Header
Header
Article

Intraoperative Optical and Fluorescence Imaging of Blood Flo… : Plastic and Reconstructive Surgery


Insufficient blood flow causes mastectomy skin flap necrosis in 5 to 30 percent of cases.1–3 Current methods for assessing skin flap viability include clinical judgment and intraoperative fluorescence angiography. Fluorescence angiography with the injection of indocyanine green has shown high sensitivities (90 to 100 percent) but moderate specificities (72 to 50 percent) in predicting mastectomy skin flap necrosis.4,5 Limitations with fluorescence angiography include allergic reaction, short time-window for observation (indocyanine green plasma half-life is approximately 3 minutes),6 high cost for equipment and supplies,6 and lack of continuous use in preoperative and postoperative settings.5,7

We previously reported a noncontact diffuse correlation spectroscopy for measuring local blood flow levels in mastectomy skin flaps and head/neck free muscle flaps.8,9 Preliminary results demonstrated high sensitivity (100 percent) and specificity (81 percent) in predicting mastectomy skin flap necrosis,8 although the spectroscopic measurements (not imaging) did not provide spatial resolution to determine ischemic tissue volumes for excision. We recently developed an innovative, noncontact, camera-based speckle contrast diffuse correlation tomography method for high-density imaging of blood flow distributions in deep tissues (up to approximately 10 mm).10–19 Case studies with speckle contrast diffuse correlation tomography were reported for three-dimensional imaging of blood flow distributions in brain, wound/burn tissues, and mastectomy skin flaps.10–18 In this preliminary study, a small group of patients undergoing mastectomies were imaged sequentially by dye-free speckle contrast diffuse correlation tomography and dye-based commercial fluorescence angiography (SPY-PHI; Novadaq/Stryker, Kalamazoo, Mich.) to identify ischemic tissues in mastectomy skin flaps. We hypothesized that speckle contrast diffuse correlation tomography and indocyanine green–fluorescence angiography measurements of blood flow distributions in mastectomy skin flaps are consistent.

PATIENTS AND METHODS

With the University of Kentucky Institutional Review Board approval, 11 patients participated in this nonblinded, observational study. Women undergoing skin-sparing or nipple-sparing mastectomies and immediate one- or two-stage implant-based immediate reconstruction were included. Exclusion criteria were delayed or autologous tissue reconstruction, age younger than 18 years, and allergy to indocyanine green.

The upgraded speckle contrast diffuse correlation tomography method was used to image blood flow distributions immediately following mastectomy while the incision was temporarily closed with staples (Fig. 1). Indocyanine green was then injected intravenously and its perfusion in the mastectomy skin flap was imaged by a handheld SPY-PHI probe. Breast reconstruction was performed by a single plastic surgeon (L.W.). Additional excision of ischemic tissues was performed in some patients based on the SPY-PHI imaging. The primary clinical endpoint was the development of mastectomy skin flap necrosis.

Fig. 1.:

Intraoperative imaging of mastectomy skin flaps. (Left) The upgraded speckle contrast diffuse correlation tomography (scDCT) method for noncontact three-dimensional imaging of blood flow distributions in mastectomy skin flaps. A new scientific complementary metal-oxide-semiconductor (sCMOS) camera was used to measure both blood flow distribution (by speckle contrast diffuse correlation tomography) and tissue surface geometry (by photometric stereo technique). This new scientific complementary metal-oxide-semiconductor camera has a larger pixel quantity (2048 × 2048), faster frame rate (30/second), and higher quantum efficiency (>50 percent at 800 nm), compared to our previous electron-multiplying charge-coupled device camera (pixels, 1004 × 1002; frame rate, 9/second; quantum- efficiency, 40 percent at 800 nm (Cascade 1K; Photometrics, Tucson, Ariz.). Movements of red blood cells in the measured tissue volume (“banana-shape”) produced continuous laser speckle fluctuations on the tissue surface, which were captured by the scientific complementary metal-oxide-semiconductor camera. These boundary data from multiple sources (e.g., S1, S2) and multiple pixel-windows/detectors (e.g., D1, D2) were input into the finite-element method–based NIRFAST program for three-dimensional reconstruction of blood flow distributions in the selected region of interest. (Right) The speckle contrast diffuse correlation tomography and SPY-PHI measurements during mastectomy in the operating room. NIR, near infrared; LEDs, light-emitting diodes.

In speckle contrast diffuse correlation tomography,10–18 a galvo mirror remotely delivered coherent point near-infrared light to multiple source positions in a selected region of interest and an electron-multiplying charge-coupled device camera (Cascade 1K; Photometrics, Tucson, Ariz.) measured spatial diffuse speckle contrasts on the tissue boundary [See Video (online), which shows our speckle contrast diffuse correlation tomography instrument and experimental setup for imaging a breast mannequin.] An advanced photometric stereo technique was integrated into the speckle contrast diffuse correlation tomography method to obtain breast surface geometry for better image reconstruction.13 For this study, the speckle contrast diffuse correlation tomography was upgraded with a new scientific complementary metal-oxide-semiconductor camera (ORCA-Flash4.0; Hamamatsu Photonics, Hamamatsu, Japan) to improve imaging sensitivity and temporal and spatial resolution. The visible light-emitting diodes for photometric stereo technique illumination were replaced with near-infrared light-emitting diodes (850 nm, Luxeon Star light-emitting diodes) to adapt for the long-pass filter (>750 nm) used for speckle contrast diffuse correlation tomography measurements, thereby reducing the operation time.

{“href”:”Single Video Player”,”role”:”media-player-id”,”content-type”:”play-in-place”,”position”:”float”,”orientation”:”portrait”,”label”:”Video.”,”caption”:”This video shows our speckle contrast diffuse correlation tomography instrument and experimental setup for imaging a breast mannequin.”,”object-id”:[{“pub-id-type”:”doi”,”id”:””},{“pub-id-type”:”other”,”content-type”:”media-stream-id”,”id”:”1_q8g059ta”},{“pub-id-type”:”other”,”content-type”:”media-source”,”id”:”Kaltura”}]}

The data acquisition sequence included the following: (1) selecting the region of interest of 80 × 80 mm2, (2) obtaining tissue surface geometry by photometric stereo technique, and (3) scanning point light to 9 × 9 source positions on the region of interest and recording 10 images at each source position. Red blood cell movements in the measured tissue volume (“banana-shape”) (Fig. 1, right) produced spatial laser speckle fluctuations on the tissue surface, which were captured by the scientific complementary metal-oxide-semiconductor camera with an exposure time of 2 msec. The spatial speckle contrast was quantified over a selected window of 7 × 7 pixels by simply calculating the ratio of standard deviation to mean intensity over these 49 pixels. To improve signal-to-noise ratio, speckle contrast data were averaged over 3 × 3 adjacent pixel windows (as a detector) and across 10 images at each source position. In total, 9 × 9 source positions and 41 × 41 detectors were evenly distributed in the region of interest. These boundary data from multiple source-detector pairs were input into a modified finite-element method–based NIRFAST program for three-dimensional image reconstruction of blood flow distributions.20,21

Because reconstructed ischemic areas have irregular shapes, an innovative contour-based algorithm was created to compare three-dimensional images of blood flow distribution and two-dimensional maps of indocyanine green perfusion. (See Document, Supplemental Digital Content 1, which describes our innovative contour-based algorithm for comparing three-dimensional images of blood flow distribution by speckle contrast diffuse correlation tomography and two-dimensional maps of indocyanine green perfusion by SPY-PHI, https://links.lww.com/PRS/F194. See Figure, Supplemental Digital Content 2, which illustrates the methodologic flowchart for the comparison of speckle contrast diffuse correlation tomography and SPY-PHI measurements. Text boxes with different colors represent multiple steps for data processing: red boxes indicate speckle contrast diffuse correlation tomography; purple boxes indicate SPY; and a green box indicates the comparison, https://links.lww.com/PRS/F195.) Briefly, an ischemic area with the lowest blood flow determined by speckle contrast diffuse correlation tomography was registered on the indocyanine green perfusion map. Four cubes with varied volumes centering this ischemic area were selected for comparisons. Eight contours were generated based on eight evenly distributed blood flow levels across the minimal and maximal values inside each individual cube.

Pearson correlations were calculated using MATLAB (MathWorks, Natick, Mass.) to investigate relationships between the speckle contrast diffuse correlation tomography and SPY-PHI measurements. Post hoc power analysis was performed using Python. A value of p < 0.05 was considered significant.

RESULTS

Eleven female patients were studied, and only one patient (patient 8) developed wound dehiscence and implant exposure after breast reconstruction (Fig. 2). [See Table, Supplemental Digital Content 3, which demonstrates patient demographics and surgical procedures (n = 11), https://links.lww.com/PRS/F196. See Figure, Supplemental Digital Content 4, which illustrates the steps and corresponding results to compare the SPY-PHI (above) and speckle contrast diffuse correlation tomography (below) measurement results from two patients (patients 11 and 2), respectively, https://links.lww.com/PRS/F197.] Table 1 summarizes comparison results in eight contours (C1 through C8) from the four selected areas (from 10 × 10 mm2 to 40 × 40 mm2) inside the region of interest of 80 × 80 mm2 over 11 patients. Significant correlations were observed in all eight contours from the selected area of 10 × 10 mm2 (r ≥ 0.78; p < 0.004) and five contours (C1 through C5) from the selected area of 20 × 20 mm2 (r ≥ 0.73; p < 0.02). Considering the range of Pearson r (0.61 to 0.87) with values of p < 0.05 as the correlated group (Table 1), the estimated power range was 0.41 to 0.85 for differentiating the two groups of correlated and noncorrelated measurements.


Table 1. -
Pearson Correlations between the Speckle Contrast Diffuse Correlation Tomography and SPY-PHI Measurements (n = 11)











Contours 10 × 10 mm2 20 × 20 mm2 30 × 30 mm2 40 × 40 mm2

r

p

r

p

r

p

r

p
C1 0.83* < 2 × 10−3* 0.75* <9 × 10−3* 0.42 <2 × 10−1 0.41 <3 × 10−1
C2 0.87* <4 × 10−4* 0.81* <3 × 10−3* 0.55 <8 × 10−2 0.53 <9 × 10−2
C3 0.86* <7 × 10−4* 0.87* <5 × 10−4* 0.54 <9 × 10−2 0.54 <9 × 10−2
C4 0.86* <6 × 10−4* 0.81* <3 × 10−3* 0.61* <5 × 10−2* 0.65* <4 × 10−2*
C5 0.85* <1 × 10−3* 0.73* <2 × 10−2* 0.58 <6 × 10−2 0.63* <4 × 10−2*
C6 0.81* <3 × 10−3* 0.58 <6 × 10−2 0.39 <3 × 10−1 0.44 <2 × 10−1
C7 0.80* <3 × 10−3* 0.51 <2 × 10−1 0.38 <3 × 10−1 0.41 <3 × 10−1
C8 0.78* <4 × 10−3* 0.47 <2 × 10−1 0.40 <3 × 10−1 0.45 <2 × 10−1

*Statistically significant (p < 0.05).

F2
Fig. 2.:

Steps and corresponding results to compare speckle contrast diffuse correlation tomography and SPY-PHI measurements in patient 11. (Above, left) A square area of 20 × 20 mm2 was superimposed on top of the original indocyanine green map obtained by SPY-PHI at the ischemic region with the lowest blood flow value detected by speckle contrast diffuse correlation tomography. The dashed red ellipse shows high-intensity perfusions as artifacts attributable to indocyanine green augmentations. The dashed yellow box represents the selected region of interest of 80 × 80 mm2 for speckle contrast diffuse correlation tomography. (Above, right) The area of 20 × 20 mm2 was segmented into eight regions/contours based on indocyanine green perfusion levels. (Below, left) A three-dimensional view of blood flow distribution reconstructed by speckle contrast diffuse correlation tomography. A cube of 20 × 20 × 20 mm3 was selected at the ischemic area with the lowest blood flow. (Below, right) The cube of 20 × 20 × 20 mm3 was segmented into eight volumes/contours based on blood flow levels. Only four contours (C2, C4, C6, and C8) of eight are shown to facilitate better illustration.

DISCUSSION

This preliminary study on a limited number of subjects tested the hypothesis that speckle contrast diffuse correlation tomography and indocyanine green–fluorescence angiography measurements of blood flow distributions in mastectomy skin flaps are consistent. Significant correlations were observed between the two measurements of blood flow distributions in ischemic areas with the lowest blood flow values. These ischemic areas are the most important regions to determine during surgery because compromised tissues may be excised to minimize mastectomy skin flap necrosis. Fluorescence angiography is a dye-based, expensive device (up to $250,000 for equipment plus hundreds of dollars for dye/supplies) for two-dimensional mapping of indocyanine green perfusion.6 By contrast, our inexpensive speckle contrast diffuse correlation tomography device (approximately $40,000) does not require dye injection (no supply cost) and can be used continuously and frequently for three-dimensional imaging of blood flow distributions at different depths.

We recognize limitations of this preliminary study. The small number of patients impacts the statistical power. Also, we are not comparing our results to other noninvasive imaging techniques that are being developed and might be superior to indocyanine green in terms of costs and accuracy (e.g., laser speckle contrast imaging, near-infrared diffuse optical tomography, thermal imaging).22–25 The region of interest of 80 × 80 mm2 used may not be enough for imaging larger breasts. In the future, we may combine multiple regional images to cover the entire flap area or improve speckle contrast diffuse correlation tomography design to enlarge the region of interest. We are also developing fast reconstruction methods with parallel computations by the graphics processing units for near real-time image reconstruction.26–28 With further optimization and validation in large populations, our innovative speckle contrast diffuse correlation tomography may ultimately be used as an inexpensive imaging tool for perioperative assessments of skin flap viability to prevent mastectomy skin flap necrosis.

REFERENCES

1. Robertson SA, Jeevaratnam JA, Agrawal A, Cutress RI. Mastectomy skin flap necrosis: Challenges and solutions. Breast Cancer (Dove Med Press). 2017;9:141–152.

2. Antony AK, Mehrara BM, McCarthy CM, et al. Salvage of tissue expander in the setting of mastectomy flap necrosis: A 13-year experience using timed excision with continued expansion. Plast Reconstr Surg. 2009;124:356–363.

3. Nykiel M, Sayid Z, Wong R, Lee GK. Management of mastectomy skin flap necrosis in autologous breast reconstruction. Ann Plast Surg. 2014;72(Suppl 1):S31–S34.

4. Munabi NC, Olorunnipa OB, Goltsman D, Rohde CH, Ascherman JA. The ability of intra-operative perfusion mapping with laser-assisted indocyanine green angiography to predict mastectomy flap necrosis in breast reconstruction: A prospective trial. J Plast Reconstr Aesthet Surg. 2014;67:449–455.

5. Phillips BT, Lanier ST, Conkling N, et al. Intraoperative perfusion techniques can accurately predict mastectomy skin flap necrosis in breast reconstruction: Results of a prospective trial. Plast Reconstr Surg. 2012;129:778e–788e.

6. Griffiths M, Chae MP, Rozen WM. Indocyanine green-based fluorescent angiography in breast reconstruction. Gland Surg. 2016;5:133–149.

7. Duggal CS, Madni T, Losken A. An outcome analysis of intraoperative angiography for postmastectomy breast reconstruction. Aesthet Surg J. 2014;34:61–65.

8. Agochukwu NB, Huang C, Zhao M, et al. A novel noncontact diffuse correlation spectroscopy device for assessing blood flow in mastectomy skin flaps: A prospective study in patients undergoing prosthesis-based reconstruction. Plast Reconstr Surg. 2017;140:26–31.

9. Huang C, Radabaugh JP, Aouad RK, et al. Noncontact diffuse optical assessment of blood flow changes in head and neck free tissue transfer flaps. J Biomed Opt. 2015;20:075008.

10. Yu G, Lin Y, Huang C inventors; University of Kentucky Research Foundation, assignee. Noncontact three-dimensional diffuse optical imaging of deep tissue blood flow distribution. US patent 9,861,319 B2. January 9, 2018.

11. Huang C, Irwin D, Lin Y, et al. Speckle contrast diffuse correlation tomography of complex turbid medium flow. Med Phys. 2015;42:4000–4006.

12. Huang C, Irwin D, Zhao M, et al. Noncontact 3-D speckle contrast diffuse correlation tomography of tissue blood flow distribution. IEEE Trans Med Imaging. 2017;36:2068–2076.

13. Mazdeyasna S, Huang C, Zhao M, et al. Noncontact speckle contrast diffuse correlation tomography of blood flow distributions in tissues with arbitrary geometries. J Biomed Optics. 2018;23:096005.

14. Huang C, Mazdeyasna S, Chen L, et al. Noninvasive noncontact speckle contrast diffuse correlation tomography of cerebral blood flow in rats. Neuroimage. 2019;198:160–169.

15. Mazdeyasna S, Huang C, Zhao M, et al. Noninvasive noncontact 3D optical imaging of blood flow distributions in animals and humans. Paper presented at: The 18th IEEE International Symposium on Signal Processing and Information Technology; December 6–8, 2018; Louisville, Ky.

16. Bonaroti A, DeCoster RC, Mazdeyasna S, Huang C, Yu G, Wong L. The role of intraoperative laser speckle imaging in reducing postoperative complications in breast reconstruction. Plast Reconstr Surg. 2019;144:933e–934e.

17. Zhao M, Mazdeyasna S, Huang C, et al. Noncontact speckle contrast diffuse correlation tomography of blood flow distributions in burn wounds: A preliminary study. Mil Med. 2020;185(Suppl 1):82–87.

18. Abu Jawdeh EG, Huang C, Mazdeyasna S, et al. Noncontact optical imaging of brain hemodynamics in preterm infants: A preliminary study. Phys Med Biol. 2020;65:245009.

19. Huang C, Mazdeyasna S, Mohtasebi M, et al. Speckle contrast diffuse correlation tomography of cerebral blood flow in perinatal disease model of neonatal piglets. J Biophotonics. 2021;14:e202000366.

20. Dehghani H, Eames ME, Yalavarthy PK, et al. Near infrared optical tomography using NIRFAST: Algorithm for numerical model and image reconstruction. Commun Numer Methods Eng. 2008;25:711–732.

21. Lin Y, Huang C, Irwin D, He L, Shang Y, Yu G. Three-dimensional flow contrast imaging of deep tissue using noncontact diffuse correlation tomography. Appl Phys Lett. 2014;104:121103.

22. To C, Rees-Lee JE, Gush RJ, et al. Intraoperative tissue perfusion measurement by laser speckle imaging: A potential aid for reducing postoperative complications in free flap breast reconstruction. Plast Reconstr Surg. 2019;143:287e–292e.

23. Whitaker IS, Pratt GF, Rozen WM, et al. Near infrared spectroscopy for monitoring flap viability following breast reconstruction. J Reconstr Microsurg. 2012;28:149–154.

24. Hill WF, Webb C, Monument M, McKinnon G, Hayward V, Temple-Oberle C. Intraoperative near-infrared spectroscopy correlates with skin flap necrosis: A prospective cohort study. Plast Reconstr Surg Glob Open. 2020;8:e2742.

25. Tenorio X, Mahajan AL, Wettstein R, Harder Y, Pawlovski M, Pittet B. Early detection of flap failure using a new thermographic device. J Surg Res. 2009;151:15–21.

26. Wu X, Eggebrecht AT, Ferradal SL, Culver JP, Dehghani H. Quantitative evaluation of atlas-based high-density diffuse optical tomography for imaging of the human visual cortex. Biomed Opt Express. 2014;5:3882–3900.

27. Wu X, Eggebrecht AT, Ferradal SL, Culver JP, Dehghani H. Fast and efficient image reconstruction for high density diffuse optical imaging of the human brain. Biomed Opt Express. 2015;6:4567–4584.

28. Doulgerakis M, Eggebrecht A, Wojtkiewicz S, Culver J, Dehghani H. Toward real-time diffuse optical tomography: Accelerating light propagation modeling employing parallel computing on GPU and CPU. J Biomed Opt. 2017;22:1–11.



Source link

Back to top button