Clinical profiles and intraoperative identification of complex glands in stage I lung adenocarcinoma

Authors:

Abstract
This study aimed to investigate the potential of complex glandular patterns (CGP) in lymph node micrometastasis (LNMM), and to determine the clinical beneficiaries in stage I lung adenocarcinoma (LUAD) with CGP. Meanwhile, the feasibility of detecting CGP on frozen section was also evaluated. We retrospectively analyzed the clinicopathological characteristics of 848 pathologic-stage I LUADs. A logistic regression model and a Cox proportional-hazards model were conducted to define the risk factors for LNMM and survival respectively. Furthermore, five pathologists reviewed frozen sections of 100 LUADs independently. The logistic regression model indicated that CGP [odds ratio 3.89, 95% confidence interval (CI) 2.46–6.15; P < 0.001] was an independent predictor of the presence of LNMM. Subgroup analysis revealed that CGP-present/LNMM-positive LUAD had the highest risk of both loco-regional and distant recurrence. Moreover, adequate lymphadenectomy [recurrence-free survival: hazard ratio (HR) 0.61, 95% CI 0.40–0.95; P = 0.028; overall survival: HR 0.64, 95% CI 0.41–0.99; P = 0.043) and adjuvant chemotherapy (recurrence-free survival: HR 0.30, 95% CI 0.18–0.52; P < 0.001; overall survival: HR 0.33, 95% CI 0.19–0.57; P < 0.001) brought survival benefits to CGP-present patients, especially to CGP-present/LNMM-positive subgroup. Across the five pathologists, sensitivity ranged from 59% to 68% and specificity ranged from 79% to 83%, with moderate diagnostic agreement and high interobserver agreement for detecting CGP on frozen section. LNMM was more frequently observed in stage I LUAD with CGP. Adequate lymphadenectomy and adjuvant chemotherapy were associated with improved survival in CGP-present patients, especially in CGP-present/LNMM-positive subgroup. Additionally, it’s feasible to identify CGP on frozen section intraoperatively.
Funding Information
  • Medical Technology Program of Suzhou National Hi-Tech District (2019Q015)

This publication has 24 references indexed in Scilit: