paediatrics Brussels 17

558

Merchant et al.

International Journal of Radiation Oncology Biology Physics

Table 4 Statistical significance of dose-volume intervals on longitudinal cognitive scores after craniospinal irradiation

Cut point V25 Gy Cut point V55 Gy V < 2 Gy V > 25 Gy V < 35 Gy V > 35 Gy V < 45 Gy V > 45 Gy V < 55 Gy V > 55 Gy Cut point V35 Gy Cut point V45 Gy

Volume

Test

Brain

EIQ

n.s

.0079

n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.

.0027

n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.

.0140 .0010 .0252 .0153 .0032 .0177 n.s.

n.s. n.s. n.s. n.s. n.s.

.0185

Math

n.s.

< .0001 < .0001

< .0001 < .0001

n.s. n.s. n.s. n.s. n.s. n.s. n.s.

Reading n.s. Spelling n.s.

.0003 .0077 .0001

.0001 .0028

Supratentorial

EIQ

n.s. n.s.

Math

< .0001 < .0001 < .0001

.0403 .0051 .0024 .0018

Reading n.s. Spelling n.s.

< .0001

.0002

.0486 .0483

n.s.

Infratentorial

EIQ

n.s. n.s.

n.s.

n.s.

.0144

.0002

Math

.0002

.0003

< .0001 < .0001 < .0001

n.s. n.s. n.s. n.s.

< .0001 < .0001 < .0001

Reading n.s. Spelling n.s.

< .0001 < .0001

< .0001 < .0001

.0309 .0152

Temporal Left

EIQ

n.s. n.s.

.0362 .0001

.0010

.0016

.0167

Math

< .0001 < .0001 < .0001 < .0001 < .0001 < .0001 .0006

< .0001 < .0001

.0590 .0063 .0033

n.s. n.s. n.s.

Reading n.s. Spelling n.s.

< .0001 < .0001

.0035 .0003

Temporal Right EIQ

n.s. n.s.

.0413 .0004

.0094

.0203

n.s. n.s.

.0019

Math

< .0001 < .0001

n.s.

Reading n.s. Spelling n.s.

< .0001 < .0001

.0282 .0234

.0201

.0006 n.s. Abbreviations: EIQ Z estimated intelligence quotient; Math Z WIAT math scores; Reading Z WIAT reading scores; Spelling Z WIAT spelling scores. Other abbreviations as in Table 2 .

it had a significant impact on longitudinal scores in all models ( Table 3 ). The P values are included in Table 3 to show the presence or absence of an association and the relative significance. Appendix 1 (available online) in- cludes the full models.

The only exception was for EIQ. At the high-dose cut points of 45 Gy and 55 Gy, the higher-dose term was smaller than the lower-dose term and retained statistical significance ( Table 4 ). The P values are included in Table 4 to show the presence or absence of an association and the relative significance. Appendix 2 includes the full models.

TD 50/5 for below-average IQ and academic achievement according to mean normal tissue dose

TD 50/5 for below-average IQ and academic achievement according to radiation dose intervals

We calculated the mean dose required for a child to have a 50% risk of a below-average IQ or academic achievement test score 5 years after irradiation. The calculation was performed using age-adjusted mean dose models. The estimated mean doses are presented as iso-effect curves in Figure 1 . The larger-volume normal tissue structures required a lower dose to achieve the same effect as the smaller-volume normal tissue structures.

For the entire brain and left and right temporal lobes we calculated, according to the age of the patient at the time of RT, the threshold volumes receiving dose in excess of 25 Gy, 35 Gy, 45 Gy, and 55 Gy that would have 50% of cognitive scores falling below 85 for EIQ 5 years after RT. The results show that no additional dose to the entire brain above a specified level would be required for patients with the specified ages or younger to have a 5% probability of EIQ < 85 at 5 years: age 8 years and 25 Gy, age 12 years and 30 Gy, age 15 years and 35 Gy. For both the left and right temporal lobes these values were age 8 years and 25 Gy, age 8 years and 30 Gy, age 10 years and 35 Gy, age 12 years and 40 Gy, age 12 years and 45 Gy, and age 15 years and 50 Gy. The results show that there is a < 50% probability of an EIQ < 85 for the following combination of brain dose and age: < 25 Gy and 8 years, < 30 Gy and 12 years, and < 35 Gy and 15 years. The probability of an EIQ < 85 at 5 years is < 50% for the following combi- nations of left and right temporal lobe dose: < 25 Gy and

Impact of radiation dose intervals on longitudinal trends in cognitive scores

Using the cut points of 25 Gy and 35 Gy, the higher dose interval had a consistent and statistically significant impact on the longitudinal trend in cognitive test scores, whereas the lower dose interval did not. A similar finding was observed for the infratentorial and temporal lobe volumes at 45 Gy. For the other normal tissue volumes evaluated at 45 Gy and all normal tissue volumes at 55 Gy, the impact of dose was significant only when the dose interval (high or low) included the majority of the volume (data not shown).