Introduction

Bone and soft tissue sarcomas (BS/STS) represent a heterogeneous group of mesenchymal malignancies. They account for approximately 1% of adult and 15% of pediatric malignancies (Beckingsale and Shaw 2017). Overall prognosis is poor, and up to 50% of sarcoma patients develop metastatic disease (Italiano et al. 2011; Marko et al. 2016). With survival rates around 12–24 months under treatment with standard chemotherapy, prognosis remains dismal in advanced and metastatic stages (Van Glabbeke et al. 1999; Lochner et al. 2020). In the context of systemic therapy, there has only been little progress in the treatment of BS/STS in the past 3 decades. In STS, despite the current knowledge about clinical and biological differences, patients have been treated in a “one-size-fits-all” approach with anthracyclines and alkylating agents (Katz et al. 2018). Histology-tailored chemotherapy regimens have not led to improved outcomes over the standard first-line chemotherapy regimen in STS (Gronchi et al. 2020). The only FDA-approved first-line targeted therapy in STS represents tyrosine kinase inhibitors (TKI) including imatinib for gastrointestinal stroma tumors (GIST) and dermatofibrosarcoma protuberans (DeMatteo et al. 2009). Systemic therapy for BS includes conventional chemotherapeutic agents as part of subtype specific protocols (Ferrari et al. 2018; Brennan et al. 2020; Smeland et al. 2019). Drug resistance is believed to cause treatment failure in over 90% of patients with metastatic cancer, which underlines the need for additional therapy lines and novel therapeutic approaches in patients with BS/STS (Longley and Johnston 2005).

Next-generation sequencing (NGS) enables parallel sequencing of RNA and DNA from (archived) tumor tissue. It has affected diagnostic and therapeutic management in many advanced malignancies and has been increasingly used for focused biomarker screening in colorectal cancer, non-small-cell lung cancer (NSCLC), and breast cancer (Van Cutsem et al. 2016; Sosman et al. 2012; Gennari et al. 2021; Planchard et al. 2018; Chakravarty et al. 2022). Taking into account the frequency of actionable alterations and a potential therapeutic benefit, the European Society of Medical Oncology (ESMO) recommends the routine use of multigene NGS in advanced NSCLC, prostate cancer, ovarian cancer, and cholangiocarcinoma in addition to tumor-specific molecular targets (Mosele et al. 2020). Importantly, there are now six drugs FDA-approved in a histology–agnostic fashion for a total of six different biomarkers, underscoring the need for scaling of tumor profiling within sustainable structures.

In BS/STS, the role of NGS in routine clinical practice is not yet defined. Implementation of molecular tumor boards (MTB) for interdisciplinary discussion is a valid option to integrate NGS results into clinical routine. At the CCCLMU Munich, a precision oncology program was initiated in 2016. Heinrich et al. demonstrated feasibility of the program in a retrospective analysis of the first 1000 patient contacts of the program. In this single-center analysis, 41% of patients received a treatment recommendation based on comprehensive genomic profiling (CGP). Despite the high rate of treatment recommendations, only 17% of all patients received treatment based on MTB discussion, being in line with previously published reports (Heinrich et al. 2022; Tannock and Hickman 2019).

Here, we aim to determine the feasibility and impact of routine NGS and molecular tumor boards in patients with BS/STS treated in large German Sarcoma referral center. Our results indicate that a relevant proportion of patients with BS/STS harbor druggable alterations, but only a fraction ultimately receive recommended treatments. This study highlights current challenges in the clinical implementation of comprehensive genomic profiling in BS/STS.

Materials and methods

Patient selection

Eligible patients had pathologically confirmed bone or soft tissue sarcoma (BS/STS). Treatment options were discussed in a multidisciplinary team within the LMU Munich Sarcoma Center (SarKUM) between August 2016 and August 2022. If at least one of the criteria for comprehensive genomic profiling (young age, rare histopathological subtype, and limited therapeutic options) was met, patients were presented to the CCCLMU Munich Precision Oncology program after discussion within the sarcoma tumor board. Clinical, pathological, and outcomes data were extracted from our MTB database.

NGS-based genomic profiling

Comprehensive genomic profiling (CGP) via targeted NGS was performed at the LMU Munich Institute of Pathology. Patients with externally performed CGP were registered and discussed within our MTB in select cases. Starting in 2016, in-house panel sequencing was performed with Oncomine™ Focus Assay, a 52-gene panel, which was continuously replaced by the 161-gene panel Oncomine™ Comprehensive Assay (Thermo Fisher, Darmstadt, Germany) and the TruSight Oncology (TSO) 500™ Assay (Illumina, San Diego, USA). All panels allow simultaneous RNA and DNA sequencing with detection of insertions/deletions (indels), gene fusions, single-nucleotide variants (SNV), and copy number variations (CNV). Additionally, tumor mutational burden (TMB) was evaluated in select cases (Oncomine™ Tumor Mutational Load Assay, ThermoFisher Scientific). CGP was either performed on FFPE tumor tissue or liquid biopsies (e.g., cerebrospinal fluid, blood, or ascites). Material from primary tumor, metastases, or locally recurrent tumor was used. Young patients with rare sarcomas were referred to the National Center for Tumor Diseases (NCT) and German Cancer Consortium (DKTK) Molecularly Aided Stratification for Tumor Eradication (MASTER) program for whole genome/exome and transcriptome analysis in select cases (Horak et al. 2017, 2021).

Interpretation of molecular tumor board recommendations

Following molecular testing, results were individually discussed in a dedicated molecular tumor board (MTB) consisting of pathologists, medical oncologists, geneticists, and representatives from organ-specific tumor boards. To define clinical actionability of molecular alterations, a structured database query was performed. In addition, an on-site literature database was created within the LMU Munich MTB. For this study, druggable alterations were assigned to baskets based on the cellular pathway involved according to a modified classification by Horak et al. (2021): Tyrosine kinases (TK), PI3K-AKT-mTOR (PAM), DNA damage repair (DDR), RAF-MEK-ERK (RME), Cell cycle (CC), Immune evasion (IE), and IDH mutations (IDH). Further druggable alterations not associated with one of these cellular pathways were termed Other (OTH). After discussion of NGS results in our MTB, clinical implementation of MTB recommendations was discussed in a multidisciplinary team within the LMU Munich Sarcoma Center (SarKUM).

Statistical analysis

Descriptive and statistical analyses, as well as the generation of graphs were performed with IBM SPSS Statistics Version 28.0, R Version 4.2.2 and Microsoft 365 Version 2206. The comparison of the mean age of two groups was performed using the Student’s t test. Survival time was calculated from the initial diagnosis to either death or the date of last contact. Statistical significance was determined as a p value < 0.05. Linear regression was computed to calculate the association between date of registration and the number of days to MTB discussion. To measure the clinical benefit of MTB-based therapies, the progression-free survival ratio (PFSr, PFS under MTB-based therapy divided by PFS of prior systemic therapy) was calculated (Von Hoff et al. 2010). The chosen cut-off to measure a therapeutic effect was set to > 1.3 according to previous studies (Massard et al. 2017; Horak et al. 2021).

Results

Patient characteristics and molecular tumor board workflow

The clinicopathologic characteristics of the study cohort are summarized in Table 1. In total, 92 patients with histologically confirmed bone or soft tissue sarcoma (BS/STS) and presentation at our molecular tumor board (MTB, n = 89) and/or the DKTK MASTER program (n = 5) between August 2016 and August 2022 were analyzed (Fig. 1). Median age at initial diagnosis was 49 years (range 0–80 years), and 52.2% of patients were female. The median time interval between initial diagnosis and MTB discussion was 18.7 months (range 0.3–283.2 months). For patients with synchronous metastatic disease, this time interval was 7.6 months (range 0.3–85.8 months). Most common histological subtypes were liposarcoma (n = 11, 12%), undifferentiated pleomorphic sarcoma (n = 11, 12%), leiomyosarcoma (n = 9, 9.8%) and chondrosarcoma (n = 9, 9.8%). 84 sarcomas (91.4%) were classified as high grade. Metastatic disease was present in 43.5% of patients at initial diagnosis and in 84.8% of patients at date of CGP. Costs for NGS testing were covered by health insurance providers in all patients. While the number of sequenced and discussed patients increased over time (Fig. 2), turnaround time between CGP and MTB discussion decreased over the examined period (median 62.5 days in 2016/2017 vs. 33 days in 2021, β = − 6.153, p = 0.20). After exclusion of two external patients with high turnaround times, the effect became significant (β = − 5.512, p = 0.014). 82 patients (89.1%) underwent at least one local or systemic treatment before discussion of CGP results, with a median number of two previous systemic (range 0–6) and two local (range 0–10) therapies (Table 2). MTB referral was initiated by our medical oncology department in 75.8% of cases.

Table 1 Baseline characteristics
Full size table
Fig. 1
figure 1

Molecular tumor board workflow in patients with bone and soft tissue sarcomas. CGP comprehensive genomic profiling

Full size image
Fig. 2
figure 2

MTB discussions over the examined study period

Full size image
Table 2 List of patients with therapeutic change after MTB discussion
Full size table

CGP assays were performed as previously described (Heinrich et al. 2022). In most patients, FFPE tumor tissue was used for comprehensive genomic profiling. Testing by liquid biopsy was only done in two patients (2.2%). For whole exome/genome and transcriptome sequencing within the DKTK MASTER program, new biopsies were performed to obtain fresh frozen tumor samples for five patients (5.4%). NGS analysis was successful in 93.5% of patients, and insufficient quality of tumor material was the main reason for unsuccessful testing. One patient was discussed twice due to additional CGP to detect further alterations. The performed CGP and MTB workflow can be seen in Fig. 1.

Other histological subtypes include 2 intimal sarcomas, 1 sclerosing epithelioid fibrosarcoma, 1 epithelioid sarcoma, 2 clear cell sarcomas, 2 solitary fibrous tumors, 1 phosphaturic mesenchymal tumor, 2 CIC-rearranged sarcomas, 1 alveolar soft part sarcoma, 1 adenosarcoma

Molecular alterations and actionability

CGP was technically successful in 86 of 92 patients (93.5%), and molecular alterations were detected in 71 patients (82.6%). In 32 patients (37.2%), at least one druggable alteration was found. The most common alterations were mutations in TP53 (n = 20; 23.3%), followed by PIK3CA and MDM2 alterations (9.3% each). The most common druggable alterations were mutations in IDH (7% of patients) and CDK4/6 (8.1% of patients) (Figs. 3 and 4). Histological subtypes with no druggable alterations were rhabdomyosarcoma and subtypes termed as “Other” including ultra-rare STS such as clear cell sarcoma or sclerosing epithelioid fibrosarcoma (Stacchiotti et al. 2021). As previously mentioned, druggable alterations were assigned to baskets based on the cellular pathway involved according to a modified Horak classification (Horak et al. 2021). Expected disease-defining gene alterations were detected in several histological subtypes, including MDM2 and CDK4 in liposarcoma, IDH mutations in chondrosarcoma, and alterations related to tyrosine kinases in GIST (KIT, PDGFR). Of all detected alterations, 53.5% were associated with a loss and 46.5% with a gain of function. Mutations, amplifications, and fusions accounted for 64.3%, 23.6%, and 12.1% of alterations, respectively. Tumor mutational burden (TMB) was assessed in 48 patients (52.2%), and TMB was high (≥ 10 mutations per megabase) in three patients (6.3%).

Fig. 3
figure 3

Molecular alterations in BS/STS. Other: rare alterations identified only once in our study collective

Full size image
Fig. 4
figure 4

Proportion of actionable alterations in BS/STS subtypes divided into alteration baskets. TK tyrosine kinases, PAM PI3K-AKT-mTOR, DDR DNA damage repair, RME RAF-MEK-ERK, IDH isocitrate dehydrogenase, CC Cell cycle, IE immune evasion, NA no actionable alterations

Full size image

Targeted treatment recommendations

In 32 patients (37.2%), CGP detected an actionable alteration and led to a treatment recommendation in our dedicated molecular tumor board. Based on the MTB recommendation or a positive DKTK MASTER report in two cases, our sarcoma-specific tumor board recommended a targeted therapy in 15 patients (17.4%). This finally resulted in an actual treatment change at disease progression in ten patients (11.6%). The remaining five patients did not receive the recommended therapy due to death, rapid tumor progression (n = 4), or due to a therapy recommendation for an expected progression not yet occurred. The treatment group included two chondrosarcomas, one Ewing sarcoma, one osteosarcoma, two angiosarcomas, two liposarcomas, one myxofibrosarcoma, and one GIST. Mean age at initial diagnosis was lower (M = 38 years, SD = 0.05) than in the whole cohort (M = 53 years, SD = 15.53) (p = 0.046). In the treatment group, median duration from initial diagnosis to CGP was 13.6 months, and the median turnaround time was 41.5 days. Nine patients were treated with off-label therapies after approval by their respective health insurance. The patient with GIST received an FDA-approved therapy with imatinib after NGS and detection of an uncommon PDGFRA exon 12 mutation. Six patients derived clinical benefit with a progression-free survival ratio (PFSr) > 1.3, and one patient who received a first-line MTB-based therapy had a PFS of more than three months (Table 2).

Discussion

Over the last decades, progress has been slow in the treatment of bone and soft tissue sarcomas (BS/STS). While multigene NGS has led to a better understanding of the genetic landscape of sarcomas and allows better subgrouping, it does not yet impact therapeutic algorithms in daily clinical practice. Except for GIST and dermatofibrosarcoma protuberans, there are no approved therapies based on molecular targets specifically for BS/STS. In this retrospective single-center analysis, we set out to evaluate the use of comprehensive genomic profiling (CGP) to drive molecular tumor board recommendations in patients with BS/STS. Although there are several reports on single-center experiences with molecular tumor boards, the role of NGS in routine clinical practice in BS/STS is not sufficiently defined (Cote et al. 2018; Gusho et al. 2022). Current ESMO guidelines recommend molecular pathology for STS when it is required for diagnosis or when it “may have prognostic and/or predictive relevance” (Gronchi et al. 2021). For BS, ESMO guidelines remain equally vague (Strauss et al. 2021).

In our cohort, 82.6% of patients had at least one molecular alteration. Not surprisingly, TP53 mutations were found as the most common alterations (23.3% of patients), followed by PIK3CA and MDM2 alterations (9.3% each). These alterations are commonly found in BS/STS and correlate with other single-center experiences on NGS in sarcomas (Cote et al. 2018; Gusho et al. 2022; Lucchesi et al. 2018). Expectedly, most molecular alterations including TP53 mutations did not have a therapeutic consequence. While TP53 mutations are frequently found in tumor sequencing, they currently do not carry therapeutic relevance and are often associated with poor prognosis in several tumors including breast cancer, Ewing sarcoma, or colorectal cancer (Andrikopoulou et al. 2021; Huemer et al. 2018; Tirode et al. 2014; Hu et al. 2021). Isocitrate dehydrogenase (IDH) mutations were found in more than half of patients with chondrosarcoma, being consistent with previous studies (Amary et al. 2011). The efficacy of IDH inhibitors was analyzed in several studies including a phase I trial demonstrating a median PFS of 5.6 months for advanced chondrosarcoma patients (Tap et al. 2020). Interestingly, no druggable alterations were found in our group of ultra-rare sarcomas, such as sclerosing epithelioid fibrosarcomas, intimal sarcomas, or CIC-rearranged sarcomas. In addition to the low number of actionable alterations in translocation-associated sarcomas (Lucchesi et al. 2018), possible explanations could lie in the commercial panels adopted in our study and lack of clinical trials for these tumor entities, which limits the discovery of potential therapeutic targets. Specific studies including CGP in this ultra-rare group are needed for better diagnosis and therapy.

High tumor mutational burden was identified as a predictive parameter for response to immunotherapy in multiple cancer types (Marabelle et al. 2019; Yarchoan et al. 2017). This led to the FDA approval of the PD-1 inhibitor pembrolizumab for TMB-high (TMB, ≥ 10 mutations per megabase) solid tumors. In our cohort, only three patients (6.3% of assessed patients) met this condition, which is consistent with previous literature on BS/STS (Yarchoan et al. 2019). As only a minority of sarcoma patients qualify as TMB-high, the PEMBROSARC trial suggested the immune-active tumor microenvironment with tertiary lymphoid structures as an alternative predictive parameter of response to immunotherapy (NCT02406781).

37.2% of patients harbored druggable alterations leading to a therapeutic recommendation of the MTB and/or the DKTK MASTER program (n = 5). This rate of druggable alterations is in line with previous studies on comprehensive tumor profiling in sarcomas. Lucchesi et al. demonstrated at least one targetable alteration in 41% of patients with soft tissue sarcomas (Lucchesi et al. 2018). Gusho et al. discovered clinically actionable alterations in 47.1% of patients with BS/STS, and Cote et al. demonstrated a druggable alteration in around 40% of patients (Gusho et al. 2022; Cote et al. 2018). In the DKTK MASTER cohort, Horak et al. presented a rationale for targeted therapies beyond current guidelines in 86.9% of patients with rare cancers including bone and soft tissue sarcomas and 31.8% were treated accordingly. In our cohort, all patients (n = 5) included in the DKTK MASTER program received a therapeutic recommendation based on their sequencing results. In the end, only one patient received a molecularly targeted therapy after discussion in this program. Due to the identification of an alternative transcript (ATI) on ALK, the patient with osteosarcoma underwent a targeted therapy with the ALK-inhibitor crizotinib, which would likely not have been identified in our in-house panel testing (Schoch et al. 2020). This emphasizes the value of a combination of whole exome, genome, and transcriptome sequencing in soft tissue sarcomas and could represent a valid alternative to in-house panel NGS testing, which might not sufficiently reflect the diverse biology of BS/STS.

In contrast to other studies, the rate of therapeutic implementation in our cohort is high (31.3% of patients with actionable alterations). In other studies, MTB recommendations were followed through in 3–31.8% of patients with druggable alterations (Rodler et al. 2021; Heinrich et al. 2022; Dorman et al. 2023; Gusho et al. 2022; Sultova et al. 2021; Horak et al. 2021). Compared to the overall cohort, the ten patients treated according to their MTB recommendation were younger (mean 38 vs. 53 years) possibly indicating a selection bias toward patients with a good performance status. Two patients received the MTB-based therapy in addition to conventional chemotherapy, emphasizing the value of personalized combination therapies to combat drug resistance. With seven of ten patients (70%) benefiting from MTB-based therapy in terms of disease stabilization for more than three months and six of eight patients (75%) deriving a clinical benefit defined as a PFSr > 1.3, our results should motivate for routine use of NGS testing for sarcoma patients. In comparison, Horak et al. demonstrated a PFSr > 1.3 under treatment with molecularly targeted therapies in 35% of soft tissue sarcoma patients, and specific histological subtypes such as synovial sarcoma or GIST (PFSr > 1.3 in 50% of patients, respectively) were among the entities associated with the most significant therapeutic benefits. In their study, cases with exceptional clinical benefit were often associated with detection of RNA-based biomarkers and treatment with TKIs. In contrast, targeted therapies appeared to be mostly ineffective in bone sarcoma patients (PFSr > 1.3 in 6% of patients) (Horak et al. 2021). Due to the low sample size in our analysis, possible selection bias and variety of histological subtypes, it is currently difficult to draw conclusions on our high rate of therapeutic changes and benefits. While several studies suggest that NGS has a significant impact in aiding diagnosis and selecting targeted therapies in sarcomas, clinical trials such as the MULTISARC trial (NCT03784014) are needed to determine the final role of comprehensive genomic profiling in sarcomas.

One of our main findings is the latency between initial diagnosis and NGS analysis: the median time from initial diagnosis to MTB discussion was 18.7 months (range 0.3–283.2 months). For patients with synchronous metastatic disease, this time interval was 7.6 months (range 0.3–85.8 months). 84.8% of patients demonstrated metastatic disease at initiation of CGP compared to 43.5% of patients at initial diagnosis. In contrast, Rodler et al. observed that genitourinary cancer patients were mainly included at primary presentation with metastatic disease (Rodler et al. 2021). In pancreatic cancer, a cancer with a notoriously bad prognosis, Dorman et al. observed a median time of 5.5 months between initial diagnosis and discussion in the MTB, owing to the fact that most patients with metastatic pancreatic cancer undergo tumor profiling at initial presentation (Dorman et al. 2023). Regarding the dismal prognosis of BS/STS patients with metastatic disease, the long interval between initial diagnosis and discussion in the MTB raises the question whether comprehensive genomic profiling will positively impact the outcome of patients at later stages of their respective disease. In fact, it has been shown that most patients benefit from targeted agents rather early in the course of their disease (Subbiah and Kurzrock 2018; Westin and Kurzrock 2012). A growing number of molecularly guided clinical (basket) trials are being conducted across the globe, often with limited treatment slots available in a given disease. As such, early and broad integration of biomarker testing helps to identify patients that could potentially benefit from innovative treatments within a clinical trial (Mateo et al. 2022). This assumption is supported by the fact that patients with dedifferentiated liposarcoma characteristically carry MDM2 (and CDK4/6) amplifications (Binh et al. 2005) and novel MDM2 inhibitors are tested in the first-line setting (NCT05218499).

Limitations of our study are the relatively small and very heterogeneous cohort of patients, which remains a typical finding in rare cancers such as bone and soft tissue sarcomas. Our study however reflects the experience of one of the largest sarcoma centers in Germany and Europe. To improve evidence for the routine use of NGS, a structured registry and follow-up program has been initiated at our institution which will allow further evaluation of the benefit of the CCCLMU Munich molecular tumor board (Heinrich et al. 2022).

Conclusion

With comprehensive genomic profiling becoming a cornerstone in the management of patients with advanced malignancies and the growing number of molecularly guided treatment options, we anticipate that an increasing number of patients with bone and soft tissue sarcomas will derive benefit from the concept of precision oncology. In fact, our experience shows that a significant number of patients with bone and soft tissue sarcomas benefit from targeted treatment options when treated at a large-volume center and in the setting of a dedicated precision oncology program.