(P 55) RELATIVE WASH-IN RATE ON DYNAMIC CONTRAST-ENHANCED MRI AS A NEW PROGNOSTIC FACTOR IN PATIENTS WITH OSTEOSARCOMA; A POSSIBLE TOOL FOR PERSONALIZED SARCOMA-CARE?

Objective: Poor histological response to neoadjuvant chemotherapy is associated with impaired prognosis in osteosarcoma. Although considered the gold standard for evaluating response to chemotherapy, histological response assessment has limitations, including sampling bias due to the examination of only one slab of tumor tissue, high inter-observer variability, and the use of a binary cut-off (≥10% viable tumor cells indicate poor response). In our previous work (to be published) we found that relative wash-in rate (rWIR) as determined by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) correlates with histological response. In this study, we aim to explore the potential of this novel radiological biomarker as prognostic factor for clinical outcome such as event-free survival (EFS).

Methods: Patients with high-grade skeletal osteosarcoma, treated with curative intent, between 2005 and 2022 at University Hospital Ghent (UHG) and Leiden University Medical Center (LUMC) were retrospectively included. All patients underwent DCE-MRI before neoadjuvant treatment and during or after receiving the second round of neoadjuvant chemotherapy, prior to surgical resection. For image analysis, the largest tumor slab was manually segmented in the DCE-MR images. The rWIR was calculated by dividing the wash-in rate on baseline imaging, defined as the maximum percentual intensity rise per second, by the wash-in rate post-chemotherapy. A rWIR < 2.3 was considered a radiological poor and rWIR ≥ 2.3 a good response based on our previous report. A previously published population study (Evenhuis et al, Cancers 2021) in 402 patients was used to identify relevant prognostic factors for EFS in Cox-proportional hazards regression modelling. EFS was defined as the time from resection to first event which consisted of local recurrence (LR), progression of metastasis, new metastasis, or death. First, multivariate Cox regression was used to estimate the effect of known clinical prognostic factors and histological response on EFS. Subsequently, this was repeated using rWIR of DCE-MRI as prognostic factor instead of histological response. EFS was estimated with Kaplan-Meier's methodology.

Results: In total, 86 patients were included with a median age of 18 years (IQR 14-31) and a median follow-up time of 11.9 years (95%CI 10.9-12.9) in which 37 events occurred. In 46 (53%) patients histological response was classified as poor while in 45 (52%) radiological response was classified as poor. In 19 patients (22%) histological and radiological classifications were discordant. Evenhuis et al. identified age group, tumor size, and metastasis at presentation as significant covariates of response for EFS estimation. When combined with histological response in Cox-regression analysis, only the presence of distant metastasis at presentation [Hazard ratio (HR) 3.58 (95%CI 1.63-7.89)], and not histological response, [HR 1.69 (95%CI 0.84-3.40)], was significantly associated with poor EFS. When using rWIR instead of histological response as risk factor in the Cox model, the presence of distant metastasis at presentation, [HR 3.63 (95%CI 1.65-8.01)] and rWIR < 2.3, [HR 2.64 (95%CI 1.29-5.43)] were associated with poor EFS. Patients with rWIR < 2.3, had a 2-year EFS of 53% (95%CI 38-67%) vs. 83% (95%CI 71-94%) in patients with good radiological response (rWIR≥2.3).

Conclusion: The new biomarker, rWIR < 2.3, is associated with poor EFS in a multivariable analysis where the most important covariates were incorporated in a Cox model. In this cohort rWIR was associated with EFS when corrected for covariates while histological response was not. Advantages of rWIR as a response biomarker relative to histology are that it can be estimated pre-operatively instead of postoperatively and the whole tumor volume can be assessed. These findings suggest rWIR could potentially be a tool for individualized patient-care.