Temozolomide and olaparib for the treatment of an ARID1A-mutated uterine carcinosarcoma: A case report Authors: Chrissy Liu, Yasmin Abozenah, Thomas J. Rutherford, Vaagn Andikyan, Adrian Kohut, Matthew L. Anderson Published: 2026 ABSTRACT BackgroundPathogenic ARID1A mutations are common in uterine carcinosarcoma (UCS) and may confer vulnerability to the combination of temozolomide (TMZ) and a poly (ADP-ribose) polymerase inhibitor (PARPi) through synthetic lethality. Clinical experience with this regimen in gynecologic cancers is limited and has not been reported in UCS.CaseA 54-year-old woman with stage IIIB UCS harboring a pathogenic ARID1A mutation developed isolated cavitating pulmonary metastases shortly after completing chemoradiation, surgery, and adjuvant carboplatin/paclitaxel with dostarlimab. She received off-label temozolomide and olaparib on a 21-day cycle. After four cycles, she achieved a complete radiographic response and has remained in clinical remission with normal CA-125 and negative circulating tumor DNA. Treatment was complicated by reversible grade 3–4 thrombocytopenia.ConclusionThis case suggests that TMZ/PARPi may produce durable responses in biomarker-selected UCS and supports prospective evaluation of this regimen in genomically characterized, ARID1A-defined gynecologic cancers. Introduction The AT-rich interactive domain-containing protein 1A (ARID1A) is a human tumor suppressor frequently mutated in gynecologic cancers. Encoded at 1p36.11, ARID1A functions as a key subunit of the SWI/SNF (switch/sucrose nonfermenting chromatin remodeling) complex, an ATP-dependent chromatin remodeling assembly that regulates transcriptional activity. Consistent with its observed biologic roles, pathogenic loss-of-function ARID1A mutations result in dysregulated cell proliferation, metabolic reprogramming and an epithelial to mesenchymal transition (Wu et al., 2014; Kearns et al., 2025). Other functions for ARID1A potentially relevant to cancer have also been identified, including its ability to regulate adaptive immune responses and suppress the activity of STING Type-I interferon pathways important for anti-tumor immunity (Maxwell et al., 2024). As many as 57% of ovarian clear cell carcinomas, 39% of well-differentiated endometrioid endometrial adenocarcinomas and 17% of uterine carcinosarcomas potentially harbor a pathogenic ARID1A mutation (Morgan et al., 2024). Recent preclinical data suggest that it may be possible to clinically leverage the loss of normal ARID1A function found in many cancers for therapeutic benefit by combining temozolomide (TMZ) with a small molecule inhibitor for poly (ADP-ribose) polymerase, such as olaparib. As an alkylating agent, exposure to TMZ induces DNA lesions typically repaired through base excision. However, because base excision repair (BER) is impaired by ARID1A dysfunction, ARID1A-deficient cells must rely on alternate pathways that depend on poly (ADP-ribose) polymerase (PARP) for survival. It has been hypothesized that exposure to a PARP inhibitor (PARPi) under these circumstances results in double-stranded DNA breaks, and replication stress that cannot be corrected (Morgan et al., 2024). Pre-clinical data demonstrate that the combination of TMZ and olaparib markedly enhances replication fork collapse and durable tumor suppression in xenograft models, while ARID1A-mutant tumors exhibit limited sensitivity to either agent alone (Yu et al., 2023). The clinical efficacy of this approach has been recently evaluated as part of a multicenter phase II trial that enrolled patients with advanced uterine leiomyosarcoma (uLMS). Investigators for this study observed an overall response rate of 27% (6 out of 22 patients) (Ingham et al., 2023). Furthermore, the combination of olaparib and temozolomide was deemed to have met the prespecified primary endpoint for meaningful clinical benefit. In an exploratory analysis, median progression-free survival was found to be significantly better (11.2 vs 5.4 months, p = 0.05) for the subset of uLMS patients characterized as homologous recombination deficient (HRD by the RAD51 foci assay). This benefit was observed despite the fact that subjects received an average of three prior chemotherapy regimens. Despite encouraging early results, clinical experience with treating gynecologic and other cancers with the combination of TMZ and PARPi remains limited. In this report, we present the outcomes for a patient with recurrent uterine carcinosarcoma harboring an ARID1A mutation successfully treated with this combination therapy, illustrating its potential activity in this aggressive subtype of endometrial cancer. Case presentation A 54-year-old woman, G1P1, was diagnosed with stage IIIB uterine carcinosarcoma (UCS; FIGO 2009). Owing to extent of disease, she was initially treated with whole pelvic radiation therapy (WPRT; total dose 50 Gy) with weekly cisplatin (40 mg/m2, NTE 70 mg). She tolerated therapy well and completed all fractions of pelvic radiotherapy within a window of 36 days. Her sixth dose of weekly cisplatin was held due to neutropenia (ANC = 700 cells/μl), which subsequently recovered without additional intervention. Approximately 6 weeks after completing WPRT, she underwent total abdominal hysterectomy and bilateral salpingo-oophorectomy. Histologic evaluation of specimens collected with this surgery confirmed a diagnosis of uterine carcinosarcoma. Surgical margins were negative for metastatic disease. Somatic molecular profiling revealed a microsatellite stable (MSS) cancer harboring pathogenic mutations in ARID1A (Exon 7, p.G801fs, VAF 69%), KRAS (Exon 2, p.G12V, VAF 35%), PIK3CA (Exon 2, p.I102del, VAF 37%) and PTEN (exon 5, p.A86fs, VAF 77%). Pathogenic mutations were also observed in DICER1 (p.D1709N, VAF 37%) and FANCL (Exon 11, c40-841delGT, VAF 38%). Tumor mutation burden (TMB) was scored as low with no significant genomic loss of heterozygosity (LOH). By immunohistochemistry, expression of MLH1, MSH2, MSH6 and PMS2 were all deemed intact; estrogen receptor (ER) expression was assessed as negative (0%). A commercially available phenotypic signature for HRD (Caris, Inc) was reported as negative. Following recovery from her hysterectomy, the patient received three cycles of adjuvant therapy, consisting of carboplatin (AUC-5), paclitaxel (175 mg/m2), and dostarlimab (500 mg). Dostarlimab was discontinued after cycle 3 due to a persistent rash. After the third cycle, her CA-125 remained within normal range (13.3 U/mL). Computed tomography of the chest, abdomen, and pelvis showed no evidence of disease, and circulating tumor DNA (ctDNA; Signatera) was negative. Adjuvant therapy was completed approximately 8 months after initial diagnosis. One month after the completion of adjuvant therapy, chest radiographs revealed multiple new cavitating pulmonary nodules as large as 1.5-cm (Fig. 1a, c). No other clinical or radiographic evidence of recurrent disease was identified. CA-125 was 17.2 U/mL and tumor-informed ctDNA was reported as negative. Biopsy of the pulmonary lesions was considered to obtain histologic confirmation of recurrent disease; however, the patient declined the procedure. The patient had no associated fever, cough, or dyspnea, no leukocytosis, and no radiographic findings suggestive of pneumonia.Fig. 1Computed tomography of the patient's chest before and after temozolomide and olaparib. (a) Axial image before treatment demonstrates a cavitating pulmonary nodule in the left lower lobe (green circle). (b) Axial image following 4 months of treatment demonstrates resolution of the previously observed pulmonary lesion. (c) Coronal image before treatment demonstrates the cavitating pulmonary nodule in the left lower lobe. (d) Coronal image following 4 months of treatment demonstrates resolution of the previously observed pulmonary lesion.Fig. 1 Computed tomography of the patient's chest before and after temozolomide and olaparib. (a) Axial image before treatment demonstrates a cavitating pulmonary nodule in the left lower lobe (green circle). (b) Axial image following 4 months of treatment demonstrates resolution of the previously observed pulmonary lesion. (c) Coronal image before treatment demonstrates the cavitating pulmonary nodule in the left lower lobe. (d) Coronal image following 4 months of treatment demonstrates resolution of the previously observed pulmonary lesion. After reviewing potential therapeutic options with the patient, she elected to continue treatment with an off-label regimen consisting of temozolomide (145 mg PO daily, days 1–7) and olaparib (200 mg PO twice daily, days 1–7) administered on a 21-day cycle. Treatment was complicated by a single episode of severe thrombocytopenia (33,000/μL) following her third cycle. This resolved spontaneously, but necessitated a 6-week hold, after which, the patient was able to resume treatment. CT imaging performed following the fourth treatment cycle revealed a complete radiographic response (Fig. 1b, d); ctDNA remained negative and serum CA-125 levels were within normal limits throughout the course of treatment. The decision was made to withhold additional therapy. Since that time (∼ 14 months), the patient has remained in clinical remission for at least 14 months with no limitations on her daily activities. During this time, serum levels of CA-125 have remained normal and ctDNA testing has been negative. A detailed synopsis of the patient's clinical course is summarized in Fig. 2.Fig. 2Clinical timeline of diagnosis, treatment, recurrence, and response to temozolomide and olaparib in a patient with ARID1A-mutated uterine carcinosarcoma. Treatment was complicated by thrombocytopenia but resulted in a complete radiographic and clinical response which has persisted for at least 14 months.Fig. 2 Clinical timeline of diagnosis, treatment, recurrence, and response to temozolomide and olaparib in a patient with ARID1A-mutated uterine carcinosarcoma. Treatment was complicated by thrombocytopenia but resulted in a complete radiographic and clinical response which has persisted for at least 14 months. Discussion At least two prospective clinical trials have evaluated the efficacy of TMZ/olaparib in cohorts of heavily pre-treated patients of uLMS. In the first of these trials, a prospective single arm phase II assessment, TMZ/olaparib met the threshold established for clinical efficacy (Ingham et al., 2023). However, a second phase II/III trial (Alliance-A092104 / NCT05432791) enrolling 74 uLMS patients who progressed on ≥2 prior systemic therapies (n = 37 per arm) failed to demonstrate an advantage when median progression free survival when patients treated with this combination were compared to those treated with pazopanib or trabectedin (3.2 mos vs. 5.5 mos, respectively. Notably, this trial enrolled a biomarker-unselected population and was ultimately closed for futility (Van Tine et al., 2025). Since that time, no other studies have evaluated the potential efficacy of a TMZ/PARPi regimen in gynecologic cancers. Our experience suggests that the combination of TMZ/olaparib may in fact confer clinical benefit for appropriately selected, biomarker defined populations of patients with UCS and possibly other gynecologic cancers. Our observation is noteworthy, as pathogenic ARID1A mutations are frequently observed in endometrial cancer, including uterine carcinosarcoma. Thus, this combination could provide an effective option and address an important unmet clinical need. It is noteworthy that neither trial examining the efficacy of TMZ/olaparib in uLMS specifically evaluated the relationship between clinical responses and pathogenic ARID1A mutation. An important question with regards to this patient's experience is the degree to which the unique genomic composition of her cancer contributed to its clinical response. In glioblastoma, promoter methylation leading to decreased expression of O-6-methylguanine-DNA methyltransferase (MGMT) plays a well-recognized role in determining clinical response to TMZ. Although MGMT promoter methylation testing was not specifically performed for this patient's cancer, results of whole exome profiling (WES), suggest that MGMT expression is relatively low (27th percentile) compared to other histologically similar cancers. It is conceivable that low levels of MGMT expression may have contributed to her favorable clinical outcome. It is also possible that other pathogenic mutations identified in this patient's cancer played a role in determining its clinical response. For example, a pathogenic mutation was observed in FANCL, a component of the Fanconi anemia (FA) core complex. Although FANCL actively contributes to DNA damage repair via homologous recombination, the clinical significance of the FANCL mutation observed in this patient is unclear (Yao et al., 2025; van de Kooij et al., 2024). It is conceivable that the FANCL mutation observed in this patient's cancer could contribute to HR deficiency and, in turn, its response to TMZ/olaparib. However, a commercially available phenotypic assay for homologous combination characterized the patient's tumor proficient for homologous recombination. These conflicting results can be potentially reconciled by the fact that homologous recombination defects may not be fully captured by current HRD assays (Yao et al., 2025; van de Kooij et al., 2024; Telli et al., 2016; Hoppe et al., 2018). It is similarly difficult to solely assess the contribution of other genomic mutations that have potentially contributed to the clinical response observed in this patient. Because this patient's tumor harbored multiple pathogenic alterations, including PIK3CA, PTEN, DICER1, and FANCL mutations, the observed clinical response cannot be attributed to ARID1A deficiency alone. To our knowledge, clinical data evaluating TMZ/PARPi specifically in patients with ARID1A-mutated tumors lacking these co-occurring alterations are not currently available. Nevertheless, preclinical studies using isogenic ARID1A-deficient and ARID1A-wild-type models have demonstrated greater sensitivity of ARID1A-deficient cells to combined TMZ/PARPi, (Yu et al., 2023). These observations support a direct role for ARID1A loss in mediating sensitivity to this combination and as well as the use of pathogenic mutations in ARID1A as a potential predictive biomarker. Moving forward, these concerns and the complexity of the underlying biology highlight the importance of evaluating TMZ/olaparib across a larger number of biomarker-defined patients with genomically well-characterized cancers. A noteworthy feature of this case was the persistence of negative tumor-informed ctDNA and a normal serum CA-125 level despite the development of radiographically apparent pulmonary lesions. This raises the question of whether the observed pulmonary lesions were actually malignant in nature. Unfortunately, the patient declined confirmatory biopsy and other tumor markers such as CA-125 and/or ctDNA were negative during this phase of her treatment. The lack of additional test results, despite the cavity nature of the observed lesions, does not preclude the malignant nature for several reasons. First, although tumor-informed ctDNA has demonstrated prognostic utility in uterine malignancies, its sensitivity is not absolute (Recio et al., 2024). Furthermore, the anatomic site of metastatic disease may influence ctDNA detection; studies in other solid tumors have demonstrated lower ctDNA levels in patients with lung-only metastases (Bando et al., 2022). Perhaps most importantly, because ctDNA testing was not performed prior to initiating treatment, it remains unclear whether ctDNA was a reliable marker of disease status in this individual. Furthermore, Similarly, although CA-125 may be useful for longitudinal disease monitoring in uterine carcinosarcoma, a normal CA-125 level does not exclude recurrent disease (Ross et al., 2019). These limitations underscore the importance of validating the clinical efficacy of the regimen used to manage this patient across a larger number of subjects with clearly confirmed metastatic disease. Given the relatively acceptable toxicities of both agents, as well as their ready availability to clinicians, our experience with this combination would suggest that prospective trials to better evaluate the efficacy of this potential therapeutic combination in appropriately biomarker-selected populations of patients are warranted. CRediT authorship contribution statement Chrissy Liu: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Yasmin Abozenah: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Thomas J. Rutherford: Writing – review & editing, Formal analysis. Vaagn Andikyan: Writing – review & editing, Formal analysis. Adrian Kohut: Writing – review & editing, Formal analysis. Matthew L. Anderson: Writing – review & editing, Formal analysis, Conceptualization. Patient consent statement Written informed consent was obtained from the patient for publication of this report and any accompanying images. Documentation is available for the editorial office upon request. Ethics declaration Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed. This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. Ethics committee approval was not required under relevant laws and institutional guidelines. Per institutional policy, single subject case reports do not require IRB review. This research follows the CARE guidelines and the CARE checklist. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.