1183PD - Risk stratification model for resected squamous cell lung cancer (R-SQLC) patients (pts) according to clinical and pathological factors

Date 27 September 2014
Event ESMO 2014
Session NSCLC early stage, SCLC and other thoracic malignancies
Topics Non-Small-Cell Lung Cancer, Early Stage
Pathology/Molecular Biology
Presenter Emilio Bria
Citation Annals of Oncology (2014) 25 (suppl_4): iv409-iv416. 10.1093/annonc/mdu347
Authors E. Bria1, S. Pilotto1, U. Peretti1, M. Milella2, F. Facciolo3, S. Novello4, A. Marchetti5, L. Crinò6, S. Kinspergher1, A. Santo1, M. Brunelli7, I. Sperduti8, M. Chilosi7, A. Scarpa7, G. Tortora9
  • 1Medical Oncology, Azienda Ospedaliera Universitaria Integrata Verona-"Borgo Roma", 37134 - Verona/IT
  • 2Divisione Di Oncologia Medica A, Regina Elena National Cancer Institute, 00144 - Roma/IT
  • 3Thoracic Surgery, Istituto Regina Elena, 00144 - Roma/IT
  • 4Thoracic Oncology Unit, Azienda Ospedaliero-Universitaria ASOU San Luigi Gonzaga, 10043 - Orbassano/IT
  • 5University-foundation, Center of Predictive Molecular, 00000 - Chieti/IT
  • 6Oncologia Medica, Ospedale S. Maria della Misericordia, 06156 - S. Andrea delle Fratte/IT
  • 7Patologia E Diagnostica, Anatomia Patologica, Azienda Ospedaliera Universitaria Integrata Verona-Borgo Roma, 37134 - Verona/IT
  • 8Biostatistics, Regina Elena National Cancer Institute, 00144 - Roma/IT
  • 9Oncologia Medica, Azienda Ospedaliera Universitaria Integrata Verona-"Borgo Roma", 37134 - Verona/IT

 

Abstract

Aim

The aim of this preliminary analysis (AIRC-MFAG project 14282) was to define a risk classification for R-SQLC on the basis of the combination of clinical and pathological predictors, to provide a practical tool for a better pts' selection from a prognostic perspective.

Methods

Clinical and pathological data were retrospectively correlated to disease-free-, cancer-specific-, and overall-survival (DFS/CSS/OS) using a Cox model. Individual patient probability (IPP) was estimated by logistic equation. A continuous score to identify risk-classes was derived according to the model ratios and dichotomized according to prognosis with the ROC analysis.

Results

Data from 573 pts from 4 different institutions were gathered. Pts characteristics: median age: 68 years; male/female: 387/89; tumor (T)-size 1-2/3-4: 352/118; Nodes 0/ > 0: 339/139; stage I-II/III-IV: 371/99. Hazard Ratios (with 95% confidence intervals and p-values) of the multivariate analysis are shown in the table:

DFS CSS OS
Age 1.58 (1.14-2.18), p = 0.005 Not significant 2.17 (1.48-3.17), p < 0.001
T-size 1.75 (1.22-2.51), p = 0.002 2.26 (1.40-3.66), p = 0.001 2.12 (1.40-3.21), p < 0.001
Nodes 2.27 (1.57-3.27), p < 0.001 2.93 (1.79-4.80), p < 0.001 2.59 (1.70-3.96), p < 0.001
Grading 1.41 (1.03-1.94), p = 0.033 1.45 (0.94-2.23), p = 0.08 1.65 (1.13-2.40), p = 0.008

Multivariate model predicted IPP with high prognostic accuracy (0.67 for DFS). On the basis of the ROC-derived cut-off, a 2-class model differentiated low-, and high-risk pts for 3-yrs DFS (32.4% and 21.8%, p < 0.0001), CSS (84.4 and 44.3%, p < 0.0001), and OS (77.3 and 38.8%, p < 0.0001). A 3-class model differentiated low-, intermediate-, and high-risk pts for 3-yrs DFS (64.6%, 39.8%, and 21.8%, p < 0.0001), CSS (84.4%, 55.4%, and 30.9%, p < 0.0001), and OS (77.3%, 47.9%, and 27.2%, p < 0.0001), The prognostic power of both models was maintained at 5 years.

Conclusions

A risk classification system comprising the commonly adopted clinical and pathological parameters (age, tumor size, nodes and grading) accurately separates R-SQLC pts into different risk classes. The project is ongoing to integrate the model with investigational molecular predictors

Disclosure

All authors have declared no conflicts of interest.