ESTRO 2020 Abstract book

S76 ESTRO 2020

simple dose reduction to the hippocampus in the context of PCI will be beneficial. This underlines the need for randomized studies and cautions against the implementation of techniques that “should” work without evidence that they actually work. SP-0145 For the motion -consider brain metastases management as a whole A. Tallet 1 , V. Favrel 1 , L. Varela-Cagetti 1 1 paoli-Calmettes, Radiation Oncology, Marseille, France Abstract text Due to its neurotoxicity profile, WBRT has given way this last decade to SRS and new systemic molecules allowing for WBRT postponing or avoidance. However, in patients with symptomatic multiple brain metastases (BMs) not suitable for SRS (due to a large total volume), and unable to wait for systemic treatment efficiency, WBRT is still relevant. WBRT-related neurotoxicity has been suggested to particularly affect memory and at a lesser extent executive functioning. Preservation of the sub-granular zone of the hippocampal dentate gyrus (hosting neural stem cells - NSCs) has been thought to be able to prevent memory decline, thanks to the ability of NSCs to give rise to new neurons supposed to replace damaged ones. Several studies investigated both the risk of BM presence and the rate of BM recurrence in the HA area. The percentage of BMs within a 5-mm envelope around the hippocampi was found to range from 0.8% to 4.4%, corresponding to a mean percentage of patients with BM within the HA area of less than 10%. Studies investigating the rate of BM recurrence in the HA area found it to range from 0.5% to 7.6%. Three randomized phase II and III studies and one meta- analysis presented at the 61st ASTRO meeting showed significant cognitive preservation from the HA-WBRT. Yang et al. reported on the first 46 BM patients randomized to HA-WBRT or WBRT in a single blind randomized controlled trial. Neurocognitive function (NCF) was assessed through a standardized neurocognitive test battery, with both patients and examiners blinded to the allocated treatment. The primary endpoint was the difference in changes of HVLT-R delayed recall at 4 months between HA- WBRT and WBRT. First differences were observed at 6 months, with a mean change of HVLT-R total recall of +2 ± 1.26 vs. -1 ± 0.97, ( p =0.08), and of HVLT-R recognition of 0.6 ± 0.54 vs. -1.4 ± 0.62, ( p =0.03) in patients receiving HA-WBRT and WBRT, respectively. The phase III PREMER study compared HA-PCI to PCI in 118 SCLC patients, with NCF as primary end-point, assessed by the Free and Cued Selective Reminding Test. HA-PCI significantly prevented 3- and 6-months memory disorders (rate of free delayed recall drop: 21.7 vs 5.1%; p 0.01; 32.6 vs 7.3%; p 0.008; 18.5 vs 3.8%; p 0.09 at 3, 6, and 12 months, respectively, in PCI vs PCI-HA arm. The NRG Oncology CC001 phase III trial randomized 518 BM patients to WBRT + Memantine (M) or HA-WBRT+M. Both multidimensional cognitive tests and patient-reported symptoms (PRS) were recorded. HA was associated with a 26%-reduction of cognitive decline (adjusted HR=0.74, 95% CI: 0.58-0.95, p=0.02). Time to NCF failure was significantly longer in the HA-WBRT arm. PRS also favored HA. Toxicity, overall survival, and intracranial progression were similar between the treatment arms. Lehrer et al. conducted a meta-analysis of five randomized prospective trials to quantify and compare the cognitive benefits derived from SRS, WBRT + M, and HA-WBRT, relative to WBRT. Three to four-months cognitive failure rates were 53.2% (95%CI: 48.5%-57.8%), 43.7% (95%CI: 37.7%-49.8%), 33.3% (95%CI: 28.6%-52%), 20.9% (95%CI: 15.7%-26.7%), in patients treated with WBRT, WBRT + M, HA-WBRT, and SRS, respectively. SRS, WBRT + M, and HA-WBRT had all statistically lower rates of cognitive failure than WBRT. Despite a large trend toward the use of SRS in patients with BMs, WBRT still retains value in patients unfitted for SRS

Debate: This house believes that hippocampus avoiding whole brain irradiation is the current standard of care

SP-0143 For the motion: Hippocampal Avoidance: Proving Causation from Bench to Bedside to Beam G. Vinai 1 1 Northwestern Medicine Proton Center and Cancer Center Warrenville, Radiation Oncology, Chicago, USA Abstract text This presentation will review 1) preclinical and clinical evidence supporting the cognition-specific radiosensitivity of hippocampal neurogenesis and 2) practice-changing evidence from NRG Oncology CC001 establishing hippocampal avoidance during whole-brain radiotherapy for brain metastases as a standard of care. SP-0144 Against the motion: Hippocampus avoidance in PCI: Blinded by the light D. De Ruysscher 1 1 MAASTRO Clinic, Radiation Oncology, Maastricht, The Netherlands Abstract text PCI has become the standard of care for all stages of small cell lung cancer (SCLC) for it improves the overall survival (OS). In NSCLC, PCI also reduces the incidence of BM significantly (HR 0.38), but here no effect on the OS (Witlox, WCLC 2019). PCI leads to a decline in neuro- cognitive functioning, without affecting the HRQoL (Witlox, Radiother Oncol 2019). It should be noted that the reporting of side effects in PCI is subject to a major bias between patients and doctors/ nurses (De Ruysscher, J Clin Oncol 2018). Because the neuro-cognitive decline is the most prominent in tests that relate to the functioning of the hippocampus such as the HVLT-R and because in retrospective series decreasing the radiation dose to the hippocampus was associated with less neuro-cognitive decline, hippocampus-sparing PCI (HA-PCI) gained popularity. In some centers, HA-PCI was introduced without randomized evidence and without data about the safety. Lowering the dose to even a small part of the brain may indeed result in more BM. A randomized Dutch- Flemish study (Belderbos, ESTRO 2019) in which patients with all stages of SCLC were randomized after platinum- etoposide chemotherapy or concurrent chemo-radiation between regular PCI (25 Gy/ 10 fractions). Patients underwent a contrast-enhanced brain MRI at the time of randomization and in the follow-up, including at 4 months post-(HA)-PCI. Patients with BM were not eligible for the primary endpoint in order to exclude interference between BM and neuro-cognition. HA-PCI did not show any effect on the primary endpoint, the HVLT-R 4 months after therapy. Approximately 28 % of patients in both arms of the trial had a decline of 5 points or more. The results were similar 8 months after therapy. The OS, PFS and incidence of BM were similar in both groups, supporting the safety of HA-PCI. Even with (HA)-PCI, about 18 % of patients did have BM after 2 years, even though they were thoroughly staged with an MRI of the brain at the time of PCI. In the NRG CC001 phase III trial (Gondi V, ASTRO 2019), 518 patients with BM were randomized between WBRT and memantine, which is hardly used in Europe, vs. HA-WBRT and memantine. The primary endpoint was any neurocognitive test decline at any moment in time. In NRG CC001, the main difference at 4 months was first seen in the executive functioning (p= 0.01), and encoding (p=0.049) and consolidation (p=0.02) at six months. However, executive functioning does not have a high specificity for hippocampal functioning. In NRG CC001 it is not clear if patients that had intracranial progression were still tested. The systemic treatment may also differ from our study. It is therefore far from established that just a