Abstract
To explore the visible near-infrared spectroscopic (VNIRS) characteristics of intracerebral hematoma, and provide experimental basis for hematoma localization and residual detection in hypertensive intracerebral hemorrhage (HICH) surgery. Using VNIRS, spectral data of cerebral hematoma and cortex were collected during HICH craniotomy, and characteristic spectra were matched with paired-sample T-test. A partial least squares (PLS) quantitative model for cerebral hematoma spectra was established. The reflectance of cerebral hematoma spectra in the 500-800nm band was lower than that of the cortex, and there were statistically significant differences in the 510, 565, and 630nm bands (P < 0.05). The calibration correlation coefficient of the PLS quantitative model for cerebral hematoma spectra was R2= 0.988, the cross-validation correlation coefficient was R2cv=0.982, the root mean square error of calibration was RMSEC=0.101, the root mean square error of cross-validation was RMSEV=0.122, the external validation correlation coefficient was CORRELATION=0.902, and the root mean square error of prediction was RMSEP=0.426, indicating that the model had high fitting degree and good predictive ability. VNIRS as a noninvasive, real-time and portable analysis technology, can be used for real-time detection of hematoma during HICH surgery, and provide reliable basis for hematoma localization and residual detection.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call