Background: Several studies have shown the prognostic value of desmoplasia for lung adenocarcinomas. The authors evaluated the density and extent of desmoplasia by modifying the scar grade, as well as the prognostic impact on patient survival.
Methods: Modified scar grade was defined as follows: Grade 1, no desmoplasia; Grade 2, sparse desmoplastic reaction; Grade 3, dense desmoplastic reaction with diameter of 10 mm or less; Grade 4, dense desmoplastic reaction with diameter exceeding 10 mm. In addition, the prognostic impact of conventional histologic factors and modified scar grade was analyzed in 239 cases of small peripheral lung adenocarcinoma (maximum dimension, </= 30 mm) for which long-term follow-up data were available.
Results: The 5 and 10-year survival rates according to the modified scar grade were 100% and 100% for Grade 1 lung adenocarcinoma (n = 29); 91.7% and 83.7% for Grade 2 (n = 61); 67.6% and 52.7% for Grade 3 (n = 78); and 50.0% and 37.5% for Grade 4 (n = 71), respectively. A significant difference in patient survival was found between Grade 1 or 2 versus Grade 3 or 4 (P < 0.0001, by log rank test). Multivariate analysis showed that modified scar grade was an independent prognostic factor (P = 0.0176), as were pathologic stage (P = 0.0293), lymph node metastasis (P = 0.0191), lymphatic permeation (P = 0.0022), and pleural involvement (P = 0.0452). Modified scar grade also had a significant impact on survival in various subsets of patients, including those with pathologic Stage IA disease, patients with tumors of diameter 20 mm or less, or patients with mixed subtype tumors with a bronchioloalveolar component.
Conclusions: Modified scar grade is a useful prognostic factor in patients with small lung adenocarcinomas. Tumors with a sparse fibroblastic reaction (modified scar Grade 2) may represent early invasive cancers or invasive cancers with low malignant potential, which should be distinguished from frankly invasive cancers (modified scar Grade 3 or 4).
Copyright 2002 American Cancer Society.DOI 10.1002/cncr.11006