초록접수 현황
| 14F-252 | 구연 발표 |
Survival Prediction Model after Surgery of NSCLC
이해원, 강문철, 백희종, 박종호
원자력의학원 흉부외과
Background : The prediction of prognosis has been essential for programing cancer treatment strategies. In NSCLC, the TNM staging system is a worldwide standard form and there have been considerate improvements in its accuracy. However, the TNM staging system still has many limitations. So we propose a hazard ratio scoring system based on survival prediction model as an alternative prognosis prediction method which can be used for specific clinical purposes.
Methods : Survival prediction model was developed and validated using clinical data of 1,750 patients who received surgery for NSCLC. Patients were randomly divided into the training set (n=1,000) and validation set (n=750). Nomogram predictions were calculated for each patient in training set by using Cox proportional hazard model and compared with actual survival. Validation set was used for evaluating prediction model and AJCC staging.
Results : The median overall survival was 55.6 months for the training set and 58.1 months for validation set. In the prediction of 5-year survival, the nomogram showed an area under the receiver operating characteristic curve of 0.78 (95% CI, 0.76-0.79) in the training set. The validation set showed a good discrimination with an AUC of 0.75 (95% CI, 0.82-0.89). AJCC TNM staging system showed an AUC of 0.67 (95% CI, 0.70-0.79) with a training set.
Conclusion : The prediction model we proposed could represent the prognosis more accurately than AJCC TNM staging system. This survival prediction model cannot replace the AJCC TNM system, however it can be used for establishment of treatment strategies based on specific treatment methods, biomarkers and new clinical parameters.
Methods : Survival prediction model was developed and validated using clinical data of 1,750 patients who received surgery for NSCLC. Patients were randomly divided into the training set (n=1,000) and validation set (n=750). Nomogram predictions were calculated for each patient in training set by using Cox proportional hazard model and compared with actual survival. Validation set was used for evaluating prediction model and AJCC staging.
Results : The median overall survival was 55.6 months for the training set and 58.1 months for validation set. In the prediction of 5-year survival, the nomogram showed an area under the receiver operating characteristic curve of 0.78 (95% CI, 0.76-0.79) in the training set. The validation set showed a good discrimination with an AUC of 0.75 (95% CI, 0.82-0.89). AJCC TNM staging system showed an AUC of 0.67 (95% CI, 0.70-0.79) with a training set.
Conclusion : The prediction model we proposed could represent the prognosis more accurately than AJCC TNM staging system. This survival prediction model cannot replace the AJCC TNM system, however it can be used for establishment of treatment strategies based on specific treatment methods, biomarkers and new clinical parameters.
책임저자: 박종호
원자력의학원 흉부외과
연락처 : 이해원, Tel: 02-970-1281 , E-mail : thorax@kirams.re.kr
