Tumor Control Probability Predicts the Fate of Multiple Metastatic Brain Tumors

Abstract

Over 50,000 cases of metastatic brain tumors (MBTs) have been treated with gamma knife (GK) radiosurgery. However, optimal patient selection and treatment factors are still controversial in the cases with over five lesions. Tumor control probability (TCP) may be useful in these situations. From 1992 to 1999, 284 patients with metastatic brain tumor were treated by GK. Within that group, 32 patients (18 male, 14 female) had over five lesions. Their ages ranged from 26 to 77 years with a median age of 56 years. The number of lesions per patient ranged from 5 to 56 (average 11.5). The total volume of the lesions ranged from 1.8 to 111.8cm3 (average 21.8 cm3). Marginal dose ranged from 8 to 22 Gy (average 15.3 Gy). Normal tissue complication probability (NTCP) ranged from 0.0 to 31% (average 8.7%) and TCP ranged from 0.0 to 83% (average 19.6%). Eleven patients had undergone previous craniotomy. Three were irradiated previously and 6 were given concomitant whole-brain irradiation. Eleven patients had additional extracranial metastatic lesions. Survival period ranged from 0.2 to 26.2 months (median 5 months). Using Cox’s proportional hazard model, only TCP correlated with the survival times (STs) (p = 0.007). Two patients whose TCP values were over 75% but NTCP values were over 18% survived 5 and 10 months, respectively. These values were very low compared to the expected STs estimated from very high TCP values. The equation, ST = 118.5 ×(0.3 - NTCP)×TCP + 3.2, may explain this phenomenon. ST and (0.3 - NTCP)×TCP were somewhat well correlated (R = 0.67, p