Search

Safely maximizing extent of resection has become the central goal in glioma surgery. Especially in eloquent cortex, the goal of maximal resection is balanced with neurological risk. As new technologies emerge in the field of neurosurgery, the standards for maximal safe resection have been elevated. Fluorescence-guided surgery, intraoperative magnetic resonance imaging, and microscopic imaging methods are among the most well-validated tools available to enhance the level of accuracy and safety in glioma surgery. Each technology uses a different characteristic of glioma tissue to identify and differentiate tumor tissue from normal brain and is most effective in the context of anatomic, connectomic, and neurophysiologic context. While each tool is able to enhance resection, multiple modalities are often used in conjunction to achieve maximal safe resection. This paper reviews the mechanism and utility of the major adjuncts available for use in glioma surgery, especially in tumors within eloquent areas, and puts forth the foundation for a unified approach to how leverage currently available technology to ensure maximal safe resection.

Background Laser interstitial thermal therapy (LITT) is becoming an increasingly popular technique for the treatment of brain lesions. More minimally invasive that open craniotomy for lesion resection, LITT may be more appropriate for lesions that are harder to access through an open approach, deeper lesions, and for patients who may not tolerate open surgery. Methods A search of the current primary literature on LITT for brain lesions on PubMed was performed. These studies were reviewed and updates on the radiological, pathological, and long-term outcomes after LITT for brain metastases, primary brain tumors, and radiation necrosis as well as common complications are included. Results Larger extent of ablation and LITT as frontline treatment were potential predictors of favorable progression-free and overall survival for primary brain tumors. In brain metastases, larger extent of ablation was more significantly associated with survival benefit, whereas tumor size was a possible predictor. The most common complications after LITT are transient and permanent weakness, cerebral edema, hemorrhage, seizures, and hyponatremia. Conclusions Although the current literature is limited by small sample sizes and primarily retrospective studies, LITT is a safe and effective treatment for brain lesions in the correct patient population.

This is an interim report on a first in human Phase I dose escalation trial of the combination of two adenoviral vectors expressing HSV1-TK or Flt3L for the treatment of newly diagnosed, resectable malignant gliomas. Lack of dendritic cells from the brain precludes anti-glioma immune responses. We combined tumor cytotoxicity (Ad-HSV1TK) with recruitment of dendritic cells to gliomas (Ad-Flt3L) to induce anti-glioma immunity. In experimental models this treatment induces powerful cytotoxic CD8 and CD4 T-dependent anti-glioma immunity, immunological memory, and the capacity to recognize neo-antigens. The trial was approved through a FDA-IND, and all institutional cttees. Treatment was administered intraoperatively following complete glioma resection in newly diagnosed tumors. The trial consisted of vector dose escalation, starting at 1x10^9 v.p., and increasing to 1x10^11 v.p. of each vector, through 6 cohorts of 3 patients each. Two cycles of 14 days of valacyclovir were administered to activate HSV1-TK cytotoxicity. Cycle 1 starts on Day 1–3 post surgery for 14 days, and Cycle 2 on Week 8–12. Standard radiation, i.e., 60 Gy in 2 Gy fractions over 6 weeks, with concurrent temozolomide, was followed by cyclic temozolomide. Examination of tumor samples at primary resection and first recurrence show an increase in the infiltration of inflammatory cells. The experimental treatment was well tolerated. An MTD was not reached. There were approx. 248 AEs, and 26 SAEs; these were not linked to treatment. As secondary outcome, median survival of contemporary controls was 604 days, and median survival of trial patients was 742 days. Our results show for the first time that reprogramming of the host’s brain immune system to recognize gliomas reveals a new approach for the treatment of highly malignant brain tumors. Clinical trial information: NCT01811992.

Intraoperative diagnosis is essential in the surgical management of brain tumors. Conventional intraoperative histology is time- and labor-intensive, depletes scant specimens and is prone to artifact. Stimulated Raman histology (SRH) uses the intrinsic biochemical properties of fresh, unprocessed surgical specimens to provide label-free digital histologic images. We have previously shown that SRH is an ideal imaging modality to implement machine-learning strategies for tissue diagnosis due to the robust histochemical information encoded in pixel data. Here, we develop and validate a deep convolutional neural network (CNN) for rapid and standardized diagnosis of fresh brain tumor specimens with minimal tissue requirement using SRH. For CNN training, 153,000 400x400μm SRH fields of view (FOVs) from 70 patients across 12 common diagnostic classes of CNS tumors were used. Our deep neural network was developed from the GoogleNet InceptionV3 CNN architecture, which includes 23.8 million trainable parameters. Model validation was completed on 1,635 FOVs from 30 patients. The model was evaluated on its ability to differentiate 1) nonlesional versus lesional tissue, 2) glial versus nonglial tumors and 3) provide the correct histopathologic diagnosis. The CNN differentiated lesional from non-lesional FOVs with 98.4% accuracy, glial from non-glial tumor FOVs with 96.4% accuracy and classified FOVs into correct histopathologic subtype with 83.5% accuracy. By compiling FOV data from each specimen, the CNN predicted accurate diagnosis in 96.7% (29/30) of cases in the validation set. The only error by the CNN was a misclassification of a WHO grade III ependymoma as a WHO grade IV glioblastoma. We present the first application of a deep CNN to intraoperative diagnosis of brain tumors using SRH. Our study demonstrates the feasibility of applying deep machine learning to brain tumor diagnosis and justifies more extensive study on combining automated diagnosis with SRH to improve the surgical care of brain tumor patients.

INTRODUCTION: Accurate intraoperative histopathologic diagnosis is essential for optimal surgical management of pediatric brain tumors. Standard intraoperative histopathology suffers from intensive tissue preparation and processing artifact. Stimulated Raman histology (SRH) uses the intrinsic biochemical properties of fresh surgical specimens to provide label-free histologic images. We aimed to evaluate the ability of SRH to reveal histopathologic features of pediatric brain tumors for accurate intraoperative diagnosis. METHODS: Pediatric and young adult brain tumor patients (≤26 years) were prospectively enrolled over 20 months. Fresh surgical specimens were imaged intraoperatively using the first clinical fiber-laser-based SRH microscope. Reviewers evaluated SRH images for histopathologic features and diagnostic accuracy was calculated. RESULTS: A total of 23 patients were enrolled and histopathologic diagnoses included pilocytic astrocytoma(8), ganglioglioma(5), medulloblastoma(4), primitive neuroectodermal tumor(2), ependymoma(2), high-grade glioma(1) and germinoma(1). SRH provided rapid histologic images with sub-micron resolution for all specimens without tissue processing artifact. A virtual hematoxylin and eosin color scheme was developed to recreate standard light microscopy conditions. Diagnostic histopathologic features (e.g. pilocytic processes, small round blue cell morphology, etc.) were evaluated in 1,617 400μm x 400μm fields of view for the above listed tumor types. Reviewers identified diagnostic features in 34/35 (97%) fresh surgical specimens using SRH. Additionally, SRH provided unique intraoperative histologic data on immune cell infiltration not detected on standard light microscopy that improved diagnostic accuracy of tumor grade. High stimulated Raman signal at 2845 cm-1 wavenumber, corresponding to CH2 symmetric stretching mode, highlights phagocytosed lipid droplets in tumor-associated macrophages. Lipid-laden macrophage infiltration predicted higher tumor grade with 99% accuracy (AUC: 0.988). CONCLUSION: SRS microscopy provides rapid, high-resolution, intraoperative images for accurate diagnosis and grading of common pediatric brain tumors. Further investigation in a larger pediatric cohort is currently underway.