Juan Carlos Martinez Gutierrez, MD

Sarraj A, Mlynash M, Heit J, Pujara D, Lansberg M, Marks M, Albers GW. Clinical Outcomes and Identification of Patients With Persistent Penumbral Profiles Beyond 24 Hours From Last Known Well: Analysis From DEFUSE 3. Stroke. 2020;52:838–849.

Late window (6-24 hour) reperfusion with endovascular thrombectomy (EVT) has become standard of care for emergent large vessel occlusions (LVOs). Expanding indications for this highly effective intervention is the next frontier in acute stroke. While very late window (>24 hour) thrombectomy case reports and series are encouraging, there is still a paucity of high-quality evidence. A recent post hoc analysis of DEFUSE 3 by Saraj et al. aims to identify and characterize the putative candidates for very late window reperfusion through practical radiographic metrics.

In DEFUSE 3, 24 hour infarct volume between EVT and medical management (MM) patients was not significantly different despite the clear clinical benefit of reperfusion. The authors hypothesize that patients without complete reperfusion retain a persistent penumbra that can be lost over time, causing progression to larger final infarct volumes, thus explaining their poor outcomes. To identify these patients at continued risk beyond the 24 hour mark, the authors coined the novel Persistent Penumbra Index (PPI), a ratio of DWI over Tmax >6s volume. Penumbral profiles or PPI>1 are cases with potentially salvageable tissue compared to nonpenumbral profiles or PPI≤1.

144 patients were included, 75 in the EVT and 69 in the MM arm. Of these, 77.8% had non-penumbral (64% EVT, 36% MM) and 22.2% had penumbral profile (9% EVT, 91% MM). Non-penumbral patients were younger, had higher rates of EVT, larger cores at presentation despite similar upfront National Institutes of Health Stroke Score (NIHSS) , and had higher reduction in follow up NIHSS (24 hr, discharge and 90 day). Penumbral profile patients had, not surprisingly, clinical-radiographic mismatch at 24 hours and had higher rate of neurological deterioration or lack of improvement in NIHSS at follow up. Consequently, patients with non-penumbral profiles had better functional outcomes (90-day modified Rankin Score [mRS] 0-2, 39 vs 9% p=0.002; and 90-day mRS shift p<0.001).

Most importantly perhaps is that 42% of the MM patients had persistent perfusion deficits at 24 hours and one third continued up to 60 hours from last seen well. With only 3 of these 32 patients with penumbral profile achieving functional independence, there is undoubtedly an unmet need. Whether these very late window patients would benefit from delayed reperfusion remains unanswered.

Limitations of this study include a lack of complete delayed imaging beyond 24 hours to confirm progression of infarct in penumbral profile cases, lack of comparison between 24-hour CT perfusion core with MRI-DWI volumes, no directed analysis to only MM arm, and lack of generalizability given stringent DEFUSE 3 selection criteria. Nevertheless, this article provides a fascinating look into important patient selection considerations for future very late window reperfusion and neuroprotection trials.