Yield of Screening for Long-Term Complications Using the Children's Oncology Group Long-Term Follow-Up Guidelines

Abstract

PurposeThe Children's Oncology Group Long-Term Follow-Up (COG-LTFU) Guidelines use consensus-based recommendations for exposure-driven, risk-based screening for early detection of long-term complications in childhood cancer survivors. However, the yield from these recommendations is not known.MethodsSurvivors underwent COG-LTFU Guideline–directed screening. Yield was classified as negligible/negative (< 1%), intermediate (≥ 1% to < 10%), or high (≥ 10%). For long-term complications with high yield, logistic regression was used to identify subgroups more likely to screen positive.ResultsOver the course of 1,188 clinic visits, 370 childhood cancer survivors (53% male; 47% Hispanic; 69% leukemia/lymphoma survivors; median age at diagnosis, 11.1 years [range, 0.3 to 21.9 years]; time from diagnosis, 10.5 years [range, 5 to 55.8 years]) underwent 4,992 screening tests. High-yield tests included thyroid function (hypothyroidism, 10.1%), audiometry (hearing loss, 22.6%), dual-energy x-ray absorptiometry scans (low bone mineral density [BMD], 23.2%), serum ferritin (iron overload, 24.0%), and pulmonary function testing/chest x-ray (pulmonary dysfunction, 84.1%). Regression analysis failed to identify subgroups more likely to result in high screening yield, with the exception of low BMD (2.5-fold increased risk for males [P = .04]; 3.3-fold increased risk for nonobese survivors [P = .01]). Screening tests with negligible/negative (< 1%) yield included complete blood counts (therapy-related leukemia), dipstick urinalysis for proteinuria and serum blood urea nitrogen/creatinine (glomerular defects), microscopic urinalysis for hematuria (hemorrhagic cystitis, bladder cancer), ECG (anthracycline-related conduction disorder), and hepatitis B and HIV serology.ConclusionScreening tests with a high yield are appropriate for risk groups targeted for screening by the COG-LTFU Guidelines. Elimination of screening tests with negligible/negative yield should be given consideration.