Understanding biological activity, tumor response and pseudoprogression in a phase-IIb study of MDNA55 in adults with recurrent or progressive glioblastoma (GB)

5581 Background MDNA55, a novel Interleukin-4 empowered cytokine fused to Pseudomonas exotoxin, binds to IL-4 receptor (IL-4R) overexpressed by GB and immunosuppressive cells of the tumor microenvironment. Understanding biological effects of MDNA55 and defining their time course requires multi-modal imaging. Consecutive increases in tumor size defined by Response Assessment in Neuro-Oncology (RANO) can result in premature withdrawal of subjects prior to evaluation of clinical benefit. Standard tumor response tools need optimizing for localized immunotherapies to capture delayed treatment benefit. Methods MDNA55-05 is a 52 subject, open-label, non-randomized, study of intratumoral delivery of ≤ 240 μg MDNA55 via ≤ 4 catheters in a single treatment in GB at 1st or 2nd recurrence, tumors ≤ 4 cm. Multimodality MRI determines tissue response and disease status using RANO-based criteria, combining contrast-enhanced and perfusion imaging. Results 27 subjects, median age 51 (35 - 77) received 18 – 180 mg MDNA55. Prior treatments included surgery (N = 26), radiation (N = 25) and temozolomide (N = 25). Review of some early post-treatment MRIs show extensive changes to enhancement which could increase for up to 120 days then take another 6 months to resolve. Multi-modal MRI helps differentiate changes due to progression from local tissue reactions (pseudoprogression). Response patterns seen include early progression, early onset response and delayed onset response. Withdrawal of subjects without differentiating pseudoprogression from true progression can preclude the ability to fully assess therapeutic effect. A refined MDNA55 evaluation regimen consisting of multi-modal MRI and/or biopsies has been implemented to optimize the chance for increased therapeutic benefit. Conclusions These findings show biologic activity of MDNA55 in recurrent GB with appearances on imaging suggestive of disease control in some subjects. We illustrate how supportive diagnostic modalities are required for accurate response assessments and argue the importance of refining overall response tools according to treatment type.

Duke Scholars

Author John Howard Sampson Neurosurgery