Virtual Library
Start Your Search
R. Shah
Author of
-
+
P1.08 - Poster Session with Presenters Present (ID 460)
- Event: WCLC 2016
- Type: Poster Presenters Present
- Track: Surgery
- Presentations: 1
- Moderators:
- Coordinates: 12/05/2016, 14:30 - 15:45, Hall B (Poster Area)
-
+
P1.08-022 - Risk Stratification Model to Predict Survival Following Surgical Resection for Lung Cancer Using Pathological Variables (ID 4495)
14:30 - 14:30 | Author(s): R. Shah
- Abstract
Background:
The risk of lung cancer recurrence remains a significant problem following curative-intent treatment. Novel methods of calculating this risk may have potential benefits in defining adjuvant strategies and stratifying the intensity of surveillance programs. The aim of this study was to identify factors at surgical resection of NSCLC that influenced survival in attempt to develop a probability model to predict mortality.
Methods:
Pathological variables were recorded from 1311 patients undergoing surgical resection for NSCLC from 2011 to 2014 at a tertiary UK lung cancer centre. Pathological variables analysed included T-stage, N-stage, adequacy of intra-operative lymph node sampling, pleural invasion, lymphovascular invasion, extracapsular spread, histological sub-typing, extent of surgery, grade of differentiation and R status (residual disease). Survival data was obtained from national death registries and logistic regression was used to develop a probability model to predict mortality.
Results:
Table 1. Pathological predictors of survival 1 year post surgery for NSCLC Figure 1 Using the probabilities from the logistic regression model to predict one year mortality gives an AUC of 0.741. If a probability of 0.144 is used to predict whether a patient will die within one year of surgery, sensitivity is 70.0% (119/170), specificity is 67.3% (625/929), PPV is 28.1% (119/423) and NPV is 92.5% (625/676).
Conclusion:
Survival post-curative intent surgery for NSCLC is based on multiple pathological factors as described above. Further analysis of these factors will be performed in the future to determine a risk stratification model to predict patients with low versus high risk mortality post surgery. Whilst indications for adjuvant therapy are well documented, the optimal surveillance regime is not as clear. Given the heterogenous group of patients receiving surgery for NSCLC, a predictive model may be useful in determining optimal surveillance strategies.
