Preoperative FDG Pet SUVmax Predicts Survival Independent Of Clinical But Not Pathologic TNM Staging Of NSCLC
Back to 86th Annual Meeting
Back to Program Outline
Objectives: Preoperative PET SUVmax predicts survival after resection of NSCLC but the extent to which SUV is independent of other clinical and pathologic prognostic variables has not been defined. The optimal SUV value for stratifying patients by prognosis has not been established.
Methods: We performed a retrospective review of patients who had both PET imaging and a R0 resection for NSCLC without induction or adjuvant therapy at one institution during the period of 1/1/00-12/31/04. Associations between survival and tumor histology, pathologic TNM stage and pathologic maximal tumor diameter were tested using log-rank test and association between survival and SUVmax of the primary site of disease evaluated with Cox regression. Estimation of optimal values for stratification by maximal chi-square method. Multivariate modeling to identify independent prognostic factors was performed using Cox regression.
Results: 501 patients (236 men:265 women) met study criteria. Median followup was 18.2 months. Resection was pneumonectomy in 18, bilobectomy in 20, lobectomy in 347, segmentectomy in 50 and wedge resection in 67. Histology was adenocarcinoma in 342, squamous carcinoma in 109, and other non-small cell carcinoma in 50. Pathologic stage was IA in 254, IB in 139, IIA in 20, IIB in 33, IIIA in 26, IIIB in 16, and IV in 13.
In an analysis using the median values for tumor size (2.5 cm) and SUV (4.4), SUV was an independent predictor of survival (p=0.03) after adjusting for pathologic tumor size (p =0.02) and histology (p<0.01). The optimal stratification value for SUV was 4.3 and for tumor size 3.3 cm. SUV was not an independent predictor of survival (p=0.09) after adjusting for pathologic TNM stage (stage IA vs IB vs II-IV p<0.01) and for histology (p <0.01).
Combining SUV, size and histology identifies three significant pre-operative prognostic categories:
Good: Adeno, SUV<4.4 and size<2.5
Poor: Large cell/sarcomatoid histology, SUV>4.4 and size>2.5
Intermediate: All other patients
Combining histology, and pathologic TNM stage identifies three statistically significant post-operative prognostic categories:
Good: Adeno, Stage IA
Poor: Large cell/sarcomatoid, Stage IB-IV
Intermediate: All other patients
15% of patients will change prognostic groups after resection, primarily from a good to an intermediate prognosis.
Conclusions: 1. PET SUV was an independent prognostic factor from clinical variables available preoperatively in this large cohort of patients. PET may be useful in selecting patients for induction therapy. 2. The optimal values for preoperative stratification by a combination of histology, SUV and tumor size are identified. 3. PET SUV does not add to the prognostic significance of post-operative pathologic TNM staging.
Back to 86th Annual Meeting
Back to Program Outline
