Study case flow diagram for WES of translational specimens from patients enrolled in GOG281. One hundred thirty-four samples underwent successful WES, of which 112 were also evaluable for pERK via IHC. Additionally, 36 samples were evaluable for pERK IHC that did not have matching WES.

Purpose: Low-grade serous ovarian carcinoma (LGSOC) is a distinct form of ovarian cancer characterized by younger patient age and relative chemoresistance. The GOG281/LOGS trial (NCT02101788) investigated the efficacy of the MEK inhibitor trametinib compared with physician's choice standard-of-care (SOC) in patients with LGSOC with persistent/recurrent disease. The study demonstrated significantly improved progression-free survival (PFS) in the trametinib-treated arm. Experimental Design: Two hundred and sixty patients with recurrent/persistent LGSOC were enrolled and randomly assigned in GOG281. We performed molecular analysis of 170 patients with available tumor specimens, comprising whole-exome sequencing and phospho-ERK (pERK) IHC, to identify biomarkers of clinical benefit from trametinib. The demographics of the translational cohort (n = 170) were comparable with those of the total trial cohort. Results: High tumor pERK expression (greater than the median histoscore of 140) was associated with significantly prolonged PFS with trametinib treatment versus SOC (median 20.1 vs. 5.6 months, log-rank P < 0.0001; test for interaction P = 0.023). Tumors harboring canonical RAS-RAF-MAPK mutations (KRAS/BRAF/NRAS: 44/134, 32.8% of cases) had a higher response rate to trametinib (50.0% vs. 8.3%; Barnard's P = 0.0004; test for interaction P = 0.054), but KRAS/BRAF/NRAS status was not predictive of prolonged PFS (test for interaction P = 0.719). KRAS amplification (n = 5 without KRAS/NRAS/BRAF mutation) and mutation of MAPK-associated genes (n = 25 without KRAS/NRAS/BRAF mutation or KRAS copy number gain) expanded the number of cases with identifiable MAPK defects to 55.2%, but consideration of these events did not improve the discrimination of trametinib responders. Chr1p loss (49% of cases) was associated with lower pERK expression (P = 0.021). Conclusions: This exploratory analysis suggests that pERK expression and mutation of KRAS/BRAF/NRAS are candidate biomarkers of improved PFS and response to trametinib, respectively.

Impact of pERK status and KRAS/BRAF/NRAS mutation status on GOG281 patient outcome. A, PFS of GOG281/LOGS patients with low-pERK tumors (pERK ≤ 140) in trametinib versus SOC arms. B, PFS of GOG281/LOGS patients with high-pERK tumors (pERK > 140) in trametinib versus SOC arms. C, PFS of GOG281/LOGS patients with KRAS/BRAF/NRAS WT tumors in trametinib versus SOC arms. D, PFS of GOG281/LOGS patients with KRAS/BRAF/NRAS-mutant tumors in trametinib versus SOC arms. Labeled HR refer to comparison of trametinib versus SOC arm.

Demographics of translational research population.

Unsupervised hierarchical clustering of chromosome arm-level copy-number alterations across biospecimens from patients enrolled in GOG281/LOGS.

Supplementary Figures and Supplementary Figure Legends

Molecular landscape of tumor samples from patients enrolled in GOG281/LOGS. MAPK-associated, MAPK-associated genes as defined in the Gene Ontology Term GO:0000165; Chr1pq-aberrant, chr1p-loss with concurrent chr1q-gain.

pERK and other potential biomarkers of therapy response in LGSOC. A, pERK IHC examples demonstrating negative (0), weak (1+), moderate (2+), and intense (3+) positivity for histoscore calculation. Scale bars, 50 µm. B, Response to trametinib (top) and physician’s choice SOC (bottom) according to KRAS/BRAF/NRAS mutation status (left), pERK status (middle), and chr1 abnormalities (right). C, pERK histoscore according to KRAS/BRAF/NRAS mutation type. Labels specify the numbers within each group that were evaluable for pERK expression levels. D, pERK histoscore by chr1p-loss status. E, Frequency of pERK status between chr1p-loss and -intact groups. Chr1pq-aberrant, chr1p-loss with concurrent chr1q-gain.

Low-grade serous carcinoma of the ovary (LGSOC) is relatively resistant to chemotherapy. Given its biological parallels to hormone receptor-positive breast cancer, including responsiveness to anti-estrogen therapies, we conducted a pilot phase II study to assess clinical benefit of neoadjuvant fulvestrant and abemaciclib in women with advanced, unresectable LGSOC (NCT03531645). Imaging assessments were performed every 8 weeks until resectable. The primary endpoint was clinical benefit rate (CBR). Exploratory objectives included evaluation of safety profile, accessing biomarkers related to clinical benefit, and description of tumor phenotypic changes. Fifteen patients were enrolled and evaluable for efficacy. CBR was 100%, with 1/15 patients (7%) achieving complete response (CR), 8/15 patients (53%) achieving partial response (PR), and 6/15 (40%) exhibiting stable disease (SD). Interval cytoreductive surgery (ICS) was performed in 9 patients (60%), with 7 (78%) achieving complete or optimal resection. Fourteen tumor samples underwent transcriptomic and proteomic profiling. Tumors from long-term survivors showed significantly higher baseline expression of cell-cycle-related programs and estrogen signaling-related genes, both suppressed upon treatment. Neoadjuvant fulvestrant and abemaciclib was well tolerated, achieved high response rates, and enabled optimal surgical resection in most patients. Tumor proliferative activity and estrogen signaling dependency may predict therapy response.

More than half of all gynecological cancers are classified as rare (annual incidence <6 per 100,000), and present significant challenges in diagnosis, management, and research. Rare cancers collectively comprise more than 20% of new cancer diagnoses and exceed the burden of individual common cancers. Their rarity complicates evidence-based guideline development, clinical trial design, and drug access, and is exacerbated by variations in epidemiology, histology, and biological behavior. Key controversies include the need for centralized pathological reviews and harmonized diagnostic criteria. Recent World Health Organization classification updates, such as the redefinition of ovarian and fallopian tube cancers, illustrate the impact of evolving guidelines on epidemiology and patient management. Variations in the classification among pathologists and limited access to molecular diagnostics further hinder effective management. Multidisciplinary care in expert centers improves outcomes; however, significant geographic and resource disparities persist. National and international collaborations including European Reference Network for rare adult solid tumors, Gynecologic Cancer Intergroup, Gynecologic Oncology Group, Asia-Pacific Gynecologic Oncology Trials Group, and European Network for Gynaecological Oncology Trials have made strides in standardizing care and advancing research. Novel trial designs, such as basket and umbrella trials, alongside synthetic control arms, are essential for addressing the small sample sizes typical of rare tumors. Emerging consortia, such as International Ovarian Tumor Tissue Analysis Consortium and International Consortium for Low-grade Serous Ovarian Cancer, provide robust platforms for translational research and biomarker validation. However, challenges remain in fostering cross-border collaboration, streamlining regulatory pathways, and ensuring equitable access to trials and therapy. To optimize outcomes, a comprehensive approach that integrates centralized care, innovative trial designs, and international networks is imperative. This paradigm fosters the harmonization of care, accelerates translational research, and bridges the gap between scientific innovation and patient benefits.

Low grade serous ovarian carcinoma is a rare epithelial ovarian cancer that occurs more frequently in younger women. RNA sequencing (RNAseq) has been playing a pivotal role in understanding the molecular pathogenesis of the low-grade ovarian serous carcinoma (LGSOC). The quantification of gene expression with enough sequencing depth can be fairly accurate. However, the current short-read RNAseq approach is not very accurate in measuring individual transcript activity. This is because multiple transcripts from the same gene share high sequence similarity, which complicates the transcript mapping. As the accuracy of long-read sequencing technology is improving, long-read full transcript RNA sequencing is becoming an important tool for interrogating alternative transcripts and the discovery of gene fusion transcripts. In this study, we compare short-read RNAseq and long-read RNAseq from the same LGSOC patient samples. The RNA-Seq analysis for long-reads were performed by mapping sequencing reads to annotated reference human genome GrCh38 with minimap2. First, all annotated transcripts or genes were extracted including any annotated splice variants. Subsequently, the reads were mapped against all the transcripts as well as to the whole genome using minimap2. The distributing and counting the reads across genes and transcripts were calculated. We found that the expression level of transcripts in short-read in general is higher than that that of long-read. It is because short reads frequently could not be accurately mapped to the correct alternative transcripts from the same genes. When a read mapped to multiple transcripts, the short reads were randomly assigned to a transcript from the same gene. As a result, more than 100, 000 transcripts had at least one assigned read in the short-read RNAseq analysis, but less than 50, 000 transcripts had at least one assigned read in long-read RNAseq analysis. Long read with greater read-depth had better quantification accuracy for detecting specific transcripts. To identify potential gene fusions, the raw Oxford Nanopore long-read RNA-seq fastq files were corrected and polished with high sequencing quality short-read RNA-seq data by racon assembly tool. Those long-reads with unaligned ends of reads from a mapping generated by RNA-Seq analysis were re-mapped to the reference using the unaligned ends of reads. In this study, we identified several gene fusions including a gene fusion between NSUN4 and FAAH genes with at least seven long reads covering the breakpoint. FAAH encodes the enzyme fatty acid amide hydrolase and has been shown to be a metastasis suppressor in breast cancer previously. It is possible that the truncated FAAH protein because of gene fusion is involved in the metastasis of the LGSOC that we have analyzed. This could be another mechanism involved in the progression of LGSOC. Additional analysis of the long-read full transcript sequencing of LGSOC is in progress. Kwong-Kwok Wong, Yvonne Tsang, David M. Gershenson. Analysis of low-grade serous ovarian cancer by long-read full length transcripts sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4081.

OBJECTIVES:To evaluate survival outcomes among patients with adult-type granulosa cell tumors who have telomerase reverse transcriptase (TERT) promoter mutations. METHODS:This is a retrospective cohort study using the MD Anderson Rare Gynecologic Malignancy Registry. Patients with adult granulosa cell tumors who underwent molecular testing for TERT promoter and FOXL2 c.C402G mutations were included. We used descriptive statistics to compare demographic and clinical variables and estimated progression-free and overall survival with Kaplan-Meier curves. Cox proportional hazards regression and log-rank tests were employed for comparisons, with multivariable analyses adjusting for various factors. RESULTS:Among 70 patients, 28 (40%) had TERT+ tumors. The median age at diagnosis was 40 years (range 12-71) for TERT- patients and 46 years (range 25-76) for TERT+ patients. At diagnosis, 22 (63%) of 35 TERT- patients were stage I, 10 (29%) stage II, and 3 (9%) stage III, while in the TERT+ group, 17/23 (74%) were stage I, 3 (13%) stage II, and 3 (13%) stage II. Univariable analysis showed no difference in time from diagnosis to first recurrence (p=0.19) and from first recurrence to second recurrence (p=0.24) based on tumor TERT status. The median time from first to second recurrence in the TERT- group was 27.3 months (95% CI 14.1 to 40.0) and in the TERT+ group was 14.8 months (95% CI 8.1 to 21.0). There was no observed difference in overall survival between the groups (HR=0.53; 95% CI 0.19 to 1.45; p=0.21).Multivariable analysis adjusting for age at diagnosis, TERT promoter mutation status, systemic chemotherapy, and stage demonstrated a significant difference in progression-free survival based on TERT mutation status (HR=2.89; 95% CI 1.32 to 6.36). CONCLUSIONS:After adjustment for covariates, patients with adult granulosa cell tumors and TERT+ tumors had shorter progression-free survival after first recurrence. TERT promoter mutations may identify a subset of patients with recurrent adult granulosa cell tumors and less favorable outcomes.

S1. Heatmap of unsupervised hierarchical clustering of the top 1000 most variable genes.S2. Violin plots showing the expression levels of genes previously identified as differentially expressed genes between primary and recurrent tumors in Haltia et al., 2020.S3. Gene set enrichment analysis performed with GO terms and KEGG pathways showing that primarily immune-related and hormone-regulated gene sets expression are altered between primary and recurrent aGCTs.S4. Gene set enrichment analysis results showing that top enriched gene sets are significantly enriched in recurrent tumors in comparison with primary tumors.S5. Boxplot showing the fraction of each noncancerous cell type identified by CIBERSORTx (A) and xCell (B) in primary and recurrent tumor samples.S6. Heatmap showing the correlation between immune cells fractions.

Clinical Characteristics by Patient