Considerations for Clinical Trial Design in Relapsed and Refractory Osteosarcoma: An FDA Symposium

Abstract

Osteosarcoma is the most common primary malignant bone tumor in children and adolescents, with a peak incidence coinciding with the pubertal growth spurt. Metastatic disease occurs in approximately 10 to 20% of newly-diagnosed patients, most commonly in the lung. While 5-year overall survival (OS) for patients with localized disease at diagnosis is approximately 70% after standard upfront treatment of multi-agent chemotherapy and surgical resection, patients with metastatic or recurrent disease have a 5-year OS of approximately 30% in the pediatric age group. Currently, there is no standard treatment or FDA-approved therapy for relapsed and refractory osteosarcoma. Progress in identifying promising new treatments has been limited by its complex disease biology and rarity, as well as unique challenges of clinical trial design including unreliability of objective response rate (ORR) as a predictor of drug activity. Given these challenges and the unmet need for new therapies, the FDA Oncology Center of Excellence hosted an educational symposium, “Current Challenges in Clinical Trial Design for Relapsed and Refractory Osteosarcoma,” in May 2023. During this mini-symposium, patient advocates, regulators from the FDA, and academic thought leaders in the field of pediatric sarcoma discussed challenges in clinical trial design and implementation. Achieving progress in relapsed and refractory osteosarcoma will require intentional collaboration among all stakeholders to identify trial designs that are acceptable to patients and will provide sufficient evidence of efficacy and safety to support a marketing application. Herein, we summarize the key points and future directions discussed at this meeting.

Introduction/Background

Osteosarcoma is the most common malignant bone tumor in children and adolescents, with an incidence of approximately 800–900 new cases per year in the United States (1). The peak age of incidence is between 10 to14 years of age and 15 to 19 years of age in females and males, respectively, which coincides with the pubertal growth spurt. Standard upfront treatment consists of multi-agent chemotherapy with high-dose methotrexate, doxorubicin and cisplatin (MAP), and maximum feasible surgical resection of all lesions for patients with both localized and metastatic disease (1, 2). Metastatic disease occurs in approximately 10 to 20% of newly diagnosed patients, most commonly in the lung. While 5-year overall survival (OS) for patients with localized disease at diagnosis is approximately 70%, patients with metastatic or recurrent disease have a dismal prognosis, with a 5-year OS of approximately 30% in the pediatric age group. Death most commonly occurs due to lung metastases (2).

The FDA approved methotrexate for the treatment of osteosarcoma in 1953, and doxorubicin for metastatic bone sarcoma in 1974; no therapies have since been approved for this indication. Currently, there is no standard treatment for relapsed and refractory osteosarcoma and no FDA-approved therapies for recurrent disease. Given that relapsed osteosarcoma most commonly occurs in the lung, complete surgical resection of all pulmonary lesions, if feasible, is the cornerstone of treatment for recurrent osteosarcoma and offers the best chance of long-term survival (3, 4). Chemotherapy regimens used in this setting include ifosfamide and etoposide, high-dose ifosfamide alone, cyclophosphamide and topotecan, and gemcitabine and docetaxel (3). Studies of multitargeted kinase inhibitors such as regorafenib, cabozantinib and lenvatinib have more recently demonstrated a signal of activity in relapsed or refractory osteosarcoma as measured by 4-month progression-free survival (PFS) rate compared to either placebo or historical controls (5, 6, 7). However, objective response rates (ORRs) are low and the majority of patients with stabilization of disease develop disease progression after a few months (8).

Osteosarcoma biology presents unique challenges for identification of effective new therapies, as well as for clinical trial design and implementation. The genomic landscape of osteosarcoma is heterogeneous and is characterized by relatively high rates of structural variations (SVs), which are generally difficult to target therapeutically, and lower rates of single nucleotide variations (SNVs) (8, 9). Osteosarcoma tumors also produce malignant osteoid, a bony matrix that does not resolve on imaging with systemic therapies; therefore, ORR per Response Evaluation Criteria in Solid Tumors (RECIST) may not accurately reflect drug activity (10). Because surgical resection is recommended for recurrent disease when feasible, many patients may not have measurable disease at trial entry, further limiting the utility of tumor-based endpoints such as PFS and ORR. A randomized controlled trial with OS as the primary endpoint is the gold standard for clinical trial design, but roadblocks to implementing this type of trial include rarity of the disease and patient concerns regarding randomization.

Given these challenges and the persistent unmet need for new therapies for patients with recurrent and metastatic osteosarcoma, the FDA Oncology Center of Excellence hosted an educational symposium, “Current Challenges in Clinical Trial Design for Relapsed and Refractory Osteosarcoma,” in May 2023. During this mini-symposium, patient advocates, regulators from the FDA, and academic thought leaders in the field of pediatric sarcoma gathered to discuss challenges in clinical trial design and implementation, appropriate endpoints, and the role of randomization in relapsed and refractory osteosarcoma. Herein, we summarize the key points and future directions discussed at this meeting.

Identifying promising drugs in the relapsed/refractory setting

Expedient identification of promising drugs for the treatment of relapsed or refractory osteosarcoma emerged as both a major challenge to drug development and a high priority during the mini-symposium. Despite consistent clinical and pre-clinical research efforts, very few drugs have shown efficacy signals which warrant moving forward into larger trials designed to support a marketing application. The unique biologic features of osteosarcoma, including heterogeneity within and between tumors, complexity of the genome, lack of common druggable targets, and immune-suppressive tumor microenvironment, present difficulties for drug development. In addition, drugs that have shown efficacy in animal models have largely failed to show similar efficacy in the clinic (8).

Given that osteosarcoma is a rare tumor, many molecularly targeted drugs under investigation in this setting were originally developed to treat other tumor types. Because of the molecular heterogeneity of osteosarcoma, efficacy signals of an active drug may be difficult to identify in an unselected population. To address this issue, trial designs with different arms for patients with specific molecular subtypes, such as umbrella trials, may be considered.

During the meeting, FDA stated that it may be helpful to consider the recommendations and information in FDA’s Guidance for Industry, entitled, “Expedited Programs for Serious Conditions” at each step of the development pathway (11). Expedited programs are designed to speed the availability of new drugs to patients with serious conditions, particularly those with an unmet medical need. Expedited programs provide opportunities for frequent interactions with the FDA review teams and expedited review at the time of NDA or BLA submission. Fast Track Designation is available for products in early stages of development which demonstrate the potential to address an unmet medical need. While other expedited programs require clinical data, Fast Track Designation may be granted based on activity in a relevant nonclinical model.

Endpoints

Tumor-based endpoints

Efficacy for anti-cancer agents is generally demonstrated by objective tumor measures or survival. ORR is an endpoint unique to oncology that can be assessed in a single-arm trial and is also a direct measure of the effect of an intervention because tumors generally do not regress spontaneously. However, due to the bony matrix deposition by osteosarcoma tumors, response to therapy is not always evident on imaging. Thus, ORR may not reliably reflect the activity of a given drug. However, as progressive disease by RECIST is associated with inferior outcomes, PFS or event-free survival (EFS) may be more appropriate endpoints for earlier, signal-finding studies (10). While acknowledging that a randomized controlled trial with OS as the primary endpoint is the most appropriate design for a registrational trial, academic thought leaders have identified an EFS benchmark of 4 months in patients with measurable disease not amenable to resection as potentially useful in earlier studies to identify drugs which merit further evaluation. For the 25% of patients with recurrent osteosarcoma isolated to the lung parenchyma that undergo surgery and have no radiologic evidence of disease post-operatively, a benchmark EFS of 12 months has been proposed as a reasonable measure to identify promising therapies for clinical trials investigating adjuvant treatment for patients with completely resected disease at trial entry (12).

Using these historical benchmarks, multi-targeted kinase inhibitors have demonstrated signals of activity in recurrent osteosarcoma and selected multi-TKIs have been moved into larger randomized trials in combination with chemotherapy. The multi-TKI regorafenib is included in the NCCN guidelines as second-line therapy for treatment of recurrent or refractory osteosarcoma (Category 1) based on an improvement in PFS with regorafenib compared to placebo in a randomized trial (5, 6). Cabozantinib has also shown activity in the treatment of patients with relapsed or refractory osteosarcoma in two non-randomized, single-arm studies enrolling pediatric patients, the CABONE study (7) and the Children’s Oncology Group study ADVL1622 (13). The results of these studies led to initiation of the ongoing Children’s Oncology Group study AOST2032, a randomized study evaluating cabozantinib in combination with MAP chemotherapy in patients with newly-diagnosed osteosarcoma.

As described in FDA guidance, single-arm trials do not adequately characterize time-to-event endpoints, and therefore benchmark EFS rate would not be an appropriate primary efficacy endpoint for a trial designed to support a marketing application; however, benchmark EFS of 4 months for patients with measurable disease and 12 months for patients with completely resected disease appropriate for earlier, can be useful in signal-finding trials in recurrent and refractory osteosarcoma (14). Given the rarity of the disease and need for efficient identification of effective therapies, use of 4- and 12-month benchmark EFS rates in single-arm trials offer an additional opportunity to identify active agents in an efficient manner. Promising results from early studies can be discussed in a product-specific meeting with FDA to determine whether they can provide support for a designation under one of FDA’s expedited programs for serious conditions (11, 15).

Use of clinical outcome assessment endpoints

Assessments of patient-reported outcomes can provide supportive evidence of the clinical impact of treatment and a better understanding of the tolerability of a study therapy. Participants discussed integration of clinical outcome assessment endpoints that may be particularly relevant to patients with osteosarcoma. Patient advocates participating in the meeting reported that complications from systemic treatment and surgery such as hospitalizations, infections, and thromboembolic events were among the most significant quality of life issues encountered during treatment of recurrent and refractory osteosarcoma.

Surgical resection of primary and metastatic lesions is a key component of treatment for osteosarcoma and is currently considered essential to achieving long-term disease control. Open thoracotomy has traditionally been the standard of care for resection of pulmonary lesions as this approach allows for palpation of nodules not detected on imaging. However, open thoracotomy is associated with increased post-operative pain and other complications compared to minimally invasive thoracoscopy (16). Current trials in patients with osteosarcoma are evaluating minimally invasive surgical techniques and surgical complications related to new systemic therapies. The ongoing COG study AOST2031 (NCT05235165) is a randomized trial evaluating open surgery (thoracotomy or sternotomy) versus minimally invasive surgery by thoracoscopy in patients with oligometastatic pulmonary disease at initial diagnosis or at first recurrence. Patient reported outcomes, specifically functional impairment of the upper extremities, pain intensity, and health-related quality of life (HRQoL), and surgical complication rates are exploratory objectives in this trial (17).

Attention to patient reported outcomes related to surgical complications in future osteosarcoma trials is vital, particularly given increasing use of multi-TKIs targeting the VEGF pathway and the toxicity profile of these agents, including wound healing complications. The ongoing COG trial AOST2032 (NCT05691478), which will evaluate cabozantinib in combination with standard MAP and surgical resection for patients with newly diagnosed osteosarcoma, is designed to assess surgical complications and delays in resumption of chemotherapy. Continued engagement with patient advocacy groups to identify relevant clinical outcome measures for patients with relapsed and metastatic osteosarcoma will facilitate development of clinical trial designs that are more patient-centered and may have utility in evaluating clinical benefit of new drugs and biologics in this setting.

Randomization

Given the challenges in interpretation of ORR in osteosarcoma, trial designs with time-to-event endpoints such as EFS and OS are needed to evaluate the clinical benefit of novel agents. OS has been considered the gold standard for trials in oncology as it is a direct measure of clinical benefit, does not require interpretation of imaging, and incorporates information about drug safety. However, for purposes of regulatory decision-making, time-to-event endpoints are not interpretable in single-arm trials, as single-arm trials do not control for potential confounders and therefore the observed outcomes cannot be definitively attributed to the effect of the drug (14). Hence, with rare exception, randomized controlled trials are required to determine the effect of an intervention on survival. Nevertheless, challenges associated with randomization, particularly in rare diseases, may necessitate flexible approaches. Although randomized, placebo-controlled trials have been conducted in the setting of relapsed osteosarcoma (e.g., regorafenib), the rarity of this disease and the concerns with use of placebo control arms present challenges for accrual. During the mini-symposium, patients and caregivers expressed that given the poor prognosis for relapsed or refractory osteosarcoma and the financial and logistical burdens associated with clinical trial participation, they would not be likely to consider participating in a randomized trial if there was a possibility of their child receiving placebo instead of the investigational product. Instead, if the treatment under investigation is already commercially available for other cancer indications, patients and caregivers may opt for off-label use of the product outside of a clinical trial. Earlier identification of promising investigational agents for the treatment of osteosarcoma and initiation of osteosarcoma-specific trials earlier in the drug development process may help to reduce off-label use of oncology products by providing patients the opportunity to receive a given drug while allowing for collection of valuable information within a clinical trial. In addition, randomized trial designs with active comparator arms, such as treatment of physician’s choice and crossover designs, were discussed as alternatives to a placebo-controlled design which patients and caregivers may find more acceptable. Dosage optimization trials, which randomize patients to different dosages of the same investigational agent, represent another opportunity to obtain randomized data while collecting important information about the benefit-risk profile of each dosage.

A recent example of a randomized trial design for patients with refractory or relapsed osteosarcoma highlights the complexity of designing and implementing randomized controlled trials for patients with relapsed and refractory osteosarcoma, and the potential introduction of bias when treatment assignments are not blinded. The OLIE trial (NCT04154189) is an international, randomized, open-label study comparing lenvatinib in combination with ifosfamide and etoposide (IE) versus ifosfamide and etoposide in children and young adults with relapsed or refractory osteosarcoma. Surgery was permitted at Week 18, after 5 cycles of treatment. Median OS was 11.9 months in the lenvatinib + IE arm and 17.4 months in the IE arm, although this study was not powered to detect differences in OS. While there was no difference in the number of patients considered to have resectable tumors between arms, more patients in the IE arm (n=11) underwent surgery than in the lenvatinib + IE arm (n=5) (18). Given the importance of surgical resection for survival in osteosarcoma, it is possible that the longer survival in the IE arm was at least partly due to the higher rate of surgical resection. While the reason for the higher rate of surgical resection in the IE arm is not entirely clear, it is possible that knowledge of the treatment arm to which patients were randomized influenced investigators’ decisions regarding whether to recommend surgery at Week 18. In the case of patients receiving lenvatinib, investigators may have been cautious to recommend surgery given the wound-healing complications seen with VEGF-targeting agents; for patients receiving IE without lenvatinib, investigators may have been more likely to recommend surgery due to perceived higher risk of relapse in the chemo-only arm.

Conclusions

The lack of improvement in survival rates for patients with relapsed or refractory osteosarcoma is due to several factors including the rarity, complex biology, and tumor heterogeneity of the disease, and absence of a reliable efficacy endpoint for use in a single-arm trial. Collaboration between the academic community, the pharmaceutical industry, patient advocates and regulatory agencies is essential for achieving progress in this challenging disease space. Thoughtful design of signal-finding clinical studies including the identification of appropriate endpoints for each stage of development is important to ensure efficient use of resources and promote timely recognition of products with the greatest potential. Although treating physicians may use results of a single arm trials to guide individual patient care decisions, benchmark EFS rate is not an appropriate endpoint for trials intended to support a marketing application for new products in osteosarcoma; however, this endpoint can be useful in earlier stages of development to identify drugs to be taken forward to larger studies and may support a designation under one of FDA’s expedited programs for serious conditions.

Future development of novel endpoints and biomarkers of response will be important for enabling earlier assessment of activity and identification of intermediate endpoints for use in trials intended to support a marketing application for accelerated approval. As randomized controlled trials are necessary for studies with EFS or OS as a primary efficacy endpoint, discussion among stakeholders regarding trial designs that are acceptable to patients should be prioritized prior to beginning a study. Such input from patient advocacy groups is crucial for identifying patient reported outcome measures which can support the primary assessment of efficacy of novel therapies. International collaboration will likely be important to support feasibility and efficiency of registrational trials in osteosarcoma given the rarity of the disease. Alignment on trial design and endpoints among international regulatory agencies can also prevent or reduce duplicative efforts. Early and frequent interactions among all interested groups throughout drug development and during the clinical trial design process is crucial to delivering promising drugs to patients as expeditiously as possible.