Quantification of tumor fluorescence during intraoperative optical cancer imaging

Abstract

Intraoperative optical cancer imaging is an emerging technology in which surgeons employ fluorophores to visualize tumors, identify tumor-positive margins and lymph nodes containing metastases. This study compares instrumentation to measure tumor fluorescence. Three imaging systems (Spectropen, Glomax, Flocam) measured and quantified fluorescent signal-to-background ratios (SBR) in vitro, murine xenografts, tissue phantoms and clinically. Evaluation criteria included the detection of small changes in fluorescence, sensitivity of signal detection at increasing depths and practicality of use. In vitro, spectroscopy was superior in detecting incremental differences in fluorescence than luminescence and digital imaging (Ln[SBR] = 6.8 ± 0.6, 2.4 ± 0.3, 2.6 ± 0.1, p = 0.0001). In fluorescent tumor cells, digital imaging measured higher SBRs than luminescence (6.1 ± 0.2 vs. 4.3 ± 0.4, p = 0.001). Spectroscopy was more sensitive than luminometry and digital imaging in identifying murine tumor fluorescence (SBR = 41.7 ± 11.5, 5.1 ± 1.8, 4.1 ± 0.9, p = 0.0001), and more sensitive than digital imaging at detecting fluorescence at increasing depths (SBR = 7.0 ± 3.4 vs. 2.4 ± 0.5, p = 0.03). Lastly, digital imaging was the most practical and least time-consuming. All methods detected incremental differences in fluorescence. Spectroscopy was the most sensitive for small changes in fluorescence. Digital imaging was the most practical considering its wide field of view, background noise filtering capability, and sensitivity to increasing depth.

Figure 1

(a) Chemical structure of ICG. (b) Absorbance spectrum of ICG. Y-axis is measured in arbitrary units. (c) Fluorescence and bright field image of a C57BL/6 mouse bearing a LLC flank tumor injected intravenously with with ICG. (d) Chemical structure of EC-17. (e) Absorbance spectrum of EC-17. Y-axis is measured in arbitrary units. (f) Fluorescence and bright field image of an excised KB tumor from a C57BL/6 mouse that has been injected intravenously with EC-17.

Figure 2

(a) Residual fluorescent tumor foci detected in the surgical bed by fluorescence imaging after macroscopic tumor resection. (b) H&E staining was performed on the tumor margin and was confirmed by a pathologist to contain tumor cells.

Figure 3

(a) Serial dilution concentrations of ICG and EC-17. (b) SBR vs ICG concentration. Y-axis is measured in arbitrary units. Error bars are reported as standard deviations (STD). (c) SBR vs. EC-17 concentrations. Dilutions were prepared with phosphate buffered saline. Y-axis is measured in arbitrary units. Error bars are reported as standard deviations (STD).

Figure 4

(a) KB cells incubated with 18.4 μM EC-17 under 200× magnification fluoresce upon excitation by 490 nm light. (b) KB cells incubated with 18.4 μM EC-17. Black wells containing increasing logarithmic values of cells were imaged, and the signal was quantified using the luminometer and ROI software. The pseudocolor map shows decreasing areas of detectable fluorescence, and some glare is present in all wells. Y-axis is measured in arbitrary units. Error bars are reported as standard deviations (STD). (c) RCC10 cells incubated 18.4 μM EC-17. Black wells containing increasing logarithmic values of cells were imaged, and the signal was quantified using the luminometer and ROI software. Y-axis is measured in arbitrary units. Error bars are reported as standard deviations (STD).

Figure 5

(a) Brightfield, fluorescent, and pseudocolor images of LLC flank tumors from BL6 mice injected with ICG. (b) Signal-to-background ratio of TC1 murine flank tumors imaged with ICG. Error bars are reported as standard deviations (STD). (c) Signal-to-background ratio of LLC murine flank tumors imaged with ICG. Error bars are reported as standard deviations (STD). (d) Signal-to-background ratio of KB murine flank tumors imaged with EC17. Error bars are reported as standard deviations (STD). (e) Signal-to-noise ratio of IGROV murine flank tumors imaged with EC17. Error bars are reported as standard deviations (STD). (f) CT scan, bisected nodule of human adenocarcinoma patient, bright field, and fluorescent,. (g) Signal-to-background ratio of human adenocarcinoma imaged with ICG. Error bars are reported as standard deviations (STD).

Figure 6

(a) Fluorescent images of 3 cm ICG balloon phantoms submerged under various depths of liquefied butter. The Flocam ROI SBR data was gathered from these images. (b) Flocam ROI SBR vs. depth of tumor. ICG phantom sizes range from 1–5 cm. (c) Spectropen SBR vs. depth of tumor. ICG phantom sizes ranges from 1 cm–5 cm.