Background Detection of glioma recurrence remains a challenge in modern neuro-oncology. Noninvasive radiographic imaging is unable to definitively differentiate true recurrence versus pseudoprogression. Even in biopsied tissue, it can be challenging to differentiate recurrent tumor and treatment effect. We hypothesized that intraoperative stimulated Raman histology (SRH) and deep neural networks can be used to improve the intraoperative detection of glioma recurrence. Methods We used fiber-laser-based SRH, a label-free, non-consumptive, high-resolution microscopy method (<60 secs per 1 x 1 mm2) to image a cohort of patients (n = 35) with suspected recurrent gliomas who underwent biopsy or resection. The SRH images were then used to train a convolutional neural network (CNN) and develop an inference algorithm to detect viable recurrent glioma. Following network training, the performance of the CNN was tested for diagnostic accuracy in a retrospective cohort (n = 48). Results Using patch-level CNN predictions, the inference algorithm returned a single Bernoulli distribution for the probability of tumor recurrence for each surgical specimen or patient. The external SRH validation dataset consisted of 48 patients (recurrent, 30; pseudoprogression, 18), and we achieved a diagnostic accuracy of 95.8%. Conclusion SRH with CNN-based diagnosis can be used to improve the intraoperative detection of glioma recurrence in near-real time. Our results provide insight into how optical imaging and computer vision can be combined to augment conventional diagnostic methods and improve the quality of specimen sampling at glioma recurrence.

Higher 24-hour DBP and greater nighttime systolic dipping were significantly associated with improved cognitive function. Future studies should examine whether low 24-hour DBP and lack of nighttime systolic dipping predict future cognitive impairment.

Objective:To test the hypothesis that brain injury is more common and varied in ECMO patients than radiographically observed, we described neuropathology findings of ECMO decedents and associated clinical factors from three institutions.Methods:We conducted a retrospective multi-center observational study of brain autopsies from adult ECMO recipients. Pathology findings were examined for correlation with demographics, clinical data, ECMO characteristics, and outcomes.Results:Forty-three subjects (n=13 female; median age=47 years) received autopsies after undergoing ECMO for ARDS (n=14), cardiogenic shock (n=14), and cardiac arrest (n=15). Median duration of ECMO was 140 hours, most decedents (n=40) received anticoagulants, 60% (n=26) underwent VA ECMO, and 40% (n=17) underwent VV ECMO. Neuropathology was found in 35 decedents (81%), including microhemorrhages (37%), macrohemorrhages (35%), infarctions (47%), and hypoxic-ischemic brain injury (n=17, 40%). Most pathology occurred in frontal neocortices (n=43 occurrences), basal ganglia (n=33), and cerebellum (n=26). Decedents with hemorrhage were older (median age 57 vs. 38, p=0.01), those with hypoxic brain injury had higher Sequential Organ Failure Assessment scores (8.0 vs. 2.0, p=0.04), and those with infarction had lower peak PaCO2 (53 vs. 61 mmHg, p=0.04). Six of nine patients with normal neuroimaging results were found to have pathology on autopsy. The majority underwent withdrawal of life-sustaining therapy (n=32, 74%), and two of eight patients with normal brain autopsy underwent withdrawal of life-sustaining therapy for suspected neurologic injury.Conclusion:Neuropathological findings after ECMO are common, varied, and associated with various clinical factors. Further study on underlying mechanisms is warranted and may guide ECMO management.