Translator Disclaimer
1 November 2006 Compression-induced changes in the physiological state of the breast as observed through frequency domain photon migration measurements
Author Affiliations +
Abstract
We use optical spectroscopy to characterize the influence of mammographic-like compression on the physiology of the breast. We note a reduction in total hemoglobin content, tissue oxygen saturation, and optical scattering under compression. We also note a hyperemic effect during repeated compression cycles. By modeling the time course of the tissue oxygen saturation, we are able to obtain estimates for the volumetric blood flow (1.64±0.6 mL/100 mL/min) and the oxygen consumption (1.97±0.6 μmol/100 mL/min) of compressed breast tissue. These values are comparable to estimates obtained from previously published positron emission tomography (PET) measurements. We conclude that compression-induced changes in breast physiological properties are significant and should be accounted for when performing optical breast imaging. Additionally, the dynamic characteristics of the changes in breast physiological parameters, together with the ability to probe the tissue metabolic state, may prove useful for breast cancer detection.

1.

Introduction

Optical imaging is a rapidly evolving imaging modality with potential clinical impact for detecting and diagnosing breast cancer. 1, 2, 3, 4 Diffuse optical spectroscopy provides quantitative functional information based on the absorption and scattering properties of tissue that complements well high-resolution anatomical information obtained from radiographic methods. Tissue parameters derived from optical measurements include the concentration of oxy and deoxyhemoglobin ( HbR and HbO ), lipid and water, as well as the tissue hemoglobin oxygen saturation (StO2) . These parameters correlate with cancer pathological markers such as angiogenesis, hypoxia, and edema.5

Our group has developed a multimodality instrument that performs simultaneous diffuse optical tomography with x-ray tomosynthesis, and has obtained some promising initial results in imaging angiogenesis related to large breast tumors.6 One notable feature of this system is that optical imaging is performed under mammographic compression. Jiang 7 studied the effect of applying circumferential compression on the breast and found increases in both the total hemoglobin concentration and optical scattering while water content decreased. However, due to significant differences in measurement geometry as well as much lower applied pressures, it is not obvious how their results apply to our imaging system. Thus, we performed a pilot study to assess the change in the physiological parameters of the breast under mammographic-like compression and to also look for dynamic features that could allow us to enhance the diagnostic ability of our optical imaging system. For this purpose, we have recruited 56 healthy volunteers and we have measured the concentration of the hemoglobin species at various compression levels using a frequency domain spectrometer (ISS, Incorporated, Champaign, Illinois) in a same-side reflective configuration. We have also developed an algorithm for extracting the tissue oxygen consumption and volumetric blood flow from the transient changes in hemoglobin oxygen saturation, and we include data for select volunteers. We believe these quantities could become useful biomarkers for the presence of breast cancer.

2.

Materials and Methods

2.1.

Experimental Design

Figure 1 shows a schematic of the breast compression spectroscopy system. A computer-controlled translation stage applies between 0 and 12lbs of compression to the subject’s breast by means of two plastic compression plates. The upper plate is attached to the translation stage by means of two strain gages to allow force measurement, while the lower plate is table mounted on a support, and contains the various light delivery and collection fibers.

Fig. 1

Experimental setup. A computer-controlled translation stage applies compression to a subject’s breasts while frequency domain optical measurements are acquired through one source and three detector fibers. Force transducers are integrated into the compression frame for monitoring.

064016_1_010606jbo1.jpg

The near-infrared measurements were performed with a frequency-domain tissue spectrometer (model 96208, ISS, Champaign, Illinois). This instrument uses four parallel photomultiplier tube detectors that are time shared by eight multiplexed laser diodes emitting at 635, 670, 691, 752, 758, 782, 811, and 831nm , respectively. The frequency of intensity modulation is 110MHz , and heterodyne detection is performed with a cross-correlation frequency of 5kHz . A complete acquisition cycle over the eight wavelengths is completed every 80ms . The laser diodes and the photomultiplier tubes are all coupled to fiber optics. The eight individual illumination fibers, each 400μm in internal diameter, are arranged into a fiber bundle having a rectangular cross section of 3.5×2.0mm2 . The collecting circular fiber bundles are 3.0mm in internal diameter. The optical fibers are placed in contact with the breast by means of a plastic plate. As shown in Fig. 2 , the tips of the illuminating and collecting fiber bundles are arranged along a line, 1.5cm from the edge of the plate, with three collecting fiber bundles at distances of 1.5, 2.0, and 2.6cm from the single illuminating bundle.

Fig. 2

Optical probe. The detector fiber bundles (D) are placed 1.5, 2.0, and 2.6cm away from the source fiber bundle (S, individual fibers shown as tiny circles), respectively. All fibers are arranged in a line parallel to the edge of the bottom plate of the compression system, approximately 1.5cm away from the chest wall.

064016_1_010606jbo2.jpg

2.2.

Human Subject Protocol

Our Institutional Review Board (IRB) approved protocol called for the volunteer’s breast to be placed in contact with the lower plate of the compression system with her chest wall resting against the edge of the plate. The upper plate is then brought down until it just makes contact with the breast. Baseline optical properties are recorded for 30 to 90s , then 3lbs of force are applied and the position is held for 30s , after which the force is increased to 6lbs and the optical properties are acquired for another 30 to 90s . The upper plate is then moved up until all compression has been released and the entire cycle is repeated a total of 3 to 5 times. If the subject agreed, a four-step cycle was also used, containing 0-, 3-, 6-, and 12-lbs force plateaus (holding for 10s at each force level) to more closely simulate the effects of mammographic compression (generally quoted to equal approximately 20lbs of force). We chose these lower compression levels instead of full mammographic compression to improve subject compliance/accrual. Subjects were selected among healthy volunteers, more than 21 years old, with no open wounds on breast, breast implants, or breast biopsy within 3 months prior to participation in the study. The age range was 22 to 60 years old with an average of 30 years of age and standard deviation of 10 years.

2.3.

Spectroscopy Data Processing

The tissue optical absorption (μa) and scattering (μs) coefficients are derived from the frequency domain measurements using a multidistance fitting algorithm as described in Refs. 8, 9. The dependence on the source detector separation of the measured DC and AC components (IDC,IAC) as well as the phase (φ) of the diffusely reflected light is predicted by diffusion theory as follows:9

Eq. 1

ln(r2IDC)=rSDC(μa,μs)+CDC,

Eq. 2

ln(r2IAC)=rSAC(μa,μs)+CAC,

Eq. 3

φ=rSφ(μa,μs)+Cφ,
where SDC , SAC , and Sφ are the slopes characterizing the relationship between the source-detector separation r and ln(r2IDC) , ln(r2IAC) , and φ , respectively. CDC , CAC , and Cφ are constants independent of r . Again, from diffusion theory (if r3μaμs1 ):8

Eq. 4

μa=ω2ct(SφSACSACSφ),

Eq. 5

μs=SAC2Sφ23μaμa,
where ω is the modulation frequency of the light sources and ct is the speed of light in tissue. The μa(λ) calculated from Eq. 4 are then corrected for water absorption: μaw(λ)=μa(λ)Cwaterμwater(λ) . We assumed a value of 0.3 for the water fraction in breast tissue (Cwater) .5, 10, 11

To determine the oxy- (HbO2) and deoxyhemoglobin (HbR) concentrations, we assume that μaw(λ)=CHbO2μa,HbO2(λ)+CHbRμa,HbR(λ) , or in matrix form:

Eq. 6

μaw=μa,HbC,
where μaw=(μaw,λmin;;μaw,λmax) is the vector containing the water-corrected absorption coefficient values, C=(cHbO2,cHbR) is a vector reflecting the concentration of the hemoglobin species, and μa,Hb=(μa,HbO2,λmin,μa,HbR,λmin;;μa,HbO2,λmax,μa,HbR,λmax) is a matrix containing the absorption spectra of the hemoglobin species at the measurement wavelengths. Finally, by inverting Eq. 6 we obtain:

Eq. 7

C=(μa,HbT*μa,Hb)1*μa,HbT*μaw.

2.4.

Oxygen Consumption and Blood Flow Estimation

Every molecule of hemoglobin carries four molecules of oxygen, thus a mass balance for oxygenated hemoglobin (HbO2) within the measurement volume can be written as follows:

Eq. 8

dNHbO2dt=OC4+F[HbT]bloodSa,O2F[Hbt]bloodSv,O2,
where NHbO2 is the molar amount of oxygenated hemoglobin inside the measurement volume, OC is the oxygen consumption (in moles O2second ) within the measurement volume, F is the volumetric blood flow into the measurement volume (in L/second), [HbT]blood is the total blood hemoglobin concentration in the bood circulating through the tissue (i.e., as opposed to [HbT]tissue which is the total hemoglobin concentration in the entire tissue volume sampled by the diffuse photons), Sa,O2 is the oxygen saturation of arterial blood (incoming), and Sv,O2 is the oxygen saturation of venous blood (outgoing). Here we assumed that blood inflow is equal to blood outflow (i.e., constant blood volume—this assumption is not always valid, but it applies to a significant number of our measurements). Dividing Eq. 8 by the tissue volume sampled by the spectrometer Vtissue , we obtain:

Eq. 9

d[HbO2]dt=OC4Vtissue+FVtissue[HbT]blood(Sa,O2Sv,O2).

Assuming the measured tissue oxygen saturation SO2 (=[HbO2][HbT]tissue) is the average of Sa,O2 and Sv,O2 , and further noting that [HbT]tissue remains constant, we can rewrite Eq. 9 as:

Eq. 10

dSO2dt=OC4[HbT]tissueVtissue +2FbloodVtissue[HbT]blood[HbT]tissue(Sa,O2SO2).

The general solution to Eq. 10 is:

Eq. 11

SO2(t)=Cexp(2Fblood[HbT]blood[HbT]tissueVtissuet)OC8Fblood[HbT]blood+Sa,O2,
where C is an integration constant. Subject to the initial condition that SO2(t=0) is equal to the measured tissue SO2 at t=0 (SO2,init) , the particular solution to Eq. 10 can be written as:

Eq. 12

SO2(t)=Sa,O2(OCVtissue)8(FbloodVtissue)[HbT]blood +(SO2,initSa,O2+(OCVtissue)8(FbloodVtissue)[HbT]blood)exp(2Fblood[HbT]bloodVtissue[HbT]tissuet).

By fitting Eq. 12 to time-resolved tissue oxygenation measurements, the parameters (OCVtissue) and (FbloodVtissue) can be estimated (oxygen consumption and blood flow per unit tissue volume). Before this can be accomplished, however, assumptions for the values of Sa,O2 and [HbT]blood need to be made. In our analysis, we assumed Sa,O2=0.98 , which is generally regarded as a normal value for arterial blood oxygenation in a healthy person. The choice of a value for [HbT]blood is more complex. Normal systemic hemoglobin concentration values are 120 to 160gL for females.12 However, the blood in the region probed by our spectrometer is dominantly contained in capillary vessels, whose hematocrit values are known to be 2 to 5 times lower than the systemic hematocrit13 (due to the Fahraeus effect). For the purposes of this study, we assumed a value of 140gL for systemic hemoglobin concentration and a three times dilution factor in the capillaries, resulting in a value of 0.72mM for [HbT]blood .

3.

Results

Figure 3 shows a representative recording of the total hemoglobin content acquired during two consecutive compression cycles for a 30 year old woman with a body mass index (BMI) of 25. The shading in the figure indicates the force plateaus. There is a notable drop in total hemoglobin concentration, most of which occurs in going from 0 to 3lbs of force. As shown in Table 1 , when averaged over the entire patient set, a drop of 13.5±10% in HbT (mean ± standard deviation) is noted to correspond to applying 3lbs of force, with a further drop of 3.9±4% resulting from increasing compression to the 6-lb level. For five subjects that agreed to be tested at 12lbs of compression, an extra 5% decrease in HbT was observed in going from 6 to 12lbs of force.

Fig. 3

Sample total hemoglobin concentration ([HbT]) recording. The shading indicates the compression plateaus. The [HbT] decreases proportionally to the amount of force applied.

064016_1_010606jbo3.jpg

Table 1

Changes in total hemoglobin concentration averaged over all measurements. Only five subjects tested at 12lbs of force.

Compression (lbs)Relative blood volume changeStandard deviation
03 13.5% 10%
36 3.9% 4%
612 5% (Insufficient samples)

Figure 4 shows a total hemoglobin recording illustrating another physiological effect of breast compression: hyperemia. As can be seen in the figure, upon releasing compression after the first cycle, [HbT] does not simply return to the baseline level but rather significantly exceeds it. The overshoot in [HbT] continues to increase in later cycles and in the end stabilizes at approximately 25% beyond the initial baseline [HbT]. This hyperemic effect was observed in 35% of the subjects, with an average HbT overshoot of 11%.

Fig. 4

Example of reactive hyperemia during repeated compression cycles. The shading indicates the compression force. The peak total hemoglobin content increases from cycle 1 through cycle 4, then plateaus.

064016_1_010606jbo4.jpg

Figure 5 shows a time-resolved measurement of the hemoglobin oxygen saturation for the same subject as Fig. 3. While SO2 is relatively constant when no pressure is applied, it decreases during the compression plateaus, suggesting blood flow has been at least partially occluded and the tissue oxygen consumption exceeds oxygen delivery. Upon compression relief, SO2 returns to the precompression value, but does so more slowly than the recovery in HbT levels.

Fig. 5

Hemoglobin oxygen saturation (SO2) recording. SO2 decreases during compression due to reduced blood flow in the presence of sustained tissue oxygen consumption.

064016_1_010606jbo5.jpg

As shown in Fig. 6 , optical scattering also changes during breast compression and generally mimics the changes in [HbT], suggesting a correlation between the two physiological parameters.

Fig. 6

Recording of the reduced scattering coefficient (μs) at λ=691nm . The shading indicates the compression force. μs decreases during compression.

064016_1_010606jbo6.jpg

Finally, we used the time-resolved SO2 profile from the 6-lbs compression step for ten subjects where the compression plateau lasted for 90s to estimate the volume-normalized blood flow and oxygen consumption. Figure 7 shows an example fit to the measured SO2(t) profile using Eq. 11. The small oscillations in SO2 are probably artifacts caused by the cardiac cycle and serve as proof that blood flow has not ceased due to compression. Table 2 summarizes the oxygen consumption and blood flow findings. Average blood flow was 1.64±0.6mL100mLmin , while average oxygen consumption was 1.97±0.6μmol100mLmin .

Fig. 7

Sample SO2(t) fit using Eq. 11. The solid line is the model prediction, while the dotted line is the experimental SO2 recording. The small oscillation in SO2 is probably caused by the cardiac cycle.

064016_1_010606jbo7.jpg

Table 2

Volume normalized blood flow and oxygen consumption for ten subjects during 90-s , 6lbs compression plateaus.

SubjectBlood flow (mL∕100mL∕min) Oxygen consumption (μmol∕100mL∕min)
11.411.90
22.382.31
32.832.71
40.640.75
51.902.52
61.802.24
71.512.34
81.001.17
91.582.09
101.351.63
Average1.641.97
Standard deviation0.640.62

4.

Discussion

In the past, concerns have been raised that optical imaging under mammographic compression might be affected by depletion of the blood in breast tissues. From our results we can estimate that at least 50 to 70% of the initial blood volume should still be present in tissue at mammographic compression (20lbs) , which should allow for effective optical tomography. In fact, we could speculate that a higher percentage of the blood volume would be retained inside a tumor in comparison with the surrounding tissue because of the higher interstitial pressure of malignant lesions;14 thus compression could actually enhance the tumor/normal tissue contrast.

A hyperemic effect was noted in many of the subjects, in which the total hemoglobin concentration rose above its initial level on release of compression. Only very weak correlations could be found between the strength of the hyperemic effect and the subject’s body mass index, age, or day of menstrual cycle, but it is possible that the strength or time scale of hyperemia will be different in the presence of tumors/lesions.

As mentioned in the previous section, the hemoglobin oxygen saturation steadily decreases in all subjects during the compression plateaus. By modeling this decrease for the simplified case of constant [HbT]tissue , we were able to obtain quantitative measures for the normalized blood flow (1.64±0.6mL100mLmin) and oxygen consumption (1.97±0.6μmol100mLmin) within the measurement volume. To our knowledge, these are the first absolute measurements of breast tissue oxygen consumption (OC) and volumetric blood flow (BF) by optical methods. Beaney 15 have measured the regional blood flow and oxygen consumption in patients with breast carcinoma by means of the oxygen-15 steady-state inhalation technique and positron emission tomography (PET). They noted an average BF of 3.96mL100mLmin and an average OC of 0.45mLO2100mLmin=19.8μmol100mLmin for normal breast tissue. Another PET study by Wilson 16 reported an estimated blood flow of 5.6mL100mLmin . Our values are approximately three and ten times lower, respectively, for BF and OC. We believe most of the difference is explained by the difference in measurement geometry—the PET study measured the more metabolically active central area of the breast, while our measurements are limited to the most superficial 1cm of tissue, where a significant fat content leads to lower oxygen consumption estimates,17 while at the same time the compression applied to the tissue has lead to decreased blood flow. It is worth noting that in the PET study, breast tumors were found to have 4.7 times higher BF and 1.5 times higher OC than normal breast tissue. This leads us to believe that the tissue metabolic state will prove to be a highly specific marker for breast cancer.

As shown in Fig. 8 , from the optical scattering measurements a certain degree of correlation (R2=0.45) has been found between the change in μs and the change in the total hemoglobin concentration. However, the data suggest a change of 0.12cm1 in μs for every μM of HbT, which in turn would indicate that blood is responsible for approximately half of the tissue scattering. This is not realistic and in fact is contradicted by previously published results.18 Therefore, other tissue components that co-vary with HbT during compression must account for the additional scattering changes, such as water and lipid concentration. To quantify changes in these components, the wavelength range of our spectroscopic system needs to be extended further into the infrared, a future focus for our work.

Fig. 8

Correlation between changes in [HbT] and μs . The solid line indicates the linear fit, while the star (*) symbols indicate individual measurements.

064016_1_010606jbo8.jpg

5.

Conclusions

We show that breast compression leads to changes in the total hemoglobin content as well as hemoglobin oxygen saturation and tissue scattering. From the steady decrease in hemoglobin saturation observed during compression plateaus, we are able to extract quantitative estimates of the tissue metabolic state that have the potential to aid cancer detection. While changes in breast optical properties during compression are significant, we do not expect them to impede tomographic optical imaging once taken into account. Furthermore, the compression-induced dynamic features of the optical/physiological property changes as well as the tissue oxygen consumption will likely be different in the presence of cancer pathology and should provide useful diagnostic information.

Acknowledgments

We acknowledge Dana Brooks and Dibo Ntuba for useful discussions, as well as funding under the National Institutes of Health (NIH) grants R01-CA97305 and U54-CA105480.

References

1. 

B. W. Pogue, S. P. Poplack, T. O. McBride, W. A. Wells, K. S. Osterman, U. L. Osterberg, and K. D. Paulsen, “Quantitative hemoglobin tomography with diffuse near-infrared spectroscopy: pilot results in the breast,” Radiology, 218 261 –266 (2001). 0033-8419 Google Scholar

2. 

J. C. Hebden, H. Veenstra, H. Dehghani, E. M. C. Hillman, M. Schweiger, S. R. Arridge, and D. T. Delpy, “Three-dimensional time-resolved optical tomography of a conical breast phantom,” Appl. Opt., 40 3278 –3287 (2001). 0003-6935 Google Scholar

3. 

H. Jiang, Y. Xu, N. Iftimia, J. Eggert, K. Klove, L. Baron, and L. Fajardo, “Three-dimensional optical tomographic imaging of breast in a human subject,” IEEE Trans. Med. Imaging, 20 1334 –1340 (2001). https://doi.org/10.1109/42.974928 0278-0062 Google Scholar

4. 

R. Barbour, H. Graber, Y. Pei, S. Zhong, and C. Schmitz, “Optical tomographic imaging of dynamic features of dense scattering media,” J. Opt. Soc. Am. A, 18 3018 –3036 (2001). 0740-3232 Google Scholar

5. 

N. Shah, A. E. Cerussi, D. Jakubowski, D. Hsiang, J. Butler, and B. J. Tromberg, “The role of diffuse optical spectroscopy in the clinical management of breast cancer,” Dis. Markers, 19 95 –105 (2003, 2004). 0278-0240 Google Scholar

6. 

Q. Zhang, T. J. Brukilacchio, A. Li, J. J. Stott, T. Chaves, T. Wu, M. Chorlton, E. Rafferty, R. H. Moore, D. B. Kopans, and D. A. Boas, “Coregistered tomographic x-ray and optical breast imaging: initial results,” J. Biomed. Opt., 10 024033 (2005). https://doi.org/10.1117/1.1899183 1083-3668 Google Scholar

7. 

S. D. Jiang, B. W. Pogue, K. D. Paulsen, C. Kogel, and S. P. Poplack, “In vivo near-infrared spectral detection of pressure-induced changes in breast tissue,” Opt. Lett., 28 1212 –1214 (2003). 0146-9592 Google Scholar

8. 

S. Fantini, M. A. Franceschini, J. B. Fishkin, B. Barbieri, and E. Gratton, “Quantitative determination of the absorption spectra of chromophores in scattering media: a light-emitting-diode based technique,” Appl. Opt., 33 5204 –5213 (1994). 0003-6935 Google Scholar

9. 

S. Fantini, M. A. Franceschini, and E. Gratton, “Semi-infinite-geometry boundary problem for light migration in highly scattering media: a frequency-domain study in the diffusion approximation,” J. Opt. Soc. Am. B, 11 2128 –2138 (1994). 0740-3224 Google Scholar

10. 

D. R. White, H. Q. Woodard, and S. M. Hammond, “Average soft-tissue and bone models for use in radiation dosimetry,” Br. J. Radiol., 60 907 –913 (1987). 0007-1285 Google Scholar

11. 

H. Q. Woodard and D. R. White, “The composition of body tissues,” Br. J. Radiol., 59 1209 –1219 (1986). 0007-1285 Google Scholar

12. 

R. Berkow, M. H. Beers, R. M. Bogin, and A. J. Fletcher, “Common medical tests,” The Merck Manual of Medical Information (Home Edition), 1375 –1376 Merck Research Laboratories, Whitehouse Station, NJ (1997). Google Scholar

13. 

C. Desjardins and B. R. Duling, “Microvessel hematocrit: measurement and implications for capillary oxygen transport,” Am. J. Physiol. Heart Circ. Physiol., 252 (3), H494 –H503 (1987). 0363-6135 Google Scholar

14. 

S. D. Nathanson and L. Nelson, “Interstitial fluid pressure in breast-cancer, benign breast conditions, and breast parenchyma,” Ann. Surg. Oncol., 1 (4), 333 –338 (1994). 1068-9265 Google Scholar

15. 

R. P. Beaney, A. A. Lammertsma, T. Jones, C. G. McKenzie, and K. E. Halnan, “Positron emission tomography for in-vivo measurement of regional blood flow, oxygen utilisation, and blood volume in patients with breast carcinoma,” Lancet, 8369 (1), 131 –134 (1984). 0140-6736 Google Scholar

16. 

C. B. J. H. Wilson, A. A. Lammertsma, C. G. McKenzie, K. Sikora, and T. Jones, “Measurements of blood-flow and exchanging water space in breast-tumors using positron emission tomography—a rapid and noninvasive dynamic method,” Cancer Res., 52 (6), 1592 –1597 (1992). 0008-5472 Google Scholar

17. 

M. Niwayama, T. Hamaoka, L. Lin, J. Shao, N. Kudo, C. Katoh, and K. Yamamoto, “Quantitative muscle oxygenation measurement using NIRS with correction for the influence of a fat layer: comparison of oxygen consumption rates with measurements by other techniques,” Proc. SPIE, 3910 256 –265 (2000). https://doi.org/10.1117/12.384911 0277-786X Google Scholar

18. 

L. A. Paunescu, A. Michalos, J. H. Choi, U. Wolf, M. Wolf, and E. Gratton, “In vitro correlation between reduced scattering coefficient and hemoglobin concentration of human blood determined by near-infrared spectroscopy,” Proc. SPIE, 4250 319 –326 (2001). https://doi.org/10.1117/12.434505 0277-786X Google Scholar
©(2006) Society of Photo-Optical Instrumentation Engineers (SPIE)
Stefan A. Carp, Tina Kauffman, Qianqian Fang, Elizabeth A. Rafferty, Richard H. Moore, Daniel B. Kopans, and David A. Boas "Compression-induced changes in the physiological state of the breast as observed through frequency domain photon migration measurements," Journal of Biomedical Optics 11(6), 064016 (1 November 2006). https://doi.org/10.1117/1.2397572
Published: 1 November 2006
JOURNAL ARTICLE
5 PAGES


SHARE
Advertisement
Advertisement
Back to Top