Clinical utility of circulating tumor DNA sequencing with a large panel in patients with advanced soft‐tissue sarcomas

Abstract

To the editor Soft-tissue sarcomas (STS) represent a very heterogeneous group of rare tumors including more than 100 different subtypes [1]. Surgery and neo/adjuvant radiation therapy represent the cornerstone of treatment for STS. However, despite an optimal resection of the tumor, up to 40% of patients will develop metastatic relapse and will die from the disease [1]. Doxorubicin represents the first-line standard of care for patients with advanced disease since the 1970s, despite several attempts to identify better regimens. The median overall survival (OS) of patients with metastatic disease is < 18 months and has only modestly improved over the past 20 years [2]. We and others have previously reported that next-generation sequencing (NGS) of tumor tissues allows the identification of genomic aberrations with the potential to influence and personalize therapy in up to 50% of patients with advanced STS [3, 4]. Genomic profiling of circulating tumor DNA (ctDNA) is increasingly used to tailor therapy in cancer patients. Indeed, such liquid biopsy has several advantages: non-invasiveness, reduced turnaround times for faster results, and the ability to fully capture the landscape of tumor heterogeneity [5]. The aims of the present study were to investigate the impact of ctDNA profiling in a large cohort of patients with advanced STS included in two prospective precision medicine studies and to decipher the ctDNA molecular landscape of sarcoma. Between December 2020 and August 2021, 98 patients with advanced STS were included in two ongoing institutional molecular profiling studies (Bergonie Institute Profiling, BIP: NCT02534649; Gustave Roussy Profiling, STING: NCT04932525). Their characteristics are described in Supplementary Table S1. Genomic analysis (ctDNA and tissue when available) was performed by using the United States Food and Drug Administration (FDA) approved Foundation One Liquid CDx Assay (324 genes, tumor mutational burden [TMB], microsatellite instability status) [5]. All the results were discussed weekly among a multidisciplinary molecular tumor board (MTB). Actionable targets were classified using the European Society of Medical Oncology Scale for Clinical Actionability of Molecular Targets (ESCAT) [6], and matched therapies were proposed based on the molecular profiling results. The amount of ctDNA at diagnosis correlates with survival in patients with localized Ewing sarcoma and osteosarcoma [6]. Interestingly, we found that a tumor fraction ≥10% was significantly associated with short median OS in our series (13 months [95% CI, 0-26.3 months] vs. 22 months [95% CI, 13.9-30.1 months], P = 0.021) (Supplementary Figure S1). ctDNA NGS allowed the identification of at least 1 genomic aberration in 93 cases (94.9%). A total of 624 aberrations were identified in 186 genes. The proportion of patients with aberrations was higher in those suffering from complex genomics sarcomas (49.2%) than in those suffering from translocation-related sarcomas (21.6%). The distribution of the number of aberrations was also significantly different according to the sarcoma genomics group (complex genomics sarcomas, translocation-related sarcomas, and other sarcomas; χ2 = 0.032, P = 0.048). The 5 genes most frequently altered in the entire cohort were tumor protein 53 (TP53), ataxia telangiectasia mutated (ATM), tuberous sclerosis 1 (TSC1), checkpoint kinase 2 (CHEK2), and tyrosine-protein kinase (KIT). A total of 289 alterations were identified among all the genes explored in the NGS panel: 256 missense mutations, including substitutions or short indels plus splice variants, 27 truncation mutations/nonsense mutations, 1 gene amplification, 3 gene homozygous deletions, and 2 gene rearrangements. The 26 genes most frequently altered are represented in Supplementary Figures S2-S3. At least one actionable aberration that could be used to tailor therapy was identified in 31 patients (32.2%): 14 (31.1%) in the complex genomics sarcoma group, 5 (28.3%) in the translocation-related sarcomas group, and 12 (34.2%) in the other sarcoma group (P = 0.842). Targetable genes are presented in Supplementary Figure S4. Overall, MTB recommended a therapy matched to the genomic profile for 22 patients (22.4%). The 2 main reasons for not recommending a therapy in patients with an actionable alteration were: no clinical trial available for the patient (n = 6); the matched therapy was previously received by the patient (n = 2). Ten patients (10.2%) were treated with molecular targeted therapy (Supplementary Table S2, Supplementary Figure S5). The objective response and disease control (objective response and stable disease) rates as per RECIST 1.1 were 20.0% and 40.0%, respectively (Figure 1). The median progression-free survival and OS were 12 weeks and 20 weeks, respectively (Supplementary Figure S6). FIGURE 1 Objective responses observed in sarcoma patients who received therapy matched to actionable alterations identified through ctDNA sequencing. (A-B) Partial response was observed 6 weeks (B) after treatment onset with mTOR inhibitor in a 63-year-old female patient with advanced leiomyosarcoma harbouring a PTEN deletion. (C-D) Partial response was observed 6 weeks after treatment (D) onset in a 73-year-old male patient with advanced angiosarcoma harbouring a HRAS and KRAS mutation and treated with a combination of HRAS and MEK inhibitors. Abbreviations: ctDNA, circulating tumor DNA; mTOR, mammalian target of rapamycin; PTEN, Phosphatase and TENsin homolog; HRAS, Harvey Rat sarcoma virus; MEK, mitogen-activated kinase. Six patients were found to have a high TMB (> 16 MB). The MTB recommended treatment with an immune checkpoint inhibitor (ICI) for 4 patients (4.1%); 2 patients refused the recommended treatment. Two patients (2.1%) received an ICI. The best response was progressive disease in both. The two other patients did not receive ICI as per the physician's decision. Forty patients also underwent NGS of tissue samples besides ctDNA profiling. Tissue profiling was unsuccessful for 7 (17.5%) patients. No actionable alteration was found in both ctDNA and tissue profiles for 17 (42.5%) patients. The numbers of actionable aberrations in tissue and ctDNA profiles were similar in 5 (12.5%) patients. The number of actionable alterations was higher in tissue than in ctDNA for 10 (25.0%) patients. The 15 actionable alterations observed in tissue but not in ctDNA are listed in Supplementary Table S3: 10 of them (66.7%) were copy number variants (CNVs). The number of actionable aberrations was higher in ctDNA than in tissue for 1 (2.5%) patient. Tissue profiling allowed the identification of an ESCAT I, IIIA and IV aberration not presented in ctDNA for 1, 4 and 7 patients, respectively, whereas ctDNA profiling allowed the identification of an actionable aberration (ESCAT IIIA) not presented in tissue for 1 patient. Thus far, data on ctDNA in STS have been very limited. In patients with localized Ewing sarcoma and osteosarcoma, the amount of ctDNA at diagnosis correlates with survival [7]. Monitoring ctDNA can also have clinical implications particularly to evaluate response to treatment and disease progression [8]. In patients with gastrointestinal stromal tumor, ctDNA and tissue sequencing results were similar, suggesting that liquid biopsy represents a robust approach for tumor genotyping and treatment tailoring [8]. In the present study, we found that 66.7% of ctDNA findings were also identified in tissue profiles. We also found more alterations in tissue than in ctDNA profiling, with the majority (66.7%) being CNVs. This result is in line with a previous study showing lower sensitivity of ctDNA profiling in identifying CNVs [10]. Five patients also had a single nucleotide variant in tissue which was not found through ctDNA profiling. Among them, 3 patients had no detectable ctDNA. ctDNA profiling represents an alternative tool for detecting actionable alterations in STS but is less sensitive than tissue sequencing to detect CNVs (losses or amplifications). We have previously shown that NGS of tissue allowed the identification of actionable alterations in up to 41% of patients with advanced STS [5]. We demonstrated the capacity of ctDNA profiling to detect actionable aberrations in up to 38% of patients with STS. Future technological development would increase the sensitivity of this approach for CNVs even more. Conceptualization, Antoine Italiano; methodology, Julie Blanchi, Sofiane Taleb, Arnaud Bayle, Melissa Alame, Yechan Laizet, Paul Dubos; investigation, BV, EK, MT, MS, AI; resources, Antoine Italiano; writing—original draft preparation, Julie Blanchi, Antoine Italiano; writing—review and editing, Julie Blanchi, Sofiane Taleb, Antoine Italiano; supervision, Antoine Italiano; project administration, Antoine Italiano. All authors have read and agreed to the published version of the manuscript. The authors have nothing to report. AI received research grant and honoraria from ROCHE, BAYER, MSD, ASTRAZENECA, MERCK, PHARMAMAR, BMS, PARTHENON, CHUGAI, NOVARTIS. The other authors have no competing interests. FONDATION MSDAVENIR (HEART grant). The study was approved by the Institutional Review Board of Institut Bergonié (IRB 2-323Z). All patients signed informed consent. Not applicable. Source data are available on request to the corresponding author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts