Faculty Bibliography

PURPOSE/OBJECTIVE:Patients with KRAS-mutant non-small cell lung cancer (NSCLC) with brain metastases (BM) have a poor prognosis. Adagrasib (MRTX849), a potent oral small molecule KRASG12C inhibitor, irreversibly and selectively binds KRASG12C, locking it in its inactive state. Adagrasib has been optimized for favorable pharmacokinetic (PK) properties, including long half-life (~24 hours), extensive tissue distribution, dose-dependent PK, and central nervous system penetration, however BM-specific anti-tumor activity of KRASG12C inhibitors remains to be fully characterized. EXPERIMENTAL DESIGN/METHODS:A retrospective database query identified patients with KRAS-mutant NSCLC to understand their propensity to develop BM. Preclinical studies assessed physiochemical and PK properties of adagrasib. Mice bearing intracranial KRASG12C-mutant NSCLC xenografts (LU99-Luc/H23-Luc/LU65-Luc) were treated with clinically relevant adagrasib doses and levels of adagrasib in plasma, cerebrospinal fluid (CSF), and brain were determined along with anti-tumor activity. Preliminary clinical data were collected from 2 patients with NSCLC with untreated BM who had received adagrasib 600 mg BID in the Phase 1b cohort of the KRYSTAL-1 trial; CSF was collected, adagrasib concentrations measured, and anti-tumor activity in BM evaluated. RESULTS:Patients with KRAS-mutant NSCLC demonstrated high propensity to develop BM ({greater than or equal to}40%). Adagrasib penetrated into CSF and demonstrated tumor regression and extended survival in multiple preclinical BM models. In 2 patients with NSCLC and untreated BM, CSF concentrations of adagrasib measured above the target cellular IC50. Both patients demonstrated corresponding BM regression, supporting potential clinical activity of adagrasib in the brain. CONCLUSIONS:These data support further development of adagrasib in patients with KRASG12C-mutant NSCLC with untreated BM.

Importance/UNASSIGNED:Neuropsychiatric symptoms have been reported as a prominent feature of postacute sequelae of COVID-19 (PASC), with common symptoms that include cognitive impairment, sleep difficulties, depression, posttraumatic stress, and substance use disorders. A primary challenge of parsing PASC epidemiology and pathophysiology is the lack of a standard definition of the syndrome, and little is known regarding mechanisms of neuropsychiatric PASC. Observations/UNASSIGNED:Rates of symptom prevalence vary, but at least 1 PASC neuropsychiatric symptom has been reported in as many as 90% of patients 6 months after COVID-19 hospitalization and in approximately 25% of nonhospitalized adults with COVID-19. Mechanisms of neuropsychiatric sequelae of COVID-19 are still being elucidated. They may include static brain injury accrued during acute COVID-19, neurodegeneration triggered by secondary effects of acute COVID-19, autoimmune mechanisms with chronic inflammation, viral persistence in tissue reservoirs, or reactivation of other latent viruses. Despite rapidly emerging data, many gaps in knowledge persist related to the variable definitions of PASC, lack of standardized phenotyping or biomarkers, variability in virus genotypes, ascertainment biases, and limited accounting for social determinants of health and pandemic-related stressors. Conclusions and Relevance/UNASSIGNED:Growing data support a high prevalence of PASC neuropsychiatric symptoms, but the current literature is heterogeneous with variable assessments of critical epidemiological factors. By enrolling large patient samples and conducting state-of-the-art assessments, the Researching COVID to Enhance Recovery (RECOVER), a multicenter research initiative funded by the National Institutes of Health, will help clarify PASC epidemiology, pathophysiology, and mechanisms of injury, as well as identify targets for therapeutic intervention.

BACKGROUND:Abstraction of critical data from unstructured radiologic reports using natural language processing (NLP) is a powerful tool to automate the detection of important clinical features and enhance research efforts. We present a set of NLP approaches to identify critical findings in patients with acute ischemic stroke from radiology reports of computed tomography (CT) and magnetic resonance imaging (MRI). METHODS:We trained machine learning classifiers to identify categorical outcomes of edema, midline shift (MLS), hemorrhagic transformation, and parenchymal hematoma, as well as rule-based systems (RBS) to identify intraventricular hemorrhage (IVH) and continuous MLS measurements within CT/MRI reports. Using a derivation cohort of 2289 reports from 550 individuals with acute middle cerebral artery territory ischemic strokes, we externally validated our models on reports from a separate institution as well as from patients with ischemic strokes in any vascular territory. RESULTS:In all data sets, a deep neural network with pretrained biomedical word embeddings (BioClinicalBERT) achieved the highest discrimination performance for binary prediction of edema (area under precision recall curve [AUPRC] > 0.94), MLS (AUPRC > 0.98), hemorrhagic conversion (AUPRC > 0.89), and parenchymal hematoma (AUPRC > 0.76). BioClinicalBERT outperformed lasso regression (p < 0.001) for all outcomes except parenchymal hematoma (p = 0.755). Tailored RBS for IVH and continuous MLS outperformed BioClinicalBERT (p < 0.001) and linear regression, respectively (p < 0.001). CONCLUSIONS:Our study demonstrates robust performance and external validity of a core NLP tool kit for identifying both categorical and continuous outcomes of ischemic stroke from unstructured radiographic text data. Medically tailored NLP methods have multiple important big data applications, including scalable electronic phenotyping, augmentation of clinical risk prediction models, and facilitation of automatic alert systems in the hospital setting.

Embedded performance validity tests (PVTs) are key components of neuropsychological evaluations. However, most are memory-based and may be less useful in the assessment of attention-deficit/hyperactivity disorder (ADHD). Four non-memory-based validity indices derived from processing speed and executive functioning measures commonly included in ADHD evaluations, namely Verbal Fluency (VF) and the Trail Making Test (TMT), were cross-validated using the Rey 15-Item Test (RFIT) Recall and Recall/Recognition as memory-based comparison measures. This consecutive case series included data from 416 demographically-diverse adults who underwent outpatient neuropsychological evaluation for ADHD. Validity classifications were established, with ≤1 PVT failure of five independent criterion PVTs as indicative of valid performance (374 valid performers/42 invalid performers). Among the statistically significant validity indicators, TMT-A and TMT-B T-scores (AUCs = .707-.723) had acceptable classification accuracy ranges and sensitivities ranging from 29%-36% (≥89% specificity). RFIT Recall/Recognition produced similar results as TMT-B T-score with 42% sensitivity/90% specificity, but with lower classification accuracy. In evaluating adult ADHD, VF and TMT embedded PVTs demonstrated comparable sensitivity and specificity values to those found in other clinical populations but necessitated alternate cut-scores. Results also support use of RFIT Recall/Recognition over the standard RFIT Recall as a PVT for adult ADHD evaluations.

This proceedings article presents actionable research targets on the basis of the presentations and discussions at the 2nd Curing Coma National Institutes of Health (NIH) symposium held from May 3 to May 5, 2021. Here, we summarize the background, research priorities, panel discussions, and deliverables discussed during the symposium across six major domains related to disorders of consciousness. The six domains include (1) Biology of Coma, (2) Coma Database, (3) Neuroprognostication, (4) Care of Comatose Patients, (5) Early Clinical Trials, and (6) Long-term Recovery. Following the 1st Curing Coma NIH virtual symposium held on September 9 to September 10, 2020, six workgroups, each consisting of field experts in respective domains, were formed and tasked with identifying gaps and developing key priorities and deliverables to advance the mission of the Curing Coma Campaign. The highly interactive and inspiring presentations and panel discussions during the 3-day virtual NIH symposium identified several action items for the Curing Coma Campaign mission, which we summarize in this article.

OBJECTIVE:Accreditation Council for Graduate Medical Education (ACGME)-accredited epilepsy fellowships, like other ACGME accredited training programs, use Milestones to establish learning objectives and to evaluate how well trainees are achieving these goals. The ACGME began developing the second iteration of the Milestones 6 years ago, and these are now being adapted to all specialties. Here, we describe the process by which Epilepsy Milestones 2.0 were developed and summarize them. METHODS:A work group of nine board-certified, adult and pediatric epileptologists reviewed Epilepsy Milestones 1.0 and revised them using a modified Delphi approach. RESULTS:The new Milestones share structural changes with all other specialties, including a clearer stepwise progression in professional development and the harmonized Milestones that address competencies common to all medical fields. Much of the epilepsy-specific content remains the same, although a major addition is a set of Milestones focused on reading and interpreting electroencephalograms (EEGs), which the old Milestones lacked. Epilepsy Milestones 2.0 includes a Supplemental Guide to help program directors implement the new Milestones. Together, Epilepsy Milestones 2.0 and the Supplemental Guide recognize advances in epilepsy, including stereo-EEG, neurostimulation, genetics, and safety in epilepsy monitoring units. SIGNIFICANCE:Epilepsy Milestones 2.0 address the shortcomings of the old Milestones and should facilitate the assessment of epilepsy fellowships and fellows by program directors, faculty, and fellows themselves.

There is an urgent need for an in-depth and systematic assessment of a wide range of predictive factors related to populations most at risk for delaying and refusing COVID-19 vaccination as cases of the disease surge across the United States. Many studies have assessed a limited number of general sociodemographic and health-related factors related to low vaccination rates. Machine learning methods were used to assess the association of 151 social and health-related risk factors derived from the American Community Survey 2019 and the Centers for Disease Control and Prevention (CDC) BRFSS with the response variables of vaccination rates and unvaccinated counts in 1,555 ZIP Codes in California. The performance of various analytical models was evaluated according to their ability to regress between predictive variables and vaccination levels. Machine learning modeling identified the Gradient Boosting Regressor (GBR) as the predictive model with a higher percentage of the explained variance than the variance identified through linear and generalized regression models. A set of 20 variables explained 72.90% of the variability of unvaccinated counts among ZIP Codes in California. ZIP Codes were shown to be a more meaningful geo-local unit of analysis than county-level assessments. Modeling vaccination rates was not as effective as modeling unvaccinated counts. The public health utility of this model provides for the analysis of state and local conditions related to COVID-19 vaccination use and future public health problems and pandemics.

Orthostatic hypotension is an unusually large decrease in blood pressure on standing that increases the risk of adverse outcomes even when asymptomatic. Improvements in haemodynamic profiling with continuous blood pressure measurements have uncovered four major subtypes: initial orthostatic hypotension, delayed blood pressure recovery, classic orthostatic hypotension, and delayed orthostatic hypotension. Clinical presentations are varied and range from cognitive slowing with hypotensive unawareness or unexplained falls to classic presyncope and syncope. Establishing whether symptoms are due to orthostatic hypotension requires careful history taking, a thorough physical examination, and supine and upright blood pressure measurements. Management and prognosis vary according to the underlying cause, with the main distinction being whether orthostatic hypotension is neurogenic or non-neurogenic. Neurogenic orthostatic hypotension might be the earliest clinical manifestation of Parkinson's disease or related synucleinopathies, and often coincides with supine hypertension. The emerging variety of clinical presentations advocates a stepwise, individualised, and primarily non-pharmacological approach to the management of orthostatic hypotension. Such an approach could include the cessation of blood pressure lowering drugs, adoption of lifestyle measures (eg, counterpressure manoeuvres), and treatment with pharmacological agents in selected cases.

Clinical practice guidelines (CPGs) are statements that provide evidence-based recommendations aimed at optimizing patient care. However, many other documents are often published as "guidelines" when they are not; these documents, although also important in clinical practice, are usually not systematically produced following rigorous processes linking the evidence to the recommendations. Specifically, the International League Against Epilepsy (ILAE) guideline development toolkit aims to ensure that high-quality CPGs are developed to fill knowledge gaps and optimize the management of epilepsy. In addition to adhering to key methodological processes, guideline developers need to consider that effective CPGs should lead to improvements in clinical processes of care and health care outcomes. This requires monitoring the effectiveness of epilepsy-related CPGs and interventions to remove the barriers to epilepsy CPG implementation. This article provides an overview of what distinguishes quality CPGs from other documents and discusses their benefits and limitations. We summarize the recently revised ILAE CPG development process and elaborate on the barriers and facilitators to guideline dissemination, implementation, and adaptation.