2586 - Early Apparent Diffusion Coefficient (ADC) Changes during Concurrent Chemoradiation (CRT): An Imaging Biomarker for Recurrence Prediction in Glioblastoma

D. M. Palhares¹, L. S. P. Lawrence², H. Soliman¹, J. Stewart¹, S. D. Myrehaug¹, H. Chen¹, M. E. Ruschin¹, J. Detsky¹, D. Dinakaran¹, P. Maralani³, C. L. Tseng¹, A. Sahgal¹, and A. Lau⁴; ¹Department of Radiation Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada, ³Department of Medical Imaging, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada, ⁴Department of Physical Sciences, Sunnybrook Research Institute, University of Toronto, Toronto, ON, Canada

Purpose/Objective(s): Diffusion-weighted imaging (DWI) enables in vivo diffusivity measurement through the ADC. Aggressive hypercellular tumours exhibit restricted water diffusion and low-ADC values. We hypothesized that areas of glioblastoma that have a high likelihood of recurrence would exhibit smaller ADC change during concurrent CRT than those areas responsive to treatment. Our study aims to compare ADC changes within the areas of the gross tumour volume (GTV) that developed recurrence versus those that remained recurrence-free. Materials/

Methods: We reviewed a prospectively collected cohort of patients with glioblastoma imaged between Dec 2017 and Apr 2021 with DWI at planning (Fx0), fraction 10 (Fx10), fraction 20 (Fx20), and 1 month after a standard 6-week course of concurrent CRT (P1M). The GTV was contoured at all time points and included the surgical cavity and any residual enhancing tumour. The contrast-enhancing recurrence was contoured at the first magnetic resonance imaging (MR) timepoint showing progression per RANO 2.0 criteria. The intersection of the GTV and the recurrence volume was labelled resistant-GTV (R-GTV), while the GTV that did not intersect with the recurrence was labelled sensitive-GTV (S-GTV). Patients who did not experience failure within the GTV and had a minimum follow-up of 36 months were included in the analysis, and the entire GTV was labelled S-GTV. The interior of the surgical cavity was contoured at each time point and excluded from measurement of ADC values to avoid bias. Absolute ADC values and ADC changes (%) at each time point relative to Fx0 were compared between R-GTV and S-GTV using one-sided Wilcoxon rank-sum tests. Temporal changes were assessed using a linear mixed-effects model (fixed effects: region, time; random effects: subject).
Results: A total of 51 patients were included (median age 56y, range 20-69). The median absolute ADC values for R-GTV and S-GTV were 0.91 (IQR: 0.84-1.08) vs 0.94 (IQR: 0.87-1.10) at Fx0 (P=0.358), 1.02 (IQR: 0.88-1.18) vs 1.17 (IQR: 0.96-1.35) at Fx10 (P=0.011), 1.09 (IQR: 0.95-1.22) vs 1.16 (IQR: 1.00-1.44) at Fx20 (P=0.058), and 1.19 (IQR: 1.07-1.33) vs 1.37 (IQR: 1.17-1.52) at P1M (P=0.008), respectively. The median relative ADC change for R-GTV and S-GTV was 0.9% (IQR: -2.3-20.4) vs 15.3% (IQR: 7.5-28.5) at Fx10 (p=0.006), 11.2% (IQR: 2.7-27.4) vs 23.1% (IQR: 11.8-27.8) at Fx20 (p= 0.048), and 22.5% (IQR: 10.1-39.0) vs 33.3% (IQR: 21.1-50.2) at P1M (p= 0.034), respectively. The linear mixed-effects model revealed a significant difference in relative ADC changes between R-GTV and S-GTV (p<0.001).
Conclusion: Early ADC change during CRT is an imaging biomarker for treatment response and recurrence prediction in glioblastoma. Regions exhibiting smaller ADC changes during CRT indicate potential sites of recurrence, suggesting utility for MR-guided biologically adapted radiation clinical trials.