Is There a Right Way to Interpret SPY? Normalization of Indocyanine Green Angiography Readings in a Burn Model

Abstract

Sir: The applications of indocyanine green dye angiography in plastic and reconstructive surgery continue to expand. LifeCell SPY-Elite (LifeCell Corp., Branchburg, N.J.) has allowed for the intraoperative assessment of blood flow within mastectomy skin flaps and burn wounds. Kanuri and colleagues recently performed a fiscal analysis of the intraoperative use of indocyanine green dye angiography for mastectomy reconstructive flaps in high-risk patients such as smokers, the obese, and diabetics, suggesting its utility for these patients.1 However, challenges remain in defining consistent and effective criteria for interpreting these measurements. In this communication, we suggest that normalization of SPY-Elite measurements to uninjured tissue within the field of view improves measurement precision compared with reading raw absolute values as provided by the device’s onboard software. We demonstrate this technique by using data from prior work evaluating burn interspace survival in a validated porcine hot comb burn model.2 SPY-Elite was used to retrospectively characterize the interspaces between standardized full-thickness burns in female Yorkshire swine.3 Interspace viability was evaluated histopathologically 3 weeks after injury for evidence of horizontal burn progression. Interspaces that were found to have adnexal components or viable epidermal layer were considered viable, whereas skin lacking these components was deemed nonviable. Indocyanine green dye angiography measurements were taken 1, 4, 24, and 48 hours after injury. Each field was evaluated by SPY-Elite for absolute value (measured in arbitrary perfusion units) (Fig. 1, left), and normalized as a percentage of the highest value measured in uninjured skin within that frame (Fig. 1, right). Each frame was renormalized in this same method. Statistical analysis between absolute values and normalized values in viable and nonviable interspaces was performed by means of t test using GraphPad Prism 6.0 (GraphPad Software, Inc., LaJolla, Calif.).Fig. 1: Reading of image showing absolute values (left) and normalized values (right) of blood flow using the LifeCell SPY-Q analysis software. Normalization values were obtained by finding the highest blood flow measurement on the uninjured tissue within the same captured field as the target wound. Each time the wound was imaged, the normalization was recalculated based on these criteria.Comparing absolute values showed no significant differences between viable and nonviable interspaces, attributable in part to increased variability between animals (Fig. 2, left). When interspace values were normalized to the maximal perfusion in normal skin, thereby accounting for intrinsic variation in the vascularity and systemic state of the animal, perfusion levels showed significance (p < 0.05) between viable and nonviable tissues at 1, 4, and 48 hours after injury.3 Our group postulates that several factors relate to the increased variability of absolute perfusion values. Differences in the vascularity and the plasma composition of each animal may skew absolute value measurements. Plasma turbidity, a background fluorescence related to indocyanine green dye binding to a nonalbumin fraction of plasma proteins, is also organism-dependent, and contributes to subject variability.4 Clinically, absolute values may vary because of different patient demographics and systemic comorbidities such as obesity, smoking, and diabetes. Normalization would account for these variables and should result in more consistent indocyanine green dye angiography reading.Fig. 2: Perfusion levels are reported for 1, 4, 24, and 48 hours after burn injury in viable and nonviable tissue. Results are shown for absolute values (left) as arbitrary perfusion units (apu) and for normalized values (right) as percentages of normal perfusion. (*p < 0.05; **p < 0.01).A persistent obstacle preventing reliable interpretation of indocyanine green dye angiography is the precision of the measurement. In this communication, we describe a method of standardizing and normalizing quantitative readings provided by the LifeCell SPY-Elite. We have used this technique with success when studying the blood flow over time within Integra artificial skin (Integra LifeSciences, Plainsboro, N.J.).5 It will be further validated when used in upcoming randomized prospective studies on burn and breast reconstruction patients, and ongoing basic science studies. DISCLOSURE No author has any personal financial interest in the study presented. The authors have received research funds from the LifeCell Corporation unrelated to the data presented herein; that funding was intended for studying applications of the SPY-Elite device in reconstructive surgery. Mitchell S. Fourman, M.D., M.Phil. University of Pittsburgh Medical Center Pittsburgh, Pa. Robert P. Gersch, Ph.D. Hospital of the University of Pennsylvania Philadelphia, Pa. Heather A. Levites, M.D. Stony Brook University Medical Center Stony Brook, N.Y. Brett T. Phillips, M.D., M.B.A. Duke University Medical Center Durham, N.C. Duc T. Bui, M.D. Stony Brook University Medical Center Stony Brook, N.Y.