.
End-to-end automation of realistic healthcare operations stresses three capabilities underrepresented in current benchmarks: policy density, decisions must be grounded in a large library of medical, insurance, and operational rules; Multi-role composition: a single task requires the agent to play multiple roles with handoffs; and multilateral interaction: intermediate workflow steps are multi-turn dialogs, such as peer-to-peer review and patient outreach. We introduce χ-Bench, a benchmark of long-horizon healthcare workflows across three domains: provider prior authorization, payer utilization management, and care management. Each task hands the agent a clinical case in a high-fidelity simulator of 20 healthcare apps exposed via 87 MCP tools, which it must drive to a terminal status through tool calls and writing the role's artifacts, guided by a 1,290+ document managed-care operations handbook skill. Across 30 agent harness/models configurations, the best agent resolves only 28.0
Abstract Background and aims The DISTALS randomized trial evaluates the safety and effectiveness of the TIGERTRIEVER13 Revascularization Device, a radially-adjustable stent-retriever designed for deployment in distal medium vessel occlusions (DMVO). Early (24h) outcome results were presented at ISC (February-2026); full, results, including 3-month outcomes, will be presented at ESOC-2026. Methods DISTALS is a US/Europe, multicenter, prospective, randomized, blinded endpoint, FDA/IDE trial in patients ineligible for IV thrombolysis. Trial objective is to evaluate Tigertriever13 safety/effectiveness in restoring blood flow in DMVO patients presenting within 24h of onset. The primary outcome is successful reperfusion (24h/CTP) without symptomatic intracranial hemorrhage (sICH). Clinical efficacy/safety secondary outcomes include 90d global disability (mRS shift), the EQ-5D, MoCA, mortality and any ICH. Results Between May 2022 and December 2025, 149 patients were enrolled, 31 lead-in and 118 randomized. Mean age was 69 (±11); 39%-women; baseline NIHSS 7.3 (±4.5). Occlusion locations were: non-/co-dominant M2-41%; M3-25%; PCA-25%; ACA-9%. LKWT-to-randomization was 511 minutes. The primary endpoint, successful reperfusion without sICH, favored the treatment arm, 86.3% vs. 27.7%, p<0.001. sICH events were 1.6% vs 0% (p=1), ITT, with 0 sICH in randomized patients treated with Tigertriever13. The last patient 3-month follow-up will be completed in March 2026, allowing secondary outcomes results to be presented at ESOC. Conclusions Early outcome results of the DISTALS trial demonstrate that the Tigertriever13 is a safe and effective clot removal treatment strategy for DMVO patients. Clinical endpoint results will provide important additional insights into treatment outcomes. Conflict of interest
Background Since the inception of spinal radiosurgery decades ago, the factors influencing spinal cord recovery have remained unclear. Elucidating these factors would aid in reirradiation planning and delivery. We describe modeling reirradiation normal tissue complication probabilities (NTCP) from compiled published data, akin to the collaborative Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) and High Dose per Fraction, Hypofractionated Treatment Effects in the Clinic (HyTEC) projects.Purpose Simple dose-time recovery factors for the spinal cord serve as a basis to propose the Reirradiation Treatment Effects in the Clinic (ReTEC) project. ReTEC aims to define human organ recovery factors to convert the total composite dose to an equivalent physical dose compatible with the de novo therapy dose tolerance limits. Additionally, we present a system for automating analyses as well as clinical implementation of radiation dose-time recovery.Methods Published literature was queried for stereotactic voxel-wise composite reirradiation plans involving the spinal cord, but none were found. Hence, minimum reporting criteria were used instead: spinal cord maximum dose (Dmax) of each course, time interval between courses, and number of fractions or biologically effective dose, per patient or in sufficiently homogeneous groups of patients. Recovery factors and logistic model parameters were fitted with maximum likelihood techniques.Results An NTCP model was derived from 13 papers with data from 282 lesions, including six myelopathy cases. The model suggested spinal cord recovery of more than 50% in less than 1 year, which is faster and more complete than current clinical practice, but with wide confidence intervals (CI) (bootstrap 95% CI: 31%-95% when the time interval was 1 year).Conclusion Due to limited data, we recommend remaining within current standard guidelines until more comprehensive analyses, optimally under the ReTEC initiative, refute or validate these estimates. ReTEC would require composite plans accounting for differing Dmax locations in each course. Nevertheless, this proof-of-concept Dmax model provides insight and serves as a basis to propose ReTEC.