Faculty Bibliography

Cancer management remains fragmented across its continuum, from late-stage diagnosis and salvage therapies to non-personalized surveillance. Here, we present Oncoformer, a unified multimodal transformer model trained on the China Oncology Multimodal Prediction and Surveillance Study (COMPASS) cohort (3.67 million individuals, 17.7 million clinical visits) and validated on independent external cohorts, including the UK Biobank. Oncoformer integrates longitudinal electronic health records with chest X-ray imaging to address multiple clinical tasks: pan-cancer diagnosis (area under the receiver operating characteristic curve [AUROC] = 0.956), future cancer prediction up to 1 year before diagnosis (AUROC = 0.869), tumor stage inference (mean AUROC > 0.90), patient-specific treatment-response forecasting, and recurrence-free survival stratification across ten cancer types (all p < 0.01). Staging predictions were independently validated against postoperative pathological endpoints and shown to converge on core cancer genomic pathways. By translating routine clinical data into a dynamic view of cancer evolution, Oncoformer provides a framework for risk-informed cancer prediction and treatment stratification using routine clinical data.

Extracranial-intracranial (EC-IC) bypass surgery remains an essential tool in the management of complex cerebrovascular disease, including chronic-ischemic hypoperfused territory as well as intracranial aneurysms requiring parent vessel sacrifice. Reliable postoperative surveillance of graft patency and hemodynamic function is central to long-term success. Sonographic imaging has evolved over four decades from indirect extracranial Doppler waveform analysis through intraoperative microvascular Doppler, quantitative duplex ultrasonography, and transcranial Doppler assessment of intracranial hemodynamics to the contemporary development of trans-sonolucent cranioplasty ultrasonography (TCUS). TCUS, enabled by the replacement of the autologous bone flap with an acoustically transparent polymethyl methacrylate (PMMA) cranioplasty, permits direct, real-time visualization of the anastomosis and surrounding intracranial vasculature. This can be performed at the bedside and in the outpatient setting, without radiation or contrast. Current evidence, drawn largely from small series and early multicenter experience, positions TCUS as a reliable and potentially cost-effective complement to digital subtraction angiography for serial bypass surveillance, though prospective validation remains needed. This review traces the full arc of sonographic bypass monitoring and addresses current evidence, technical considerations, established quantitative parameters, and directions for future investigation.

Study DesignSystematic review and meta-analysis.ObjectivesWe sought to evaluate the association between pre-treatment symptom duration and outcomes in patients undergoing radiotherapy (RT)/surgery for metastatic epidural spinal cord compression (MESCC).MethodsA systematic review included publications evaluating the association between pre-treatment symptom duration and outcomes after RT/surgery in adults with MESCC. Primary exposure was pre-treatment symptom duration. Outcomes were motor-recovery, ambulation, survival and local control. Pooled-effect-estimates were calculated.Results Of 4639 studies, 37 met the inclusion criteria (26-RT,11-surgery). RT: All studies defined symptom duration as time from motor-weakness onset to RT. Longer symptom duration was associated with improved motor-recovery (Pooled-effect-estimate=2.08, 95%CI:1.68-2.58,p<0.001) and decreased mortality-risk (improved-survival)((Pooled-effect-estimate=0.72, 95%CI:0.69-0.76,p<0.001). Although longer symptom duration was consistently associated with better ambulation and lower local recurrence, few studies precluded meta-analysis. Surgery: Symptom duration was defined as time from neurological-deficit onset to surgery in 7/11 studies; three-studies used ambulatory status, and one-study used both. Longer symptom duration was associated with increased risk-of-death (worse-survival)(Pooled-effect-estimate=1.28, 95%CI:0.54-3.03,p=0.575), though statistically insignificantly. Meta-analysis for motor-recovery wasn't feasible, but most studies found longer symptom duration worsened motor-recovery, while ambulation findings were inconsistent.ConclusionSymptom duration was associated with differing outcome patterns by treatment modality. In RT cohorts, longer symptom duration was associated with improved survival and motor recovery. In surgical cohorts, it trended toward worse survival, though this did not reach statistical significance. Most surgical studies suggested an inverse association between symptom duration and motor recovery. These findings are exploratory, and should be interpreted in context of treatment selection-bias and between-cohort heterogeneity.

Diagnostic MRI evaluation of temporal lobe epilepsy (TLE) depends on the subjective visual interpretation of MRI images. These interpretations could be enhanced by quantitative artificial intelligence (AI) support tools. Humans often make sequential and conditional decisions during their radiological interpretations, such as whether an abnormality is present and, if present, characterizing the abnormality. It is not known whether it is superior to train AI to treat every decision separately in a similar step-wise manner or to train a model holistically on all decisions simultaneously. Here, we analysed three large epilepsy MRI datasets [n = 3676, 2320 people with epilepsy and 1356 healthy controls (HC)] to perform two tasks: (i) establish the presence of a TLE pattern on MRI and (ii) determine TLE pattern lateralization. We compared Step-wise models that independently classify TLE versus HC and lateralize patients as left TLE (L-TLE) or right TLE (R-TLE), against a simultaneous model trained to distinguish all three classes in a single step. To do this, 3D volumetric T1-weighted images were input into an EfficientNetV2 model multiple times to ensure reproducibility of results. Class prediction, model classification confidence and saliency maps were output for interpretability. Step-wise models outperformed the Simultaneous model on both tasks (both Ps < 0.001), with an average ∼2.8% accuracy increase for discriminating HC from TLE and an average 12.7% accuracy increase for distinguishing L-TLE from R-TLE. For both the Step-wise and Simultaneous models, important features discriminating TLE from HC included the known TLE limbic pattern involving the hippocampus, parahippocampal cortical regions, cingulate cortex and lateral temporal regions. However, there was less concordance between the Step-wise and Simultaneous models for the L-TLE versus R-TLE task (all Fisher's Zs > 10.5, Ps < 0.001); the Step-wise model focused less on subcortical regions such as the thalamus and hippocampus and focused more on distributed cortical pathology. Across the two Step-wise models, 95.1% of TLE patients had accurate classifications in either HC versus TLE and/or L-TLE versus R-TLE tasks. These results included 69.6% of patients being both correctly labelled as TLE and lateralized, 13.9% being correctly labelled TLE but lateralized incorrectly and 11.6% being lateralized correctly but not detected as TLE. These findings provide evidence that diagnostic tasks with simpler, Step-wise AI models may enhance diagnostic performance and interpretability in clinical workflows. Future AI clinical support tools can leverage this step-wise approach in the early identification of TLE-related structural patterns, supporting timely diagnosis and treatment decisions.

BACKGROUND:Stereotactic radiosurgery (SRS) is used in selected patients with craniopharyngioma, yet counseling and follow-up planning often rely on population-level local control rates rather than individualized expectations over time. OBJECTIVE:To develop and internally validate a multicenter survival model to predict imaging-defined time to progression after SRS for craniopharyngioma. METHODS:We analyzed a multicenter IRRF registry of SRS-treated craniopharyngioma patients. Imaging progression was defined by the overall last imaging response (PD vs. non-PD), with censoring at last imaging follow-up when progression was not observed. A Random Survival Forest (RSF) model was evaluated using 5-fold out-of-fold cross-validation. Performance was assessed using the concordance index, time-dependent AUC at 12, 24, and 60 months with bootstrap 95% confidence intervals, integrated Brier score (IBS) over 0-60 months, and risk-stratified calibration. Benchmarks included a penalized Cox model and a Kaplan-Meier baseline. RESULTS:Among 277 patients (event rate 13.0%; median imaging follow-up 57.0 months by reverse Kaplan-Meier), RSF achieved an out-of-fold C-index of 0.905. Time-dependent AUC was 0.895 (95% CI 0.828-0.959) at 12 months, 0.897 (95% CI 0.833-0.952) at 24 months, and 0.934 (95% CI 0.889-0.969) at 60 months. IBS (0-60 months) was 0.050 with favorable calibration. CONCLUSIONS:A multicenter machine learning survival model can provide individualized, well-calibrated estimates of local control over time after SRS for craniopharyngioma to support non-prescriptive decision support. CLINICAL TRIAL NUMBER/BACKGROUND:Not applicable.

Background A landmark study of 1,600 Italian-Americans in Roseto, PA, challenged the prevailing view that high saturated fat intake was a major cause of myocardial infarction (MI). Despite similar rates of cigarette smoking and obesity, and even higher levels of saturated fat consumption compared to neighboring towns, Rosetans experienced far lower MI death rates. More than 50 years later, it remains uncertain whether Roseto's residents had better heart health than the average American and, if so, what protective factors may have been responsible. Methodology We compared MI deaths in Roseto and neighboring towns to the contemporaneous Framingham Heart Study cohort matched for age and sex. Results We found no evidence that MI deaths were lower in Roseto, PA, than in Framingham, MA when controlling for age and sex. While the role of social support in health has been established in subsequent studies, methodological issues, confounding factors, and biases challenge the validity of the Roseto study. Conclusions The dramatically lower MI and MI mortality rates among males in Roseto reflect biases in sampling and comparison populations, which also impacted the contrasting Diet-Heart Hypothesis that saturated fats cause heart disease. Although social support enhances health outcomes, the Roseto study neither supported nor refuted this connection.

OBJECTIVE:To evaluate the feasibility of same-day (postoperative day 0; POD0) physical therapy (PT) following lumbar fusion and to identify factors associated with failure to participate. METHODS:This retrospective study analyzed prospectively collected data from patients undergoing single-level posterior spinal fusion (PSF), with or without anterior (ALIF) or lateral (LLIF) interbody fusion, between January and December 2024 at a single institution. A standardized POD0 PT protocol was implemented for eligible patients. Patients were categorized into two groups: successful POD0 PT (ambulatory on POD0) and unable to participate. Demographic and surgical variables were compared between groups. Reasons for inability to participate were recorded and categorized. RESULTS:Among 129 patients in whom POD0 PT was attempted, 84 (65%) successfully participated, while 45 (35%) were unable. There were no significant differences in age, sex, BMI, ASA class, operative time, estimated blood loss, or surgical approach between groups. Patients who successfully completed POD0 PT had a significantly shorter hospital length of stay compared to those who did not (3.4 ± 1.6 vs 5.8 ± 2.9 days, P < 0.001), with no differences in complication rates, discharge disposition, emergency department visits, or reoperation rates. The most common barriers to POD0 PT were postoperative pain, medical issues (e.g., orthostatic hypotension, nausea, dizziness), and anesthesia-related somnolence. Less common factors included postoperative restrictions and logistical issues such as brace availability. CONCLUSIONS:POD0 PT following lumbar fusion is feasible in the majority of patients and is associated with a shorter hospital stay without increased complications. Failure to participate was not associated with the baseline patient or surgical characteristics evaluated in this study. Instead, the most common barriers were postoperative pain, transient medical issues, and anesthesia-related somnolence, suggesting that optimization of modifiable perioperative factors may improve the implementation of POD0 PT.

PURPOSE/OBJECTIVE:Spinal metastases may progress to debilitating pain, spinal instability, and neurological deficits. Timely referral is essential, yet delays are common because patients often first present to non-spine clinicians where red flags rarely expedite referral and guidelines primarily target spine specialists. We aimed to develop a staging-based referral tool to support non-spine clinicians in recognizing progression and guiding referral urgency. METHODS:We defined the Spinal Metastasis Staging (SMS) system as four stages: SMS I, asymptomatic; SMS II, inflammatory pain; SMS III, mechanical pain and/or spinal instability; and SMS IV, neurological deficits and/or high-grade spinal cord compression. Stages were translated into a referral algorithm organized by urgency and presented as a pocket map. The tool was refined through regional and international multidisciplinary expert panels, and feasibility was evaluated in an international survey. RESULTS:Panels endorsed the four-stage SMS system and referral algorithm. Among all survey respondents (n = 120), high acceptability was reported. Among non-spine clinicians (n = 32), 94% found the tool easy to understand, 91% considered the format suitable for clinical use, and 91% anticipated improved referrals. Overall, 88% would use the tool at least occasionally, including 55% who would use it frequently or always. CONCLUSION/CONCLUSIONS:The SMS staging system and referral tool (link) was rated feasible by expert panels and survey respondents. However, only 32 of 120 survey respondents (27%) were non-spine clinicians, so findings in this group are preliminary and may overstate acceptance. The tool should be considered provisional: prospective studies are needed to validate effects on referral and patient outcomes.

BACKGROUND:The therapeutic landscape of acute ischemic stroke (AIS) has been transformed by expanding endovascular therapy (EVT) criteria. With the incorporation of MR perfusion imaging in the 2026 AHA/ASA Guidelines for the early management of patients with AIS to support EVT patient selection, there is renewed energy and focus on the clinical applications of different MR perfusion techniques in AIS. A reliable, contrast-free perfusion technique with acceptable acquisition time and capability to identify salvageable tissue and assess collateral status may be utilized in certain clinical contexts. Arterial spin labeling (ASL) is a clinically feasible technique that can be considered as a valuable modality within stroke workflows and MRI-based EVT selection protocols, particularly in patients with renal insufficiency, contrast allergy, and contrast-limited settings. Although ASL has been investigated in multiple previous studies, its systematic integration into contemporary acute stroke clinical workflows has yet to be routinely adopted and standardized practical guidance in this regard is lacking. Furthermore, ASL is not yet validated in prospective EVT-selection trials to support timely reperfusion decisions. METHODS AND PURPOSE/UNASSIGNED:This state-of-practice paper was developed on behalf of the American Society of Functional Neuroradiology (ASFNR) by an expert panel of neuroradiologists with expertise in cerebrovascular imaging and MR perfusion. We appraise the current evidence, clinical applications, implications for therapeutic decision-making, and translational barriers of ASL in AIS triage and EVT patient selection. We also discuss ASL challenges related to workforce capacity and reimbursement, and propose recommendations for future clinical validation and implementation. CONCLUSION/CONCLUSIONS:ASL may serve as a valuable adjunct within MRI-based stroke workflows by providing complementary information on ischemic penumbra and collateral status. However, validation through multicenter prospective studies for EVT patient selection, standardized acquisition protocols, and automated postprocessing quantification pipelines are needed before ASL can be routinely integrated into time-sensitive stroke workflows.

BACKGROUND/UNASSIGNED:Advances in endoscopic endonasal approaches in ventral skull base surgery have led to increasingly complex resections and reconstructions. This study investigates differences in postoperative outcomes, including postoperative cerebrospinal fluid (CSF) leak and infection, following anterior skull base surgery and reconstruction, as a function of the type of nasal packing (absorbable vs. nonabsorbable) used to bolster the skull base reconstruction. METHODS/UNASSIGNED:A retrospective chart review was performed at a single tertiary care institution. Patients who underwent ventral skull base surgery with placement of nasal packing between January 1, 2020, and December 1, 2023, were included. We included only patients of a single rhinologist (S.L.) involved. Outcome measures included postoperative CSF leaks and postoperative infection. RESULTS/UNASSIGNED: = 0.309) between these two cohorts of patients. CONCLUSION/UNASSIGNED:While no significant differences were found in the rate of postoperative CSF leak or infection, additional patient factors should be taken into consideration when choosing between absorbable and nonabsorbable nasal packing for ventral skull base reconstruction. LEVEL OF EVIDENCE/UNASSIGNED:4.