Quantification of irrigated lesion morphology using near-infrared spectroscopy

Figure 1. Combined NIRS radiofrequency ablation catheter. (a) Cartoon illustrating the concept of RF ablation catheters incorporating NIRS measurement and its salient features. (b) Finished catheter construction distal end showing embedded optical fibers for NIRS measurement at the tip electrode.

Figure 2. Exemplary real-time monitoring of lesion formation using NIRS. (a) Spectral changes in measured relative reflectance (RRel) in response to RF energy (nRF = 44, nnoRF = 44). Two main groups of changes were observed: an increase in the mean spectral amplitude or offset (black arrow) and pronounced changes in spectral morphology (blue arrows). Dashed lines (standard deviation). Bold line (mean) (b) Temporal changes in measured relative reflectance in response to RF ablation. (c) The first derivative of the RRel curve, normalized at 650 nm, to further emphasize changes in spectral shape. With RF onset, RRel near 600–700 nm decreased (black arrow), but RRel in the 740–760 nm and 840–960 nm increased. (d) Temporal changes in the first derivative of the RRel curve. (e) Corresponding lesion stained with TTC to reveal the extent of RFA damage. The maximum width and central lesion depth were derived from digitized images as shown. Black dashed lines (RF ablation onset). White dashed lines (RF ablation offset).

Figure 3. Catheter probe contact assessment ex vivo and in vivo. (a) Represents RRel spectra of ex vivo contact (red) and non-contact (blue) measurements for 10 s. Black dashes denote contact optical index (COI) wavelengths. COI is described as follows: RRel (764 nm)/RRel(730 nm). (b) Compares recorded COI between contact and non-contact. Bars are presented as mean and standard deviation (****P < 0.0001). (c,d) Repeated for in vivo demonstration.

Figure 4. Comparison between model predictions and measured lesion depth (a), width (b), and depth percentage (c) across all samples. Black circles correspond to the evaluations across the full training set which comprises the validation set (n = 75); red circles correspond to test set samples (n = 13).

Figure 5. Representative samples for real-time lesion size approximation ex vivo. (a,d) Show recorded lesion optical index (LOI1 = \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sum\nolimits_{{ \wedge_{1} }}^{{ \wedge_{2} }} {}$$\end{document}∑∧1∧2 RRel (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda$$\end{document}λ), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left[ { \wedge_{1} ,\; \wedge_{2} } \right]$$\end{document}∧1,∧2 = [600 nm, 1000 nm]) changes over time for two different size lesions. Black and red dashes denote the onset and completion. (b,e) Show the corresponding predictions of lesion depth, width, and depth percentage,…

Figure 6. In vivo deployment of NIRS-RFA catheter. (a) Shows NIRS-RFA catheter placement at the apex under fluoroscopic guidance. (b) Shows bipolar electrogram measurements indicating good catheter contact with tissue. (c) Shows corresponding temporal fluctuations of RRel spectra over the cardiac cycle. (d) Shows amplitude variations at 800 nm reflecting a similar periodicity to the bipolar electrogram. CM cardiac motion. (e) Shows the corresponding gaussian process regression model outputs with relatively stable predictions under cardiac motion of near zero values for untreated, normal tissue.

Figure 7. In vivo NIRS assessment of endocardial lesion delivery in pigs. Example measurements are shown for two cases: a transmural (Lesion A) and non-transmural (Lesion B) case. (a,b) Shows percutaneous navigation of the NIRS-RFA catheter under fluoroscopic guidance for lesion delivery near the apex (Lesion A) and at the RV free wall (Lesion B), respectively. (c) Shows final RRel spectra measured following the NIRS acquisition of lesion delivery. (d,e) Depicts time courses of gaussian process regression model predictions derived from final RRel spectra for both lesions. Black and red dashes denote the onset and completion. (f,g) Show corresponding TTC stained cross-sections annotated with actual lesion dimensions.