi, 360 received cytotoxic therapies or radiation and 232 were untreated or received targeted therapies. The growth rate of DDR CH was significantly higher among those who were exposed to PARPi (median, +2.8% increase in VAF per year) compared to untreated patients (+0.08% per year, p = 0.02) and those exposed to other cytotoxic therapies (+1% per year, p = 0.04). Conclusions: Taken together our data suggests that PARPi therapy promotes the expansion of DDR CH. Future studies should examine the potential of CH to identify individuals at high risk of tMN following PARPi therapy and to develop therapies aimed to prevent tMN in patients with CH.

Olaparib treatment in patients (pts) with platinum-sensitive relapsed (PSR) ovarian cancer (OC) by BRCA mutation (BRCAm) and homologous recombination deficiency (HRD) status: Phase II LIGHT study

6013

Background: In Study 19 (NCT00753545), olaparib capsules demonstrated improvement in progression-free survival (PFS) vs placebo in the PSR OC maintenance setting, irrespective of BRCAm status (Ledermann et al. Lancet Oncol 2014). LIGHT is the first prospective study to evaluate olaparib tablet treatment in PSR OC pts by BRCAm and HRD status. Methods: This is an open-label, non-randomized study (NCT02983799) that assessed efficacy and safety of olaparib monotherapy (300 mg BID) in pts with PSR, high-grade serous/endometrioid epithelial OC and ≥1 prior line of platinum chemotherapy. Pts were assigned to one of four cohorts: germline (g) BRCAm; somatic (s) BRCAm; HRD+ve (non-BRCAm); HRD–ve; by Myriad BRACAnalysis CDx and myChoice tests. HRD+ve was a score ≥42. Primary endpoint was objective response rate (ORR). Secondary endpoints included: disease control rate (DCR) and investigator-assessed PFS (RECIST v1.1). Primary analysis was to be ~6 months (mo) after the last pt was enrolled. Results: Data cut off was 8/27/19. Of 271 pts treated (median of 31.7 weeks [2.1–96.0]), 270 had measurable disease at baseline and were included in efficacy analyses (Table). The most common treatment-emergent adverse events (AEs) were nausea (66%) and fatigue (62%).Serious AEs and Grade ≥3 AEs were experienced by 25% and 44% of pts, respectively. AEs leading to olaparib dose interruptions, reductions and discontinuations occurred in 33%, 24% and 4% of pts, respectively. Conclusions: Olaparib treatment demonstrated activity across all cohorts. As observed in the maintenance setting, similar efficacy was seen in the gBRCAm and sBRCAm cohorts. For non-BRCAm pts, longer median PFS and higher ORR were observed in the HRD+ve cohort. Olaparib treatment was well tolerated with no new safety signals identified and a safety profile consistent with that seen in the PSR and first-line settings. Clinical trial information: NCT02983799. [Table: see text]

Targeted therapy based on germline analysis of tumor-normal sequencing (MSK-IMPACT) in a pan-cancer population

1500

Background: Tumor mutational profiling for identification of somatic alterations for targeted treatment is increasingly being performed in advanced cancer patients (pts). We sought to assess the clinical utility of germline mutation profiling for targeted therapeutic interventions in a pan-cancer patient population. Methods: All pts who had germline genetic testing through a prospective protocol via a next-generation sequencing panel (MSK-IMPACT) were identified (N=11,975) from 2015-5/2019. The medical record of pts with likely pathogenic/pathogenic germline (LP/P) alterations in genes with known therapeutic targets were reviewed to identify germline-targeted treatment either in a clinical or research setting. Results: We identified 2,043 (17.1%) pts who harbored LP/P variants in a cancer predisposition genes including 777 (6.5%) in genes with potentially targetable therapeutic implications: 416 BRCA1/2, 149 DNA mismatch repair genes (Lynch syndrome, LS), 122 ATM, 45 PALB2, 26 RAD51C/D, 7 RET, 4 TSC, 3 PTCH1, 2 ALK, 1 EGFR, 1 MET and 1 KIT. Of those with advanced disease (n=554), 45.3% received targeted therapeutic treatment (Table) including 50.9% BRCA1/2, 58.3% LS (67.4% of microsatellite-high LS cases), 41.7% PALB2, 36.8% RAD51C/D and 19.3% ATM carriers. Of patients receiving a poly (ADP-ribose) polymerase inhibitor (PARP-I) in the setting of a BRCA1/2 mutation, 55.1% had breast or ovarian cancer; however, 44.8% had other tumors, including pancreas, prostate, bile duct, gastric, wherein the drug was given in a research setting. Among PALB2 pts receiving PARP-Is, 53.3% (8/15) had breast or pancreas cancer; 46.7% had cancer of the prostate, ovary or unknown primary. Conclusions: In our pan-cancer analysis, 6.5% of pts harbored a targetable germline variant highlighting the importance of germline analysis in advanced cancer pts for selection of both FDA-approved treatments and clinical trial participation with germline-targeted therapeutics. [Table: see text]

Basket study of the oral progesterone antagonist onapristone ER in women with progesterone receptor positive (PR+) recurrent granulosa cell tumor (GCT), low-grade serous ovarian cancer (LGSOC), or endometrioid endometrial cancer (EEC)

TPS6098

Background: Onapristone extended release (ER) is a type I full progesterone antagonist that inhibits progesterone mediated PR activation and stabilizes PR association with corepressors. Onapristone has shown activity across multiple preclinical models of hormonally driven cancer. A phase I dose escalation study of onapristone ER in PR+ breast, endometrial and ovarian cancer patients found all doses tested to be well tolerated, with 50mg PO BID determined to be the recommended phase 2 dose (RP2D). GCT (98% of cases PR+), LGSOC (58% of cases PR+) and EEC (67% of cases PR+) are hormonally driven cancers which generally have poor responses to chemotherapy and limited treatment options in the recurrent setting. Methods: This is an open-label, investigator-initiated basket study of onapristone ER in patients with PR+ recurrent GCT, LGSOC, or EEC currently enrolling patients at Memorial Sloan Kettering Cancer Center in NY, USA (NCT03909152). The primary objective is to evaluate the efficacy, in terms of response rate by RECIST 1.1 criteria, within 36 weeks of treatment. Eligible patients must have received at least 1 prior line of chemotherapy, have measurable disease by RECIST 1.1 criteria, and have tumor tissue collected within 3 years prior to enrollment with PR expression ≥ 1% by IHC. Patients are allowed to have unlimited additional prior lines of chemotherapy, biologic therapy, immunotherapy or hormonal therapy. Enrolled patients are treated with onapristone ER 50mg PO BID until time of progression or intolerable toxicity. The 3 disease cohorts are currently enrolling to Stage I in parallel with expansion from stage I to stage II planned when the prespecified response criteria are met for each cohort as described in the table below. Clinical trial information: NCT03909152. [Table: see text]

Characterization of patients with multiple primary tumors

1502

Background: 17% of patients (pts) with a cancer diagnosis in the U.S. have a prior malignancy. We sought to characterize pts with multiple (≥2) primary cancers (MPC) & identify potential drivers of cancer risk to guide management. Methods: Pts prospectively consented (1/2013-2/2019) to tumor-normal sequencing via custom targeted NGS panel. A subset consented to testing of >76 germline cancer predisposition genes. IARC 2004 rules for defining MPC were applied.Age adjusted gender specific standardized incidence ratios (SIR) for cancer event combinations occurring in at least 5pt were calculated using R statistical package. Results: Of 24417 pts sequenced, 4341 had MPC (18%). (Table) 3465 (80%) had 2, 4% had >4 cancers. Cancer pairs where SIR of 2nd cancer was higher than expected included: colon-colon, prostate-pancreas, bladder-prostate in men & lung-lung, breast-pancreas, thyroid-pancreas in women. 1580 (36%) pts had germline testing; 324 (21%) had 361 pathogenic/likely pathogenic (P/LP) variants (vts). Of these, 157 (48%), 66(20%), pts had high, moderate penetrance vts. The remainder had low penetrance, recessive or vts of uncertain utility. Of pts with high penetrance vt, 132 (84%) had at least one tumor type concordant with germline findings. Conclusions: 18% of pts in this cohort had MPC. There was a significant excess over expected incidence in some cancer combinations. Of pts with germline testing, 21% had a P/LP vt, with most (69%) being high or moderate penetrance. Assessment for loss of heterozygosity in tumor & germline sequencing of the full MPC cohort is ongoing. [Table: see text]

Gastric-type adenocarcinoma of the cervix: Genomic drivers and clinical outcomes

6030

Background: Cervical Gastric-type adenocarcinoma (CGA) is a non-HPV-associated adenocarcinoma, comprising 10% of all cervical adenocarcinomas (Park et al, 2019). The optimal management approach is unclear, given that most data in advanced cervical cancer is driven by HPV-positive disease. We summarize our experience with this rare tumor type at a large cancer center. Methods: A retrospective review was performed for all women diagnosed with CGA 6/1/2002- 7/1/2019. Patients who did not follow up after a single visit were excluded. Kaplan-Meier survival analysis was performed to determine progression-free survival (PFS) and overall survival (OS) from date of diagnosis. Tumors from a subset of patients were subjected to MSK-IMPACT targeted sequencing and analysis (Zehir et al, 2017). Results: A total of 68 women were identified; 47 met inclusion criteria. The median age at diagnosis was 52 years (range 27-83). The majority of patients were white (70%), an additional 19% were Asian. The majority of patients (60%, n=28) presented with advanced disease (FIGO 2018, stage II-IV), while 40% (n=19) were Stage I. Of note, 26% (n=12) had positive pelvic lymph nodes and 13% (n=6) had ovarian metastases at time of surgical resection. For upfront treatment: 13% (n=6) had surgery alone of whom 83% had stage 1 disease, 36% (n=17) had surgery followed by adjuvant therapy, 30% (n =14) received definitive chemo-radiation (CRT). All patients with stage IV disease 15% (n=7) received chemotherapy alone. At completion of primary treatment, 19% (n=9) of patients had persistent disease. In patients who received CCRT, 65% (n=22) recurred, the majority (64%) within 12 months of completion of upfront therapy. Pelvic recurrence was the most common site (n=14, 64%). With a median follow up time of 30 months (range 1-159), the median PFS for Stage I was 34.4 months, compared to 17.5 months in patients with Stage II-IV disease (p= 0.29). Of the 24 patients that had MSK-IMPACT, the most common mutation was TP53 (n=16, 64%) followed by mutations in the RAS pathway (n=8, 33%), PIK3CA (n=3, 12.5%), STK11 (n=3, 12.5%), and ERBB2 alterations (n=2, 8.3 %). 2 (8.3%) women enrolled on a clinical trial based on their NGS results, one targeting ERBB2 and one targeting PIK3CA. Conclusions: Consistent with prior published literature, CGA is an aggressive form of cervical cancer with poor median OS in the advanced setting. With universal HPV vaccination, HPV negative cervical cancer will represent a larger percentage of newly diagnosed cancers and further research is needed to identify the optimal management approach.

A phase II randomized study of avelumab plus entinostat versus avelumab plus placebo in patients (pts) with advanced epithelial ovarian cancer (EOC)

5511

Background: Preclinical evidence suggests that combining avelumab (A), a human anti-PD-L1 monoclonal antibody, and entinostat (E), a class I selective histone deacetylase (HDAC) inhibitor, may increase tumor immunogenicity and responsivity to checkpoint inhibition. This study evaluated whether A+E would lead to improved progression free survival (PFS) vs A in pts with advanced EOC. Methods: Pts with EOC which had progressed or recurred after 1st-line platinum-based chemotherapy and received 3- 6 lines of therapy were randomized 2:1 to receive A (10 mg/kg IV Q2W) plus E (5 mg PO QW) or A plus placebo (P). Treatment continued until disease progression (PD) or unacceptable toxicity. The primary endpoint was PFS (investigator-assessed, RECIST 1.1), stratifying on the presence/absence of bulky disease (tumor ≥ 50 mm) and platinum-refractory disease. The hypothesis was that the combination would reduce the hazard of PD or death by 43%, representing a 75% improvement in median PFS. 97 events (from 120 pts) provided 90% power with 1-sided significance level of 0.10. Secondary endpoints included ORR, duration and time to response, toxicity, clinical benefit rate, and OS. Results: 126 pts were enrolled, median age 63 yrs (range 43-82), median 4 prior lines, 83% serous histology. Median PFS was 1.64 and 1.51 mos for A+E and A+P, respectively (p = 0.31; HR 0.90, 95% CI: 0.58-1.39). No significant differences in ORR (6% vs 5%), or OS (NE vs 11.3 mos) were observed. 4 pts (3%) had clear cell EOC, with no responses observed. The incidence of related adverse events (AEs) was higher in the A+E arm compared to A+P (any grade: 93% vs 78%, Grade 3/4: 41% vs 10%), and the most frequent (≥20%) related AEs with A+E were fatigue (46%), nausea (31%), diarrhea (26%), anemia (26%), and chills (20%). Grade 3/4 related AEs occurring in ≥5% with A+E were fatigue (9%), and neutropenia (8%). 47% of pts in A+E arm required E dose holds/reductions. Discontinuations due to AEs were similar between arms (21% vs 17.5% for A+E and A+P, respectively), as was duration of study therapy (median 4 and 5 cycles started). Conclusions: In pts with heavily pretreated EOC, median PFS was not prolonged when E was added to A compared to A alone and the combination resulted in greater toxicity. Clinical trial information: NCT02915523.

Response to immune checkpoint inhibition and survival in BRCA-associated recurrent ovarian cancer

2615

Background: Alterations in the DNA mismatch repair pathway increase susceptibility to immune checkpoint inhibition (ICI). Tumors with BRCA-related DNA repair defects may have increased antigenicity, which could drive response to ICI. Responses to ICI in ovarian cancer (OC) have been modest. We seek to evaluate the association of BRCA mutations with response to ICI and survival in recurrent OC. Methods: A single-center, retrospective review identified 103 women with recurrent OC and known BRCA mutation status (90 germline and 33 somatic testing) who received ICI between 1/2013-7/2018 (98 on study). Clinical characteristics and duration of ICI (Long > = 24 vs. Short < 24 weeks) were compared by BRCA status. Kaplan Meier survival analysis was used to calculate progression-free (PFS) and overall survival (OS) from start of ICI, and CoxPH models/logrank test were used to assess survival differences by BRCA status. Results: Deleterious germline (g) or somatic (s) BRCA 1/2 mutations were present in 29 (28%) women (12 g BRCA1, 9 g BRCA2, 3 s BRCA1, 5 s BRCA2, 1 g BRCA1/s BRCA1, 3 g BRCA2/s BRCA2, and 1 g BRCA2/s BRCA1). Patients (pts) with BRCA mutations had more lines of treatment prior to ICI (median 5 vs. 4, p = 0.03) and a longer time from diagnosis to ICI (median 54 vs. 38.5 months (mo), p = 0.01), but there were no significant differences in other variables including histology (86% high grade serous), stage at diagnosis (96% Stage III/IV), and platinum status (83% resistant), p > 0.05. Four pts (15%) with BRCA mutations had long duration of ICI as compared with 20 pts (27%) in those without mutations, p = 0.20. Median PFS was 2.2 mo (95% CI 1.7-2.7) in those with BRCA mutations and 3.4 mo (95% CI 2.1-4.0) in those without mutations, HR 1.22 (95% CI 0.78-1.91, p = 0.38). At a median follow-up of 23.3 mo, median OS was 21.3 mo (95% CI 13.7-31.8) in those with BRCA mutations and 19.8 mo (95% CI 13.8-25.3) in those without. This was not significantly different, HR 1.00 (95% CI 0.54-1.87, p = 0.99), after adjustment for prior lines and time from diagnosis to ICI. Conclusions: In our study of heavily pretreated OC pts receiving ICI, BRCA 1/2 mutations were not associated with improved response or survival. These findings should be validated in larger studies.

Adavosertib with chemotherapy (CT) in patients (pts) with platinum-resistant ovarian cancer (PPROC): An open label, four-arm, phase II study

5513

Background: Adavosertib (AZD1775; A), a highly selective WEE1 inhibitor, demonstrated activity and tolerability in combination with carboplatin (C) in primary PROC. This study (NCT02272790) assessed the objective response rate (ORR) and safety of A in PROC. Methods: Pts with recurrent RECIST v1.1 measurable PROC received A with C, gemcitabine (G), weekly paclitaxel (P), or pegylated liposomal doxorubicin (PLD) in 3- (C) or 4-week (G, P, PLD) cycles (Table). Tumor assessments were performed every 2 cycles until disease progression. Primary objective: ORR; other objectives: disease control rate (DCR), progression-free survival (PFS) and safety. Results: In the 94 pts treated (median treatment duration 3 months; range 0–16 months), outcomes were greatest with A (weeks [W]1–3) + C (Table), with ORR of 67% and median PFS (mPFS) of 10.1 months for this cohort. Most common grade ≥3 treatment-emergent adverse events (TEAEs) are shown in the Table, with hematologic toxicity most notable with A (W1–3) + C. TEAEs led to A dose interruptions, reductions and discontinuations in 63%, 30% and 13% of the whole cohort, respectively. A possible positive relationship between CCNE1 amplification and response warrants further investigation. Conclusions: A shows preliminary efficacy when combined with CT. Pts receiving A (W1–3) + C showed greatest benefit. The increased but not unexpected hematologic toxicity is a challenge and could be further studied to optimize the dose schedule and supportive medications. Clinical trial information: NCT02272790. [Table: see text]

A phase II trial of durvalumab with or without tremelimumab in patients with persistent or recurrent endometrial carcinoma and endometrial carcinosarcoma

5582

Background: Monoclonal antibodies Durvalumab (D) and Tremelimumab (T) inhibit binding of programmed cell death ligand 1 (PDL1) to PD1 and inhibit activation of cytotoxic T-lymphocyte-associated protein 4 (CTLA4), respectively, resulting in improved tumor immunosurveillance. There is rationale to study D and DT based on recent genomic and tumor microenvironment evaluation of endometrial cancer (EC). Methods: Eligible patients (pts) were randomized to D or DT. Pts received D 1500 mg intravenously (IV) every 4 weeks (wks). DT therapy pts received D 1500 mg IV every 4 wks and T 75 mg IV every 4 wks for 4 cycles, followed by D 1500 mg IV every 4 wks until progression or unacceptable toxicities. Pts were stratified by histology with 10 carcinosarcoma or MSI-H EC pts per arm. Efficacy assessments were every 8 wks and treatment related adverse events (TRAEs) were assessed per CTCAE v.4.03. The primary endpoint was overall response rate (ORR) by RECIST v1.1. Descriptive statistics and 90% one sided CI are reported. Progression free survival (PFS) rate at 24 wks (PFS24wks) was estimated by Kaplan Meier method. Results: At planned interim analysis, 56 pts were enrolled (28 per arm). 15 pts: carcinosarcoma, 15 pts: endometrioid (3: Gr1), 14 pts: serous, and 12 pts: other histology. 5(9%) pts: MSI-H, 48(86%) pts: microsatellite stable (MSS), 3(5%): unknown. 2 pts were excluded due to early death. 27 pts per arm were evaluable for efficacy. In the D arm: 1 pt had complete response (CR)(MSS) and 3 pts had a partial response (PR) (2:MSS, 1:MSI-H) with an ORR of 14.8% (CI: 6.6-100%). The median PFS was 7.6 wks and PFS24wks was 13.3% (CI 6.1-100%). Median duration of response (DOR) was 16 wks in the D arm. In the DT arm, 2 pts achieved CR (1:MSI-H, 1:MSS) and 1 had PR (MSS). The ORR was 11.1% (CI: 4.2-100%). Median PFS was 8.1 wks, PFS24wks was 18.5% (CI 10.1-100%) and DOR was 8 wks. Grade 3 TRAEs occurred in 2 (7%) pts in D and 9 (32%) pts in DT. Grade 4 TRAEs occurred 1 (4%) pt in D and 3 (11%) pts in DT. 2 pts discontinued due to a TRAE. Most common TRAEs in total were fatigue (23%), diarrhea (20%), nausea (14%), vomiting (13%) and pruritis (11%). Conclusions: D and DT show modest activity in EC. No new safety signals were identified. Second stage accrual is ongoing. Clinical trial information: NCT03015129.

Tumor testing in men with prostate cancer to predict for germline DNA-damage repair mutations

229

Background: Current guidelines recommend genetic testing for patients (pts) who have BRCA1/2 mutations on tumor-only testing, where germline is not subtracted. The clinical utility of this approach, and possible inclusion of other DDR genes associated with cancer susceptibility, has not been examined in men with prostate cancer. Methods: Pts with mostly advanced prostate cancer were prospectively enrolled to a matched tumor-germline DNA sequencing protocol and consented for disclosure of germline results. Germline analysis was done with an institutional, CLIA-certified next generation sequencing (NGS) platform (MSK-IMPACT) and analyzed for likely pathogenic or pathogenic germline mutations in at least 76 cancer susceptibility genes. Clinical data was retrieved from the medical record. We report on the frequencies in the germline and in the tumor of a subset of DDR genes. Results: 1243 men had analysis of both germline and tumor. Median age 64 (range 35-90). 12% had a second malignancy and 40% reported a relative with prostate cancer. 19% were of Ashkenazi Jewish (AJ) descent. 330 (27%) had any mutation in BRCA1/2, ATM, CHEK2, PALB2, RAD51C, RAD51D, MSH2, MSH6, MLH1, PMS2. 127 (10%) had a germline mutation, of which 36% were AJ founder mutations. For each gene, the percentage of mutations found in germline and tumor is shown in the table. Conclusions: Of prostate cancer pts found to have any DDR mutation on tumor-germline testing, 29% had a germline mutation. Proportion of germline mutations was highest for PALB2, CHEK2 and BRCA2. These findings not only support germline testing when BRCA1/2 mutations are found tumor-only sequencing, but also support germline testing when other DDR mutations are seen. Clinical trial information: NCT01775072. [Table: see text]

Somatic tumor profiling of DNA mismatch repair (MMR) deficient endometrial cancers (EC)

e17121

Background: Approximately 20% of EC have loss of MMR ( MSH2, MSH6, MLH1, PMS2) protein expression by immunohistochemistry (IHC). The majority have somatic MLH1/PMS2 loss, driven by MLH1 promoter hypermethylation. For the remaining patients (pts), germline testing for Lynch Syndrome (LS) is recommended. However, half do not have a corresponding germline mut. This is considered “Lynch-like syndrome” (LLS) & clinical management is challenging. We sought to determine if tumor profiling could identify somatic mut potentially underpinning loss of protein expression. Methods: Per institutional standard, all EC, regardless of age or family history undergo reflex LS screening with IHC for MMR protein expression. Pts consented to IRB approved protocols. Tumor-normal sequencing was performed via custom next-generation sequencing panel (MSK-IMPACT). Electronic medical records were reviewed. Results: 16 pt have completed tumor sequencing, median age 53 (35-83), 6 (38%) < 50 yrs at diagnosis. 2 had personal history of additional cancer (DCIS, ovary), none had first degree relative with colon or EC. A mix of EC histologies was represented: 10 endometrioid (all grades), 2 clear cell, 4 mixed. There were no serous cancers. There were median 58 mut (9-546), 14 (88%) had hyper or ultra mutated EC. 5 EC were driven by somatic POLE mut (3 known hotspot, 1 likely pathogenic), all ultra-mutated phenotype, resulting in multiple somatic MSH6 muts with isolated IHC MSH6 loss. 4 EC had MSH2/MSH6 IHC loss with corresponding double somatic mut in MSH2. 5 had one somatic mut corresponding to the MMR protein loss, assessment of LOH in these cases is pending. Two cases are unexplained: 40 yo with IHC MLH1 loss, 47mut; 69 yo, IHC MSH6 loss, 12 mut. In this cohort of LLS, somatic muts were frequently observed in ARID1A (13,81%), PTEN (10,63%) & PIK3CA(9,56%), in keeping with non serous histologies. Conclusions: In line with our prior report that pt with LLS had benign personal & family cancer histories compared with LS pts, we have identified that in 56% of LLS EC either POLE mut or double somatic MMR mut likely underpins the MMR IHC loss. As such, in these LLS EC cases, somatic tumor profiling may help to rule out LS. Further testing is ongoing to increase cohort size.

A phase 2 study to assess olaparib by homologous recombination deficiency status in patients with platinum-sensitive, relapsed, ovarian, fallopian tube, or primary peritoneal cancer

TPS5606

Background: The poly (ADP-ribose) polymerase (PARP) inhibitor olaparib is approved for treatment of patients (pts) with germline BRCA mutations ( BRCAm) and advanced ovarian cancer (OC). BRCA mutations are genetic alterations leading to homologous recombination deficiency (HRD) and tumor susceptibility to DNA-damaging agents, including PARP inhibitors. Loss of genetic heterozygosity, telomeric-allelic imbalance, or large-scale state transitions may identify additional pts who could benefit from PARP inhibitor therapy. Methods: LIGHT is a non-randomized, open-label, phase 2 study to assess the efficacy and safety of olaparib in patient cohorts identified by different HRD genetic tests (NCT02983799). Patients will have platinum-sensitive (progression > 6 mo after the end of the last platinum-based chemotherapy), relapsed, high-grade epithelial ovarian, fallopian tube, or primary peritoneal cancer; ECOG performance status 0–1; and ≥2 prior lines of platinum-based chemotherapy for OC. Key exclusion criteria include prior PARP inhibitor treatment; concomitant use of potent CYP3A4/5 inhibitors or inducers; and symptomatic uncontrolled brain metastases. Patients will be enrolled into 4 cohorts of 30 pts each based upon BRCAm or tumor HRD status determined by genetic test results: germline BRCAm; somatic BRCAm; MyChoice HRD-positive and wildtype BRCAm; MyChoice HRD-negative and wildtype BRCAm. All pts will receive 300 mg olaparib tablets twice daily until disease progression (RECIST v1.1) or unacceptable toxicity. The primary endpoint is investigator-assessed objective response rate (ORR) according to RECIST v1.1 criteria. The maximum precision of the ORR is approximately ±18.7% for 30 pts. Secondary endpoints include duration of response, cancer antigen-125 response rate, disease control rate, progression-free survival, time to any progression, overall survival, and homologous recombination repair gene panel mutation status related to clinical outcome. Planned enrollment of 120 pts at sites in the United States was initiated in December 2016. Clinical trial information: NCT02983799.

Somatic mutations in homologous recombination pathway genes in ovarian cancer

5545

Background: 15% of patients with OC have a germline BRCA 1 / 2 mut & 6% of OC have a somatic BRCA 1/ 2 mut. Inhibitors of poly ADP ribose polymerase (PARPi) are approved in this setting. Data is evolving as to activity of PARPi in BRCA wild type OC with somatic homologous recombination (HR) gene deficiency. We sought to determine the prevalence of somatic alterations in HR genes in an unselected cohort of epithelial OC of any histologic type. Methods: From 03/2014-10/2016 patients consented to an IRB approved protocol for tumor-normal sequencing, via a custom next-generation sequencing panel (MSK-IMPACT) with return of results for tumor mut only. Muts in 14 HR genes ( ATM, BARD1, BRCA1, BRCA2, BRIP1, CHEK1, CHEK2, FANCA, MRE11, NBN, PALB2, PTEN, RAD51D, RAD51C) and p53were assessed. Statistics were analyzed using GraphPad Prism v. 6. Results: 260 tumors were sequenced; 156 (60%) high grade serous (HGS), 34 (13%) low grade serous (LGS), 34 (13%) clear cell, 10 (4%) mucinous, 8 (3%) endometrioid, 18 (7%) mixed/other histology. 151 (97%) of HGS and 24 (23%) of the remaining tumors harbored p53 mut. 48 (18%) somatic HR mut were identified. (Table 1) 26 HGS (17%) tumors and 22 (21%) of the remaining tumors harbored a HR gene mut (HGS vs other histology, p=0.4). Notably, 8 (24%) of clear cell cancers demonstrated HR gene muts (vs HGS, p=0.3). 18 (38%) of the identified muts in the overall cohort were in BRCA 1/2. There were no somatic HR gene muts identified in the mucinous OC, which were characterized by frequent p53 (90%) and KRAS(60%) muts. Conclusions: In this cohort of OC, 17% of tumors harbor somatic HR gene muts. The prevalence of HR somatic muts was similar in HGS vs other histologies, with the exception of mucinous OC. BRCA1/2 muts predominated, however 60% of identified muts were in other genes. Accrual is ongoing to increase histologic cohort sizes. Broad HR gene somatic mutational profiling may identify a wider cohort of OC patients with potential to benefit from PARPi therapy. [Table: see text]

Multi gene panel testing in unselected patients (pts) with endometrial cancer (EC)

e17119

Background: Inherited mutations (muts) in Lynch Syndrome genes (LS) & PTEN are associated with EC. The prevalence of other cancer predisposition genes is unclear. The majority of studies have selected pts by age, family history or specific tumor features. We sought the prevalence of cancer predisposition genes in unselected pts attending for surgical consultation. Methods: 03/2016-10/2016, pts with new EC diagnosis were offered to consent to an IRB approved protocol. Tumor-normal sequencing, was performed via a custom next-generation sequencing panel (MSK-IMPACT) with return of results for 76 cancer predisposition genes. Per institutional standard, all ECs undergo reflex screening for LS with IHC for mismatch repair proteins (MMR P). Results:77 pts consented, median age 60 (27-84), median BMI 27 (16-66), 27% Ashkenazi Jewish (AJ) descent. Tumors: 56 (73%) stage 1, remainder stage 3 or 4, majority (52, 68%) endometrioid histology, of which 31 (60%) grade 1. 15 pathogenic germline variants were identified in 14 pts (18%) including 3 (4%) in LS genes (2 MSH6, 1 MLH1) with corresponding abnormal MMR P. One pt with a known BRCA1 mutation, without prior cancer, with prior risk reducing salpingo-oophorectomy had stage III grade 3 endometrioid EC at 47 yo, tumor LOH at BRCA1 was identified. Of the 4 pts with high-penetrance muts, 3 met criteria for genetic testing for the implicated gene due to personal/family cancer history, 1 pt with MSH6 mutation was identified via absent MMR P only. The remaining 11 pathogenic variants were incremental findings in moderate penetrance ( CHEK2 I157T, MRE11, ATM, APC I1307K, MUTYH) or autosomal recessive genes ( MUTYH, RECQL4, ATM). 5 pt had moderate penetrance variants in known AJ founder muts. Conclusions: In this cohort of unselected EC pts the prevalence of LS was as expected & reflex IHC screening captured all pts appropriately. While all high-penetrance muts were captured by clinical criteria, the incremental identification of moderate penetrance muts in these largely early stage/ low risk EC pts may alter personal & at-risk family member breast & colon cancer screening recommendations. Continuing accrual will reveal the extent to which additional high penetrance genes are seen in unselected EC pts.

A phase II trial of enzalutamide in patients with androgen receptor positive (AR+) ovarian, primary peritoneal or fallopian tube cancer and one, two, or three prior therapies

TPS5610

Background: Approximately 75% of women with epithelial ovarian cancer (OC) present with advanced disease. Most of these women will ultimately recur and require life-long treatment for their cancer. Well tolerated therapies for treatment in the recurrent setting are needed. The AR is expressed in greater than 60% of cases of OC and is more prevalent than the estrogen or progesterone receptor. All past clinical studies of AR inhibition in OC have focused on unselected populations of heavily pretreated women; however preclinical data suggests that AR expression decreases in OC cells with increasing lines of therapy. Enzalutamide is a small molecule androgen receptor-antagonist that is FDA approved for treatment of prostate cancer and is currently being investigated as treatment for breast and ovarian cancer. Methods: This is a phase II, single-institution trial of enzalutamide 160mg po QD in patients with AR+ ovarian, fallopian tube or primary peritoneal cancer. Eligible patients must be found to have greater than or equal to 5% AR staining by IHC on FFPE tumor tissue and been treated with only 1,2 or 3 prior cytotoxic therapies. Patients must have RECIST 1.1 defined measurable disease. Enrolled patients are treated with enzalutamide until progression of disease or unacceptable toxicity. The primary endpoint is to estimate the proportion of women who achieve a complete or partial response by RECIST 1.1 criteria or survive progression free for at least 6 months. Secondary objectives include the retrieval of optional tumor biopsies at time of progression to evaluate the effect of enzalutamide on AR expression and to observe the effect of enzalutamide on serum testosterone and estradiol levels. This study will enroll 58 patients at Memorial Sloan Kettering Cancer Center and its regional sites. The study utilizes a safety lead-in phase and a two-stage design. The first patient enrolled in April 2015. The safety lead-in phase has been completed. The prespecified activity goal for the first stage was met; second stage accrual began in October 2016. Thus far, 35 patients have initiated treatment. Clinical trial information: NCT01974765.

A phase I dose-escalation study of intraperitoneal (IP) cisplatin, IV/IP paclitaxel, IV bevacizumab, and oral olaparib for newly diagnosed adenxal carcinoma

5572

Background: IP cisplatin (Cis) plus IV/IP paclitaxel (Tax) is a standard therapy for optimally debulked adnexal cancer. We previously demonstrated the feasibility of combining bevacizumab (Bev) with this IV/IP regimen. In this study IP Cis, IV/IP Tax, and IV Bev are combined with olaparib (Ola) as front-line therapy. Methods: Patients with newly diagnosed adnexal (ovarian, fallopian tube, or primary peritoneal) carcinoma, acceptable organ function, and KPS ≥ 70% are eligible. Patients receive 6 cycles of chemotherapy plus Bev: Tax 135 mg/m2 IV over 3 hours on Day 1, Bev 15 mg/kg IV on Day 1 (starting cycle 2), Cis 75 mg/m2 IP on Day 2, Tax 60 mg/m2 IP on Day 8. Bev is continued every 3 weeks for 21 treatments after chemotherapy is complete. In addition, Ola is given at escalating doses (50/100/200mg tabs BID) on Days 2-8 during cycles 1-6, and then 300mg BID Daily during cycles 7-22. The primary objective is to evaluate the MTD and safety/tolerability of Ola when combined with IP Cis, IV Bev and IV/IP Tax using a 3+3 dose escalation scheme. Results: Seventeen women have been treated [median age 57 (47-73)]: 8 in cohort 1 (50mg), 3 in cohort 2 (100mg) and 6 in cohort 3 (200mg). Thirteen (76%) completed all 6 cycles of IV/IP cis/tax; 2 (12%) experienced IP port malfunction (both were removed and replaced); 2 (12%) switched from IP Cis to IV carboplatin due to nephrotoxicity (via ATN and/or OCT-2 inhibition). Grade 3/4 toxicities have included: neutropenia (50%), hyperglycemia (12.5%), leukopenia (12.5%), anemia (18.8%), fatigue (12.5%), and lymphopenia (31.3%). There were 2 occurrences of related grade 3 small bowel obstructions (12.5%), during cycles 2 and 7, respectively. There were no perforations or fistulae. Maintenance therapy with Bev + Ola was generally well tolerated. Conclusions: The addition of Ola to this IV/IP regimen appears to be feasible. Ola may increase the risk of creatinine elevation and myelotoxicity. The MTD of intermittent dosing of Ola tabs concurrent with chemotherapy appears to be 200mg BID, while maintenance bev + full-dose ola at 300mg BID continuous appears feasible following IV/IP. Updated results will be presented. Clinical trial information: NCT02121990.

Is chemotherapy always required for cancer in pregnancy? An observational study

BACKGROUND

Cancer in pregnancy is relatively rare, but the incidence is increasing. Several studies show that cytotoxic agents are safe to use in pregnancy from the second trimester onwards.

AIMS

This study assesses the maternal and foetal outcomes of cancers diagnosed during pregnancy. In particular, it focuses on a subset of women who elected to defer systemic chemotherapy until after delivery. This study examines if all cancers need to be treated during pregnancy or if, in certain cases, treatment can be safely deferred until after full-term delivery.

METHODS

This is a retrospective observational study of women diagnosed with cancer during pregnancy in an Irish cancer centre over a 27-year period. All women diagnosed with cancer during pregnancy who were referred to the medical oncology department for consideration of chemotherapy were included in this study. Medical and pharmacy records were extensively reviewed.

RESULTS

Twenty-five women were diagnosed with cancer in pregnancy and referred to medical oncology for consideration of systemic chemotherapy. Sixteen women (64%) commenced chemotherapy during pregnancy, seven women (28%) did not receive chemotherapy while pregnant, but commenced treatment immediately after delivery, and two (8%) did not receive any systemic chemotherapy at all. Of the seven women who commenced chemotherapy after delivery, six (85.7%) were diagnosed before 30/40 gestation. There were three cases of Hodgkin's lymphoma, two breast cancers and one ovarian cancer. After a median follow-up of 12 years, all six mothers remain disease-free.

CONCLUSIONS

This study identified a select cohort of patients that did not receive chemotherapy during pregnancy. There were no adverse outcomes to mothers due to delayed treatment.

Perioperative chemotherapy in the treatment of osteosarcoma: a 26-year single institution review

Background

Chemotherapy in the multimodality treatment of osteosarcoma has improved survival. Reported outcomes on adult patients are limited. Poor necrosis rates post neoadjuvant chemotherapy (NAC) is considered an adverse prognostic factor and attempts have been made to improve survival in this group.

Patients and methods

Adult and young adult patients diagnosed with osteosarcoma between January 1986 and August 2012 were retrospectively reviewed. Patients identified were stratified according to stage (localised or metastatic) and age (≤40 and >40 years). Event free survival (EFS) and overall survival (OS) outcomes were determined. In patients with localised disease ≤40 years, survival was assessed according to necrosis rates post NAC (<90 and ≥90%). NAC consisted of two cycles of methotrexate alternating with doxorubicin/cisplatin (MAP) followed by definitive surgery. Those with ≥90% tumour necrosis continued on MAP. Patients with <90% necrosis received ifosfamide and etoposide (IE) post operatively.

Results

A total of 108 patients were reviewed and 97 were included. Median age was 23 years (range 16–75) and 70% of patients were male. Five year EFS and OS across all groups was 57% and 63% respectively. Of the patients with localised disease (N = 81), 5-year overall survival (OS), with a median follow up of 7 years (2–26) was 70% (p < 0.0001). Patients aged 16–40 (N = 68) with localised osteosarcoma had a significantly improved 5-year OS (74%) compared to those >40 years (N = 13) (42%) (p = 0.004). Of the 68 patients with localised osteosarcoma ≤40 years, 62 were evaluated according to necrosis rates post MAP. In 33 patients who achieved ≥90% necrosis and continued MAP, 5-year OS was 82%. In 29 patients who had <90% tumour necrosis and received adjuvant IE, 5-year OS was 68% (p = 0.15). Multivariate analysis confirmed age and stage as prognostic factors but not poor necrosis rates in our treated population.

Conclusions

Long-term survival outcomes in a predominantly adult Irish population are similar to large reported trials. Age and stage at diagnosis are prognostic. Postoperative ifosfamide/etoposide alone in patients with poor necrosis rates is a feasible regimen, but its role in the adjuvant setting remains uncertain.

Electronic supplementary material

The online version of this article (doi:10.1186/s13569-015-0032-0) contains supplementary material, which is available to authorized users.

Abstract P2-16-12: An exploratory analysis of the role of dasatinib in preventing progression of disease in bone in patients with metastatic breast cancer

Introduction: The role of dasatinib, an oral SRC inhibitor is being explored for the treatment of metastatic breast cancer. In a phase I study, we previously established that the combination of dasatinib and weekly paclitaxel was feasible. The activity of this combination is currently being explored in an ongoing phase II trial. Since Src kinase has a major role in osteoclast function and dasatinib has established anabolic and anti-resorptive effects in bone in vitro, we hypothesized that patients receiving this combination would have good control of osseous metastases and primarily develop progression of disease in sites other than bone.

Patients and methods: Patients were included in this analysis if they participated in the phase I or II metastatic breast cancer studies and received dasatinib at or above the recommended phase II dose of 120mg with paclitaxel (80mg/m2 day 1 and 8 of each 21day cycle). Patients who discontinued therapy for reasons other than progression were excluded. Per protocol, patients were required to discontinue bisphosphonates or other bone modulating agents for the first 8 weeks of study due to the potential for hypocalcaemia. Thereafter, they were permitted to receive these agents at the discretion of their treating physician. Patients provided serum samples for correlative studies. Assessment of N-telopeptide of type 1 collagen (NTX), a product of mature bone collagen that reflects bone specific resorption, is planned.

Results: The median age of the 24 patients who met criteria for analysis was 50y (37 - 66y). Of these, 15 (63%) had ER+ disease, and 24 (100%) were negative for human epidermal growth factor receptor (HER2). At study entry, 17 (71%) patients had bone involvement. Following the initial eight week moratorium, 7 (29%) patients received a bisphosphonate or rank ligand inhibitor during treatment with dasatinib + paclitaxel. Patients received a median 2 months (range 1-23) of dasatinib + paclitaxel therapy. To date, 3 (13%) continue on therapy, and 21 (88%) have had progression of disease. Among patients who progressed, 18 (86%) have progressed in visceral sites and only 3 (14%) progressed in bone. Analyses of serum NTX levels are ongoing and will be compared by site of progression.

Conclusion: The potential role of serum NTX as a predictive biomarker of benefit from dasatinib and paclitaxel is being explored and updated results will be presented.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-16-12.

Biological subtypes of breast cancer: current concepts and implications for recurrence patterns

Breast cancer is comprised of a number of complex and heterogeneous subtypes with differing clinical behavior and outcomes. In recent years, significant advances have been made in discerning the molecular drivers of this disease, and characterizing distinct subtypes of breast cancer based on gene expression profiles. These advances have begun to translate into greater individualization of treatment for patients. Although these advances have shaped our understanding of the underlying biology of breast cancer, most clinical decisions are currently based on tumor expression of the estrogen receptor (ER), progesterone receptor (PR) and the human epidermal growth factor receptor 2 (HER2). These biomarkers have prognostic and predictive significance in breast cancer and have important implications for tumor growth and metastatic patterns. In this review, we focus on the three broad phenotypes of breast cancer used in clinical practice; ER/PR positive, HER2 positive and triple negative breast cancer (TNBC), which is characterized by lack of expression of ER, PR and HER2. We discuss the influence of these tumor-related factors as well as histological subtype, on the potential for breast cancer recurrence and patterns of disease spread.

Switching chemotherapy in adult osteosarcoma patients with poor necrosis rates post neoadjuvant methotrexate, cisplatin, and doxorubicin (MAP)

10530

Background: Chemotherapy in the treatment of osteosarcoma has improved 5 year overall survival (OS) from 20% with surgery alone to 60-70%. However, poor tumor necrosis following neoadjuvant chemotherapy (NAC) is associated with decreased survival, therefore strategies to improve outcomes are required for these patients. Methods: Records from all adult patients diagnosed with osteosarcoma between 1986 and 2012 were retrospectively reviewed. Patients were stratified according to age at diagnosis (<40yrs and >40yrs), stage (localised or metastatic) and tumor necrosis post NAC (<90% and >90%). All patients received 2 cycles of methotrexate alternating with cisplatin/doxorubicin (MAP) preoperatively. Following surgery, patients with >90% tumor necrosis continued MAP whilst those with <90% necrosis switched to 4 cycles of ifosfamide and etoposide (IE). Results: 105 patients were identified and 98 who received systemic chemotherapy were included. Median age was 23yrs (Range 15-75yrs); 68% of patients were male. Limb sparing surgery was performed in 76% of applicable patients. Of the patients with localised disease (N=85), 5 year OS, with a median follow up of 8 years (1-26 yrs) was 68% (p=0.002). Patients <40 yrs with localised disease had a 5yr OS of 71% (N=73) compared to 40% in those >40 yrs (N=12) (p=0.05). 2/13 patients with metastatic disease at diagnosis are disease free >10 years post diagnosis. 65 of 73 patients with localised disease < 40 yrs had histology reviewed post neoadjuvant MAP. 34/65 (52%) had >90% tumor necrosis and continued on MAP, 5 yr OS 79%, 31 patients (48%) had <90% necrosis and received adjuvant IE, 5 yr OS 68% (P=0.10). Conclusions: Age and stage are important prognostic factors in patients with osteosarcoma treated with chemotherapy and surgery. Historically, patients with <90% tumor necrosis post NAC are considered to have a poorer prognosis. Switching from MAP to IE is an appropriate salvage regimen in such patients and appears to improve long term survival.

Cancer in pregnancy: To treat or not?

e12533

Background: Cancer in pregnancy accounts for ~1 in 1,000 pregnancies. Studies show that cytotoxic agents are safe from the second trimester. Long-term follow up has not shown increased malformations or malignancies in children exposed to chemotherapy in utero. There is no evidence of worse outcomes among women diagnosed in pregnancy. Methods: We retrospectively identified women diagnosed with cancer in pregnancy over a 25-year period. Medical records were reviewed for demographics, diagnosis, gestation, timing of treatment and outcomes. We assessed if all cancers need to be treated in pregnancy or if treatment could be safely deferred to allow normal delivery. Results: Twenty-five women were diagnosed with cancer in pregnancy and referred to medical oncology. Of 25 women, 16 (64%) received chemotherapy during pregnancy. These included 13 cases of breast cancer, one Ewing’s sarcoma, one ovarian cancer, and one small cell of cervix. All 16 women received doxorubicin and cyclophosphamide. There were 15 live births and no abnormalities seen in children who received chemotherapy in utero. At a median follow-up of 6 years 11 mothers (69%) are disease free and 4 (25%) have recurrent disease. Of nine mothers who did not receive chemotherapy in pregnancy, seven received chemotherapy immediately post-partum. Six (86%) were diagnosed in early pregnancy (median gestation 13 weeks). There were three cases of Hodgkin lymphoma, two breast cancers, and one ovarian cancer. At a median follow-up of 12 years, all mothers remain disease free. There were no abnormalities seen in these children. Conclusions: We did not identify any adverse outcomes in mothers or infants exposed to chemotherapy during pregnancy. We identified a cohort of patients that do not need immediate treatment during pregnancy. In selected cases, it is safe and appropriate to delay chemotherapy until delivery of the baby. There were no adverse outcomes to mothers due to delayed treatment and no adverse outcomes to babies not exposed to chemotherapy in utero. A multi-disciplinary team is essential to individualize treatment planning. [Table: see text]

Single-agent bevacizumab for recurrent high-grade glioma

e13019

Background: High grade gliomas(HGG) are aggressive primary brain tumors. Most patients relapse following adjuvant therapy and treatment options are limited. Novel therapies have impacted little on overall survival (OS). Bevacizumab is approved for use following recurrence. We aimed to evaluate the extent of clinical benefit derived from this treatment, as defined by decreased steroid usage and median OS. Methods: We retrospectively reviewed medical records of consecutive patients with recurrent HGG treated with single agent Bevacizumab from May 2009 to Dec 2012 at our institution. We recorded patient demographics, histological features, therapeutic interventions, OS, steroid use and radiological response. Results: 29 patients with recurrent glioma were identified, 19 male and 10 female, with a median age at diagnosis of 47 years (16-72). All patients developed tumour recurrence following various combinations of multimodal therapies including 1-3 partial debulking surgeries, radiotherapy and chemotherapy. The median time from diagnosis of HGG to commencement of bevacizumab was 15 months (6-42). Patients received a median of 7 cycles (1-22) of therapy on a variable dosing schedule. 4 patients are alive, 2 continue on Bevacizumab. 19 patients were on steroid therapy when commencing therapy. There was a significant reduction in mean daily steroid use during therapy (7.65mg vs 3.97mg, p = 0.012). However, daily steroid use ultimately increased in some patients following treatment failure. Interval imaging appearances improved in 6 cases, deteriorated in 15, were stable in 4 and 1 patient had a mixed response. 1 patient was not re-imaged due to deterioration in performance status on therapy. The median OS was 5.5 months (0.5-16 months) from commencement of bevacizumab. Conclusions: The management of patients with recurrent HGG is challenging and prognosis remains dismal. In our unselected cohort, Bevacizumab therapy resulted in temporary or sustained steroid reduction in a majority. However, only a small number of patients demonstrated an objective radiological response and median OS was poor. Tumour response was variable and the addition of a predictive biomarker to establish which patients are likely to benefit from treatment may be helpful.

Breast cancer in Irish families with Lynch syndrome

1543

Background: Breast cancer is not among the recognised malignant manifestations of Lynch Syndrome which include colorectal, endometrial, gastric, ovarian and upper urinary tract tumours. In this study we report the prevalence of breast cancer in Irish Lynch Syndrome families and determine immunohistochemical (IHC) expression of mismatch repair proteins (MMR) in available breast cancer tissue. Methods: Breast cancer prevalence was determined among Lynch Syndrome kindreds from two institutions in Ireland. One kindred was omitted due to a biallelic MMR and BRCA1 mutation.The clinicopathological data that was collected on breast cancer cases included age of onset, morphology, and hormone receptor status, and a genotype phenotype correlation was investigated. Immunohistochemical staining was performed for MLH1, MSH2, MSH6, and PMS2 on all available breast cancer tissue from affected individuals. Results: The distribution of MMR mutations seen in sixteen pedigrees were as follows; MLH1 (n=5), MSH2 (7), MSH6 (3), PMS2 (1). Sixty cases of colorectal cancer and 14 cases of endometrial cancer were seen. Seven breast cancers (5 invasive ductal and 2 invasive lobular cancers) and 1 case of ductal carcinoma in situ were reported in 7 pedigrees. This compared with 4 cases of prostate cancer. Of the 7 LS kindreds containing breast cancer, 6 MSH2 mutations and 1 MSH6 mutations were identified. Median age of breast cancer diagnosis was 49 years (range 38-57). Hormone receptor status is available on 3 breast cancer cases at time of abstract submission; all were ER positive and HER 2 negative. All cases had grade 2 or 3 tumours. 5 samples were available for IHC evaluation. 3 out of 5 cases showed loss of MMR expression, all showed loss of MSH2 and MSH6 expression. One of the two cases with normal IHC expression in breast tissue belonged to a kindred where 3 siblings with colorectal cancer and documented deleterious mutations demonstrated no IHC loss. Conclusions: Breast cancer occurred at an early age and was more common than prostate cancer in Irish Lynch Syndrome pedigrees. All identified breast cancer were in kindreds with MSH2 or MSH6 mutations. Enhanced breast cancer screening may be warranted in certain Lynch Syndrome kindreds.