Comparison of cerebral oxygen saturation values determined by time-resolved spectroscopy and spatially resolved spectroscopy in pediatric patients with congenital heart disease.

PURPOSE: Time-resolved spectroscopy (TRS) has recently become clinically available and offers superior measurement accuracy over conventional spatially resolved spectroscopy (SRS). This study evaluated the feasibility of transitioning from the SRS-based INVOS 5100C to the TRS-based tNIRS-1 by comparing cerebral oxygen saturation measurements taken during pediatric cardiac surgery.

METHODS

We retrospectively analyzed data from 149 children (< 6 years) who had undergone cardiac surgery. We compared simultaneously measured cerebral tissue oxygen saturation (StO₂) (tNIRS-1) and regional cerebral oxygen saturation (rSO₂) (INVOS 5100C) after anesthesia induction. Both values were assessed against reference cerebral oxygen saturation (REF CX). Clinical decisions based on StO₂ and rSO₂ were also compared to current reference thresholds.

RESULTS

The Wilcoxon matched-pairs signed-rank test showed significantly lower StO₂ than rSO₂ (54.8 [48.9-61.0] vs. 69 [61-75], p < 0.0001); this observation was consistent across all age groups and cyanotic statuses. Bland-Altman analysis confirmed non-interchangeability (mean bias: - 13.1; limits: - 31.7 to 5.5). Compared with REF CX, rSO₂ more closely reflected true cerebral oxygenation (mean bias: - 5.7; limits: - 22.5 to 11.2) than StO₂ (mean bias: - 19.3; limits: - 36.5 to - 2.1). Most patients were classified as normal by rSO₂ but borderline by StO₂.

CONCLUSION

A transition from INVOS 5100C to tNIRS-1 is not currently feasible due to clinically significant discrepancies. Therefore, it may be appropriate to either select the device based on the specific measurement objectives or use both devices in a complementary manner.