Intraoperative diagnosis is essential for providing safe and effective care during cancer surgery 1. The existing workflow for intraoperative diagnosis based on hematoxylin and eosin-staining of processed tissue is time-, resource-, and labor-intensive 2,3. Moreover, interpretation of intraoperative histologic images is dependent on a contracting, unevenly distributed pathology workforce 4. Here, we report a parallel workflow that combines stimulated Raman histology (SRH) 5-7 , a label-free optical imaging method, and deep convolutional neural networks (CNN) to predict diagnosis at the bedside in near real-time in an automated fashion. Specifically, our CNN, trained on over 2.5 million SRH images, predicts brain tumor diagnosis in the operating room in under 150 seconds, an order of magnitude faster than conventional techniques (e.g., 20-30 minutes) 2. In a multicenter, prospective clinical trial (n = 278) we demonstrated that CNN-based diagnosis of SRH images was non-inferior to pathologist-based interpretation of conventional histologic images (overall accuracy, 94.6% vs. 93.9%). Our CNN learned a hierarchy of recognizable histologic feature representations to classify the major histopathologic classes of brain tumors. Additionally, we implemented a semantic segmentation method to identify tumor infiltrated, diagnostic regions within SRH images. These results demonstrate how intraoperative cancer diagnosis can be streamlined, creating a complimentary pathway for tissue diagnosis that is independent of a traditional pathology laboratory.

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.

Diffusion within the extracellular and perivascular spaces of the brain plays an important role in biological processes, therapeutic delivery, and clearance mechanisms within the central nervous system. Recently, ultrasound has been used to enhance the dispersion of locally administered molecules and particles within the brain, but ultrasound-mediated effects on the brain parenchyma remain poorly understood. We combined an electron microscopy-based ultrastructural analysis with high-resolution tracking of non-adhesive nanoparticles in order to probe changes in the extracellular and perivascular spaces of the brain following a non-destructive pulsed ultrasound regimen known to alter diffusivity in other tissues. Freshly obtained rat brain neocortical slices underwent sham treatment or pulsed, low intensity ultrasound for 5 minutes at 1 MHz. Transmission electron microscopy revealed intact cells and blood vessels and evidence of enlarged spaces, particularly adjacent to blood vessels, in ultrasound-treated brain slices. Additionally, ultrasound significantly increased the diffusion rate of 100 nm, 200 nm, and 500 nm nanoparticles that were injected into the brain slices, while 2000 nm particles were unaffected. In ultrasound-treated slices, 91.6% of the 100 nm particles, 20.7% of the 200 nm particles, 13.8% of the 500 nm particles, and 0% of the 2000 nm particles exhibited diffusive motion. Thus, pulsed ultrasound can have meaningful structural effects on the brain extracellular and perivascular spaces without evidence of tissue disruption.Keywords: Ultrasound, Extracellular space, Nanoparticle, Diffusion

Cerebrospinal fluid (CSF) leak is a complication of endoscopic endonasal pituitary adenoma resection. Previous studies examining complications of pituitary adenoma resection have not examined associations of an exhaustive list of clinical and financial variables with CSF leak. We designed a retrospective analysis of 334 consecutive patients that underwent endoscopic endonasal pituitary adenoma resection at a single institution over 5 years, analyzing associations between CSF leak and demographic data, operative data, comorbidities, clinical complications and outcomes, costs, charges, and payments. Of the 20 preoperative variables studied, none were positively associated with CSF leak in between-groups comparison, although multivariate analysis revealed an association with a history of radiation to the skull base (odds ratio [OR], 8.67; 95% confidence interval [CI], 0.94–57.03; p < 0.05). CSF leak was associated with a significantly higher rate of postoperative diabetes insipidus (Δ = 33.4%, p = 0.040) and increased length of stay after operation in between-groups comparison. Multivariate analysis on postoperative variables revealed significant associations between CSF leak and intracerebral hemorrhage (OR, 17.44; 95% CI, 0.65–275.3; p < 0.05) and postoperative intracranial infection (OR, 28.73; 95% CI, 2.04–438.7; p < 0.05). Also, CSF leak was associated with significantly higher costs (Δ = $15,643, p < 0.05) and hospital charges (Δ = $46,026, p < 0.05). Operating room time, room and board, and supplies and implants were the strongest cost drivers. This study highlights the difficulty of utilizing preoperative variables to predict CSF leak, the clinical complications and outcomes of leak, and the financial subcategories that drive the costs, charges, and payments associated with this complication.