Total Body Irradiation Is Associated with Higher Prevalence of Cerebral Microbleeds in Haematopoietic Stem Cell Transplant Survivors

Abstract

Introduction Myeloablative conditioning (MAC) for haematopoietic stem cell transplant (HSCT) is associated with a wide range of toxicities. Cognitive impairment (CI) is a common late effect occurring in over 50% of HSCT survivors. We previously reported an increased incidence of cerebral microbleeds (CMBs) - haemosiderin deposits, appearing as hypointense lesions on magnetic resonance imaging (MRI) - in patients with ITP. No healthy controls in this cohort had CMBs (Cooper, 2020). CMBs can be caused by a range of factors, including cranial irradiation with 18 Gy or higher. Studies in patients receiving cranial irradiation for cerebral tumours have correlated CMBs with cognitive impairment. The presence of multiple CMBs has also been independently linked to an increased risk of dementia, stroke and intracerebral haemorrhage. TBI doses used in HSCT rarely exceed 15 Gy. Although this dose is smaller than reported in previous studies linking irradiation to CMBs, it is delivered at a time of considerable physiological stress. To date, there are no studies investigating the effect of TBI on the prevalence of CMBs. Aims Given the established association of CMBs with cranial radiotherapy, this study aims to assess the impact of TBI-based MAC on the prevalence of CMBs, as well as their association with cognitive impairment. Method We recruited HSCT survivors, at least 3 years post-HSCT. Patients who received TBI as part of their MAC (TBI group) and those who received purely chemotherapeutic MAC (control group) were included. We excluded patients with a known history of cranial irradiation, uncontrolled hypertension or insulin-dependent diabetes. All participants underwent a Susceptibility Weighted Imaging (SWI) MRI and CANTAB Insight cognitive testing. The number of CMBs was reported by a neuro-radiologist blinded to the patient data. Mann-Whitney U test was used to assess the location shift in CMB numbers between groups and Fisher's exact test was used to compare the prevalence of at least one or at least three CMBs. We used Spearman's rank correlation to assess the association between CMB number and CI. Results A total of 15 patients were recruited: 9 in the TBI group and 6 in control. The median patient age was 42 years: 41 years in the TBI group (range: 32-59 years) and 43.5 years in control (range: 27-62 years) with the median time since transplant being 66 months: 81 months in TBI group (range: 41-149 years) and 43.5 years in control (range: 40-72 years). All patients had a primary diagnosis of either AML (TBI: 5, control: 4), ALL (TBI: 2, control: 1) or CML (TBI: 2, control: 1). 8 patients received conditioning with cyclophosphamide-TBI and 1 with etoposide-TBI (median TBI dose 13.2 Gy; range: 12-14.4 Gy). 4 patients in the control group received thiotepa-busulfan-fludarabine and 2 received busulfan-cyclophosphamide. The median time to platelet engraftment was 18 days: 18 days in the TBI group (range: 13-73 days) and 21.5 days in control (range: 16-44 days). 5 patients in the TBI group and 4 in the control group had a history of chronic GVHD. None had disease relapse or more than one transplant. 9 of 15 patients had at least one CMB; 7 in TBI group (78%) compared to 2 in control (33%) (p=0.119). Although this was not significantly different, patients following TBI were more likely to have a higher number of CMBs: 4 patients (44%) in the TBI group had 3 or more CMBs, compared to none in the control (p=0.092). Median number of CMBs was 2 in TBI group (range: 0-9) and 0 in control (range: 0-2) (p=0.05) (Fig.1). Older age at HSCT was positively correlated with CMB number (ρ=0.5, p=0.06) with the correlation reaching statistical significance in an analysis of the TBI group only (ρ=0.78, p=0.01) (Fig.2). 7 patients (47%) showed impairment in processing speed and 6 (40%) in executive function. However, CMB number was not significantly correlated with the number of cognitively impaired domains on CANTAB insight testing (p=0.535). Conclusions The study provides the first evidence that low radiation doses, delivered through TBI, may increase CMB burden in HSCT survivors. The clinical significance of CMBs in this population remains unclear and our study did not demonstrate an association with impaired cognitive function. Given the complexity of peri-transplant physiology, further studies with larger cohorts would benefit from robust case-control matching and a longer follow-up with baseline cognitive assessment. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal