Annals of Surgical Oncology

, Volume 21, Issue 1, pp 100–106

Intraoperative Imaging of Nipple Perfusion Patterns and Ischemic Complications in Nipple-Sparing Mastectomies


    • Department of SurgeryStanford University School of Medicine
  • Monica Dua
    • Department of SurgeryStanford University School of Medicine
  • Anne Kieryn
    • Department of SurgeryStanford University School of Medicine
  • John Paro
    • Department of SurgeryStanford University School of Medicine
  • Douglas Morrison
    • Department of SurgeryStanford University School of Medicine
  • David Kahn
    • Department of SurgeryStanford University School of Medicine
  • Shannon Meyer
    • Department of SurgeryStanford University School of Medicine
  • Geoffrey Gurtner
    • Department of SurgeryStanford University School of Medicine
Breast Oncology

DOI: 10.1245/s10434-013-3214-0

Cite this article as:
Wapnir, I., Dua, M., Kieryn, A. et al. Ann Surg Oncol (2014) 21: 100. doi:10.1245/s10434-013-3214-0



Nipple-sparing mastectomies (NSM) have gained acceptance in the field of breast oncology. Ischemic complications involving the nipple–areolar complex (NAC) occur in 3–37 % of cases. Skin perfusion can be monitored intraoperatively using indocyanine green (IC-GREEN™, ICG) and a specialized infrared camera–computer system (SPY Elite™). The blood flow pattern to the breast skin and the NAC were evaluated and a classification scheme was developed.


Preincision baseline and postmastectomy skin perfusion studies were performed intraoperatively using 3 mL of ICG. The pattern of arterial blood inflow was classified according to whether perfusion appeared to originate predominantly from the underlying breast tissue (V1), the surrounding skin (V2), or a combination of V1 and V2 (V3). Ischemia, resection, or delayed complications of NAC were recorded.


Thirty-nine breasts were interrogated. Seven (18 %) demonstrated a V1 pattern, 18 (46 %) a V2 pattern, and 14 (36 %) a V3 pattern. Seven (18 %) NACs were removed; six intraoperatively and the seventh in a delayed fashion. Notably, five of the seven resected NACs had a V1 pattern. Overall, 71 % of all V1 cases demonstrated profound ischemic changes by intraoperative clinical judgment and SPY imaging. The rates of resection of the NAC differed significantly between perfusion patterns (Fisher’s exact test, p = 0.0003).


Three perfusion patterns for the NAC are defined. The V1 pattern had the highest rate of NAC ischemia in NSM. Imaging NAC and skin perfusion during NSMs is a useful adjunctive tool with potential to direct placement of mastectomy incisions and minimize ischemic complications.

During the past decade, nipple-sparing mastectomies (NSM) have gained acceptance in the field of breast surgical oncology and are offered today to a continuously growing number of breast cancer patients.15 Its use is being extended to women who present with multicentric lesions, tumors in closer proximity to the nipple–areolar complex (NAC), or following neoadjuvant chemotherapy.6,7 Inadequate skin perfusion to the NAC is the Achilles heel of this operation, as illustrated in Fig. A (supplementary material) of a woman who experienced significant ischemia after a nipple-sparing mastectomy via an inframammary fold (IMF) incision.

Moyer et al.8 reported partial nipple necrosis in 15 patients (37.5 %) from a relatively young population with a mean age of 42 years. Moreover, they found that perfusion deficits were greatest for incisions that combined circumareolar and radial approaches and lowest when a vertical infra-areolar incision was used, similar to the experience of Boneti et al.9 Partial or complete necrosis of the NAC has been reported to be lower in other series, ranging from 3.5 to 13 %.6,10

Acute or subacute ischemia can lead to wound infection, wound dehiscence, or implant exposure. Other complications, such as superficial eschar formation or epidermolysis may produce dilemmas in clinical judgment during the early postoperative period. Although these borderline ischemic events are largely transient, they can portend long-term sequelae on the appearance of the NAC resulting in shrinkage, discoloration of areola or effacement of the nipple.

Tissue perfusion can be monitored intraoperatively via a novel optical imaging device that relies on indocyanine green (IC-GREEN™, ICG) and a specialized infrared camera–computer system (SPY Elite™). ICG is a water soluble tricarbocyanine dye with a peak spectral absorption at 800–810 nm when dissolved in blood. It rapidly binds to plasma proteins and is excreted into bile unchanged. IC-GREEN™ received approval in the United States to measure cardiac output and liver function in 1959 and later for ophthalmic angiography in 1975.11 Increasingly, it is used to monitor myocutaneous flap perfusion and assess mastectomy flap ischemia.12 Intraoperative imaging can help surgeons identify devascularized tissue over visual inspection alone. Improved assessment of skin perfusion should lower postoperative complications and potentially enhance cosmetic outcomes.

Our surgical team began using the SPY Elite™ imaging system to assess preincision baseline and intraoperative skin perfusion in women undergoing mastectomies with immediate reconstruction. Early on, we observed by visual inspection, ischemic changes in the NAC after removal of the breast in two patients undergoing nipple-sparing mastectomies. This correlated with absent circulation on the SPY Elite™ imaging, prompting us to go back and reexamine the baseline studies. Blood flow to the NAC in both cases was derived predominantly from the underlying breast tissue with little contribution from the surrounding skin. This experience led us to systematically obtain baseline imaging studies in order to visualize the vascular anatomy of the skin flaps. We retrospectively reviewed those studies, analyzing the cine recordings filed in the SPY Elite™ computer system. The purpose of this report is to describe patterns of perfusion to the NAC and correlate these novel observations to ischemic findings on mastectomy flaps.


NSM and immediate reconstruction was offered to patients whose tumors did not involve the skin or NAC, those who had achieved good response to neoadjuvant chemotherapy, and to women with localized in-breast recurrences. Ablative operations were performed by one surgical oncologist (IW), and all but six breasts were reconstructed by a single plastic surgeon (GG). Patients underwent tissue-expander reconstruction except for one who received a DIEP myocutaneous flap. Radial, periareolar (partial, 180° or less) with radial extension, and inframammary fold (IMF) incisions were used. The placement of incisions was predicated on the location of the tumor, need for overlying skin excision, and in some cases, consideration of blood vessel location. Our surgical technique involves circumferential dissection along the subcutaneous plane. Dissections along the subareolar and subnipple plane were performed sharply, whereas the remainder of the flap was dissected with PEAK PlasmaBlade (Medtronic Inc., Minneapolis, MN) or standard Bovie electrocautery. The surgical margins were to the clavicle, sternal border, IMF, and midaxillary line. Intercostal perforators were not spared in the course of dissection and nipple ducts were routinely cored out and submitted separately for histopathological examination.

Skin perfusion was evaluated using the SPY Elite™ imaging system (Novadaq Technologies, Inc., Ontario, Canada) after induction of general anesthesia and before mastectomy incision. The procedure involves intravenous administration of 3 mL of IC-GREEN™ (Pulsion Medical Systems SE, Germany; 2.5 gm/mL) followed by a 10-mL saline flush. Video recording begins immediately after injection and continues over 180 s (seconds). Blood flow in the skin is represented as fluorescence or whitish color on the infrared camera screen versus dark gray or black when there is no filling. Intraoperative resection of ischemic skin or NAC was based on clinical appearance of the skin and complete absence of fluorescence in an area of the mastectomy flap. NAC perfusion was classified retrospectively by three observers (IW, AK, SM) into one of three circulatory patterns based on whether the arterial-arteriolar filling originated predominantly from the underlying breast tissue (V1), the surrounding skin (V2), or a combination of breast and surrounding skin (V3). Postmastectomy skin perfusion was compared to baseline assessments using the SPY Elite™ software to analyze the relative skin fluorescence in 4 or 5 regions of interests (ROIs). The ROIs were set at the nipple, right and left areolar edge, along the sternal border, and one additional selected location. ROIs did not overlie a vessel trajectory. Statistical analysis comparing perfusion patterns to NAC outcomes was completed with Fisher’s exact test.


Cine data was evaluable in a total of 39 breasts from 24 patients. The average age was 46.9 (range 33–69) years. Fifteen of these patients had bilateral procedures. Thirteen breasts were removed prophylactically and another five as contralateral procedures to a known cancer in carriers of BRCA 1/2 mutations. Four patients received neoadjuvant chemotherapy before NSM; two had prior breast irradiation for breast cancer and a third who underwent bilateral mastectomies for bilateral breast cancer received childhood chest wall irradiation.

On baseline perfusion studies arterial-arteriolar filling was observed within approximately 20–25 s after ICG injection (Fig. 1a). An intermediary phase of diffuse skin fluorescence follows before there is filling of veins at approximately 30–45 s (Fig. 1b). After resection of underlying breast tissue, imaging sequences typically show lower filling and decreased overall fluorescence in the skin flaps.
Fig. 1

Physiologic blood flow to nipple–areolar complex and nipple perfusion patterns. Blood supply to the nipple–areolar complex was judged on the distribution of fluorescence captured in the initial frames during the baseline studies. a Normal arterial blood flow phase demonstrating V2 pattern, inflow from surrounding skin into nipple–areolar complex; b venous phase; cV1 inflow predominantly from underlying breast tissue; dV3 inflow mixed pattern, combination of V1 and V2

Seven of 39 (18 %) breasts demonstrated a V1 perfusion pattern in the NAC, 18 (46 %) a V2 pattern, and 14 (36 %) a V3 pattern (Table 1). Figure 1a, c, d depict the three classification schemes. Figure B (supplementary video) demonstrates a V2 arterial inflow pattern followed by the venous phase at approximately 40 s. One-third of bilateral cases had discordant patterns between breasts, all involving V2 and V3 classifications. Overall, radial-lateral and IMF incisions predominated with 17 and 13 cases respectively (Table 1). None of the seven patients with V1 perfusion were operated on through an IMF incision. Of the 32 breasts with V2 and V3 classifications, IMF incisions were used in 13 cases, radial incision in 13 cases, and periareolar with radial extension in 6 cases.
Table 1

Perfusion pattern and outcome

Perfusion pattern baseline

Intraoperative post-Mx SPY imaging of NAC


Breast RT

Intraoperative resection of nipple

Postoperative < 45 days

Outcome deficits

N (%)



Periareolar ± radial extension




N (%)

Epidermolysis/partial skin flap necrosis

Necrosis NAC/delayed removal


V1 7 (18)







5 (71)




V2 18 (46)







1 (11)



3 hypopigmentation

V3 14 (36)










4 hypopigmentation

Mx mastectomy, IMF inframammary fold, NAC nipple–areolar complex

aPatient with bilateral breast cancer

Severe ischemia or no blood flow was observed intraoperatively in the NAC of 7 (18 %) cases and decreased perfusion in another 14 (36 %) cases (Table 1). Clinical judgment aided by SPY Elite™ imaging led to the removal of 6 NACs. Among these were two patients who had received standard doses of whole breast irradiation for a primary breast cancer in the past. In addition, a seventh case developed progressive ischemia and had the NAC removed in conjunction with partial resection of the mastectomy flap 15 days later. Transient ischemic postoperative changes were observed in only 10 (26 %) of the 14 cases identified above; manifested as either epidermolysis, dusky appearance of the NAC or skin eschar. One patient (depicted in Fig. 2) underwent delayed resection of flap necrosis 3 weeks postoperatively along with debridement of the tip of the nipple. These areas correlated with a marked decrease in fluorescence observed on SPY Elite™ images after breast resection. Nearly all remaining patients regained normal appearance of the NAC over the first 2 months, except for subtle focal areolar hypopigmentation in seven cases.
Fig. 2

V3 perfusion pattern with ischemic area. Premenopausal patient who failed attempt at breast conserving surgery. Mastectomy was performed by extending laterally upper-central quadrant lumpectomy incision. (a) Relative skin fluorescence was measured sequentially at points indicated in picture: B1 nipple, B2 left areolar edge, B3 right areolar edge, B4 parasternal skin. The corresponding areas in post-mastectomy imaging are referred to as M1-4. Decreased blood flow is noted in the NAC (M1) and ischemic area zone (aqua line) resulted in eschar formation, infection and loss of tissue expander. Again, a slight increase in fluorescence is registered in parasternal ROI after mastectomy

We found significant evidence that the rate of NAC loss differed by perfusion classification (p = 0.0003). Although the V1 perfusion pattern represented only 18 % of our total; 71 % of these cases had the NAC removed intraoperatively for ischemia based on the absence of fluorescence/decreased blood flow and clinical judgment. Two remaining V1 cases had reversible ischemic changes that resolved postoperatively; one had epidermolysis involving the tip of the nipple and the other epidermolysis involving the lower half of the NAC. The latter was the only patient reconstructed via a myocutaneous flap and successfully regained normal appearance of the NAC during a 4-week period. In contrast, when perfusion to the NAC originated predominantly from the surrounding skin, the rate of NAC resection was zero for patients with a V3 pattern and 11 % for patients with a V2 pattern.

In Figs. 2, 3 and 4, preincision and postmastectomy graphs depicting skin fluorescence for are shown. The values of relative fluorescence remained most stable in the parasternal ROI (B4 → M4) areas. In each respective patient, an approximate four-fold drop in fluorescence is observed in pre- and postmastectomy recordings overlying the tip of the nipple or areolar marker areas. However, in spite of this dramatic decrease in blood flow, the NAC retained its viability and was preserved in most V2/V3 cases.
Fig. 3

V2 perfusion pattern. Nipple-sparing mastectomy performed via a radial incision on a patient who underwent neoadjuvant chemotherapy. Post-mastectomy imaging (a) indicated a compensatory increase in blood flow in right periareolar ROI compared to baseline study. The parasternal skin suffered the least amount of change in perfusion after mastectomy while perfusion to tip of nipple exhibited the most dramatic decrease in blood flow. In spite of these changes, 3 week postoperative picture (b) shows slight superficial epidermolysis and overall good appearance of skin flaps
Fig. 4

V1 perfusion pattern in a patient with multiple prior biopsy scars. Intraoperative imaging identifies central skin flap ischemia involving nipple–areolar complex and surrounding skin. a Preoperative photograph depicting multiple preexisting surgical scars. A peri-areolar incision with radial extension was used for attempted nipple-sparing mastectomy. b Relative skin fluorescence was measured sequentially at points indicated in picture. c Pre-mastectomy or baseline values are shown in graph and designated as B1, B2, B3 and B4. d Postnipple-sparing mastectomy values are graphed and designated as M1, M2, M3 and M4. e Corresponding post-mastectomy perfusion deficit of NAC shown in a still photograph at approximately, approximately 120 s post ICG administration

Figure 4 graphically demonstrates the observed changes in fluorescence before and after mastectomy in a patient with multiple prior incisions and a V1 perfusion pattern at baseline imaging study. The postmastectomy intraoperative photograph clearly shows a perfusion deficit that persisted over the 170 s imaging sequence of the NAC and surrounding skin. Additional examples of changes in relative fluorescence are graphically represented in V3 and V2 perfusion pattern cases (Figs. 2, 3, respectively). In the former, the atypical location of the mastectomy incision, driven by a central-upper quadrant lumpectomy scar, seemingly contributed to compromised blood flow in superior aspect of the NAC and adjacent skin flap.


We describe a practical technique for mapping blood supply to the skin and NAC in NSM cases. This analysis describes the perfusion pattern of the NAC and respective relative changes on postmastectomy imaging as observed by SPY Elite™ imaging. These preliminary observations suggest that when the predominant blood supply is derived from the underlying breast tissue (V1), the NAC may be at greater risk of ischemia compared to the V2 and V3 circulation patterns. The circulatory anatomy of the NAC has been previously studied in cadavers.13 The breast is perfused via the internal mammary, lateral thoracic, and anterior/posterior intercostal arteries but only the internal mammary artery supplies the NAC most consistently. Even though the internal mammary perforators were routinely divided during the mastectomy procedure, the fluorescence in the B4/M4 parasternal ROI decreased the least.

Radiation therapy is a predisposing factor for ischemic complications. In a series of 1,001 nipple-sparing mastectomies at the European Institute of Oncology, intraoperative delivery of 16 Gy electron beam radiotherapy to the NAC was associated with partial or total necrosis of the NAC 9 % of cases.10 We chose to include in our series patients who received prior breast conserving therapy (n = 2) or chest wall irradiation (n = 2). Had these four cases been excluded, the rate of NAC ischemia and intraoperative resection may have decreased from 18 to 8.5 %, which would make it more comparable to other contemporary series.7

One of the potential benefits of mapping blood flow pattern to the NAC is to serve as a guide in the placement of incisions to minimize interference with the blood supply. Moyer et al. reported the use of circumareolar-radial extension incisions resulted in nipple necrosis in 75 % of cases but decreased to 27–33 % with radial or IMF incisions.8 Similarly, Regolo and colleagues reported on their experience with 102 NSMs during which their surgical technique was modified from a periareolar skin incision in the first 32 patients to a lateral skin incision or the IMF in the latter 70 patients.14 With this shift in incision placement, NAC vascular complications were reduced to 2.8 % from their initial high rate of ischemic complications at 59.7 %, concluding that periareolar incisions impair collateral flow. This is in agreement with other reports and our observations, in patients with V1 NAC perfusion pattern in whom periareolar incisions may disrupt any secondary circumferential inflow to the NAC.15

The limitations of this study are the small sample size and retrospective classification of the nipple perfusion patterns. Our numbers are too small to comment on an interaction between incision type, intrinsic baseline perfusion pattern, and resection probability. The reliability of predicting irreversible tissue necrosis based on the degree of decreased perfusion to the NAC or skin flaps requires further study. Although randomizing by incision types might be a logical next step, other considerations, such as tumor location, location of prior biopsy scars, and breast size, are confounding factors that interfere with this type of study design. To expand on these results, we are conducting a prospective, randomized trial testing the utility of ICG-guided SPY Elite™ imaging and/or clinical judgment to determine how the measured relative fluorescence can predict skin and NAC necrosis.

Our results seem to indicate that the anatomic configuration of blood flow to the NAC may be helpful to determining ischemia. We initially piloted this technology during breast reconstruction cases and showed that it markedly diminished postoperative complications through the identification of ischemic areas.12 The current report adds a new perspective to the subject of ischemic complications in NSM. Noninvasive tissue angiography is a useful and practical tool that is easy to implement and should complement clinical judgment. It provides surgeons with a unique view of breast skin and NAC perfusion, which can be factored into the placement of mastectomy incisions. The classification scheme of NAC perfusion provides surgeons with a framework to preoperatively detect high risk perfusion patterns and modify an operative approach.

Nipple-areolar sparing mastectomies are challenging, especially when dealing with larger breasts and redundant skin envelopes. The findings discussed demonstrate that both oncologic and reconstructive surgeons may be able to use SPY Elite™ technology as an important adjunct to adjust real-time intraoperative decision making. Reducing the number of postoperative ischemic complications is a key element to serve the increasing demand for NSMs and provide improved surgical outcomes.


This study was funded in part by educational and research support provided by Novadaq Technologies and LifeCell to authors Geoffrey Gurtner, Irene Wapnir, Anne Kieryn, David Kahn and Shannon Meyer.

Supplementary material

10434_2013_3214_MOESM1_ESM.docx (215 kb)
Supplementary material 1 (DOCX 215 kb)
View video

Supplementary material 2 (AVI 54491 kb)

Copyright information

© Society of Surgical Oncology 2013