Table 1.
Optical sources of contrast in breast tissue.
Figure 1.
Instrumentation and measurement procedure.
A) Photo of the spectral imaging device connected to a laptop. B) Photo of the 8-channel fiber optic probe secured in an aluminum adaptor. For lumpectomy measurements, the probe was interfaced to the tissue surface via holes in the plexi-glass box. C) Cartoon of a mastectomy specimen sliced open to reveal tumor (cross-sectional view). One channel of the probe was placed on grossly benign tissue and another on grossly malignant tissue. The two sites were inked for histopathology after measurements were taken. D) For the patent blue dye phantom studies, one channel of the probe was placed in the center of the vial containing the liquid phantom.
Table 2.
Sample sizes of histologically-confirmed sites.
Figure 2.
Characterization of [patent blue dye].
A) Histogram of the extracted [patent blue dye] for the site-level data (854 sites). Maximum extracted [patent blue dye] was 72.7 µM. B) Extracted absorption coefficient spectra for 3 different fat sites with varying [patent blue dye]. C) Extinction coefficients for oxy-hemoglobin (HbO2), deoxy-hemoglobin (HbH), and β-carotene measured by Prahl [54]; and patent blue dye.
Figure 3.
Example data acquired from lumpectomy margins in our previous study
[35]. A–D) 50× bicubic interpolated images of β-carotene, β-carotene/<µs’>, THb, and THb/<µs’> from a negative margin (3.5 cm×3.5 cm) and a positive margin (3.5 cm×6.5 cm). Benign (fat and fibro-adipose tissue) and malignant (ductal carcinoma in situ – DCIS) sites are highlighted. E) Cumulative distribution functions of the pixels in the negative and positive margins from the representative images along with their confidence intervals (CI).
Table 3.
Rates of change in lumpectomy and mastectomy sites.
Figure 4.
Percent change in each tissue parameter versus time from excision.
Predicted values of percent change; calculated as the fitted rate of change divided by the absolute value of the fitted intercept, multiplied by time from excision. Data is from histologically-confirmed lumpectomy sites (not all outliers are shown). The median percent change at 30 minutes is noted for each parameter along with a p-value denoting the significance of the fitted slope coefficient.
Figure 5.
Average percent errors for Monte Carlo simulated data and phantom data.
Data is shown for a single reference “phantom” (<µa> = 3.85 cm−1, <µs’> = 6.79 cm−1 for the simulated data; <µa> = 3.02 cm−1, <µs’> = 5.81 cm−1 for the phantom data). Simulated: 216 diffuse reflectance spectra were created consisting of 3 levels of scattering (4.85, 6.68, 9.15 cm−1), THb (16.97, 31.03, 55.09 µM), and β-carotene (10.29, 16.29, 24.37 µM) and 8 levels of patent blue dye (0∶10:70 µM). Phantom: patent blue dye was titrated 12 times (0–79 µM) into a phantom consisting of 5.81 cm−1 (<µs’>), 16.69 µM (THb), and 11.23 µM (β-carotene).
Figure 6.
Optical parameters of the first measured time point.
Optical parameters of the first time point from the histologically-confirmed benign sites of mastectomies (Mast) and lumpectomies (Lump). * indicates p<0.05.
Figure 7.
Rate of change in the tissue parameters.
Rate of change (fitted values from the model) in the tissue parameters from the histologically-confirmed benign sites of mastectomies (Mast) and lumpectomies (Lump) constrained to a time window of 10 min for all sites. * indicates p<0.05.
Figure 8.
Example plots of kinetics in benign and malignant tissue.
Example plots of the tissue parameters versus time for four histologically known sites from two mastectomy patients. Symbols indicate the measured data lines are the model fits for the benign and malignant tissues.
Figure 9.
Effects of time on ex vivo spectral imaging.
50x bicubic interpolated images of a negative and positive margin from different patients, where the “initial” images of the margins were imaged at approximately the same time points post-excision. The “initial” images represent the actual data measured. The median percent change at 10, 20, and 30 minutes was applied to either the negative or positive image to show how an image would change if measured at various time points beyond the “initial” image time point. A) For β-carotene, THb, and THb/<µs’>, the negative margins have higher values and the kinetics decrease over time. Therefore, the percent change is applied to the negative margin to show decreasing contrast (worst case scenario). B) For β-carotene/<µs’> the negative margins have higher values and the kinetics increase over time. C) For <µs’> the positive margins have higher values but the kinetics decrease over time.
Table 4.
Comparison of the percent change over time versus the percent difference in benign and malignant tissue.