Timely care is a key quality indicator. Resection delayed beyond 8 weeks for early-stage non-small cell lung cancer (NSCLC) negatively impacts prognosis, yet diagnosis-to-treatment time has steadily increased. We sought to identify drivers of delayed surgical resection and generate evidence for timeliness metrics for optimal outcomes. We evaluated patient-level factors associated with delayed versus timely surgery (>8 vs ≤8 weeks after diagnosis) for clinical stage I-II NSCLC between 2009 and 2019. Factors were identified by estimating adjusted relative risks (aRRs) and 95% CIs using modified Poisson regression. Time to process-level steps from diagnosis, along with the number and combination of these steps, were examined by timeliness of surgery. Among 2567 patients, 46.0% received timely surgery. Factors associated with delayed surgery included Black (aRR, 1.14; 95% CI, 1.02-1.27) and Asian (aRR, 1.14; 95% CI, 1.00-1.29) race, distance to surgical facility of >50 miles (aRR, 1.23; 95% CI, 1.08-1.40), and greater health care use (≥25 visits, aRR, 1.72; 95% CI, 1.53-1.93). More preoperative steps increased time to surgery, but no individual step drove delays. When performed, preoperative steps occurred within the following intervals in 75% of timely surgery recipients: positron emission tomography-computed tomography: 21 days, biopsy: 20 days, pulmonary function tests: 28 days, and thoracic surgery consult: 30 days. As health system and oncologic care processes grow in complexity, timely treatment warrants greater attention. Our results from an integrated health system indicate several patient- and process-level factors contribute to delays that can be mitigated and offer preliminary benchmarks to promote timely NSCLC management.

The objective was to synthesize study-reported outcomes of transformer-based artificial intelligence systems in surgical domains and describe settings where they appear advantageous relative to nontransformer models and human benchmarks. We searched major databases: PubMed, Embase, IEEE Xplore, ScienceDirect, Google Scholar, arXiv, and Cochrane Library. Eligible studies evaluated transformer architectures in surgical/perioperative contexts (medical imaging, workflow recognition, prognosis-related modeling, or education) and reported quantitative outcomes. Because of heterogeneous tasks/metrics, we performed a domain-organized narrative synthesis. Where the same study reported transformers and nontransformers on the same dataset/metrics, we computed within-study deltas (Δ = Transformers - Nontransformers) and summarized medians and interquartile ranges alongside vote counts (T>NT/tie/T<NT). No cross-study pooling or hypothesis testing was performed. Paired comparisons favored transformers in medical imaging for 15 of 20 (75.0%) with median Δ +1.13 percentage points (interquartile range, 3.70) and in workflow recognition for 28 of 34 (82.4%) with median Δ +1.75 percentage points (interquartile range, 3.28). Prognosis had sparse paired data (n = 1; Δ +3.0 percentage points; illustrative). Education favored transformers overall in 5 of 6 (83.3%) paired comparisons, driven by surgery time prediction; diagnostic education tasks were mixed. Reported advantages were task dependent and dataset specific; gains were typically single-digit percentage points in like-for-like settings. Transformers frequently match or exceed nontransformer baselines in surgical imaging and workflow tasks, with promising, yet heterogeneously reported, signals in prognosis and education. Translation to dependable clinical/educational impact will require standardized benchmarks, external/prospective validation, transparent comparator reporting (including human baselines), and deployment studies that address real-time operating room constraints and fairness across patient and learner groups.