Faculty Bibliography

OBJECTIVE:The real-world management of patients with non-BRCA, homologous recombination repair pathway variants with increased or uncertain risks of ovarian cancer is unknown. The objective was to determine the adoption of risk-reducing salpingo-oophorectomy (RRSO) for carriers of variants with increased or uncertain risks of ovarian cancer beyond BRCA. METHODS:This was a retrospective cohort study of patients at three hospitals with non-BRCA, homologous recombination repair pathway variants with increased risk (BRIP1, RAD51C, RAD51D) and uncertain risk (ATM, BARD1, NBN, PALB2) of ovarian cancer. Outcomes of interest were adoption of RRSO and factors associated with adoption of RRSO. Wilcoxon rank-sum, chi-square, and logistic regression were performed with p < 0.05. RESULTS:Of 318 patients, 76 (24%) had pathogenic variants with increased risks of ovarian cancer (BRIP1, 45; RAD51C, 20; RAD51D, 11), and 242 (76%) had variants with uncertain risks of ovarian cancer (ATM, 145; PALB2, 69; NBN, 23; BARD1, 5). Of 64 patients eligible for RRSO by National Comprehensive Cancer Network (NCCN) criteria or family history, 31 (48%) underwent RRSO. Among eligible patients who did not undergo RRSO, 24 (73%) were not referred for gynecologic oncology consultation. Older age at testing (adjusted odds ratio [aOR] 1.08, 95% confidence interval [CI] 1.03-1.13) and referral to gynecologic oncology (aOR 33.48, CI 8.10-138.39) were associated with increased adoption of RRSO when adjusting for personal and family history of breast and ovarian cancer. CONCLUSION/CONCLUSIONS:Half of RRSO-eligible patients by NCCN criteria beyond BRCA did not undergo RRSO. Opportunities exist for improving education to increase referrals to facilitate RRSO for these patients.

OBJECTIVE:The objective of this study was to determine the proportion of patients meeting the National Comprehensive Cancer Network (NCCN)'s BRCA genetic testing criteria prior to a diagnosis of a BRCA-related cancer. METHODS:This was a cross-sectional study of patients with BRCA pathogenic variants and a diagnosis of a BRCA-related cancer. Patients were included if they had known dates of genetic testing and cancer diagnosis. NCCN criteria (version 2.2021) were applied to determine if patients met criteria for testing before a BRCA-related cancer diagnosis. The outcome of interest was the proportion of patients undergoing genetic testing following a diagnosis of a BRCA-related cancer who qualified for genetic testing based on NCCN criteria. Chi-square, Mann-Whitney U test, and logistic regression were performed with significance at p < 0.05. RESULTS:Of 270 patients with a BRCA-related cancer, 229 (85%) underwent genetic testing after a cancer diagnosis. Most patients (97%) met at least one NCCN criteria for BRCA testing; 166 (73%) of patients who were tested following a BRCA-related cancer diagnosis also met the criteria for testing by family history. Publicly insured or uninsured patients were three times more likely to undergo BRCA testing after a diagnosis of cancer (odds ratio [OR] 3.03, 95% confidence interval [CI] 1.09-8.40). Patients with a family history of pathogenic variants were more likely to undergo testing before a cancer diagnosis (OR 0.10, 95% CI 0.05-0.23). CONCLUSION/CONCLUSIONS:Most patients with BRCA-associated cancers undergo genetic testing after their cancer diagnosis. Increased education on genetic testing criteria and novel methods to improve testing are desperately needed.

Objectives UpRi is a first-in-class NaPi2b ADC with a novel scaffold-linker-payload that enables high drug-to-antibody ratio and controlled bystander effect. NaPi2b is a sodium-dependent phosphate transporter protein broadly expressed in high-grade serous ovarian cancer (HGSOC), with limited expression in healthy tissues. Interim data from the Phase 1b expansion cohort of heavily pretreated patients with recurrent HGSOC has been reported. These data demonstrated clinically meaningful activity, notably in patients with NaPi2b-high tumors (TPS>=75). Effective and well-tolerated treatments for PROC remains an unmet medical need. The standard of care, singleagent chemotherapy, has limited efficacy, significant toxicities, and short duration of response. UPLIFT was designed as a single- arm Ph2 registrational trial for UpRi monotherapy in PROC. Methods UPLIFT is enrolling patients with PROC with up to 4 prior LoT. Prior bevacizumab is required for patients with 1-2 prior LoT only; it's not required for patients with 3-4 prior LoT. Patients may enroll regardless of NaPi2b expression; <= Grade 2 peripheral neuropathy is permitted. Primary platinum refractory patients are excluded. UPLIFT will enroll ~180 patients globally, including ~100 patients with high NaPi2b expression. UpRi is dosed IV at 36 mg/ m2 up to ~80 mg dose maximum Q4W. Baseline tumor samples (fresh or archived) will be collected for central analysis of NaPi2b expression. The primary endpoint is ORR in NaPi2b-high expressing patients. The cut-off for high NaPi2b expression is TPS>=75. Secondary endpoints include ORR in the overall population, duration of response, and safety. UPLIFT is conducted in collaboration with ENGOT (ENGOT-ov67) and GOG (GOG-3048). NCT03319628 Results trialinprogress Conclusions trialinprogress

Objectives Progression-free survival (PFS) is a rational surrogate primary endpoint in ovarian cancer (OC) trials. However, PFS is subject to biases, with validity dependent upon proper methodologic assessment. Therefore, blinded independent centralized radiologic review (BICR) is often recommended. We evaluated BICR and investigator-assessed evaluation of progressive disease (PD) in the PRIMA/ ENGOT-ov26/GOG-3012 trial examining niraparib monotherapy in intent-to-treat (ITT) and homologous recombination deficient (HRd) populations. Methods In the randomized, double-blind, placebo-controlled phase 3 PRIMA/ENGOT-ov26/GOG-3012 trial, patients with newly diagnosed stage III/IV OC were assigned to receive either niraparib or placebo. The primary endpoint was PFS (per RECIST v.1.1) by BICR. Discordance between BICR and investigator assessments of PD ([#BICR reviews with unconfirmed PD assessment]/[total# investigator-triggered reviews]) was monitored throughout the study. A training intervention was developed for BICR reviewers based on PD determination in OC. Results In an initial patient subset (n=80), a 39% discordance rate was identified between BICR and investigator-assessed PD by the sponsor, most commonly due to peritoneal carcinomatosis or fluid collections arising from new non-target lesions. After reviewer intervention, final discordance rate between BICR and investigator improved to 12% and 13% for ITT (N=733) and HRd (n=373) populations, respectively (figure 1). Across the entire study population, median PFS and hazard ratios for the ITT and HRd populations were comparable between BICR and investigator (table 1). Conclusions PRIMA/ENGOT-ov26/GOG-3012 highlights the need to optimize BICR and investigator concordance using early, specialized OC-specific training to maximize trial validity. (Figure Presented)

Objectives Focusing on gynecologic cancer, we aimed to describe the frequency of positive tumor NGS results that met ESMO 2019 recommendations for germline genetic testing (GT), receipt of germline GT and positive GT results in a large cancer cohort. Methods Patients with tumor NGS within a large, urban healthcare system were retrospectively identified (Sept 2019- Feb 2022). ESMO guidelines were used to identify eligible patients with potentially actionable germline mutations on NGS (Mandelker et al. 2019). Chi-square analysis compared rates in gynecologic and all cancer patients. Results In 3796 patients undergoing tumor NGS, 357 (9.4%) had gynecologic cancers: ovarian, 177 (49.6%); endometrial, 149 (41.7%); cervical/vulvar/vaginal, 31 (8.7%). There were 454/3796 (12.0%) total cancer patients and 79/357 (22.1%) gynecologic cancer patients with at least one positive tumor result meeting ESMO guidelines for germline GT (table 1). Of eligible patients, more gynecologic cancer patients received germline GT than total cancer patients (70.9% vs 35.7%, p<0.05). Among 56 gynecologic cancer patients receiving germline GT, 40 (71.4%) had positive results: BRCA1, 12; BRCA2, 11; MUTYH, 5; BRIP1, 5; RAD51C, 2; and one each CHEK2, SDHA, MLH1, MSH2, MSH6, NF1, PMS2. Conclusions Germline GT rates in gynecologic cancer may be higher than in all cancer patients due to guidelines recommending all ovarian cancer patients get GT and all endometrial cancer patients be screened for Lynch syndrome. Nonetheless, education and reflex protocols to further improve germline GT rates are warranted as almost 30% of gynecologic cancer patients with eligible tumor results based on ESMO criteria still did not receive GT

Objectives Mismatch repair (MMR) deficiency is caused by loss of expression of MMR proteins, MLH1, PMS2, MSH2, and/or MSH6, that function as heterodimers (MLH1/PMS2 and MSH2/ MSH6) to mediate DNA repair. Loss of function caused by mutation or epigenetic methylation leads to defective MMR and genomic instability. MMR deficient (dMMR) tumors can respond to anti-programmed death 1 (PD-1) therapy. We report on a post-hoc analysis of ORR with loss of MMR dimers and mutation status of MMR genes in patients with dMMR endometrial cancer (EC) treated with dostarlimab. Methods GARNET is a multicenter, open-label, single-arm phase 1 study. Cohort A1 enrolled patients with dMMR advanced/recurrent EC. Patients received 500 mg dostarlimab intravenously Q3W for 4 cycles, then 1000 mg Q6W until disease progression, discontinuation, or withdrawal. MMR protein status (presence or loss) was determined by local IHC. MMR gene mutation was determined by Foundation One. MLH1 loss without MMR gene mutation was a surrogate indicator for epigenetic methylation. Results Cohort A1 included 143 patients; MMR gene mutation data was available for 101 (table 1). Cohort A1 ORR was 45.5%. 66% of patients had loss of MLH1/PMS2; ORR was 48.9%. 11.2% of patients had loss of MSH2/MSH6; ORR was 56.2%. ORR was 41.7% for MLH1 loss with MMR gene mutation and 39.4% for MLH1 loss without MMR gene mutation. Conclusions Patients with dMMR advanced/recurrent EC benefitted from dostarlimab, with no noticeable difference by dimer-pair loss or MMR gene methylation/mutation status. These data suggest route to MMR deficiency does not influence response to dostarlimab

Objectives Patterns of recurrence on PARP inhibitor maintenance therapy are unclear and may affect treatment choices for subsequent therapy, including secondary cytoreductive surgery (SCS). This analysis of PRIMA/ENGOT-OV26/GOG-3012 evaluated patterns of recurrence on niraparib maintenance therapy. Methods This post hoc subgroup analysis included 314 patients treated with niraparib maintenance monotherapy following first-line chemotherapy and who had no lesions identified by CT/MRI (or by investigator assessment) at baseline. Number and site(s) of initial recurrent lesions at the time of investigator-assessed RECIST-defined progressive disease (PD) were evaluated. Results As of the primary data cut, May 17, 2019, with a median follow-up of 13.8 months (range <1-28), 141/ 314 (45%) patients developed investigator-assessed PD, with an average 1.9 (standard deviation 0.9) lesions at PD. At the time of recurrence, 62 patients (44%) had 1 lesion, 46 (33%) had 2 lesions, 24 (17%) had 3 lesions, and 9 (6%) had 4-5 lesions. The five most common sites with >=1 lesion at PD were the peritoneum (n=45), lymph nodes (n=36), liver (n=34), other (n=26), and pelvis (n=20). Conclusions For patients who received niraparib maintenance monotherapy after first-line chemotherapy and had no lesions at baseline, <50% had recurrent disease after a median 13.8 months of follow-up and >75% of patients with recurrence progressed in 1-2 sites. Prospective evaluation is required to determine whether patients with oligoprogressive disease have improved outcomes with local therapies, like SCS, in addition to systemic therapy

Objectives Perioperative management varies for patients with bleeding disorders. We sought to determine the association between bleeding disorders and perioperative transfusions for hysterectomy patients. Methods We included patients undergoing non-emergent hysterectomy between 2014-2019 from the NSQIP, a validated, risk-adjusted database from 700 hospitals. We compared 30- day perioperative transfusions between patients with and without bleeding disorders (chronic, persistent, active hematologic disorders). Transfusions <=1day were immediate, and after >=2days were delayed. Covariates were age, race/ethnicity, preoperative anemia (hematocrit <=30%) and thrombocytopenia (platelet <100,000/mL), fibroids, endometriosis, cancer, and surgical route. Results Of 290,642 patients, 10,705 (3.7%) received perioperative transfusions (8,679 ,2.9%, immediate; 2,026, 0.7%, delayed). Of 2,687 patients with bleeding disorders, 283 (10.5%) received transfusions, compared to 10,422/287,995 (3.6%) of those without (p<0.001). For gynecologic cancer patients, 17.1% (121/707) with bleeding disorders received transfusions compared to 8.4% (3,261/38,379) of those without (p<0.001). 999 (0.3%) underwent reoperation for bleeding, and this was more likely with bleeding disorders (27/2,687, 1.0% vs. 972/287,955, 0.3%, p<0.001). Anemia (OR 11.7, CI 11.1-12.4) and bleeding disorders (OR 2.0, CI 1.7-2.3) were associated with transfusions when adjusting for age, race/ethnicity, fibroids, endometriosis, cancer, and surgical approach. The effect of bleeding disorders on transfusions persisted in the laparoscopic group (OR 2.5, CI 1.9-3.3). Bleeding disorders were associated with immediate transfusions (OR 1.7, CI 1.5-2.0) and delayed transfusions (OR 2.1, CI 1.6- 2.7). Conclusions Patients with bleeding disorders are twice as likely to receive immediate and delayed transfusions for hysterectomies, even with a laparoscopic approach. Data are needed to optimize bleeding disorders to decrease transfusion risk

Objectives Minority U.S. populations are underrepresented in oncologic clinical trials. In an urban NCI-designated Comprehensive Cancer Center, we evaluated enrollment in gynecologic cancer trials by race/ethnicity pre and post the June 2020 NCI call-to-action for increased diversity in clinical trials. Methods Enrollment data in 22 therapeutic clinical trials from January 2018-May 2022 was analyzed. Chi-square/Fisher's exact analysis compared proportions of accrued patients by race/ethnicity pre and post the NCI call-to-action. Results Over the study period 205 patients were enrolled in gynecologic cancer trials: 129 (62.9%) ovarian, 52 (25.4%) endometrial, 24 (11.7%) cervical. Overall racial/ethnic distribution of patients in clinical trials was: 112 (54.6%) non-Hispanic White, 33 (16.1%) Hispanic, 32 (15.6%) Black, 28 (13.7%) Asian. Although not statistically significant, accrual of underrepresented populations increased from pre call-to-action (45/108, 41.7%) to post (48/97, 49.5%). Significantly more Black patients were accrued post NCI call-to-action (8/108, 7.4% vs 24/97, 24.7%, p=0.001). However, significantly fewer Asian patients were accrued post NCI call-to-action (21/ 108, 19.4% vs 7/97, 7.2%, p=0.01). Conclusions Overall accrual of Black patients in gynecologic cancer clinical trials increased following the NCI call-to-action, but accrual of Asian patients decreased. Further efforts are needed to ensure all racial/ethnic groups are represented in clinical trials

Objectives Endometrial cancer (EC) is a molecularly driven disease, and prognostic and treatment paradigms are transitioning to focus on molecular subtyping. Surgical staging and the role of lymphadenectomy has similarly evolved over time, however current lymph node (LND) algorithms do not account for molecular subtyping. The objective of this study was to evaluate the association of microsatellite instability (MSI) and lymph node metastases (LNM). Methods This was a retrospective cohort study of patients undergoing surgery for EC between 2010-2021. All EC patients at our institution undergo immunohistochemistry testing for mismatch repair (MMR) proteins and next generation sequencing per clinician discretion. Sarcomas were excluded. Mutations were classified as microsatellite instability high (MSI-H) or MMR proficient (MMRp). Results 367 patients were included. Of these, 273 were MMRp and 94 were MSI-H. An average of 6.1 LND were removed and there was no difference in the average LND removed between groups (p= 0.91). LNM were identified in 8% (n=31) of the entire cohort. There was a statistically significant difference in the average LNM between MMRp and MSI-H patients (p<0.0001), with 1% (n=2) of the MMRp cohort and 30% (n=28) of the MSI-H cohort having LNM. Within the MSI-H cohort, all LNM occurred within the MMR deficient, MLH1 hypermethylated subgroup - representing a LNM rate of 41%. Conclusions There is a significant association between MSI status and LNM. Molecular classification, which is obtainable from preoperative biopsy, may be used to guide intraoperative decision making and should be evaluated in the context of sentinel LND protocols