Stature loss following skeletal irradiation for childhood cancer

A model is presented to predict adult stature in children treated successfully for cancer outside the CNS. The model is based on radiation dose in Gray adjusted for location of therapy and attained stature (GALA); ideal adult stature (IAS), assuming the patient had not developed cancer, calculated by the Roche-Wainer-Thissen (RWT) method (which uses patient stature and weight before developing cancer, and parent stature data); a femur correction if both the acetabula or heads of both femurs were irradiated (FEMUR); and sex. The model was constructed using data from 49 patients with a mean time from completion of therapy to follow-up of 8.9 years (range, 3.3 to 15.4 years). Thirteen patients received no radiotherapy. All model coefficients were highly significant (P less than .001), and the model appears to be an excellent predictor of adult stature, with a multiple correlation coefficient of 0.84 (R2 = .74) between corrected adult stature (CAS) based on the most recent follow-up stature available for the patient projected to final adult stature, compared with the model's predicted adult stature (MPAS), based only on initial data at presentation and subsequent radiation treatment. Patients who did not receive radiotherapy did not have loss of stature, ie, there was no significant difference between IAS and CAS, (P less than .71; n = 13), but patients who received radiotherapy had shorter statures than would be expected from the healthy population model (P less than .0004; n = 36). The magnitude of the loss in stature appears to be well explained by the dose and location of radiation, the stature already achieved at the time of radiotherapy, along with IAS, FEMUR, and sex. We believe this model will help clinicians to predict the growth effects of radiotherapy in children with cancer not involving the CNS.