Chordoma is a rare malignant tumor that arises from the embryonic remnants of the notochord. It almost always arises along the axial skeleton, the sacrum, skull base, and mobile spine being the most common locations (1, 2). A review of population-based studies showed an estimated incidence of only 0.08 to 0.5 cases per 100,000 people per year, demonstrating its rarity and exceptionality even within the spectrum of bone malignancies (3, 4). However, despite representing only 1% to 4% of all primary bone tumors and approximately 5% of malignant bone tumors, chordoma has a significant clinical impact and morbidity due to its often-complex anatomical location and its propensity for local invasion and recurrence (5, 6).

Although chordomas tend to grow more slowly than other tumors, their behavior usually is insidiously aggressive. Despite radical surgery and adjuvant radiotherapy (RT), local recurrence remains common, and long-term disease control is difficult to achieve (2, 7). Advanced local treatment modalities such as proton beam or carbon ion therapy can be considered in order to optimize local control and minimize collateral damage to adjacent critical structures (8, 9). However, complete resections are often not feasible, and even high-dose RT cannot always prevent relapse or may be associated with local complications. Metastatic spread is less common than in other sarcomas, but once present, it is associated with limited therapeutic options and a poor prognosis (1, 7). Conventional systemic chemotherapy (ChT) has shown negligible benefit, and to date, no systemic ChT has become an established standard of care (7, 10).

Thus, other therapeutic strategies have been explored. The most investigated therapies have been tyrosine kinase inhibitors (TKI). Among these, imatinib was assessed in a prospective phase II trial. A total of 50 Response Evaluation Criteria in Solid Tumors (RECIST)-evaluable patients with advanced chordoma treated with imatinib achieved an objective response rate (ORR) of only 2%, although approximately 70% experienced stable disease, with a clinical benefit rate (complete/partial response or stable disease ≥6 months) of 64%. The median progression-free survival (PFS) was approximately 9 months. Median overall survival (OS) has been reported as 34 months in prospective and retrospective series (10, 11). Another multitarget TKI with good results is sorafenib, evaluated in a phase II trial by the French Sarcoma Group, which demonstrated an ORR of 3.7%, a 9-month PFS rate of 73%, and a 12-month OS rate of 86.5%, showing clinically significant disease stabilization in a subgroup of patients (12). These results illustrate both the potential and limitations of TKI: whereas they can provide temporary disease control, they rarely induce durable responses, and their use remains off-label.

More recently, immune checkpoint inhibitors (ICI) have shown potential promising activity in chordoma, although their role in routine practice remains unclear. In a key report, Migliorini et al. described 3 patients with recurrent chordoma treated with ICI after failure of conventional therapies, achieving durable tumor stabilization or partial responses, and 1 patient remained progression-free for more than 12 months (13). A more recent single-patient case report detailed a patient with metastatic chordoma with a PBRM1 A1209fs mutation who was treated with pembrolizumab and achieved a PFS of 9.3 months (14). In a review of treatment-refractory chordomas, Alexander et al. identified 8 case reports where programmed death-1/programmed death-ligand 1 (PD-1/PD-L1) inhibitors (sometimes combined with TKI or vaccine approaches) were used, with variable results: some patients had partial responses or prolonged disease stabilization, but not a consistent pattern of durable responses was found (15). In addition, a recent analysis of immune checkpoint expression and tumor immune microenvironment in chordomas revealed upregulation of PD-1, cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), and other immune checkpoint genes, and selective enrichment of memory T cells and macrophages, suggesting a biological rationale for their therapeutic targeting (16).

Furthermore, a recent phase I multicenter clinical trial evaluated the anti–PD-L1 monoclonal antibody FAZ053, alone or combined with spartalizumab (anti–PD-1), in patients with advanced malignancies, including a chordoma cohort. In that expansion cohort, FAZ053 monotherapy achieved an ORR of 30% (1 complete and 2 partial responses) and disease control in 70% of patients, with a manageable toxicity profile. In addition, biomarker analyses showed increased CD8+ infiltration and immune gene upregulation after treatment (17), providing prospective evidence of meaningful clinical activity of PD-(L)1 blockade in chordoma. There is also growing interest in combination immunomodulatory strategies: one report described the use of AdAPT-001 (a viral vector that delivers a TGF-β trap) to sensitize a recurrent chordoma to checkpoint inhibition, with evidence of radiographic tumor shrinkage (18). Taken together, these early experiences support the hypothesis that immunotherapy can achieve tumor control in a subset of chordomas. Definitive proof of its benefit, biomarkers of response, and optimal combination strategies remain to be established.

Given the challenges and the lack of solid evidence for systemic treatments, the management of advanced and metastatic chordoma continues to rely on multidisciplinary expertise, individualized decision-making, and participation in clinical trials, the latter of which are scarce. Therefore, we conducted a retrospective, real-world study in order to achieve a deeper understanding of treatment outcomes in clinical practice, including response patterns, recurrence dynamics, and survival.

Our single-center series reflects real-world, non-standardized treatment sequencing in a rare disease, shaped by availability of clinical trials, compassionate use access, and evolving evidence. The clinical characteristics of our cohort are consistent with those previously reported in other prospective and retrospective studies at referral centers. Patients were generally around 60 years old, conventional histology predominated, and the sacrococcygeal site was the most common primary site. Although all patients underwent surgery for the primary lesion and more than half received adjuvant RT, relapse was widespread, with most patients experiencing multiple local or systemic recurrences requiring therefore a systemic approach. This pattern is consistent with the indolent but recurrent behavior of chordoma and explains why many patients eventually exhaust local treatment options and require systemic therapy.

From a systemic treatment perspective, our cohort shows the therapeutic diversity of chordoma treatment, mainly due to the lack of standard therapeutic guidelines with robust evidence. Most patients received imatinib as first-line treatment, following the limited evidence available, while a few initiated ICI therapy. One patient received initial treatment with anthracycline-based ChT after an initial misdiagnosis of soft tissue sarcoma, illustrating how histological ambiguity can influence management in rare malignancies. Nearly half of the patients received more than 1 line of systemic therapy, including rechallenges with ICI and sequential transitions between TKI and ICI. Furthermore, 1 patient received sorafenib after immunotherapy. Treatment allocation was not randomized, and patients exposed to ICI likely represent a selected subgroup treated at later disease stages, which may introduce selection bias in subgroup comparisons. This heterogeneity is an example of the daily clinical reality in infrequent tumors without established standard treatment algorithms. Detailed and standardized toxicity data could not be systematically analyzed due to the long study period and the transition from paper-based to electronic medical records, which limited the reliability and completeness of adverse event reporting.

Despite this diversity, the observed therapeutic outcomes were in line with previously reported data. Objective responses occurred with both drug classes. The median overall PFS in our series was 12.3 months, slightly longer than the approximately 9 months reported in the pivotal phase II study with imatinib, while the median OS was long, exceeding 10 years. Moreover, our data adds to the heterogeneous but growing evidence of chordomas treated with ICI from case reports and clinical trials. Partial responses with anti-PD(L)1 drugs, as seen here, were also documented by Migliorini et al. (13) and in the recent phase I FAZ053 trial (17). The median PFS observed in our ICI-exposed patients is consistent with previously published data reported with pembrolizumab monotherapy (13, 14). These findings should be interpreted with caution, as numerically longer survival observed in the ICI-exposed subgroup may reflect treatment activity but is also likely influenced by selection bias, given that patients receiving immunotherapy within clinical trials are typically highly selected and may have more favorable prognostic features. Overall, these results suggest that systemic therapy may influence the disease course in a subset of patients.

Several limitations of this study should be acknowledged. Heterogeneity in systemic treatment exposure, including differences in agents used, treatment sequencing, and lines of therapy, introduces potential confounding. Given the retrospective design and small sample size, subgroup analyses are descriptive and hypothesis-generating. OS should be interpreted with caution, as it was measured from the date of diagnosis, whereas subgroup PFS was defined from the initiation of the subgroup-defining systemic therapy, resulting in different time-zero definitions. Together with the small sample size, the limited number of death events, and the indolent natural history of chordoma, this precludes meaningful comparative interpretation of OS between subgroups. In addition, competing risks such as non–cancer-related death may influence OS outcomes. Accordingly, OS estimates are presented to provide longitudinal clinical context and should be interpreted as descriptive and hypothesis-generating rather than reflective of treatment-related effects. The definition of the ICI-exposed subgroup is time-dependent and may be affected by immortal time bias. In addition, molecular profiling data were not systematically available, and radiological response assessments were retrospectively extracted from routine clinical reports without centralized or blinded review. These limitations underscore that the findings of this study are descriptive rather than confirmatory.

This study adds to the existing literature by providing single-center real-world data on systemic therapy outcomes in advanced chordoma across a long historical period, including both TKI and ICI exposure within the same institutional cohort. Unlike clinical trials or pooled retrospective analyses, our series reflects routine treatment sequencing, patient selection, and follow-up in daily clinical practice. These data are intended to complement, rather than directly compare with, previously published studies and should be interpreted as hypothesis-generating for future research.

Our study describes the clinical reality of chordoma in a referral center for sarcomas: a malignancy with a slow but persistent course, with frequent recurrences despite local treatment with surgery and RT, and non-standardized systemic therapy. TKI remain a rational first-line option for indolent disease, often achieving prolonged disease stabilization, while ICI may represent a promising therapeutic strategy capable of inducing meaningful responses in selected patients. Given the absence of standardized treatment sequencing and predictive biomarkers, prospective multicenter collaboration is needed to harmonize therapeutic strategies, refine patient selection, and define the role of immunotherapy in this rare and challenging disease.