Citations

Objective Lennox–Gastaut syndrome (LGS) is a severe treatment‐resistant epilepsy. Although vagus nerve stimulation (VNS) and deep brain stimulation (DBS) are established therapies, comparative data and the impact of dual neuromodulation remain limited. We compared seizure outcomes across VNS, DBS, and combined DBS + VNS therapy in LGS. Methods This retrospective cohort study at Mayo Clinic (2005–2024) included LGS patients treated with VNS and/or DBS . Seizure frequency was assessed at 6, 12, and 24 months post‐stimulation. Responder was defined as 50% or more seizure reduction. Results Forty patients were included (42% female, median, 150 seizures/month). Eleven patients underwent sequential VNS followed by DBS and contributed data to 2 treatment epochs, yielding 3 treatment groups: VNS alone (n = 29), DBS alone (n = 12), and DBS + VNS (n = 10). At 24 months, median seizure reduction rate was comparable between VNS (50%; 95% confidence interval [ CI], 44–75%) and the overall DBS cohort (59%, 95% CI, 33–86%) ( p = 0.54). However, a significant difference emerged when comparing the 3 treatment arms ( p = 0.027). Specifically, DBS + VNS therapy yielded significantly greater median reduction rate (81%, 95% CI, 60–99%) than DBS alone (43%, 95% CI, 0–58%) ( p adj = 0.027) or VNS alone (50%, 95% CI, 44.4–75.0%) ( p adj = 0.048). Four patients attained sustained seizure freedom above 1 year at last follow‐up. Interpretation VNS and DBS provide comparable seizure reduction in LGS over a 2‐year follow‐up period. DBS + VNS was associated with further improvements in seizure reduction in this small, retrospective dataset. This possibility warrants further investigation. ANN NEUROL 2026

Artificial intelligence (AI) is now applied across oncology, including imaging, pathology, therapy selection, toxicity assessment, and survival modeling, and large language models (LLMs) have been adopted particularly quickly. Reported physician use of AI roughly doubled between 2023 and 2026. The frameworks used to evaluate these tools have not advanced at the same rate. Many are available, but each was designed for a single stage of the AI lifecycle, and none directly answers the question most relevant to a treating clinician: whether a given tool is appropriate for the individual patient under care. We reviewed the frameworks in current use, including TRIPOD+AI, PROBAST+AI, CLAIM, SPIRIT-AI, CONSORT-AI, DECIDE-AI, MINIMAR, and CREMLS, together with the oncology-specific ESMO EBAI and the newer LLM-specific guidance (TRIPOD-LLM, MI-CLEAR-LLM, CHART, and ESMO ELCAP). Each addresses part of the evaluation problem, but none is sufficient on its own for oncology, where the standard of care changes rapidly, assays drift, biomarker-defined subgroups are small, and a model validated in one period may perform poorly in the next. LLMs introduce additional challenges, including sensitivity to prompting, undisclosed vendor updates, behavioral drift, and hallucination rates approaching 50% on clinical quality-assurance tasks. The predictive-model frameworks were not designed to capture these failure modes. The central argument of this review is that the frameworks the field requires already exist; what is missing is their mandatory adoption. We propose that journals and regulators move from recommending these frameworks to requiring their use, and we outline a cancer-aware evaluation pathway together with the specific responsibilities of authors, reviewers, journals, and regulators in implementing it.

1617 Background: Large language models (LLMs) are increasingly evaluated for oncology decision support, yet guideline concordance and hallucinations remain variable across cancer types and tasks. Retrieval-augmented generation (RAG) may improve safety by constraining the evidence base used during generation. Prior oncology LLM studies report discordance with guideline-based care, but few describe a reproducible health-system “vetting mechanism” that couples guideline-only retrieval with clinician-scored safety rubrics across multiple oncology domains. Methods: We assembled 316 de-identified oncology cases in a tumor-board template: medical oncology (breast, GI, GU, leukemia; n=216), gynecologic oncology (n=50), and neuro-oncology/CNS metastatic cancer (n=50). Each case was run on three systems: baseline GPT-4 (O1), NCCN-guideline anchored RAG (O2; guideline-constrained retrieval), and a literature-anchored system (OpenEvidence; O3). Oncologists independently scored outputs using a modified Generative Performance Score (mGPS; −1 to +1) combining guideline concordance with hallucination penalties; readability/rationality was scored on a 1–5 Likert scale. Inter-rater reliability was assessed (κ =.63). This workflow was designed as a pre-deployment model credentialing step for health-system governance (source policy → case bank → clinician-scored acceptance criteria). Results: Across all 316 cases (case-averaged), O2 achieved the highest mean total mGPS (0.75) compared with O1 (0.43) and O3 (0.59). In the medical oncology cohort (n=216), O2 outperformed O1 and O3 (mean mGPS 0.72 vs 0.35 and 0.54), with higher guideline concordance (0.85 vs 0.68 and 0.75) and lower hallucination penalties (−0.09 vs −0.25 and −0.19); readability was also highest for O2 (4.17 vs 3.22 and 3.86). In gynecologic oncology (n=50), O2 again performed best (mGPS 0.827) relative to O3 (0.697) and O1 (0.647). In the CNS cohort (n=50), mean composite mGPS favored O2 (0.82) over O3 (0.68) and O1 (0.55), while readability remained stable across models (3.55–3.73). Severe hallucination/unsafe recommendation rate: O1 [26%], O2 [5%], O3 [17%] (pre-specified definition; e.g., major fabricated citation or guideline-discordant harmful recommendation). Conclusions: In a 316-case, multi-domain clinician-scored evaluation, limiting retrieval to oncology guideline sources improved safety-aligned performance relative to unconstrained generation or literature-only anchoring. A guideline-constrained RAG configuration paired with clinician-scored rubrics and tail-risk reporting can serve as a practical health-system vetting mechanism for governance, re-validation, and change control of LLM decision support, including in resource-limited settings where reliable guideline-consistent recommendations are essential.

Study Design. Systematic review and meta-analysis. Objective. To assess how operative order affects clinical outcomes and complications in patients with concomitant lumbar spinal and hip pathologies. Summary of Background Data. Concomitant lumbar spinal and hip pathologies are becoming increasingly prevalent in the growing population. Determining the optimal operative order for surgical treatment remains challenging, particularly when comparing spine arthrodesis followed by hip arthroplasty (SAHA) with hip arthroplasty followed by spine arthrodesis (HASA). Prior studies report conflicting results regarding outcomes and complications. Methods. A comprehensive literature search was performed to identify studies reporting on postoperative outcomes and complications in patients with concomitant lumbar spinal and hip pathologies. Random-effects model meta-analysis of pooled outcomes was performed comparing operative order groups (SAHA and HASA), with postoperative complications and revisions being our primary endpoints. Results. A total of 22 studies, yielding 161,326 patients who received surgical treatment for concomitant lumbar spinal and hip pathologies, were included. The majority of patients underwent SAHA (pooled estimate: 88.4%, 95% CI: 76.9–96.3%, P =0.01). Females comprised 56.1% (95% CI: 49.2–62.8%) of the study population. The mean age was 70.6±6.4 years, and the mean follow-up duration was 2.6±2.1 years. The mean number of treated levels was 2.1±1.4. No significant differences were noted between the SAHA and HASA groups with respect to complications, including hip dislocation ( P =0.7), mechanical loosening ( P =0.8), periprosthetic fractures ( P =0.7), deep venous thrombosis (DVT) ( P =0.8), and infection ( P =0.9). Revisions were required in 3.6% of patients in the SAHA group (95% CI: 2.5–4.9%) and 2.6% of patients in the HASA group (95% CI: 0.5–6.2%, P =0.5). Conclusion. Among patients surgically treated for concomitant lumbar spinal and hip pathologies, SAHA was significantly more commonly performed than HASA. Our analysis shows that postoperative complications and revisions did not significantly differ depending on operative order.

Abstract Guideline-concordant breast cancer treatment selection is increasingly complex due to rapid evolution in biomarker-driven therapies. Large language models (LLMs), such as GPT-4, offer a potential tool to support clinical decision-making. However, hallucinations and deviation from evidence-based guidelines remain significant concerns. This study evaluated whether retrieval-augmented generation (RAG) using National Comprehensive Cancer Network (NCCN) guidelines improves GPT-4 performance in treatment recommendation tasks. We developed a retrieval-augmented GPT-4 agent (RAG-GPT) by indexing 2025 NCCN Breast Cancer Guidelines into a vector database using sentence embeddings. Forty clinical vignettes were created from published case reports to reflect diverse scenarios, including HER2-low, triple-negative, and BRCA-mutated disease across all stages. Two models—a baseline GPT-4 (without augmentation) and the RAG-GPT—were prompted with structured templates requesting treatment plans, rationale, clinical trial options, and citations.Outputs were independently scored by two blinded oncology-trained reviewers using the modified Generative AI Performance Score (mG-PS). The mG-PS scored outputs for 2 categories, guideline concordance (Gold standard = 1.0, Acceptable = 0.5, Non-concordant = 0.0) and hallucination errors (Severe = -1.0, Moderate = -0.5, Mild = -0.25, None = 0.0). Guideline Concordance was deemed based on NCCN guidelines and what an oncologist would due based on using the guidelines for each case. Hallucinations are a term used for falsehoods in Artificial Intelligence, like made-up facts or studies, incorrect information, improper doses, or inappropriate or wrong treatments. To ensure consistency and minimize inter-rater variability, reviewers were provided with comprehensive scoring guidelines and detailed instructions for each evaluation metric. Scores were normalized between -1.0 and 1.0 for improved evaluation of tools. Readability and rationale clarity were rated on a 5-point Likert scale. Inter-rater reliability was assessed using Cohen’s kappa. The RAG-GPT demonstrated a significantly higher mean mG-PS score (0.599 ± 0.102, 95% CI: 0.498-0.702) compared to the baseline GPT-4, which not only scored lower but had a negative mean (-0.21 ± 0.143, 95% CI: -0.353 to -0.067). For the mG-PS Cohen’s Kappa test resulted in a score of .69 with substantial aggreament. An independent sample t-test showed a difference (t(78) = 29.1, p < 0.0001). Severe hallucinations were also absent in RAG-GPT outputs (0/40) but occurred in 8/40 baseline GPT-4 outputs. The evaluation of readability and rationality also showed notable differences, with the augmented LLM scoring significantly higher (M = 4.35, SD = 0.66) compared to the naïve LLM (M = 3.05, SD = 1.06). An independent samples t-test confirmed this difference t(65.3) = 6.58, p < 0.0001, further supporting the benefit of augmented Large Language models. Cohen’s Kappa test resulted in a score of .76 for a substantial agreement for readability and rational clarity scale. Limitations include small sample size; further fine-tuning and larger real-world validation studies remain necessary. Linking GPT-4 to NCCN guidelines via retrieval-augmented generation significantly improves treatment accuracy and reduces hallucination severity in breast cancer care scenarios. While further refinement is needed, the reduced incidence of severe hallucinations, increase in guideline-concordant recommendations, and enhanced readability findings support further development and prospective validation of guideline-anchored LLMs for clinical use. Citation Format: C. Yost, A. Aseem, B. Callas, S. Monick, L. Bonilla, R. Nguyen, R. Galamaga, S. Osborn, Y. Kumar, N. Ertz-Archambault. Retrieval-augmented GPT-4 improves NCCN-concordant breast cancer treatment recommendations [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS3-06-29.

OBJECTIVE Isthmic spondylolisthesis impacts up to 11.5% of the general population. For patients who have failed conservative therapy, decompression and fusion or decompression alone are options for surgical management. However, there remains debate as to whether reduction of mobile spondylolisthesis is integral to good patient outcomes. The present systematic review and meta-analysis aimed to address this by comparing clinical and radiological outcomes between patients treated with fusion in situ and those undergoing fusion with reduction of isthmic spondylolisthesis. METHODS The Scopus, EMBASE, Medline, and Cochrane databases were systematically queried on the basis of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to identify studies comparing treatment outcomes in patients with isthmic spondylolisthesis after in situ fusion or fusion with spondylolisthesis reduction. Endpoints of interest included Oswestry Disability Index (ODI) score, hospitalization duration, operative morbidity, and rates of surgical revision. Data from individual studies were aggregated using random-effects models to assess combined effects. RESULTS Of 277 identified studies, 7 studies were included with an aggregate 308 patients (143 underwent fusion in situ and 165 underwent reduction). Patients who underwent reduction for isthmic lumbar spondylolisthesis had significantly lower rates of pseudarthrosis at last follow-up (OR 0.22, p = 0.03). Estimated blood loss, operative time, hospitalization duration, change in ODI score, change in back pain (visual analog scale [VAS]), and reoperation rate did not differ significantly between groups. CONCLUSIONS For patients with isthmic spondylolisthesis, the present data suggest that fusion with reduction of spondylolisthesis may lower rates of pseudarthrosis as compared to in situ fusion. However, operative morbidity and improvement in patient-reported outcomes were similar between strategies.

OBJECTIVE Surgical site infection (SSI) is a leading cause of morbidity following posterior spine surgery. The optimal choice of skin closure technique depending on patient- and procedure-specific characteristics remains to be elucidated. METHODS The Neurosurgery Enterprise Registry (NER) at Mayo Clinic was queried for patients who underwent posterior spine surgery from 2017 to 2023. Propensity matching was performed to match patients on the basis of the development of SSI and baseline characteristics, including age, sex, race, and comorbidities. Perioperative outcomes were analyzed using data from the NER and medical records. RESULTS A total of 18,842 cases of posterior spine surgery were identified in the NER, and SSI was observed in 154 cases (0.8%). Of these patients, 308 patients were included in the 1:1 propensity score–matched cohort analysis, with 154 patients included in each group. Sutures were used in 224 patients (72.7%) and staples in 84 patients (27.3%) (p < 0.01). Female patients comprised 45.1% of the sutures group and 46.4% of the staples group (p = 0.8). The mean ± SD age was 59.8 ± 15.4 years in the sutures group and 60.8 ± 11.9 years in the staples group (p = 0.6). Fusion was performed in 27.2% of patients in the sutures group and 44% in the staples group (p < 0.01). Multilevel fusion comprised 68.9% of fusion cases in the sutures group and 67.6% in the staples group (p = 0.8). Comorbidities, such as diabetes (p = 0.6), hypertension (p = 0.1), and the use of disease-modifying antirheumatic drugs (DMARDs) (p = 0.5) and immunosuppressants (p = 0.2), did not differ between groups. Univariate analysis performed in the propensity score–matched cohort showed that SSI was observed in 53.6% of patients in the sutures group and 40.5% in the staples group (p = 0.04). No significant differences were noted regarding 30-day (p = 0.3), 90-day (p = 0.2), and 1-year (p = 0.3) readmissions and 30-day (p = 0.7), 90-day (p = 0.8), and 1-year (p = 0.8) reoperations. On multivariable logistic regression, SSI was not significantly associated with the choice of skin closure technique (p = 0.3). CONCLUSIONS After adjustment for spinal fusion, number of fused levels, the use of immunosuppressants, and other risk factors, SSI development was not significantly associated with the use of sutures versus staples following posterior spine surgery in our institution. Additionally, no significant differences were observed regarding baseline characteristics and other perioperative outcomes. The authors’ analysis shows that skin closure technique did not significantly affect SSI rates and other outcomes in these patients.

OBJECTIVE In this study, the authors aimed to establish a supervised machine learning (ML) model based on multiple tree-based algorithms to predict long-term biochemical outcomes and intervention-free survival (IFS) after endonasal transsphenoidal surgery (ETS) in patients with Cushing’s disease (CD). METHODS The medical records of patients who underwent ETS for CD between 2013 and 2023 were reviewed. Data were collected on the patient’s baseline characteristics, intervention details, histopathology, surgical outcomes, and postoperative endocrine functions. The study’s primary outcome was IFS, and the therapeutic outcomes were labeled as "under control" or "treatment failure," depending on whether additional therapeutic interventions after primary ETS were required. The decision tree and random forest classifiers were trained and tested to predict long-term IFS based on unseen data, using an 80/20 cohort split. RESULTS Data from 150 patients, with a median follow-up period of 56 months, were extracted. In the cohort, 42 (28%) patients required additional intervention for persistent or recurrent CD. Consequently, the IFS rates following ETS alone were 83% at 3 years and 78% at 5 years. Multivariable Cox proportional hazards analysis demonstrated that a smaller tumor diameter that could be detected by MRI (hazard ratio 0.95, 95% CI 0.90–0.99; p = 0.047) was significantly associated with greater IFS. However, the lack of tumor detection on MRI was a poor predictor. The ML-based model using a decision tree model displayed 91% accuracy (95% CI 0.70–0.94, sensitivity 87.0%, specificity 89.0%) in predicting IFS in the unseen test dataset. Random forest analysis revealed that tumor size (mean minimal depth 1.67), Knosp grade (1.75), patient age (1.80), and BMI (1.99) were the four most significant predictors of long-term IFS. CONCLUSIONS The ML algorithm could predict long-term postoperative endocrinological remission in CD with high accuracy, indicating that prognosis may vary not only with previously reported factors such as tumor size, Knosp grade, gross-total resection, and patient age but also with BMI. The decision tree flowchart could potentially stratify patients with CD before ETS, allowing for the selection of personalized treatment options and thereby assisting in determining treatment plans for these patients. This ML model may lead to a deeper understanding of the complex mechanisms of CD by uncovering patterns embedded within the data.