The evolving role of immunotherapy in ovarian cancer

Ana Carolina Veneziani : Conceptualization; methodology; writing—original draft; writing—review and editing. Laura Venegas : Conceptualization; methodology; writing—original draft; writing—review and editing. Amit M. Oza : Conceptualization; supervision; writing—review and editing.

Future immunotherapy in ovarian cancer is moving toward a more integrated and multimodal approach designed to overcome resistance and improve antitumor activity. Beyond PD‐1/PD‐L1 blockade, emerging strategies include targeting alternative immune checkpoints, such as TIGIT (T‐cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine‐based motif domains), LAG‐3 (lymphocyte‐activation gene 3), TIM‐3 (T‐cell immunoglobulin and mucin domain–containing 3), and PVRIG (poliovirus receptor‐related immunoglobulin), which help restore immune activation by reducing T‐cell suppression and improving the immune response. 107 At the same time, research is expanding beyond conventional T‐cell–centered approaches to include NK cells, invariant NKT cells, γδ T cells, and engineered macrophages, alongside bispecific T‐cell engagers and neoantigen‐based vaccines that may strengthen tumor recognition and cytotoxicity. 108 , 109 A major area of focus is also the TME, in which combination strategies incorporating chemotherapy, radiotherapy, and cell‐cycle checkpoint inhibitors may induce immunogenic cell death, promote antigen release, and create a more favorable setting for immune response. 93 In parallel, efforts to reverse immune suppression through depletion or reprogramming of Tregs, MDSCs, and TAMs are gaining interest, as are approaches to normalize tumor vasculature and counteract hypoxia. 110 ADCs are also emerging as potential immunomodulatory partners rather than purely cytotoxic agents. Depending on target expression and payload, ADCs may enhance antigen release, promote immunogenic cell death, increase T‐cell infiltration, and potentially synergize with immunotherapy. These effects require prospective validation. Finally, artificial intelligence and machine‐learning approaches may support biomarker discovery by integrating clinical, pathologic, radiomic, spatial, and molecular data to improve prediction of response and resistance. Together, these advances support a future in which ovarian cancer immunotherapy becomes more personalized, combination‐based, and biologically targeted.

Clinical studies have evaluated ICIs both as monotherapy and in combination with chemotherapy, PARP inhibitors, and antiangiogenic agents. Although some combinations have demonstrated encouraging activity, the outcomes of these studies have varied (Table  1 ). In the context of first‐line and recurrent ovarian cancer, which includes both platinum‐sensitive and platinum‐resistant disease, numerous programed death protein 1 (PD‐1)/programmed death‐ligand 1 (PD‐L1) inhibitors have been extensively studied, either as single agents or in combination. Immune checkpoint inhibitors in ovarian cancer. 1 mg/kg: 21% ORR, PFS 2.2 months, OS 18 months 2 mg/kg: 20% ORR, PFS 2.5 months, OS 8 months Abbreviations: ±, with or without; APGOT, Asia‐Pacific Gynecologic Oncology Trials Group; Atezo, atezolizumab; Bev, bevacizumab; BRCA m, BRCA mutation; BRCA wt, BRCA wild type; CD, chemotherapy plus durvalumab; CI, confidence interval; CP, carboplatin and paclitaxel; CPS, combined positive score; CR, complete response; D, durvalumab; D‐N, durvalumab and niraparib; ENGOT, European Network of Gynecological Oncologic Trial Groups; EOC, epithelial ovarian cancer; EORTC, European Organization for Research and Treatment of Cancer; GOG, Gynecologic Oncology Group; HR, hazard ratio; irRECIST, Immune‐Related Response Evaluation Criteria in Solid Tumors; ITT, intent to treat; N, niraparib; NE, not evaluable; non‐t BRCA m, without a tumor BRCA 1/ BRCA 2 mutation; NS, nonsignificant; O, olaparib; OC, ovarian cancer; ORR, objective response rate; OS, overall survival; PD‐L1, programmed death‐ligand 1; PFI, platinum‐free interval; PFS, progression‐free survival; PLD, pegylated liposomal doxorubicin; PR, partial response; PROC, recurrent, platinum‐resistant ovarian cancer; PSOC, platinum‐sensitive ovarian cancer; SD, stable disease; TFI, treatment‐free interval. Mechanistically, chemotherapy may enhance antitumor immune responses by increasing tumor immunogenicity, promoting antigen presentation, and reducing immunosuppressive factors within the tumor microenvironment (TME). 39 Cytotoxic T‐lymphocyte–associated protein 4 (CTLA‐4) negatively regulates T‐cell priming by competing with the B7:CD28 interaction, which serves as a costimulatory signal during the early phases of immune activation. 40 Vascular endothelial growth factor (VEGF) contributes to abnormal vasculature, hypoxia, impaired immune trafficking, and recruitment of immunosuppressive cell populations, providing a plausible mechanism by which bevacizumab could enhance immune responsiveness. 29 Finally, PARP inhibition induces DNA double‐strand breaks, leading to accumulation of cytosolic DNA and activation of the cGAS–STING pathway, which, in turn, promotes type I interferon signaling while also upregulating PD‐L1 expression. 41 Several phase 3 trials have evaluated ICIs in combination with chemotherapy or PARP inhibitors in the first‐line setting (IMagyn050, JAVELIN 100, European Network of Gynecological Oncologic Trial Groups [ENGOT] and Gynecologic Oncology Group [GOG] trial ENGOT‐OV43/GOG‐3036/KEYLYNK‐001, FIRST/ENGOT‐OV44, DUO‐O, and ATHENA‐COMBO). IMagyn050, a phase 3 trial involving patients with newly diagnosed ovarian cancer who had gross residual disease after primary surgery or those eligible for neoadjuvant chemotherapy, reported a median progression‐free survival (PFS) of 19.5 months with atezolizumab versus 18.4 months with placebo (hazard ratio [HR], 0.92; 95% confidence interval [CI], 0.79–1.07; stratified log‐rank p  = .28) in the intent‐to‐treat (ITT) population (ClinicalTrials.gov identifier NCT03038100 ). In the PD‐L1–positive group, the median PFS was 20.8 months for atezolizumab compared with 18.5 months for placebo (HR, 0.80; 95% CI, 0.65–0.99; p  = .038). Although the results in the ITT population did not achieve statistical significance, additional exploratory analyses suggested a potential benefit in a subgroup of 20% of patients who exhibited PD‐L1 expression in ≥5% of immune cells. 20 , 42 The phase 3 JAVELIN 100 trial compared standard platinum‐based chemotherapy followed by avelumab maintenance, chemotherapy plus avelumab followed by avelumab maintenance, and chemotherapy followed by observation (ClinicalTrials.gov identifier NCT02718417 ). The primary endpoint of PFS was not met. The median PFS was 16.8 months in the avelumab maintenance group, 18.1 months in the avelumab combination group, and not reached in the control group. At a planned interim analysis, an independent data‐monitoring committee determined that neither avelumab arm would demonstrate a PFS benefit over chemotherapy alone, and the trial ultimately was terminated early because of futility. 19 In newly diagnosed, BRCA wild‐type, advanced ovarian cancer, the randomized phase 3 ENGOT‐OV43/GOG‐3036/KEYLYNK‐001 study did not demonstrate a significant improvement in PFS with the addition of pembrolizumab to first‐line chemotherapy followed by olaparib maintenance (ClinicalTrials.gov identifier NCT03740165 ). Bevacizumab was permitted at investigator discretion. A potential signal was observed in patients with high PD‐L1 expression (combined positive score [CPS], ≥10). 17 However, the absence of homologous recombination deficiency (HRD) stratification and the lack of an olaparib‐alone maintenance comparator arm make it difficult to isolate the contribution of pembrolizumab or to define the subgroup most likely to benefit. FIRST/ENGOT‐OV44 evaluated dostarlimab combined with first‐line platinum‐based chemotherapy, followed by maintenance therapy with both dostarlimab and niraparib, in newly diagnosed ovarian cancer regardless of BRCA or HRD status (ClinicalTrials.gov identifier NCT03602859 ). Patients were randomized to platinum‐based chemotherapy with niraparib maintenance alone or platinum‐based chemotherapy plus dostarlimab followed by maintenance with both agents. The combination arm showed a statistically significant, but clinically modest, improvement in PFS compared with niraparib alone (median PFS, 20.6 vs. 19.2 months; HR, 0.85; 95% CI, 0.73–0.99; p  = .0351), and no overall survival (OS) benefit was observed. Subgroup analyses indicated no differential PFS benefit in the PD‐L1–positive (Tumor Area Positivity, ≥5%) population. Approximately one half of the patients also received concurrent bevacizumab. 16 The phase 3 DUO‐O trial compared bevacizumab, durvalumab, and olaparib versus bevacizumab with or without durvalumab after platinum‐based chemotherapy (ClinicalTrials.gov identifier NCT03737643 ). The PFS for the triple maintenance regimen was 24.2 months for bevacizumab and durvalumab it was 20.6 months, and for bevacizumab alone it was 19.3 months. However, the lack of an arm with PARP inhibitor and bevacizumab impedes conclusions regarding the contribution of each agent to bevacizumab's effect. 15 , 43 In contrast, in ATHENA‐COMBO, the addition of nivolumab to rucaparib in newly diagnosed ovarian cancer did not improve PFS; indeed, it was associated with numerically inferior PFS relative to rucaparib alone (15.0 vs. 20.2 months; HR, 1.3; 95% CI, 1.1–1.5; ClinicalTrials.gov identifier NCT03522246 ). 18 The evidence across the KEYLYNK‐001, FIRST, DUO‐O, and ATHENA‐COMBO trials suggests that the PFS benefits observed in the ICI plus PARP inhibitor combination trials are largely driven by the PARP inhibitor, with immunotherapy contributing minimally and possibly detrimentally in the first‐line setting. In platinum‐sensitive, recurrent ovarian cancer, ICI combinations have shown biologic plausibility and some activity in selected early phase studies, but randomized data have not consistently demonstrated a clinically meaningful benefit. Notably, the strongest signals derive from single‐arm studies enriched for favorable prognostic features (favorable histologic subtype, less pretreated population, PARP‐naive and BRCA ‐mutated disease), whereas the randomized ATALANTE and ANITA trials (ClinicalTrials.gov identifiers NCT02891824 and NCT03598270 , respectively) were negative; the biologic rationale for these combinations thus has not been confirmed in controlled data. In a single‐arm phase 2 trial ( n  = 38; 20 platinum‐sensitive, 18 platinum‐resistant), nivolumab plus bevacizumab demonstrated higher activity in platinum‐sensitive ovarian cancer (PSOC) than in platinum‐resistant ovarian cancer (PROC; ORR, 40% vs. 16%), with a median PFS of 12 and 7 months, respectively. In the platinum‐sensitive cohort, histologic subtypes included 69% serous, 20% adenocarcinoma, 5% carcinosarcoma, and 10% clear cell. In addition, 85% of patients had BRCA wild‐type disease, and 90% had received up to two prior lines of therapy. The proportion of patients with increased PD‐L1 expression by CPS was lower in the platinum‐sensitive population at 39%, with an ORR of 42%. In contrast, in the platinum‐resistant group, a CPS ≥1 was observed in 83% of patients, but the ORR was only 6.7%, which contradicts expectations. 33 The MEDIOLA phase 2 trial (ClinicalTrials.gov identifier NCT02734004 ) included PARP‐naive patients and evaluated the combination of olaparib and durvalumab, which showed differential outcomes based on BRCA status: The ORR was 71.9% in BRCA ‐mutated patients ( n  = 51) and 34.4% in BRCA wild‐type patients ( n  = 32), with a PFS of 11.1 and 5.5 months, respectively. 22 In the phase 3 ATALANTE trial, patients with recurrent PSOC received platinum‐based chemotherapy with bevacizumab plus either atezolizumab or placebo, followed by maintenance. The study did not meet its co‐primary PFS endpoints in either the ITT or PD‐L1–positive populations. Eighteen percent of patients had received prior PARP maintenance, and 11.7% were BRCA ‐mutated. HRD was not reported. Subgroup analysis indicated better outcomes with atezolizumab among patients who had a platinum‐free interval >12 months, no prior bevacizumab treatment, tumor size <5 cm, and cancer antigen 125 (CA 125) levels <100 U/mL. These results warrant cautious interpretation because these factors are not specific to immunotherapy alone. 23 , 44 Further follow‐up demonstrated an exploratory OS signal in the ITT population, with a median OS of 35.8 months for atezolizumab versus 30.6 months with placebo (HR, 0.79; 95% CI, 0.65–0.96); PD‐L1 status was not predictive of benefit. 45 Preclinical observations indicate that PARP inhibitors upregulate PD‐L1 expression. 41 The ANITA phase 3 trial was designed to evaluate the contribution of anti–PD‐L1 therapy to platinum‐based chemotherapy and niraparib maintenance. The trial did not demonstrate that adding atezolizumab to platinum‐based chemotherapy, followed by maintenance with atezolizumab and niraparib, improved PFS compared with niraparib maintenance alone. Up to two prior lines of therapy were allowed, but prior treatment with PARP inhibitors in the recurrent setting was an exclusion criteria. Frontline PARP inhibitor use was permitted. In the study population, 15% had BRCA mutations, and 11% had received prior PARP inhibitor therapy. 21 The two randomized trials had different designs: ATALANTE added atezolizumab to platinum‐based chemotherapy plus bevacizumab, whereas ANITA combined it with chemotherapy followed by niraparib maintenance. Neither trial improved PFS. Although the final analysis of ATALANTE indicated a clinically meaningful increase in OS, the study did not meet its primary endpoint and is considered negative. 46 The role of new immunotherapy combinations in maintenance for platinum‐sensitive disease, especially after PARP inhibitor or bevacizumab maintenance, is being assessed. 47 Triple‐combination approaches targeting tumor cells, immune cells, and the TME have also been evaluated in this setting. A separate, nongermline BRCA ‐mutated MEDIOLA cohort investigated the combination of olaparib, durvalumab, and bevacizumab in recurrent PSOC ( n  = 31) and demonstrated a median OS of 31.9 months, a PFS of 14.7 months, and a disease control rate at 24 weeks of 77.4%. 48 Similarly, the OPEB‐01/APGOT‐OV4 single‐arm, phase 2 trial ( n  = 44; ClinicalTrials.gov identifier NCT03461370) evaluated maintenance therapy with olaparib, pembrolizumab, and bevacizumab in women with recurrent PSOC, mainly PARP‐naive, BRCA wild‐type. The trial reported a 6‐month PFS rate of 88.6% and a median PFS of 22.4 months. 24 Although these results are encouraging, both studies were nonrandomized and lacked comparator arms. These data are hypothesis‐generating. To date, no approved standard maintenance strategy exists for patients with relapsed PSOC who have previously received a PARP inhibitor and/or bevacizumab. It is established that patients with PSOC after PARP inhibitor progression respond poorly to platinum‐based chemotherapy, which represents a significant area for further research. Single‐agent PD‐1/PD‐L1 inhibitors have consistently demonstrated marginal activity in ovarian cancer across histologies and lines of therapy. In a phase 1b trial ( n  = 125), avelumab demonstrated an ORR of 9.6% and a median OS of 11.2 months in a heavily pretreated population. 36 Similarly, pembrolizumab in the KEYNOTE‐100 phase 2 study ( n  = 376; ClinicalTrials.gov identifier NCT02674061 ) resulted in an ORR of 8.0% and a median PFS of 2.1 months, irrespective of prior lines of therapy. 35 In another phase 2 trial focusing on PROC, nivolumab ( n  = 20) achieved an ORR of 15% and a median PFS of 3.5 months. 37 These signals did not translate into clinical benefit in the phase 3 NINJA trial (ClinicalTrials.jp identifier JapicCTI‐153004), in which nivolumab was inferior to physician's choice chemotherapy (gemcitabine or pegylated liposomal doxorubicin [PLD]), with no improvement in OS (10.1 vs. 12.1 months) and significantly worse PFS (2.0 vs. 3.8 months). 28 These findings suggest that anti–PD‐1/PD‐L1 monotherapy may have limited efficacy in ovarian cancer and does not outperform standard chemotherapy in unselected patients. Because of the limited efficacy of anti–PD‐1/PD‐L1 monotherapy, the addition of anti–CTLA‐4 therapy was explored as a strategy to enhance antitumor immunity. In the randomized phase 2 NRG GY003 trial ( n  = 100; 51 nivolumab plus ipilimumab, 49 nivolumab monotherapy; ClinicalTrials.gov identifier NCT02498600 ), nivolumab plus ipilimumab demonstrated greater activity compared with nivolumab alone in patients with recurrent disease and a platinum‐free interval <12 months. The combination achieved a higher ORR (31.4% vs. 12.2%) and longer median PFS (3.9 vs. 2.0 months; HR, 0.53; 95% CI, 0.34–0.82), but OS did not significantly differ (HR, 0.79; 95% CI, 0.44–1.42). Grade 3 adverse events were reported in 55.1% of patients who received single‐agent nivolumab compared with 66.7% in those who received the combination therapy. Notably, about one third of the patient population was considered platinum‐sensitive, including 12% with clear cell carcinoma, which exhibited a five times higher odds of response. 30 In addition, the trial lacked a chemotherapy comparator arm, and the median PFS of 3.9 months falls below what is achieved with chemotherapy and bevacizumab in PROC. Gotistobart, a novel, pH‐sensitive, anti–CTLA‐4 antibody engineered to enhance intratumoral T‐regulatory cell (Treg) depletion with reduced peripheral toxicity, has been evaluated in combination with pembrolizumab in the randomized phase 2 PRESERVE‐004/GOG‐3081 trial ( n  = 62; ClinicalTrials.gov identifier NCT05446298 ) in a heavily pretreated population, producing an ORR of 21%, a median PFS 2.2 of months, and a median OS of 18.9 months. 31 These data are preliminary and derive from an interim analysis of a phase 2 dose‐optimization study. Combinations with chemotherapy have produced inconsistent results. In another phase 2 trial, pembrolizumab plus PLD ( n  = 26) achieved an ORR of 26.1% and a PFS of 5.6 months. 32 In JAVELIN Ovarian 200 (ClinicalTrials.gov identifier NCT02580058 ), a phase 3 trial, avelumab plus PLD did not significantly improve outcomes over PLD in patients who had PROC, with no significant differences in PFS or OS. 27 The first positive study for OS in this setting was the phase 3 ENGOT‐ov65/KEYNOTE‐B96 trial (ClinicalTrials.gov identifier NCT05116189 ). That study evaluated the combination of pembrolizumab and weekly paclitaxel, with or without bevacizumab, in patients with PROC who had received up to two prior lines of therapy. Overall, 72% of the population had a CPS ≥1, 73% of patients received bevacizumab, and 40%–50% had a platinum‐free interval <3 months. At the first interim analysis, the addition of pembrolizumab led to a statistically significant improvement in PFS in both the CPS ≥1 and ITT populations, with benefit observed regardless of bevacizumab use. The second interim analysis indicated a statistically and clinically significant improvement in OS in the CPS ≥1 population with a median OS of 18.2 months versus 14.0 months (HR, 0.76; p  = .0053). 25 Metronomic chemotherapy combined with bevacizumab and an ICI represents an opportunity to augment cytotoxic lymphocyte function while potentially modulating harmful regulatory T‐cell responses and optimizing vascularization within the TME (Figure  2 ). Weekly paclitaxel remains one of the most active agents in this setting and likely serves as an effective backbone for this combination. The absolute PFS benefit remains modest, and costs and toxicities of a triple combination must be factored into clinical decision‐making. Patient subgroups that derive the greatest benefit remain to be fully defined, and correlational translational analyses are eagerly awaited. Nonetheless, the ENGOT‐ov65/KEYNOTE‐B96 trial is a landmark trial and supports this regimen as a new standard of care for patients with PD‐L1–positive PROC. Similarly, the combination of pembrolizumab, bevacizumab, and oral metronomic cyclophosphamide has shown activity in a phase 2, single‐arm study that included both platinum‐sensitive and platinum‐resistant populations. The overall ORR was 47.5%, 60.0% in the platinum‐sensitive population, and 43.3% in the platinum‐resistant population. The overall PFS was 10 months, with a median PFS of 20.2 months in the platinum‐sensitive population and 7.6 months in the platinum‐resistant population. 29 Synergistic modulation of the TME by combination therapy to enhance antitumor immunity. Schematic representation of the synergistic effects of combination therapy within the TME. Chemotherapy induces immunogenic cell death, leading to the release of neoantigens and enhanced antigen presentation by APCs, which promotes T‐cell priming and activation. Antiangiogenic therapy normalizes the tumor vasculature, facilitating improved immune cell trafficking and infiltration into the tumor. Concurrently, PD‐L1 inhibition restores T‐cell function by blocking immune checkpoint signaling, thereby enhancing cytotoxic T‐cell–mediated antitumor activity. Together, these mechanisms contribute to a more immunogenic TME and a potentiated antitumor immune response. APCs indicates antigen‐presenting cells; DC, dendritic cell; MHC I, major histocompatibility complex I; TME, tumor microenvironment; PD‐1/2, programmed cell death 1 and 2 proteins; PD‐L1, programmed death‐ligand 1; TCR, T‐cell receptor. The phase 2 European Organization for Research and Treatment of Cancer trial EORTC 1508 (EudraCT 2015‐004601‐17; ClinicalTrials.gov identifier NCT02659384 ) evaluated the combination of bevacizumab, atezolizumab, and acetylsalicylic acid in patients with PROC. Patients were randomized to receive bevacizumab alone, atezolizumab with or without acetylsalicylic acid, or the triple combination. The addition of atezolizumab to bevacizumab resulted in a numerical improvement in median PFS (4.1 vs. 2.3 months), although the primary endpoint was not met. The triple combination was well tolerated but did not confer additional efficacy. Translational analyses provided biologic insight, demonstrating an increase in TILs in the atezolizumab‐containing arms after two treatment cycles. Higher TIL infiltration was associated with a longer time to first subsequent therapy, and exploratory analyses suggested greater benefit in patients who had baseline TILs and PD‐L1–positive tumors. 26 In the TOPACIO/KEYNOTE‐162 trial ( n  = 62; ClinicalTrials.gov identifier NCT0265789), which included patients with wild‐type BRCA disease, the combination of niraparib and pembrolizumab was evaluated in recurrent ovarian cancer. This phase 1/2 trial reported an ORR of 18% and a PFS of 3.4 months, regardless of platinum sensitivity or prior bevacizumab treatment. It is important to note that these patients had not previously received PARP inhibitors. 34 Non‐high grade serous histologies, particularly OCCC and endometrioid ovarian carcinoma, are associated with endometriosis and exhibit distinct molecular and immunologic features that may render them more responsive to immunotherapy than their counterparts with high‐grade serous histology. Both subtypes display a high prevalence of ARID1A mutations, activation of the PI3K/AKT/mTOR pathway, and a cytokine‐rich microenvironment characterized by high interleukin 6 and VEGF signaling. 49 , 50 These features contribute to an immunologically dynamic tumor milieu, including increased infiltration by TILs and higher expression of immune checkpoints, which may underlie greater sensitivity to immunotherapy compared with other epithelial ovarian cancer subtypes. 51 In OCCC, single‐agent ICI studies have reported encouraging outcomes regardless of PD‐L1 status, with an ORR of 25%, a median PFS of 2.7 months, and an OS of 14 months. 52 Combination studies of nivolumab and ipilimumab have reported an ORR of 50%, with responses correlating with tumor mutation burden. 53 Lenvatinib plus pembrolizumab has also exhibited activity in patients who had pretreated OCCC, with an ORR of 30% and a median PFS of 10.9 months. 54 A recent meta‐analysis demonstrated that the ORR with ICI monotherapy is 14%, increases to 32% with dual checkpoint blockade, and reaches 33% when combined with tyrosine kinase inhibitors. 55 Endometrioid ovarian carcinoma shares overlapping molecular features with OCCC, including frequent ARID1A and PIK3CA mutations, but is further distinguished by recurrent CTNNB1 mutations, PTEN loss, and—particularly in low‐grade tumors and those arising in the context of Lynch syndrome—mismatch repair deficiency and microsatellite instability high status. The molecular overlap with endometrioid endometrial carcinoma, in which ICIs have demonstrated robust activity in the mismatch repair deficiency/microsatellite instability high subgroup, provides a strong biologic rationale for evaluating immunotherapy in this histology. Although dedicated prospective ICI trials in endometrioid ovarian carcinoma are scarce, subgroup analyses from basket and histology‐agnostic studies, as well as pembrolizumab approvals for mismatch repair deficiency/microsatellite instability high solid tumors, support its use in this molecularly selected population. Hypermutated and POLE‐mutated endometrioid tumors represent additional subsets in which ICIs may be particularly effective, mirroring observations in uterine endometrioid disease. Low‐grade serous carcinoma typically demonstrates resistance to chemotherapy. MEK inhibitor therapy has emerged as a new standard treatment, particularly for patients with KRAS mutations. 56 Endocrine therapy represents another therapeutic option. Ongoing clinical trials are investigating ubamatamab, a MUC16xCD3 bispecific antibody, administered alone or in combination with cemiplimab in this patient population. 57 In the single‐arm phase 2 NOGGO PERCEPTION study (ClinicalTrials.gov identifier NCT04575961 ), the combination of pembrolizumab and platinum‐based chemotherapy, followed by pembrolizumab maintenance, achieved the primary endpoint of 12‐month PFS (12 of 33 patients were progression‐free; median PFS, 8.4 months), indicating potential efficacy in platinum‐sensitive, recurrent, low‐grade, serous ovarian cancer. 58

The immunosuppressive nature of the TME has driven the rapid development of adoptive cell therapies designed to modulate host immune cells, enhancing their ability to target and eliminate tumor cells. Adoptive cell therapy encompasses a range of constructs, including chimeric antigen receptor (CAR) T‐cell therapy, T‐cell receptor–engineered T cells, T‐cell receptor fusion construct, TIL therapy, and natural killer (NK) cell therapy (Table  2 ). Ongoing clinical trials assessing adoptive cell therapies, including natural killer cells. Abbreviations: ±, with or without; B7‐H3, B7 homolog 3; CA 125, cancer antigen 125; CAR, chimeric antigen receptor; CAR T cells, chimeric antigen receptor T cells; CIML, cytokine‐induced memory‐like; EOC, epithelial ovarian cancer; FSHR, follicle‐stimulating hormone receptor; HGSC, high‐grade serous carcinoma; IHC, immunohistochemistry; IL, interleukin; IP, intraperitoneal; iPSC, induced pluripotent stem cells; IV, intravenous; mAbs, monoclonal antibodies; NCT, ClinicalTrials.gov identifier; NK cells, natural killer cells; NKT cell, natural killer T cell; PROC, recurrent, platinum‐resistant ovarian cancer; PRROC, recurrent, platinum‐resistant, refractory ovarian cancer; ROC, recurrent ovarian cancer; TILs, tumor‐infiltrating lymphocytes; TROP2, trophoblast cell surface antigen 2; ULN, upper limit of normal; VEGF, vascular endothelial growth factor. The application of CAR T‐cell therapy in solid tumors, including ovarian cancer, has yielded limited success to date, with numerous challenges remaining to be addressed. Major obstacles include the immunosuppressive and hypoxic TME, high tumor heterogeneity, impaired CAR T‐cell proliferation and persistence, and the limited specificity of tumor‐associated antigens (TAAs). 59 Notable TAAs expressed in ovarian cancer can be targeted such as SSEA‐4, B7‐H3, B7‐H4, HER2, MUC16, CD70, mesothelin, TAG72, claudin‐6, NYESO‐1, survivin, trophoblast cell surface antigen 2 (TROP2), melanoma‐associated antigen (MAGE)‐A4, follicle‐stimulating hormone receptor (FSHR), and folate receptor alpha (FRα). 1 , 60 , 61 , 62 Early clinical trials evaluating CAR T‐cell therapies targeting mesothelin and FRα in ovarian cancer have demonstrated an acceptable safety profile but limited clinical efficacy. An ongoing phase 1 trial is evaluating an FSHF chimeric endocrine receptor T cell administered either intraperitoneally or intravenously in patients with PROC or sex cord–stromal ovarian cancer expressing FSHR. Preliminary results have demonstrated a favorable safety profile, with no dose‐limiting toxicities. 63 In a phase 1 study, fully human antimesothelin CAR T‐cell (huCART‐meso) therapy was evaluated in 20 patients (14 with ovarian cancer). The best overall response was stable disease in 60%, with a median PFS of 12.3 weeks, and the cells were detectable in five patients at 12 months and in one patient for over 2 years, demonstrating improved persistence over earlier constructs. 64 To boost in vivo expansion and persistence, vaccine‐based strategies (peptide, nanoparticle, messenger RNA) are being combined with CAR T‐cell therapy. These approaches aim to stimulate in vivo CAR T‐cell proliferation, thereby improving therapeutic efficacy. 65 Preliminary results from the phase 1 BNT211‐01 trial ( n  = 22 in dose escalation, including patients with CLDN6‐positive ovarian cancer; ClinicalTrials.gov identifier NCT04503278 ), evaluating CLDN6‐targeted CAR‐T cells in combination with a CLDN6‐encoding CAR T‐cell–amplifying RNA vaccine, have demonstrated activity in CLDN6‐positive ovarian cancer and improved CAR T‐cell persistence, with an unconfirmed ORR of 33% and a disease control rate of 67% across evaluable patients. 66 TILs are detectable in 51%–83% of ovarian cancer tumors. 67 However, whereas the presence of TILs is essential, it alone does not guarantee clinical efficacy. 68 , 69 Despite early phase trials demonstrating disease stabilization in ovarian cancer, ORRs remain modest compared with other tumors such as melanoma. This is likely attributable to the immunosuppressive TME and TIL exhaustion. 70 , 71 For these limitations, strategies to enhance TIL function, such as an ICI during TIL expansion, are under investigation. 72 NK cells are increasingly being studied as an alternative adoptive cell therapy in ovarian cancer. Unlike T cells, NK cells do not require prior sensitization or recognition of specific antigens by major histocompatibility complex, allowing them to target a broader range of tumor cells, which is advantageous in heterogeneous tumors such as ovarian cancer. 73 NK‐cell–based therapies are being investigated in early clinical trials evaluating adoptive transfer of ex vivo expanded NK cells, combination strategies, and genetically engineered approaches, such as chimeric antigen receptor–NK cells. 74 Initial studies, including studies of peripheral blood‐derived or cord blood–derived NK cell infusions, have demonstrated acceptable safety profiles and evidence of immune activation, although clinical efficacy remains modest and heterogeneous. 75 Several challenges limit clinical translation, including the highly immunosuppressive TME, impaired NK‐cell persistence and trafficking, and lack of standardized manufacturing and treatment protocols. 76 Ongoing research is focused on optimizing NK‐cell expansion and activation, overcoming microenvironmental resistance, and identifying predictive biomarkers. There is growing interest in cancer vaccines, driven by advances in technologies such as messenger RNA vaccines and computational approaches that have enabled the identification of TAAs like CA 125, HER2/neu, FRα, mesothelin, MUC1, and p53, as well as tumor‐specific antigens, such as MAGE‐1, MAGE‐3, GAGE‐1/GAGE‐2, and BAGE. Innovative delivery platforms and combination strategies with other immunotherapies are being explored to enhance antitumor immune responses (Table  3 ). Ongoing clinical trials with vaccine/dendritic cell therapy. Abbreviations: ±, with or without; CA 125, cancer antigen 125; chemo, chemotherapy; EOC, epithelial ovarian cancer; FRα, folate receptor alpha; IGFBP‐2, insulin‐like growth factor binding protein 2; IL, interleukin 12; NACT, neoadjuvant chemotherapy; NCT, ClinicalTrials.gov identifier; NY‐ESO‐1, New York esophageal squamous cell carcinoma 1; PARP, poly(adenosine diphosphate‐ribose) polymerase; PD‐1, programmed cell death protein 1; PUMVC3‐IGFBP‐2, insulin‐like growth factor binding protein 2–targeted vaccine construct; VEGF, vascular endothelial growth factor. A meta‐analysis of 32 clinical studies, including 426 patients with recurrent ovarian cancer, reported an ORR of 4%, with a median PFS of 13 months and a median OS of 39 months. Most vaccine approaches targeted TAAs, such as CA 125, p53, HER2/neu, MUC1, mesothelin, and survivin, using platforms including peptide‐based, dendritic cell (DC), whole tumor cell, and antibody‐based vaccines. 77 Among peptide‐based approaches, a phase 1 study of maveropepimut‐S (DPX‐Survivac; n  = 47; ClinicalTrials.gov identifier NCT01416038 ), a survivin‐targeted T‐cell–activating vaccine, combined with intermittent low‐dose cyclophosphamide in patients with recurrent ovarian cancer, demonstrated immunogenicity, durable antigen‐specific T‐cell responses, and an acceptable safety profile, and it provided the rationale for subsequent combination studies with checkpoint blockade. 78 Maveropepimut‐S in combination with low‐dose cyclophosphamide and pembrolizumab demonstrated an ORR of 23% and a disease control rate of 67% in patients with advanced ovarian cancer. Survivin‐specific immune responses occurred in 62.5% of patients and were correlated with clinical benefit. 79 In line with efforts to introduce cancer vaccination earlier in the treatment course, the FLORA phase 2 trial (ClinicalTrials.gov identifier NCT04498117 ) evaluated the combination of oregovomab, an anti–CA 125 monoclonal antibody, with niraparib in patients with PSOC. Among the 10 patients enrolled, one achieved a partial response, and five had stable disease, resulting in a disease control rate of 60% at the 12‐week assessment. Because this was a single‐arm trial without a comparison group, the individual effects of oregovomab and niraparib cannot be distinguished. 80 In the maintenance setting, Gemogenovatucel‐T (an autologous tumor cell vaccine engineered to suppress furin and downstream TGFβ1/β2 activation; Gradalis, Inc.) demonstrated a favorable trend in a randomized phase 2 maintenance trial (n = 91; recurrence‐free survival, 11.5 vs. 8.4 months; HR, 0.69; 90% CI, 0.44–1.07; one‐sided p  = .078) but did not meet its primary endpoint. 81 Further follow‐up demonstrated a particular benefit in the non‐HRD and high clonal tumor mutation burden groups. 82 Similarly, the TEDOVA phase 2 trial ( n  = 185; ClinicalTrials.gov identifier NCT04713514 ), a multi‐neo‐epitope vaccine, improved PFS as a maintenance strategy in combination with pembrolizumab in the platinum‐sensitive relapse setting in patients with HLA‐A2 positivity versus best supportive care. 47 The TPIV200 multiepitope FRα peptide vaccine plus granulocyte‐macrophage colony–stimulating factor versus granulocyte‐macrophage colony–stimulating factor alone in PSOC was well tolerated in a phase 2 trial ( n  = 119) but did not improve PFS. 83 An ongoing phase 2 trial is further exploring maintenance using the UV1 vaccine, which targets telomerase, in combination with olaparib and durvalumab in patients with recurrent ovarian cancer. 84 Another emerging approach is the use of shared antigen vaccines, which target TAAs commonly expressed in multiple cancer types. These off‐the‐shelf approaches offer a cost‐effective alternative to personalized vaccines and have the potential for broader applicability. Vaccines that incorporate multiple antigens are also being explored to increase the likelihood of eliciting a robust and diverse immune response. 85 In a cohort of 21 patients with recurrent ovarian cancer, a polyvalent antigen–KLH + OPT821 vaccine plus bevacizumab demonstrated immunogenicity but did not translate into improved outcomes, underscoring that antigen‐specific responses alone may be insufficient to alter clinical outcomes. 86 DC therapy is an innovative approach aimed at enhancing the antitumor immune response in ovarian cancer by using the antigen‐presenting capabilities of DCs to stimulate T cells against TAAs. Several key studies have explored the efficacy and safety of DC vaccines in ovarian cancer. The strongest clinical signal comes from the autologous DC‐based vaccine (DCVAC) phase 2 program, in which where sequential administration after chemotherapy improved PFS in the first‐line setting ( n  = 90; SOV01; ClinicalTrials.gov identifier NCT02107937 ) and OS in the recurrent setting ( n  = 60; SOV02; ClinicalTrials.gov identifierNCT02107950). 87 A translational analysis of the SOV01 trial revealed that patients with cold tumors (low tumor mutational burden, scarce CD8‐positive T‐cell infiltration) derived the greatest benefit from DCVAC, whereas patients with hot tumors benefited more from chemotherapy alone. 88 This suggests that DC vaccines may be most valuable precisely in the patient population least likely to respond to ICIs. Oncolytic viruses have been investigated across multiple tumor types for their capacity to selectively infect and kill cancer cells. Once inside, these viruses replicate and induce direct cell lysis and immunogenic cell death. In addition to their intrinsic cytolytic activity, oncolytic viruses serve as promising platforms for delivering therapeutic transgenes—such as cytokines, ICIs, or other antitumor agents. 89 Despite encouraging preliminary results in ovarian cancer, oncolytic virotherapy remains under investigation, and its clinical utility has yet to be fully established. 90 It is emerging as a promising adjunctive strategy to enhance the efficacy of other therapeutic modalities, including chemotherapy and immunotherapy (see Table  S1 ). One potential strategy to overcome the challenges posed by the TME is the development of combination therapies aimed at enhancing the efficacy of immunotherapies. The rationale for combination approaches stems from the multifaceted nature of immune evasion mechanisms within the TME, which limit the effectiveness of single‐agent immunotherapy. 91 By combining immunotherapy with other treatments, such as targeted therapies, chemotherapies, or cell cycle checkpoint inhibitors, it may be possible to modulate the TME to a more favorable state for immune activation. 25 , 92 , 93 Integration of conventional cancer treatments, such as chemotherapy and radiotherapy, can induce immunogenic cell death, a process that enhances the release of TAAs and stimulates antigen‐presenting cells. This cascade promotes T‐cell priming and activation, ultimately amplifying the antitumor immune response. 94 The combination of chemotherapy with ICIs has already demonstrated efficacy in several solid tumors and remains an area of active investigation in ovarian cancer. 95 A challenge in ovarian cancer immunotherapy is the presence of a highly immunosuppressive TME, which impairs the function of effector immune cells and promotes tumor progression. Several strategies are being explored to counteract immune suppression and enhance the efficacy of immunotherapy. 45 One approach involves targeting key immune suppressive cell populations, including Tregs, myeloid‐derived suppressor cells (MDSCs), and M2‐polarized tumor‐associated macrophages (TAMs). Tregs play a crucial role in maintaining immune tolerance by suppressing the activity of cytotoxic T cells. 96 In ovarian cancer, increased Treg infiltration is associated with a poor prognosis and resistance to immunotherapy. Strategies to deplete Tregs or inhibit their function include low‐dose cyclophosphamide or anti‐CCR4 antibodies. 97 Similarly, MDSCs are a heterogeneous population of immature myeloid cells that exert potent immunosuppressive effects by producing arginase‐1, reactive oxygen species, and nitric oxide synthase. These factors contribute to T‐cell dysfunction and inhibit DC maturation. Targeting MDSCs through inhibitors of STAT3 or CXCR2 can help reduce their suppressive effects and improve immune checkpoint blockade efficacy. 98 Reprogramming TAMs from an M2 to an M1 phenotype is another promising strategy for modulating the TME. M2‐polarized TAMs support tumor growth by secreting pro‐angiogenic factors, such as VEGF and TGF‐β. 99 Therapeutic interventions that block CSF1R signaling or repolarize TAMs using Toll‐like receptor agonists or CD40 agonists can promote M1 polarization and enhance antitumor immunity. 100 , 101 In addition to targeting immune cells, efforts are underway to normalize the tumor vasculature and improve immune cell infiltration. 102 Abnormal tumor vasculature creates physical barriers to immune cell trafficking and contributes to hypoxia, which further enhances immune suppression. Targeting hypoxia‐inducible pathways using hypoxia‐inducible factor 1 α inhibitors or metabolic modulators, such as IDO1 (indoleamine 2,3‐dioxygenase 1) inhibitors can help counteract the immunosuppressive effects of hypoxia and improve T‐cell function. 103 Antibody–drug conjugates (ADCs) designed to selectively target DNA damage repair–related pathways can enhance tumor susceptibility to ICIs, thereby increasing T‐cell activation and overcoming immune evasion. 104 Several DNA‐damaging agents are currently being developed as ADC payloads, including camptothecin derivatives that inhibit DNA topoisomerase I, an enzyme essential for maintaining genomic stability; topoisomerase I inhibition can also trigger immune responses. 105 Furthermore, DNA damage repair–targeted therapies can induce immunogenic cell death, a process characterized by the release of damage‐associated molecular patterns such as calreticulin, adenosine triphosphate, and HMGB1 (high mobility group box 1). These damage‐associated molecular patterns serve as potent stimulators of innate immunity, promoting DC maturation and facilitating antigen cross‐presentation to T cells. 106

Immunotherapy in ovarian cancer is evolving from empirical approaches toward more precise, biology‐driven strategies. However, progress remains constrained by substantial interpatient and intratumoral heterogeneity, 1 predominantly an immunosuppressive and metabolically adverse TME, and the lack of validated predictive biomarkers. Addressing these challenges will require a deeper understanding of tumor–immune interactions, supported by advanced technologies such as spatial and single‐cell profiling to better define mechanisms of immune evasion. It is important to consider the TME within a broader systemic context because host‐related factors, including metabolism, inflammation, microbiome composition, circadian rhythms, nutritional status, sex, and age, may significantly influence immune responses and therapeutic outcomes. In parallel, stromal and metabolic components, as well as emerging modulators, such as neural signaling and nerve–tumor interactions, are increasingly recognized as contributors to immune exclusion. Innovative therapeutic approaches, including adoptive cell therapies and cancer vaccines, offer additional opportunities to personalize immunotherapy. Ultimately, meaningful advances in this field will depend on the integration of comprehensive tumor and host profiling, biomarker‐driven patient selection, and more adaptive and mechanistically informed clinical trial designs. Such approach is essential to address the complex interplay among immune, stromal, metabolic, and systemic factors that underlie therapeutic resistance in this biologically heterogeneous disease.

Ovarian cancer remains a leading cause of gynecologic cancer–related mortality worldwide, with a 5‐year relative survival rate of approximately 50%. Among its histologic subtypes, high‐grade serous ovarian carcinoma accounts for nearly 70% of cases. The remaining 30% consists of other histologic variants, including endometrioid ovarian cancer, clear cell ovarian carcinoma (OCCC), mucinous ovarian carcinoma, and low‐grade serous ovarian carcinoma. Each of these subtypes has unique molecular profiles and distinct clinical behavior. 1 The foundation of treatment in ovarian cancer remains a combination of cytoreductive surgery and platinum‐based chemotherapy. 2 A deeper understanding of the molecular vulnerabilities of ovarian cancer, coupled with advances in precision oncology, has led to the incorporation of targeted treatments, such as poly(adenosine diphosphate‐ribose) polymerase (PARP) inhibitors and antiangiogenic agents. 3 , 4 , 5 , 6 , 7 , 8 , 9 In recent years, immunotherapy has emerged as a transformative treatment modality across multiple solid tumors. In particular, immune checkpoint inhibitors (ICIs) have produced durable responses and redefined therapeutic standards in melanoma and lung cancer. 10 , 11 In ovarian cancer, however, early phase studies demonstrated limited activity, with objective response rates (ORRs) generally below 15% in recurrent disease despite prognostic indicators linking immunogenicity with outcome. Approximately one half of high‐grade serous ovarian carcinoma tumors exhibit significant intratumoral T‐cell infiltration at baseline, an independent predictor of improved outcome. 12 , 13 Slow clinical progress has led to the recognition that ovarian cancer comprises distinct immune phenotypes, described as cold , immune‐excluded , and hot tumors, defined by the presence and spatial distribution of tumor‐infiltrating lymphocytes (TILs) and the surrounding microenvironment. 14 As illustrated in Figure  1 , these phenotypes provide a conceptual framework to understand resistance to immunotherapy and to guide the development of more rational, biology‐driven therapeutic strategies. In this review, we examine the clinical evidence for immunotherapy in ovarian cancer, discuss tumor‐intrinsic and microenvironmental mechanisms of resistance, and outline priorities for more rational, biomarker‐driven development. Schematic illustrating the spectrum of tumor immune phenotypes and therapeutic strategies to reprogram the TME in ovarian cancer. ( Left ) The cold tumor microenvironment is characterized by immune cell exclusion, hypoxia, dense ECM, and nutrient deficiency and is dominated by immunosuppressive cell populations, including MDSCs, M2‐like TAMs, and Tregs. ( Right ) In contrast, hot tumors display increased immune cell infiltration, activation of cytotoxic lymphocytes, normalized vasculature, reduced ECM density, and improved metabolic conditions. ( Center ) The intermediate immune‐excluded phenotype is defined by the presence of immune cells at the tumor periphery with limited intratumoral penetration, often associated with stromal barriers, vascular dysfunction, and persistent metabolic constraints. Therapeutic strategies aimed at promoting the transition from cold or immune‐excluded tumors to hot tumors include immune checkpoint inhibitors, targeting of immunosuppressive cell populations, stromal remodeling, and metabolic modulation as well as emerging approaches involving the microbiome and neural–tumor interactions. CTL indicates cytotoxic T lymphocyte; CTLA‐4, cytotoxic T‐lymphocyte–associated protein 4; ECM, extracellular matrix; IDO1, indoleamine 2,3‐dioxygenase 1; LAG‐3, lymphocyte‐activation gene 3; MDSC, myeloid‐derived suppressor cell; NK, natural killer cell; PD‐L1, programmed death‐ligand 1; TAMs, tumor‐associated macrophages; TIM‐3, T‐cell immunoglobulin and mucin domain–containing 3; TME, tumor microenvironment; Tregs, regulatory T cells; VEGF, vascular endothelial growth factor.

Ana Carolina Veneziani reports personal/consulting or advisory fees from AbbVie and Merck Sharp & Dohme; and honoraria from AstraZeneca and Merck Sharp & Dohme outside the submitted work. Amit M. Oza reports grants/contracts from AstraZeneca Canada and Bristol Myers Squibb Canada; service as a principal investigator and on steering committees with AstraZeneca and GlaxoSmithKline; personal/consulting or advisory fees from AbbVie, AstraZeneca Canada, Bristol Myers Squibb Canada, GlaxoSmithKline, Merck Sharp & Dohme, and pharma&; support for other professional activities from Daiichi‐Sankyo Inc.; and is the chief executive officer of Ozmosis Research outside the submitted work. Laura Venegas disclosed no conflicts of interest.

Table S1