Charting a Path for Prioritization of Novel Agents for Clinical Trials in Osteosarcoma: a Report from the Children’s Oncology Group New Agents for Osteosarcoma Task Force

Abstract

Osteosarcoma is the most common bone tumor in children and young adults. Metastatic and relapsed disease confer poor prognosis, and there have been no improvements in outcomes for several decades. The disease’s biological complexity, lack of drugs developed specifically for osteosarcoma, imperfect pre-clinical models, and limits of existing clinical trial designs have contributed to lack of progress. The Children’s Oncology Group Bone Tumor Committee established the New Agents for Osteosarcoma Task Force to identify and prioritize agents for inclusion in clinical trials. The group identified multi-targeted tyrosine kinase inhibitors, immunotherapies targeting B7-H3, CD47-SIRPα inhibitors, telaglenastat, and epigenetic modifiers as the top agents of interest. Only multi-targeted tyrosine kinase inhibitors met all criteria for front-line evaluation and have already been incorporated into an upcoming phase III study concept. The task force will continue to reassess identified agents of interest as new data becomes available and evaluate novel agents using this method.

Introduction:

Osteosarcoma (OS) is the most common malignant bone tumor in children, adolescents and young adults. Prognosis is largely dependent on the presence of metastatic disease, which occurs in up to 20% of newly diagnosed patients.[1] Furthermore, roughly 30% of patients presenting initially with localized disease will relapse in distant sites. Regardless of the presence of poor prognostic indicators, treatment for pediatric high-grade OS follows a standard approach including cytotoxic chemotherapy with three agents (methotrexate, cisplatin, and doxorubicin) and surgical resection of the primary and metastatic tumors.[2–4] Despite efforts to intensify cytotoxic chemotherapy or introduce novel agents, outcomes have remained the same for forty years, with 5-year overall survival close to 70% in patients with localized disease and less than 40% with metastatic disease at presentation.[1, 3] Survival in recurrent disease is largely based on whether surgical complete remission can be achieved. Researchers face a number of challenges in designing new therapeutic strategies for OS including the disease’s complex biology, a lack of drugs developed specifically for OS, and the imperfect nature of available pre-clinical models.

Biological Complexity:

The molecular complexity of OS has contributed to a lack of improvement over nearly four decades.[5] It has long been known that the OS genome contains a high burden of structural aberrations including copy number variations, chromothripsis and kategis. [6–11] While patterns have emerged,[12] the extent of heterogeneity between tumors makes identification of the genetic changes responsible for pathogenesis difficult to determine and challenging to target therapeutically. Alterations in TP53 are nearly universal, followed by RB1 loss. ATRX mutations, along with amplifications of CCNE1/MYC/PDGFRA/CDK4 and others occur in a subset. Alterations in the PI3K/mTOR pathway are also common findings. [13, 14] Unfortunately few of these alterations are easily or universally druggable. Additionally, the role of epigenetic regulation in tumorigenesis and metastasis is poorly understood. It has recently been recognized that establishment of metastatic tumors relies in part on shifting transcriptional output and altering enhancer profiles so that metastatic cells adapt to distant microenvironments.[15, 16] A better understanding of these mechanisms may enable development of therapies for prevention or treatment of metastatic lesions.[17]

Interactions between the tumor and the host immune system likely play an important role in tumorigenesis, metastatic spread and control of minimal residual disease in OS. Infiltration of CD8+ T cells in tumors and programmed cell death ligand 1 (PD-L1) expression in the immune microenvironment have shown prognostic significance in OS, [18–20] suggesting possible susceptibility to immune-based therapies. However, therapies targeting these interactions, such as immune checkpoint inhibitors, have been disappointing in the setting of recurrent measurable disease.[21–23] The lack of success with these and other immune targeting agents may be explained by the significant heterogeneity in the immunogenomic landscape of OS, in some cases contributing to an overall immunosuppressive phenotype.[24] For example, in contrast to many adult solid cancers, OS is heavily infiltrated with immune suppressive myeloid cells and methods to perturb their function are being actively pursued.[25, 26] Successful implementation of immune based therapies will require identification of predictive biomarkers to guide appropriate patient selection, and development of rational combinations to overcome resistance mechanisms.[27] Newer approaches such as proteomic profiling identifying strongly expressed surface proteins that may be targeted via antibody-drug conjugates or cellular therapies such as CAR T-cells are exciting avenues to pursue. [28–30]

Drug Development:

Similar to other rare diseases, drug development in OS has largely relied on repurposing of agents developed for more common diseases rather than developing de novo agents specific to OS. This strategy undoubtedly confers many advantages including significantly lower cost of development, shorter time to approval, and a historical knowledge of the safety, toxicity, dosing and drug interactions. [31] However, use of exclusively repurposed agents may contribute to the slow progress toward improvement in survival.

Use of Pre-clinical Models:

In the process of evaluating new agents, researchers rely on a variety of OS models such as established cell lines and murine models including cell line xenografts, patient derived xenografts (PDX), and genetically engineered mice.[12, 32–34] However, there is no standardized approach for evaluating agents whose mechanisms target tumor-infiltrating immune cells rather than tumor cells. For evaluation of these immune-mediated modulators, it is not entirely clear whether syngeneic immune-competent murine OS models, PDXs within humanized mouse models with varying degrees of humanizing immune systems, or larger animal species are sufficient demonstration of activity prior to clinical trials in humans.

One advantage in OS is the propensity for pet dogs to spontaneously develop OS, which is biologically similar to the human form of the disease, harboring similar rates of mutations and copy number alterations in common genes.[35–39] The Comparative Oncology Trials Consortium is a national clinical trials network that assesses novel therapies in pet dogs for future use in human cancer patients.[39] Evaluation of novel therapies in dogs could inform human cancer therapy through many avenues, including exploration of dose and schedule, pharmacokinetic/pharmacodynamic relationships determined through serial blood and tumor sampling, and recognition of efficacy. Further, pet dogs provide the opportunity to study novel therapies in the context of treatment-naïve disease and in the minimal residual disease setting. The execution of these rapid (compared to human trials) and scalable clinical studies in dogs is expected to provide high-value clinical and biologic data to rationally prioritize comparative therapeutic strategies in pediatric OS patients.[17, 40]

The translational value of any of these models to the human experience is difficult to assess given the lack of recent positive clinical trials despite efficacy of agents in animal models. This is perhaps best exemplified by eribulin. This agent demonstrated a robust signal of efficacy in the murine model, [41] but the resulting phase II clinical trial failed to meet its efficacy threshold.[42] This negative trial led to re-examination of eribulin in the original PDX models and multiple additional PDX models, again demonstrating excellent efficacy in murine models.[43] The authors demonstrate a steep dose-response curve for eribulin in PDX models, underscoring the importance of robust pharmacokinetic and pharmacodynamic assessments in both the preclinical and clinical settings to ensure effective concentrations are achieved.[43] Emerging evidence also suggests that the activity of eribulin may be in early metastasis development, a mechanism that may not be adequately evaluated in current study designs aimed at testing agents in the setting of relapsed disease.[44, 45]

Clinical Trial Design:

Despite significant effort and thought, the optimal threshold for moving an agent of interest into the clinical setting, and how to test efficacy within the clinical setting remains undetermined.[17] Unlike in other tumors, use of traditional endpoints such as objective response rate measured radiographically may not reflect efficacy in OS since new bone formation limits the ability of a tumor to reduce in size.[46] Since the standard treatment for isolated pulmonary nodules includes surgical resection, many patients may not have measurable disease after relapse precluding their enrollment in conventional phase I and II trials.[47] In response to these concerns the Children’s Oncology Group (COG) suggests future trials use event free survival (EFS) rather than objective response rate as a primary endpoint. Based on a review of 7 prior ineffective COG phase II studies in OS, the group proposed using 12% at 4 months as the EFS cutoff for patients with measurable disease and 20% EFS at 1 year for those with completely resected disease as the historical benchmark to compare new agents against.[48] Based on these historical benchmarks, agents demonstrating a 4 month EFS of 40% in patients with measurable disease or 12 month EFS of 50% in patients with no evidence of disease have in some phase 2 trials been considered worthy of moving forward from the phase 2 setting. This strategy has been successful in that tyrosine kinase inhibitors (TKI), a class of agents exceeding these historical benchmarks has been prioritized for development in a first line therapy randomized study based on activity in measurable disease. However, this strategy does not fully allow for investigation of agents that may perturb subclinical metastatic OS, and further consideration into clinical trial design with metastasis prevention and minimal residual disease (MRD) early in the disease course as novel endpoints is needed.

New Agents Task Force:

Acknowledging these challenges in developing effective new therapies for OS, the COG bone tumor committee established the New Agents for Osteosarcoma Task Force.[49] Prioritization of potential agents to move forward in COG clinical trials for the treatment of high-grade OS in both up-front and relapsed settings was the purpose of this working group. Members represented a variety of disciplines including clinicians, basic scientists, veterinary oncologists, and clinical investigators. The group met monthly to evaluate new agents based on basic science, preclinical, comparative oncology and clinical evidence. The task force also focused on identifying challenges to clinical trial development in OS given its unique biology and lack of clear therapeutic targets. In this report, we describe the progress of the task force to date.

To identify new agents for task force consideration, members were asked to submit a list of candidate agents and targets for discussion, with the goal to be inclusive of any agents members felt were worthy of consideration (Table 1). Evidence including proof of the agent hitting its target, in vitro activity, in vivo activity, and clinical trial data in both human and canine subjects was reviewed. Practical aspects of agent selection were also reviewed including clinical availability of the agent, whether a pediatric phase 2 dose was known, and whether the agent was approved by the United States Food and Drug Administration (FDA). (Figure 1 and supplemental table S1) Task force members were asked to rank their top five agents for further consideration. The top agents selected based on greatest interest and achievement of set criteria are listed in Table 2. Several unique and interesting targets with potential relevance in OS pathogenesis were discussed. One of the major revelations of this endeavor was that despite identification of these interesting pathways/ targets, there were none that met all criteria to readily move forward into a clinical trial with the exception of multi-tyrosine kinase inhibitors. Here, we present the agents and targets that received the most votes by task force committee members.

TABLE 1.

Agents evaluated by task force

Drug/Target Class	Example agents/targets
Multi-targeted tyrosine kinase inhibitors	cabozantinib, sorafenib, regorafenib, apatinib, cediranib, pazopanib,
Immunotherapy Types: Chimeric antigen receptor(CAR) T cells, Bispecific T-cell Engagers, monoclonal antibodies	Targets: B7H3, HER2, GD2, LRCC5 Functional Immune Targets: PD1/PD-L1, CTLA4, CD47, Tumor microenvironment targeting: losartan+sunitinib
Immune Modulators	GM-CSF, mifamurtide, anti-CCl2, oleclumab, anti-IL1, anti-IL6, anti-IL8
Metabolic modulators	telaglenastat, metformin
Epigenetic modifiers	vorinostat, panobinostat, Aza-TdCyd, decitabine
MTOR/PI3K pathway	everolimus, RV1001, ridaforolimus, alepelisib
Notch/Gamma Secretase inhibitor	RO4929097
DNA damage/repair	Parp inhibitors: olaparib ATR inhibitors: berzosertib (VX-970)
Cell cycle cyclin dependent kinase (CDK) inhibitors	CDK 4/6 inhibitors: palbociclib (CDK 4/6) CDK 7/12 inhibitors: THZ1 (CDK 7,12)
Regulators of p53	MDM2 inhibitor:ALRN-6924
Apoptosis Inducers	Smac mimetic: birinapant

Figure 1.

Process for evaluation of new agents for osteosarcoma. Candidate agents and targets were submitted by members of the task force. Pre-clinical and clinical data were assessed using the listed criteria and task force members were asked to submit their top five agents for further consideration. Agents were then prioritized for immediate inclusion in trial, removal from consideration, or await further data. CTEP: Cancer Therapy Evaluation Program, FDA: Food and Drug Administration

TABLE 2.

Top agents evidence review

Agent or Class	Basic Science Evidence	Preclinical Evidence	Clinical Evidence	Drug Availability	Trial Design Consideration	Consensus
Multi-tyrosine Kinase Inhibitor	OS	OS	OS	Yes	Feasible in combination with cytotoxic chemotherapy, Proposed next up front study	Move forward
B7-H3 Targeting Agents	OS	OS		Yes	Agent specific considerations; combination with cytotoxic chemotherapy, possible minimal residual disease setting	Await data from phase 1 study ( NCT04483778)
Anti-CD47 (HU-5F9, ALX-148)	OS	OS	C	Yes	Combination with dinituximab	Await additional data
CB-839 (teleglenastat)	OS	OS		Yes	Combination, targeting metastatic disease/progression	Await additional data
Epigenetic modifiers (azacytidine, decitabine, TdCyd, Aza-TdCyd)	OS	OS		Yes		Await data from multiple phase 1 studies ( NCT03366116, NCT03445858, NCT03628209)

Key: - = negative data, C= evidence in cancer, OS= evidence in osteosarcoma, blank = no data

Top Agents:

Multi-targeted Tyrosine Kinase Inhibitors:

A growing body of evidence demonstrates a clear signal of activity for multitargeted TKIs in measurable OS. These small molecule inhibitors disrupt several cancer cell and tumor microenvironment (TME) functions including growth factor signaling, cell proliferation and differentiation, stromal growth, and angiogenesis. A number of TKIs have been studied in relapsed and refractory OS patients, summarized in Table 3.[21, 50–53] Of the six highlighted TKI studies, the results of four exceed the aforementioned 4 month EFS efficacy threshold.[48] One study demonstrated superior median progression free survival of regorafenib compared with placebo in a randomized controlled setting.[21] Notably, early phase trials using these agents in OS evaluated efficacy based on progression free survival rather than regression of measurable disease, which allowed for these agents to succeed when they otherwise might have failed based on common standards for early phase trials in cancer.

Table 3.

Multi-targeted tyrosine kinase inhibitor comparison

	Sorafenib51	Lenvatinib50	Regorafenib49	Regorafenib21	Cabozantinib52
N	35	30	26	22	42
Partial Response	8.6%	8%	7.7%	13.6%	12% (4–26)
4m PFS (95% CI)	46% (28–63)	33% (17–54)	62% (40–77)	44.4 N/A	N/A
6m PFS (95% CI)	N/A	N/A	N/A	N/A	33% (20–50)
Median PFS-m (95% CI)	4 (2–5)	3.4 (1.8–6.5)	4.1 (2–6.5)	3.6 (2.0–7.6)	6.7 (5.4–7.9)
Median OS (95% CI)	7 (7–8)	N/A	11.3 (5.9–23.9)	11 (4.7–26.7)	10.6 (7.4–12.5)

m- month, PFS- progression free survival, CI- confidence interval

The single agent activity of several multi-targeted TKIs demonstrated in multiple phase 2 clinical trials makes these agents an exciting prospect for future study in OS.[21, 50, 52–55] While direct comparisons of efficacy across the phase 2 trials aren’t possible, there are several attributes of the agents that influence their prioritization for incorporation onto chemotherapy backbones. Regorafenib has a black box hepatotoxicity warning and its use in patients with grade 3 or higher liver enzyme elevations should be carefully considered, limiting combination with regimens where these toxicities are expected. Sorafenib has comparatively modest efficacy, and in combination with mTOR inhibition was considered more toxic than beneficial.[56] Cabozantinib is FDA approved for medullary thyroid cancer and renal cell carcinoma with established pediatric dosing guidelines.[53] While promising single-agent activity in OS led COG to prioritize Cabozantinib for study in combination with MAP for front line OS treatment, challenges including relatively long half-life of 99 hours requiring many weeks of holding the agent in perioperative settings, and the need for a careful safety plan to determine feasibility of combining cabozantinib with cytotoxic chemotherapy.

Immunotherapy Approaches:

Several immunotherapeutic strategies are being investigated in OS both preclinically and in clinical trials. These approaches include the use of immune checkpoint inhibition, monoclonal antibodies, antibody-drug conjugates, and adoptive cellular therapy targeting the surface proteins PD-1/PD-L1, GD2, HER2, B7-H3, sFRP2, EGFR, EphA2 and others. Chimeric antigen receptor (CAR) T cells are T cells that are engineered to become tumor specific by expressing an extracellular domain derived from a monoclonal antibody specific to a tumor surface marker. The surface markers GD2, EGFR and HER2 have been the subject of completed and ongoing CAR T cell clinical trials (NCT02107963, NCT01953900, NCT03635632, NCT03618381, NCT00902044). [28]

B7-H3 is a marker expressed on the surface of a variety of tumors including OS that has shown promising results in a variety of immune therapy approaches. A phase I study of a monoclonal antibody targeting B7-H3 ( NCT02982941) has recently completed accrual. This antibody was used to create a B7-H3 targeting CAR T cell, which in murine models, leads to eradication of OS xenograft tumors and prolonged survival in a highly lethal metastatic model.[30] These results are exciting for potential future use in the adjuvant setting for patients with disease at high risk for relapse. A phase I clinical trial using this B7-H3 CAR T construct is ongoing ( NCT04483778). A B7-H3 antibody drug conjugate, m276-PBD, has also shown impressive results in the pre-clinical setting with 80% of OS xenograft mice demonstrating an objective response.[29] These data are strongly suggestive that B7-H3 is a relevant target in osteosarcoma and development of agents targeting it both alone and in combination with other effective agents should be a priority of COG Bone Tumor Committee.

Inhibitors of CD47- SIRPα

CD47 is a transmembrane protein that binds to signal receptor protein-α (SIRPα) on macrophages and dendritic cells. This interaction blocks phagocytosis through a “don’t eat me” signal and induces immunotolerance of both normal and malignant cells harboring the protein.[57] Anti-CD47 agents may restore innate immune destruction of tumor cells as well as directly induce apoptosis.[58] The level of expression of CD47 on cancer cells is associated with decreased survival, making it a prospective target for immunotherapy in the treatment of a variety of cancers.[59] CD47 has increased expression on OS cells compared to osteoblastic cell lines and neighboring normal bone tissue. In the pre-clinical setting, anti-CD47 antibodies halt tumor progression and metastasis in OS xenograft models.[60]

A number of CD47 agonists are currently available in phase I clinical trials including Hu5F9, CC-90002, TTI-621, ALX 148, SRF231, AO-176 and IBI188. [61] Hu5F9-G4 is a humanized monoclonal anti-CD47 antibody that is currently under investigation in phase I/II clinical trials in adults with advanced hematologic and solid tumors ( NCT02953509, NCT02953782, NCT03248479, NCT03558139). Another CD47 targeting agent, ALX 148, has an active phase I/II clinical trial for treatment of advanced solid tumors and lymphoma ( NCT03013218). Agonists such as Hu5F9 and ALX 148 may require synergism with other antibodies capable of engaging Fc receptors to induce antibody-dependent cell cytotoxicity (ADCC) and phagocytose tumor cells.[62] Some examples of antibodies being studied in combination with Hu5F9 are rituximab for B-cell NHL ( NCT02953509), cetuximab for solid tumors and adrenocortical carcinoma ( NCT02953782), and Avelumab for ovarian cancer ( NCT03558139). ALX 148 is being studied as monotherapy and in combination with rituximab, pembrolizumab and traztuzumab. In the treatment of OS, CD47 agonists may be more effective if used in combination with an antibody targeting a known marker on OS cells such as GD2 or B7H3. Preclinical data have shown that an anti-CD47/GD2 antibody combination results in complete reduction of pulmonary metastases in murine models. [63] CD47 agonists have yet to be studied in clinical trials in pediatric patients with OS, however, there are plans to move forward a phase I study of Hu5F9 and dinutuximab through the Cancer Immunotherapy Trials Network for pediatric patients with OS and neuroblastoma. Thus, both further preclinical optimization of combinations of CD47 agonists and emerging safety data will inform further translational directions. A promising direction may be the application of these agents to minimal residual disease states, such as after achieving surgical and radiographic remissions following relapse. This trial design has been successfully employed by the COG.

CB-839 (Telaglenastat):

The formation of metastases by OS cells is dependent on aberrant and highly efficient metabolic pathways for energy production.[64] Due to its potential to disrupt these metabolic pathways, CB-839 has been considered in the treatment of metastatic OS. CB-839 inhibits the mitochondrial enzyme glutaminase (GLS1) which disrupts the conversion of glutamine to glutamate and impedes cellular metabolism including the tricarboxylic acid cycle, oxidative respiration and amino acid synthesis.[65] Synergism with other agents, particularly drugs that result in an increased cellular reliance on glutamine metabolism may be necessary for optimal efficacy. [65]

In preclinical testing, treatment of metastatic and primary murine and canine OS cell lines with the combination of CB-839 and metformin resulted in growth inhibition and decreased mitochondrial respiration. In murine in vivo models of OS, treatment of established lung metastases with the CB-839 and metformin combination successfully resulted in decreased metastatic burden.[66] Unfortunately, high doses of metformin were required to see benefit above that observed with single agent telaglenastat, potentially impeding the potential of moving this combination forward in humans. Investigators are now testing other combinations with CB-839, such as cisplatin and gemcitabine, which may be more potent and/or offer an alternative to existing regimens with high toxicity.

CB-839 is currently in phase I/II clinical trials in adults to treat a variety of hematologic cancers and solid tumors in combination with a variety of agents. Combinations include CB-839 with pembrolizumab ( NCT04265534), sepanisertib ( NCT04250545), palbociclib ( NCT03965845), talazoparib ( NCT03875313), osimertinib ( NCT03831932), cabozantinib ( NCT03798678, NCT03428217]), carfilzomib ( NCT03798678), and temozolomide ( NCT03528642). Additional preclinical testing of CB-839 in combination with other agents used in OS recurrence in collaboration with our basic science colleagues, as mentioned above, will guide further clinical development of CB-839 in COG clinical trials.

Epigenetic Modifiers:

Azacytidine, decitabine, thio-deoxy-cytidine (TdCyd) and aza-thio-deoxy-cytidine (Aza-TdCyd) are analogs of the nucleoside cytidine that inhibit DNA methyltransferases, leading to hypomethylation.[67] In testing performed by the Pediatric Preclinical Testing Consortium, Aza-TdCyd was found to have activity in OS cell lines and in five of six OS PDX models. The median survival of mice increased nearly 4-fold when treated with Aza-TdCyd compared with control.[68] Aza-TdCyd is currently being studied in an adult phase 1 trial ( NCT03366116) for advanced solid tumors. Clinical trials using decitabine ( NCT03445858) and azacytidine ( NCT03628209) in combination with immune checkpoint inhibitors are ongoing. Histone deacetylase (HDAC) inhibitors are another class of epigenetic modifiers of interest. These agents prevent removal of acetyl groups from lysine residues, promoting chromatin remodeling and structural changes to proteins involved in transcription factor complexes.[69] Panobinostat is an HDAC inhibitor with preclinical activity in OS [70–72] and in soft tissue sarcomas in a phase 1 study.[73] This agent has not yet been tested clinically in OS. Increasingly, metastatic progression is being recognized as having an epigenetic program. [15, 16] The combination of the inability to elucidate specific and recurrent molecular changes along with the pleotropic mechanism of HDAC inhibitors means that there remains work to do in the preclinical space despite showing signals of activity. [74] HDAC inhibitors have been tested in solid tumors without tumor selection and have not shown appreciable activity and thus the path forward for translation of these inhibitors requires both more preclinical data, combinations of therapy, and different study endpoints. Specifically, the group recommended endpoints of metastasis development rather than radiographic response to best evaluate their effect on osteosarcoma progression.

Conclusions and future directions:

Despite a rigorous discussion and vetting process involving key experts in the field, the task force concluded that there were no novel agents meeting all criteria for movement to front line evaluation with the exception of multitargeted TKI’s, which have already been incorporated into a phase III study. All other top agents discussed by the task force continue to require additional preclinical and clinical evidence in OS to support further investigation. This calls for a need to develop more OS specific drugs in addition to the historical path of repurposing drugs for other indications.

Through the evaluation process, our group further revised the criteria for moving agents forward in the COG. Some of the concepts utilized were a continuation of the work of the consensus “Perspective” on OS drug development working group that met in Bethesda, MD in 2014.[17] Like the “Perspective”, our task force prioritized preclinical evidence for agents targeting metastatic progression rather than relying on evidence of regression of measurable lesions. Agents targeting pathways predominant in metastatic OS models therefore received greater consideration, such as metabolic agents (CB-839) and immune based therapies (B7-H3 CAR T cells). Incorporating rational biomarkers and/or quantitation of circulating tumor cells (CTC) for measurement of response to therapy has been suggested to aid in evaluating these agents clinically. Ultimately, more effective drug development or repurposing will require identifying biologic subsets of disease, determining mechanisms of metastasis and identifying processes for immune evasion.

Given the urgent need for novel therapies to treat metastatic and relapsed OS, the task force also emphasized the use of more Phase II studies in the relapsed setting, requiring a smaller number of participants, to assess EFS informed by defined historical benchmarks.[48] Ideally, utilizing these evolving strategies will allow for multiple simultaneous studies, investigating agents that may have optimal efficacy in different settings, such as treatment of newly diagnosed metastatic disease, prevention of metastases in patients with MRD, or prolongation of survival in patients with measurable relapsed disease. Furthermore, as new immune-modulating agents targeting various components within OS TME emerge, there is a growing need to develop a standardized systematic approach in evaluating preclinical efficacy prior to COG trial considerations. Moving forward, using these refined benchmarks, we will continue to monitor the progress of the top agents discussed here, and remain open to new therapies and strategies to treat this complex disease.

Abbreviations:

OS

Osteosarcoma

PDX

patient derived xenografts

COG

Children’s Oncology Group

EFS

event free survival

TKI

tyrosine kinase inhibitor

MRD

minimal residual disease

CTEP

Cancer Therapy Evaluation Program

TME

tumor microenvironment

MAP

methotrexate, doxorubicin, cisplatin

CAR

chimeric antigen receptor

HDAC

histone deacetylase