Timing and Utility of Relapse Surveillance after Allogeneic Hematopoietic Cell Transplantation in Children with Leukemia

- Post-transplant relapse surveillance in children with leukemia varies widely.
- The yield of surveillance is proportional to likelihood of relapse at a given time.
- By day 365, yield is low enough that surveillance is likely not justified.

The utility and optimal timing of routine bone marrow (BM) and cerebrospinal fluid (CSF) surveillance after allogeneic hematopoietic cell transplantation (alloHCT) in children with leukemia have not been previously studied. To examine the current practice concerning relapse surveillance in this population, we conducted a national survey of pediatric bone marrow transplant physicians. Sixty-two of 152 potential participants (41%) completed the survey. For acute lymphoblastic leukemia (ALL) patients, 41 physicians (66%) reported performing routine BM analysis in all such patients, 15 (24%) in some patients and 6 (10%) in no patients. Data were similar for acute myeloid leukemia (AML). Among those who do such screening in the ALL population, 11 physicians (24%) reported performing 1 BM analysis in the first year, 11 (24%) performed 2, 6 (13%) performed 3, 12 (27%) performed 4, and 5 (12%) performed 5 to 10. Data were similar for AML. The most common time point for screening in both diseases was day 100, followed closely by day 365. With respect to central nervous system (CNS) screening in ALL, 11 physicians (18%) screened all patients, 28 (45%) screened no patients, and 23 (37%) screened only patients with prior CNS disease. Use of intrathecal chemotherapy in these patients also varied, with 7 (12%) doing so in all patients, 17 (29%) only in previously CNS-positive patients, and 35 (59%) in no patients. To assess the utility of surveillance procedures, we performed a retrospective review of 108 childhood leukemia patients after alloHCT at our center. Forty-one relapses (38%) occurred with a median time to relapse of 171 days. Five (12%) occurred after day 365. Of the 36 relapses within the first year, 20 (56%) were identified by clinical suspicion, whereas 16 (44%) were identified by routine screening procedures. The percentages of patients in whom routine screening detected relapse at days 100, 180, 270, and 365, respectively, was 6.7%, 11.1%, 11.9%, and 0%. That is, by day 365, no patient (of 38) who had routine BM surveillance had evidence of relapse on analysis of the BM. Our survey confirms a lack of standardization regarding routine BM and CSF relapse surveillance after alloHCT in children with leukemia. We have demonstrated that while day 365 post-alloHCT is a very commonly utilized time point for routine screening, the yield of such screening at this time is very low, such that the performance of these procedures may not be justified at that time. Prospective collaboration among pediatric alloHCT centers may help to provide more robust evidence-based guidelines designed to maximize utility and minimize risk.