Introduction: Cerebral infarcts and associated cognitive impairments are a devastating consequence of sickle cell disease (SCD). While the underlying mechanisms are poorly understood, infarctions are thought to arise from anemia-induced microvascular perfusion abnormalities and subsequent reduced cerebrovascular reserve that is insufficient to meet tissue metabolic demands. Thus, quantification of abnormalities in microvascular cerebral blood flow (CBF) and oxygen extraction (OEF) may be useful in identifying infarct risk and monitoring therapeutic efficacy. Unfortunately, current modalities that quantify microvascular hemodynamics (e.g., PET, MRI) are prohibitively expensive, have limited availability, and require anesthesia in children <6y, making them inappropriate as routine monitoring tools. Transcranial Doppler ultrasound (TCD) is currently the standard screening tool for overt stroke risk in pediatric SCD, but it only measures blood flow velocity in the large arteries, which is a poor surrogate for microvascular perfusion in sickle cell disease. Diffuse optical spectroscopies (specifically near-infrared frequency-domain spectroscopy, FDNIRS, and diffuse correlation spectroscopy, DCS) offer a low-cost, non-invasive alternative for bedside monitoring of tissue-level OEF and CBF. We previously demonstrated that FDNIRS/DCS are sensitive to elevations in resting-state OEF and CBF in children with sickle cell disease compared to healthy controls (Lee, Neurophotonics 2019), consistent with previous studies using MRI and PET. In this feasibility study, we demonstrate these optical techniques are sensitive to altered cerebral hemodynamics in sickle patients who are 1) undergoing chronic transfusion, and 2) experiencing vaso-occlusive pain episodes (VOE).
Methods: To date, we have recruited 6 pediatric patients with sickle cell disease undergoing chronic transfusion (5 females and 1 male, 6 - 14 y, mean ± std hemoglobin change pre- to post-transfusion = 1 ± 0.8 g/dL) and 4 patients admitted to the Emergency department for VOE (2 females and 2 males, 8 - 18 y, mean±std hemoglobin on admission = 8.9 ± 1.6 g/dL). For the transfusion cohort, FDNIRS/DCS measurements were made immediately prior to the start of transfusion and again immediately upon completion. For the VOE cohort, FDNIRS/DCS measurements were made upon hospital admission. For all FDNIRS/DCS assessments, a custom sensor was manually held over right and left forehead to assess oxygen extraction fraction (OEF, %) and an index of microvascular cerebral blood flow (CBFi, cm2/s) (Lee, Neurophotonics 2019). Hemispheric results were averaged to yield a mean of each measured parameter. Total measurement time was less than 15 minutes.
Results: In the cohort undergoing chronic transfusion, one patient data was excluded due to poor DCS signal quality. Of the remaining 5 patients, OEF and CBFi decreased after transfusion by a median of -6.4% and -30.0%, respectively (Fig 1A, B). The FDNIRS-measured OEF decrease is comparable to previous results with MRI (Guilliams, Blood 2017) that quantified both cortical OEF and CBF response to transfusion in a similarly aged cohort. However, the DCS-measured CBFi decrease is more prominent than previously reported (30% vs. 9%). The enhanced sensitivity of DCS to CBF in sickle cell disease was reported in our recent study and is likely attributed to the confounding influences of hematocrit on the DCS-measured CBFi (Sathialingam, Biomed Opt Exp 2020).
In the cohort measured during VOE, one patient data was excluded due to poor FDNIRS data quality. Of the remaining 3 subjects, OEF was elevated compared to healthy controls and was on the upper range of values measured in a cohort of otherwise subjects with sickle cell disease who were without clinical complications and were measured as part of a separate study (Fig. 1C).
Conclusion: These data demonstrate how FDNIRS/DCS may be used as a simple, low-cost tool for bedside assessment of cerebral hemodynamics in non-sedated sickle children that could be used to track brain health over time, particularly during periods thought to be prone to hemodynamic instability like transfusion or VOEs. Although ~20% of data was discarded in this dataset due to improper sensor positioning leading to poor signal quality, we have recently implemented real-time quality control feedback to ensure our data passes quality criteria.
Disclosures
Lam: Sanguina, Inc: Current equity holder in private company.
Quantifying the Signal Quality of Low-cost Respiratory Effort Sensors for Sleep Apnea Monitoring