Pragmatic Study Of A Lymph Node (Ln) Collection Kit For Non-Small Cell Lung Cancer (Nsclc) Resection.

8502 Background: Surgical resection is the best curative modality for NSCLC, but overall survival (OS) rates vary with quality of pathologic (p) LN staging. We studied the impact of a pre-labeled LN collection kit on pLN staging quality, operative (OP) complications and OS. Methods: Prospective, population-based multiple baseline, staggered implementation study involving all patients undergoing curative-intent NSCLC resection in all 11 eligible hospitals (with ≥5 resections/yr) in 4 contiguous Hospital Referral Regions in E. Arkansas, N. Mississippi, W. Tennessee. After 12 months of prospective baseline observation, institutions sequentially implemented the kit in 3 balanced cohorts 3 months apart. We examined OS with Kaplan-Meier method, logrank test. Crude (HR) and adjusted Hazard Ratios (aHR) with 95% CI from Cox Models are adjusted for surgeon clustering, controlled for: age, sex, histology, tumor grade, extent of resection, pT, pM, number of comorbidities. Results: Of 1,171 OPs by 32 surgeons from 2014-2017, LN kit used in 650 (56%) OPs by 20 surgeons. Kit cases were older v non-kit cases (mean 68 v 67 years, p = .026). Race (p = .12), sex (p = .089), insurance (p = .52), clinical stage distribution (p = .18) were similar between groups. Comparing kit v non-kit cases, median OP time was 124 v 144 min (p < .0001), transfusion 9 v 5% (p = .02), all other periOP complication rates were similar, incomplete (non-R0) resection rates 3 v 4% (p = .39). LN staging quality: pNX rate 0 v 7%; no mediastinal LN 2 v 18%; attainment of NCCN quality criteria (R0 + anatomic resection + ≥1 N1 + ≥3 mediastinal LN stations) 77 v 31% (p < .0001 for all); 60-day readmission 15 v 13% (p = .35); 60 day mortality 3% v 5% (p = .02). With 20 months median follow up, 3-year OS was 80% in kit v 73% in non-kit cases (p = .005). Kit cases had > 30% reduction in both HR (.67 [.50 - .89], p = .005) and aHR (.57 [.42 - .77], p < .001). In sensitivity analyses excluding sub-lobar resections, 60-day mortality, and non-adopting surgeons (to evaluate the possible impact of higher performing surgeons using the kit), aHR ranged from .54 to .61 (all p < .013). Conclusions: A LN collection kit improves staging quality and OS without adding to morbidity of curative NSCLC resection.