Michael A. Mooney

BACKGROUND: Surgical resection of brainstem cavernous malformations (BSCMs) is challenging, and patient selection and timing of intervention remain controversial. OBJECTIVE: To evaluate the impact of surgical timing and predictors of neurological outcome after surgical resection of BSCMs. METHODS: Consecutive adult patients (≥18 years) with BSCMs undergoing surgical resection between 1985 and 2014 by the senior author (RFS) were retrospectively reviewed. Patient demographics, lesion characteristics, imaging results, surgical approach, and perioperative and long-term neurological morbidity were analyzed. RESULTS: Data were analyzed for a total of 397 adult patients (160, 40% male). On univariate analysis, a greater proportion of patients treated within 6 weeks of hemorrhage had an improved Glasgow Outcome Scale score (P = .06). On logistic regression analysis, patients treated within 6 weeks of hemorrhage experienced improved clinical outcomes (odds ratio = 1.73; 95% confidence interval = 1.06-2.83; P = .03). CONCLUSIONS: Although BSCM surgery is associated with significant perioperative morbidity and mortality, favorable long-term hemorrhage rates and symptom resolution can be achieved in a carefully selected group of patients. Overall, patients treated acutely, within 6 weeks, benefited the most from surgical intervention.

OBJECTIVE The goal of this study was to determine the interrater and intrarater reliability of the Knosp grading scale for predicting pituitary adenoma cavernous sinus (CS) involvement. METHODS Six independent raters (3 neurosurgery residents, 2 pituitary surgeons, and 1 neuroradiologist) participated in the study. Each rater scored 50 unique pituitary MRI scans (with contrast) of biopsy-proven pituitary adenoma. Reliabilities for the full scale were determined 3 ways: 1) using all 50 scans, 2) using scans with midrange scores versus end scores, and 3) using a dichotomized scale that reflects common clinical practice. The performance of resident raters was compared with that of faculty raters to assess the influence of training level on reliability. RESULTS Overall, the interrater reliability of the Knosp scale was "strong" (0.73, 95% CI 0.56-0.84). However, the percent agreement for all 6 reviewers was only 10% (26% for faculty members, 30% for residents). The reliability of the middle scores (i.e., average rated Knosp Grades 1 and 2) was "very weak" (0.18, 95% CI -0.27 to 0.56) and the percent agreement for all reviewers was only 5%. When the scale was dichotomized into tumors unlikely to have intraoperative CS involvement (Grades 0, 1, and 2) and those likely to have CS involvement (Grades 3 and 4), the reliability was "strong" (0.60, 95% CI 0.39-0.75) and the percent agreement for all raters improved to 60%. There was no significant difference in reliability between residents and faculty (residents 0.72, 95% CI 0.55-0.83 vs faculty 0.73, 95% CI 0.56-0.84). Intrarater reliability was moderate to strong and increased with the level of experience. CONCLUSIONS Although these findings suggest that the Knosp grading scale has acceptable interrater reliability overall, it raises important questions about the "very weak" reliability of the scale's middle grades. By dichotomizing the scale into clinically useful groups, the authors were able to address the poor reliability and percent agreement of the intermediate grades and to isolate the most important grades for use in surgical decision making (Grades 3 and 4). Authors of future pituitary surgery studies should consider reporting Knosp grades as dichotomized results rather than as the full scale to optimize the reliability of the scale.

BACKGROUND: A simple, reliable grading scale to better characterize nonfunctioning pituitary adenomas (NFPAs) preoperatively has potential for research and clinical applications. OBJECTIVE: To develop a grading scale from a prospective multicenter cohort of patients that accurately and reliably predicts the likelihood of gross total resection (GTR) after transsphenoidal NFPA surgery. METHODS: Extent-of-resection (EOR) data from a prospective multicenter study in transsphenoidal NFPA surgery were analyzed (TRANSSPHER study; ClinicalTrials.gov NCT02357498). Sixteen preoperative radiographic magnetic resonance imaging (MRI) tumor characteristics (eg, tumor size, invasion measures, tumor signal characteristics, and parameters impacting surgical access) were evaluated to determine EOR predictors, to calculate receiver-operating characteristic curves, and to develop a grading scale. A separate validation cohort (n = 165) was examined to assess the scale's performance and inter-rater reliability. RESULTS: Data for 222 patients from 7 centers treated by 15 surgeons were analyzed. Approximately one-fifth of patients (18.5%; 41 of 222) underwent subtotal resection (STR). Maximum tumor diameter > 40 mm; nodular tumor extension through the diaphragma into the frontal lobe, temporal lobe, posterior fossa, or ventricle; and Knosp grades 3 to 4 were identified as independent STR predictors. A grading scale (TRANSSPHER grade) based on a combination of these 3 features outperformed individual variables in predicting GTR (AUC, 0.732). In a validation cohort, the scale exhibited high sensitivity and specificity (AUC, 0.779) and strong inter-rater reliability (kappa coefficient, 0.617). CONCLUSION: This simple, reliable grading scale based on preoperative MRI characteristics can be used to better characterize NFPAs for clinical and research purposes and to predict the likelihood of achieving GTR.

Confocal laser endomicroscopy (CLE) is an advanced optical fluorescence imaging technology that has potential to increase intraoperative precision, extend resection, and tailor surgery for malignant invasive brain tumors because of its subcellular dimension resolution. Despite its promising diagnostic potential, interpreting the gray tone fluorescence images can be difficult for untrained users. CLE images can be distorted by motion artifacts, fluorescence signals out of detector dynamic range, or may be obscured by red blood cells, and thus interpreted as nondiagnostic (ND). However, just a single CLE image with a detectable pathognomonic histological tissue signature can suffice for intraoperative diagnosis. Dealing with the abundance of images from CLE is not unlike sifting through a myriad of genes, proteins, or other structural or metabolic markers to find something of commonality or uniqueness in cancer that might indicate a potential treatment scheme or target. In this review, we provide a detailed description of bioinformatical analysis methodology of CLE images that begins to assist the neurosurgeon and pathologist to rapidly connect on-the-fly intraoperative imaging, pathology, and surgical observation into a conclusionary system within the concept of theranostics. We present an overview and discuss deep learning models for automatic detection of the diagnostic CLE images and discuss various training regimes and ensemble modeling effect on power of deep learning predictive models. Two major approaches reviewed in this paper include the models that can automatically classify CLE images into diagnostic/ND, glioma/nonglioma, tumor/injury/normal categories, and models that can localize histological features on the CLE images using weakly supervised methods. We also briefly review advances in the deep learning approaches used for CLE image analysis in other organs. Significant advances in speed and precision of automated diagnostic frame selection would augment the diagnostic potential of CLE, improve operative workflow, and integration into brain tumor surgery. Such technology and bioinformatics analytics lend themselves to improved precision, personalization, and theranostics in brain tumor treatment.