Keywords
necrosis - perfusion - prevention - flap - technology
In 2016, nearly 110,000 women in the United States alone underwent breast reconstruction
by members of the American Society of Plastic Surgeons (ASPS), of which roughly 90,000
elected for tissue expander-implant or implant only techniques.[1] These rates continue to rise as breast reconstruction establishes itself as one
of the main pillars in the treatment of breast cancer patients.[2] This is in part because the benefits of reconstruction are manifold, restoring a
patient's form and sense of self in addition to providing psychosocial wellbeing and
aesthetic benefits.[2]
[3]
[4]
[5]
Nonetheless, mastectomy interrupts the axial blood supply to the skin overlying the
breast, creating a relatively hypovascular milieu and skin flaps prone to wound breakdown
and tissue necrosis in 3 to 33% of cases.[6]
[7]
[8]
[9] Particularly in prosthetic breast reconstruction, the viability of these mastectomy
skin flaps is of utmost importance to prevent implant exposure leading to explanation,
delays in adjuvant chemoradiation, and possible reconstructive failure.[10]
[11] To the plastic surgeon, mastectomy flap necrosis should represent a “never event;”
however, we lack the tools to make this the reality.
As such, predicting the risk for mastectomy flap necrosis intraoperatively has been
the subject of much research lately.[12]
[13]
[14] Techniques using fluorescein[14] or indocyanine green[13] angiography have yielded mixed results.[13]
[15]
[16] Translating intraoperative measurements of ischemia, whether quantitative or visual,
to full-thickness necrosis has proven challenging, particularly in the context of
the hemodynamic changes that occur during and after surgery. Downsides including cost,[13] device bulkiness, risk of allergic reaction,[17] and challenges interpreting output have prevented their widespread incorporation
into clinical practice, the need for a more simple and reliable method persists.
To this end, our manuscript endeavors to review the evidence on mastectomy skin flap
perfusion analysis, highlighting the benefits and downsides of the current techniques
and identifying areas of future research and development.
Breast Blood Supply
The unoperated breast has a rich blood supply from multiple arterial sources that
give rise to the perforators that perfuse the parenchyma and feed into the subdermal
plexus. Redundant perfusion via the internal mammary, lateral thoracic, intercostal,
and thoracoacromial artery perforators ([Fig. 1]) allows for the safe manipulation of the breast tissues based on any of the above
vascular pedicles. The internal mammary, or internal thoracic, perforators enter the
superomedial aspect of the breast and are estimated to provide approximately 60% of
its vascular supply.[18] Superolaterally the lateral thoracic vessels predominate and superiorly the pectoral
branch of the thoracoacromial artery. Finally, the intercostal artery provides lateral
intercostal perforators as well as anterior intercostal perforators which supply the
lateral and central/inferior aspects of the breast, respectively.
Fig. 1 Native blood supply of the breast. Artist rendering of the primary, axial blood supply
of the breast via the internal mammary, thoracoacromial, lateral thoracic, and internal
artery perforators.
The venous drainage of the breast is via two interconnecting systems. The deep system
parallels the arterial supply with venous perforators paired to their respective arteries.
The superficial system, however, is comprised of a rich subdermal venous plexus that
lies in communication with both the deep system of the breast as well as the surrounding
superficial system of the body.[19]
[20]
Mastectomy interrupts the axial perfusion of the overlying breast skin, leaving the
relatively hypovascular skin flaps dependent on random pattern perfusion and drainage
through the subdermal plexus. As such thicker mastectomy flaps with an appropriate
amount of subcutaneous fat preserve more of the tissue's blood supply and may even
reduce the risk of flap necrosis,[21] although this must be balanced with an increased risk of leaving behind breast tissue.
Both the distal flap edges and beneath the nipple-areola complex, where additional
margins are typically resected, are placed in relatively high risk for necrosis.
Risk of Flap Necrosis
The true risk of mastectomy skin flap necrosis following immediate prosthetic reconstruction
is difficult to pinpoint and a broad range of estimates have been reported within
the literature.[6]
[7]
[8]
[9] This is in part due to the variations in the definition of flap necrosis across
studies. Full-thickness necrosis is rare and occurs on the order of 3 to 9% across
studies; partial thickness necrosis rates have been reported has high as 30 to 40%.[6]
[7]
[22]
[23]
[24]
[25] Scales, including the SKIN score more precisely delineate the depth of an area of
tissue injury,[26] but no one system has been broadly adopted by the research community.
Moreover, estimates of average, population-level risk are not always relevant to the
individual patient,[27]
[28] and in the case of flap necrosis, factors such as smoking status, mastectomy technique,
flap thickness, and flap length play a considerable role in determining risk. Because
of this heterogeneity, meta-analysis of flap necrosis rates proved to be challenging
and of dubious value; to our knowledge, no such attempt has been published within
the literature.
Clinical Examination
Perhaps the oldest and most widely used method to assess mastectomy skin flaps, several
studies have demonstrated the unpredictability of clinical exam in accurately and
precisely predicting flap necrosis. Surgeons can evaluate for flap length or thinness,
uneven flaps, ecchymosis, capillary refill, color, temperature, or bleeding from skin
edges, and trim back to healthy tissue as needed. However, when decisions are based
on clinical exam alone, flap necrosis rates remain elevated between 4% to 40%.[6]
[7]
[22]
[23]
[24] Even with a “low-threshold” to excise questionable skin, necrosis rates can be as
high as 27%,[14] with multiple large series reporting specificities for clinical exam in the 10 to
30% range.[29]
[30]
[31] Not only has clinical exam been demonstrated to be inferior to more objective methods
of perfusion assessment,[15]
[29]
[32] but the degree/thickness of the associated necrosis tends to be more severe.[33] In general, a concerning exam represents a “tip of the iceberg,” as only the most
foregone of tissues will declare themselves on the table to the naked eye.
Several clinical variables, beyond the physical exam and patient comorbidities, have
also been correlated with the incidence of flap necrosis. The volume and weight of
the tissue expander applies pressure on the already tenuous flaps and can exacerbate
tissue ischemia and necrosis. The weight of the mastectomy specimen is also a useful
variable, both in and of itself, but also as a correlate for flap length and expander
volume/weight.[34] It is also important to note that clinical flap assessment is particularly challenging
in patients with darker skin as the ability to perceive color changes and capillary
refill may be obscured by high melanin concentrations.[35]
Angiography
Fluorescein
The use of fluorescein to evaluate mastectomy skin flap viability was first described
in the late 1970s by Singer and colleagues.[36] Their cohort demonstrated its merit in identifying healthy and well perfused flaps
but acknowledged its limitation in precisely delineating the extent and thickness
of necrosis in concerning ones. After the completion of mastectomy, a weight based
(10–15 mg/kg)[23]
[36] or fixed (approximately 500–1,000 mg)[14]
[37] dose of fluorescein is administered intravenously and allowed 15 minutes to circulate
throughout the flaps. A Wood's lamp or a more modern “black light,” that emits 320
to 500 nm wavelength ultraviolet light, is then turned on and used to visibly assess
the diffusion of the fluorescein throughout the tissues. Well perfused flaps will
glow yellow, whereas poorly perfused regions appear blue, and intermediate areas as
yellow with blue mottling. The dye is generally well tolerated and excreted within
the urine in 12 to 18 hours; side effects are rare and most commonly include nausea
and vomiting. Allergic reactions, typically comprised of urticaria and fever, have
been reported in 0.6% of patients, and more sever adverse including vasovagal reactions
or cardiopulmonary events are rare.[37]
[38]
[39]
Several studies[14]
[23]
[37] have built upon Singer's original data demonstrating a very high positive predictive
value for areas that fluoresce of well over 90%, but blue areas continue to underpredict
flap survival with false negative rates as high as 75%.[37] Therefore, while using fluorescein to guide skin resection certainly decreases flap
necrosis (to as low as 3% in one study[14]), as much as 80% of the resected area may actually go on to survive.[23] Surgeons are again left with clinical judgement to discern between true and false
negatives, often weighing the size and location of hypoperfusion, flap thickness,
and signs of retraction or electrocautery injury, as well as the patient's radiation
history.[37] The presence of vasoconstrictive agents, such as epinephrine from a tumescent mastectomy,[40] may also limit dye uptake and study results. The primary limitation of this technique,
and many of others described in the literature, is that they afford only a single
snapshot of perfusion even though blood flow is constantly changing in the perioperative
period due to a patient's hemodynamic state, cholinergic tone, and the insult of a
surgical procedure. Porcine data have demonstrated an increase in flap fluorescence
area of roughly 20% within 24 hours after flap elevation.[41]
No studies to date have evaluated cost, patient reported outcomes, or aesthetic results
in patients managed using intraoperative fluorescein angiography. Future prospective
randomized data will be necessary to compare algorithms using fluorescein and clinical
judgement against other techniques for flap evaluation.
Indocyanine Green
Indocyanine green (ICG) is an intravenously administered, biliary-excreted dye that
circulates throughout tissues and emits infrared light upon excitation. Compared with
fluorescein, the dye has a faster uptake into the tissues requiring only about 2 minutes
before evaluation. Molecularly, ICG is larger than fluorescein (775 vs. 375 kD) and
more likely to bind to plasma proteins, making it more likely to stay within the intravascular
compartment and resulting in a faster washout (2.5 minutes half-life vs. 23 minutes).[42]
[43] These properties not only decrease operative/evaluation time but also result in
a superior side effect profile[44]
[45] and allow for multiple injections and image captures during the same procedure.
Although rare, allergic responses, including anaphylaxis, have been reported following
ICG administration.[46] A typical dose in adults is 25 mg of ICG, with a recommended maximum dose of 2 mg/kg.[47] Other cited benefits of ICG angiography compared with fluorescein include a more
gentle learning curve that is not dependent on the operators handling of a Wood's
lamp as well as the potential for providing a quantified output ([Table 1]).[48]
Table 1
Pros and cons of different techniques
|
Indocyanine green angiography
|
Fluorescein dye angiography
|
Near-infrared spectroscopy
|
|
Pros
|
• Quantitative analysis
• Large body of literature
• Faster dye washout than fluorescein
• Visualize perfusion throughout entire flap
• More tolerable side effect profile than fluorescein
|
• Relatively cheap
• Handheld device
• Available for over 40 years
|
• Handheld device
• Relatively Cheap
• Rapid serial measurements
• Dye-free system
• Quantitative measurement
• outcome (oxygenation %)
• Easily used outside of OR
|
|
Cons
|
• Intravenous dye-based
• Serial evaluation limited by dye
• Risk of allergic reaction
• Confounded by epinephrine
• Requires specialized/bulky machinery
• Questionable cost-efficacy (approximately $650/use)
|
• User dependent results
• Intravenous dye-based
• Serial evaluation limited by dye
• Risk of allergic reaction
• Confounded by epinephrine
• Requires Wood's lamp
• Relative paucity of outcomes data
|
• Lack of clinical outcomes data
• Highly localized measurements, unable to assess entire flap
• Potentially confounded by other chromophores (melanin)
|
|
Predictive accuracy[
a
]
|
APU: 20,[54] Sen: 100, Spec 28,
APU: 15,[54] Sen: 100, Spec 51,
APU: 10,[54] Sen: 100, Spec 72,
APU: 7,[12] Sen: 88, Spec 83,
APU: 6,[12] Sen: 75, Spec 83,
APU: 3,[12] Sen: 38, Spec 83
|
Losken et al,[37] Sen: 75, Spec: 71
Phillips et al,[23] Sen: 90, Spec: 30
|
N/A
|
Abbreviations: APU, absolute perfusion units; N/A, not available; OR, operating room;
Sen, sensitivity; Spec, specificity.
a Indocyanine green data based on data using the SPY Elite system.
Several commercially available devices exist to evaluate ICG derived fluorescence
in human tissues, including: the SPY Elite System (Novadaq, San Jose, CA; [Fig. 2]), FLARE (Curadel LLC Marlborough, MA), PDENeo (Hamamatsu Photonics Hamamatsu City,
Shizuoka, Japan), and Fluobeam 800 (Fluoptics, Cambridge, MA).[48] These devices include a sensor and a source that emits infrared light in the 760
to 785 nm wavelength range that excites the ICG molecule and triggers fluorescence
in the 820 to 840 nm range.[49] The SPY and FLARE systems further provide software that is able to covert fluorescence
into a quantitative measurement. Although there are clinical or preclinical data for
each of these systems in mastectomy flap perfusion analysis; the SPY system is by
far the most extensively reported but also the most expensive with a device cost of
$250,000 and per use cost of $650.[48]
Fig. 2 Left, SPY Elite system for indocyanine green angiography. Right, ViOptix Intra.Ox
handheld tissue oximeter.
First described in humans in 1999,[50] this technology was first applied to mastectomy flap evaluation in the early 2000s
with De Lorenzi et al's proof of concept demonstrating a difference in perfusion patterns
between native nipples and those elevated with the mastectomy skin flap.[51] Newman et al retrospectively reviewed ICG fluorescence data for 20 patients undergoing
immediate tissue expander based breast reconstruction, including nine with flap necrosis,
and in hindsight demonstrated a 95% correlation between intraoperative imaging and
complications with a 100% sensitivity and 91% specificity.[52] As surgeons began incorporating this technology into their intraoperative decision
making, analyses of pre- and post-ICG implementation cohorts demonstrated a decrease
in not only mastectomy skin flap necrosis but also unplanned reoperation rates by
more than 50%.[15]
[53] Prospective comparisons to clinical exam have similarly demonstrated improved outcomes
with intraoperative laser assisted ICG angiography.[14]
[23]
Not unlike fluorescein, early adopters of ICG faced the dilemma of overly conservative
and aggressive resections threatening potentially viable tissues. Where ICG-based
systems differed, however, was in their ability to quantify perfusion. The SPY system
uses its proprietary software, the SPY-Q (Novadaq, San Jose, CA), to calculate and
report both absolute perfusion as an integer or relative perfusion as a percentage
of a selected reference area within the field. These analyses can be done either in
real time intraoperatively or post hoc. Mattison et al[54] used different thresholds to demonstrate that a higher cut-off value does not necessarily
improve sensitivity and may even come at an increased cost in specificity, positive
predictive value, and the unnecessary resection of as much as 10 cm2 or more of skin. As they search for an accurate cut-off, some authors have advocated
for the use of relative fluorescence in assessing perfusion,[55]
[56] but these measurements are more prone to interuser variability than absolute cut-offs.[12]
[55]
[57] Interestingly, these measurements appear to be confounded by smoking status, use
of epinephrine (tumescent mastectomy), and even skin pigmentation or a history of
hypertension,[55] but the relationships remain incompletely described. Sanniec et al is the only group
to date who described their algorithm for interpreting and applying these quantitative
data in clinical practice; however, no prospective randomized data exist to support
one protocol or perfusion cut-off over another.[58]
Overall, the use of ICG or fluorescein angiography does not appear to decrease mastectomy
skin loss. If anything, their use may come at the increased cost of debriding healthy
skin to ensure that all devitalized tissues are removed and to prevent the downstream
cost and morbidity associated with flap necrosis. One study by Rinker[14] compared the area of resection with ICG, fluorescein and clinical exam ([Table 2]); however, these numbers are difficult to interpret due to the discrepancies in
flap necrosis rates across the cohorts. Other studies comparing intraoperative predictions
to clinical flap necrosis within the same patients demonstrate as much as 3 to 6 cm2 of excess skin loss with ICG or 19 cm2 with fluorescein that is corroborated by the modest specificities for these techniques
([Table 2]). In the correct patient, this excess cost in tissue may be justified by the improvement
in patient outcomes and preventing delays in adjuvant chemoradiation.
Table 2
Comparisons of prospective results for skin resection area and specificity
|
Rinker 2016[14]
|
Phillips et al 2012[23]
|
Mattison et al 2016[54]
|
|
Fluorescein
|
6.2 cm2 skin resection 3% flap necrosis
|
Overpredicted necrosis by 18.86 cm2; Sens: 90%, Spec: 30%
|
|
|
ICG
|
5.4 cm2 skin resection 14% flap necrosis
|
Overpredicted necrosis by 6.57 cm2; Sens: 90%, Spec: 50%
|
Predicts 3.23 cm2 or more necrosis than exam specificity of 68.1%
|
|
Clinical exam
|
6.5 cm2 skin resection 27% flap necrosis
|
|
|
Abbreviation: ICG, indocyanine green.
Overall, ICG angiography rates in immediate breast reconstruction have been on the
rise in recent years.[59] Despite the high upfront cost of an ICG detector as well as cost per use associated
with the dye and other operative room equipment, Duggal et al also demonstrates a
$614 cost saving per patient in reducing flap necrosis rates from 44 to 25% and reoperation
rates from 14% to 6%.[15] Other studies have not replicated these findings, however, particularly in the setting
of a more modest reduction in adverse events.[13]
[59] Kanuri et al[13] performed a retrospective subgroup analysis identifying smoking, BMI > 30 kg/m2, and mastectomy weight greater than 800 g as high-risk features that result in a
net cost-savings. Further research will be necessary to better understand the cost
and benefits associated with ICG angiography and to identify the subgroup of patients
in whom this technology is most cost effective.
Spectroscopy
Light-based spectroscopy generally takes advantage of the hemoglobin molecule's ability
to absorb or reflect light at various wavelengths to provide an estimate of either
perfusion or tissue oxygenation.[60]
[61] These techniques benefit from being noninvasive in nature and have multiple applications
in different tissues and clinical scenarios.
Although limited in its application to mastectomy skin flaps, Laser Doppler flowmetry
is one such hand-held technique that has more commonly been applied to breast free
flaps.[62]
[63] In this technique, moving blood cells within the dermal plexus cause a back scattering
of photons that provides information on their speed and concentration as surrogates
for perfusion. Laser speckle contrast imaging is a similar camera-based technique
that is able to scan a large area of skin (typically 15 cm × 15 cm) and demonstrate
the reflection of light by moving red blood cells in the dermal plexus, typically
as a speckle pattern on the camera sensor giving the technique its name.[60] Polarized light has also been described to visually assess circulation in tissues
of interest. After polarized light in the hemoglobin spectrum (548 nm wavelength)
is directed toward the target, a majority of it is either reflected off its surface
in an orthogonal plane or scattered by hemoglobin molecules. Once the orthogonally
reflected light is filtered by a polarizer, the remaining, transmitted light is that
which has been scattered by hemoglobin; this is picked up by a charge coupled device
camera to provide real-time images of the microcirculation at a similar contrast level
to images obtained using fluorescent dyes.[64] Although the latter two methods provide qualitative data on perfusion, several algorithms
have been developed to provide estimates of red blood cell concentration or other
values based on pixel values from the images.[61]
[65]
Although several other similar techniques have been described, only one has been directly
applied to mastectomy skin flaps to date and will be discussed in further detail.
Near-Infrared Oximetry
Near-infrared spectroscopy samples of hemoglobin in specific areas of tissue adjacent
to a sensor, providing highly localized measurements of tissue oxygenation. It is
based on the principle that oxygenated and deoxygenated hemoglobin are chromophores
that interact with light differently. The device directs near-infrared light toward
the tissue of interest, and its intensity, as it returns to the sensor, allows for
the calculation of a relative percentage of the two molecules.[66] The ViOptix T.Ox (ViOptix, Inc., Fremont, CA) has already been demonstrated to be
specific and sensitive in the detection of postoperative vascular compromise in free
tissue transfer,[67]
[68] identifying impending complications earlier than alternatives modalities.[69] Only one study has used the T.Ox intraoperatively on mastectomy skin flaps, demonstrating
a significantly greater reduction in oxygenation in tissue that becomes necrotic compared
with healthy tissues.[70]
Several factors unrelated to tissue oxygen saturation can affect near-infrared oximetry
measurements, including instrumentation of the device, ambient light, melanin concentration,
and the presence of additional chromophores including methylene blue. Additionally,
single sensor detectors are unable to effectively distinguish the effects of absorption
versus tissue scattering on light intensity at the sensor. To this end, a more recently
FDA (Food and Drug Administration) approved device, the Intra.Ox (ViOptix, Inc., Fremont,
CA; [Fig. 2]), is designed with multiple sensors specifically to overcome these limitations for
intraoperative use. The device is wireless, battery operated, and handheld allowing
for multiple, rapid measurements throughout the procedure without the need for dye
or bulky equipment ([Table 1]). The device also provides information on the quality its reading to improve interuser
reliability. Lohman et al identified 30% oxygenation predicted tissue necrosis in
porcine fasciocutaneous flaps[66]; however, this newest generation of technology has yet to be applied in human mastectomy
skin flaps.
Future Directions
As these various technologies vie to define their role in the assessment of mastectomy
skin flap perfusion, future studies will have to demonstrate each one's potential
value to patient care, specifically how they can each improve patient outcomes and
curb medical costs. Within the issue of cost is entangled the question of the target
population. For more expensive technologies with a high per use cost (e.g., ICG angiography)
indiscriminate use may not prove feasible, and a more targeted, high-risk population
will need to be defined. But does this apply to reusable, dye-free techniques like
near-infrared spectroscopy? Is one technology clearly superior in all populations,
or should the technique for flap assessment be selected on a case-by-case basis? Both
theoretical and retrospective reviews of cost and benefit will allow us to better
understand each technology's strengths and limitations and to carve out its niche
in clinical practice.
Studies evaluating these technologies in prepectoral breast reconstruction will also
be useful given the recent resurgence of this technique. Many centers have begun to
incorporate indocyanine green angiography in their flap assessment and pocket selection
algorithm; however, the data are not yet available within the literature. In the context
of questionable mastectomy flap viability, the additional layer of vascularized muscle
may protect an expander from exposure and its downstream sequalae. In these select
cases, pocket exchange at the time of expander implant exchange may prove to be a
safer alternative.
The best time for flap assessment also remains an unanswered question. Perfusion measurements
will undoubtedly vary immediately following mastectomy to after device insertion and
even after skin closure with varying degrees of tension. The patient's hemodynamic
state is also likely to fluctuate throughout the perioperative period. Ideally, evaluation
occurs intraoperatively when surgeons have the most control over changing aspects
of the case to avoid a complication. Most studies to date have evaluated skin flaps
prior to reconstruction, and it remains to be seen what changes should be expected
after that. In general, additional stress on the flaps should be avoided after assessment
to not confound the results.
With regards to outcomes, we currently lack prospective data to validate most of the
regiments described within the literature. Different authors have put forth varying
criteria for the resection of skin based on qualitative and/or quantitative assessment
of its perfusion; which one appropriately identifies at risk tissue with the least
amount of collateral damage? How do patient factors, including skin pigmentation or
blood pressure weight into this? Also implicit to all of this is the issue of interuser
variability for each technique. With several therapeutic options available beyond
simply resecting at risk skin, including application of vasodilatory substances, such
as nitroglycerine ointment, switching from single-stage to two-stage reconstruction,
pocket selection, deflation of tissue expanders, or even delaying reconstruction altogether,
we must better understand how to interpret these new objective data and navigate the
intraoperative decision-making process.
Conclusion
Although objective assessment of tissue perfusion seems to have already delivered
on its promise to prevent complications, in the era of value-based care, further research
is necessary to do so reliably and efficiently on an even larger scale.
