SEARCH

SEARCH BY CITATION

Keywords:

  • Imaging techniques;
  • infrared;
  • kidney;
  • reperfusion

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Allograft ischemia induces delayed graft function and is correlated with increasing rates of rejection. There is not currently a way to objectively measure the effects of ischemia in real-time, nor to relate therapies combating reperfusion injury with their intended effects.

An infrared (IR) method utilizing a focal plane array detector camera was developed for real-time intraoperative IR imaging of renal allografts, and evaluated in a pilot trial to quantify perfusion in recipients of live (n = 8) and cadaveric donor (n = 5) allografts. Digital images were taken for 3–8 min postreperfusion. Image data were compared to ischemic time and allograft function to assess potential clinical relevance.

Cold ischemic time ranged from 0.5 to 29 h and was bimodally distributed between living and cadaveric donors. Renal rewarming time (RT) as determined by IR imaging correlated with cold ischemic time (p < 0.001, R2 = 0.81), and predicted the subsequent return of renal function with RT negatively correlated to the regression slopes of creatinine (p = 0.02, R2 = 0.38) and BUN (p = 0.07, R2 = 0.26).

Intraoperative IR imaging noninvasively provides clinically relevant real-time whole kidney assessment of reperfusion. This technology may aide in the objective assessment of therapies designed to limit reperfusion injury, and allow for quantitative assessment of allograft ischemic damage.


Abbreviations:
IR

infrared

RT

rewarming time

ROI

region of interest.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

It has long been appreciated that prolonged ischemia and other factors related to cadaveric donor stability significantly impair renal allograft function. It has also been shown that delayed allograft function is correlated with increasing rates of both acute and chronic rejection (1–5). Accordingly, empiric conventions for acceptable cold ischemic times and donor condition have been established to improve graft outcome based on current preservation technologies. Although the association of procurement and postreperfusion injury with adverse immune outcomes has been recognized, methods for objectively quantifying the degree of injury sustained by a kidney have been less forthcoming. Clinicians have thus relied on functional measurements over the days following transplantation to indirectly establish preservation related injury (6).

The ischemic time per se is only one of several factors correlated with renal injury. Indeed, while the rate of delayed graft function decreases substantially with decreasing cold ischemic time, some organs clearly function immediately despite ischemic times exceeding 24 h and others are slow to function after the modest periods of ischemia associated with live renal donation. Furthermore, other factors independent of ischemic time, such as donor and recipient hemodynamic stability, donor vascular disease and brain death, clearly influence the function of an allograft. Regardless, the aggregate result of procurement injury is vasoconstriction and poor parenchymal perfusion.

To date there have been no clinically applicable ways to directly assess renal parenchymal perfusion other than subjective assessment of the kidney color and turgor. There has also been no method for evaluating the effect of therapeutic or technical interventions aimed at improving perfusion. Consequently, drug development to improve renal response to ischemia has been hampered by the lack of an objective measure of graft reperfusion. Previous methods of quantification have relied upon Doppler assessment of the renal arteries or other large vessel flow (7–10). Although these methods have been useful for evaluating specific vessels, and their data have been related to parenchymal blood flow, these methods do not assess cortical blood delivery, nor do they account for physiological shunting. They also only assess small regions or single vessels. Global perfusion has not been measurable.

Recent developments in infrared (IR) technology have allowed for highly accurate measurements of temperature with accuracy to the hundredth of a degree Celsius. This optical technology has been shown to be adaptable to the operating room environment (11–15). It has the advantage of imaging entire fields and allowing for real-time assessment of specific areas and/or aggregate assessment of large areas. As temperature is directly related to cortical perfusion, this method could provide a means to objectively assess allograft reperfusion. We thus adapted a novel IR system for intraoperative use during renal transplants and asked whether high-resolution infrared imaging would be an effective method to contemporaneously quantify perfusion following renal transplantation. We find that IR technology offers a feasible, clinically relevant method for quantifying renal allograft reperfusion.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Patients and transplantation

After informed consent was obtained, 13 renal transplant recipients were enrolled in an Institutional Review Board approved trial for the evaluation of methods for monitoring renal allograft function. Recipients of both live donor (group A, n = 8) and cadaveric donor (group B, n = 5) grafts were included. Transplantation was performed using standard surgical methods. Immunosuppressive medications were individualized and were not stratified based on donor type. All kidneys were implanted into the external iliac system. There was a single renal artery in 12 cases and a dual renal artery reconstructed into a single orifice in one case. All patients were hemodynamically stable throughout their transplant procedures. The preservation method for all cases was cold University of Wisconsin solution flush followed by nonpulsatile cold storage. Cold ischemic time was calculated from the time of donor nephrectomy to the time of graft reperfusion.

Infrared imaging

An advanced digital IR camera (Infrared Focal Plane Array camera, Lockheed Martin IR Imaging Systems, Inc., Goleta, CA) was used to image local temperature gradients simultaneously across the entire transplanted kidney by passively detecting IR emission. As IR emission at the measured wavelength (3–5 µm) is directly proportional to temperature, the camera was calibrated in units of temperature. The camera has a sensitivity of 0.02 °C. One hundred images (256 × 256 pixels) were obtained at 1-s intervals and digitized at 14 bits per pixel.

The camera was attached to a ceiling mount normally used for an operating room light, allowing convenient optical access to the operative field (Figure 1), and was, except for the lens, covered with conventional surgical draping to maintain sterility. It was placed 50–60 cm above the kidney surface to achieve a field of view that fit the entire kidney (∼120 × 120 mm). Spatial resolution for individual pixels in this configuration was approximately 400 × 400 µm. No alterations of operating room temperature, light, or airflow were required.

image

Figure 1. Intraoperative set-up for real-time infrared imaging is shown. Surgery can continue while measurements are being taken.

Download figure to PowerPoint

Infrared imaging (100 images per 1.68 minute trial, 2–4 trials per patient) was performed on all patients by passive acquisition of spontaneous IR emission from the exposed tissue. Sequential digital images were taken with the plane of the IR camera's lens positioned parallel to the plane of the graft. The IR camera was focused on the renal parenchyma after the vascular anastomoses were completed. Image acquisition began 5–10 s before removal of the vascular clamps and continued for 3–8 min following graft reperfusion. The ureter was also imaged simultaneously and IR images were also acquired during repositioning of the kidney in attempts to improve perfusion. Pseudo-colored IR images were displayed on a large monitor for viewing by the surgeon (Figure 2A–C), and were evaluated in the operative theater in real time. Subjective observations were made and recorded in parallel with the IR evaluation.

image

Figure 2. Intraoperative images of a live donor kidney and its region-of-interest (ROI) temperature profiles. Kidney reperfusion started 2 s after the infrared (IR_imaging began. Images were acquired (A) at the time of arterial/venous unclamping, (B) 48 s after reperfusion, and (C) 220 s after reperfusion. The pseudocolor temperature scale is shown to the left of panel A, and is the same for panels B, and C. Notice the heterogeneity of kidney temperature in B, and the uniformity of kidney temperature in C. Note also the presence of a perfusion defect in panel C related to a finger print following graft manipulation. The yellow line on image D (a grayscale version of image A) outlines the kidney. The same image shows the locations for ROI with lowest IR signal (blue square), highest IR signal (red square), and control ROI (white square). Graph E shows temperature profiles of each of the ROIs over 250 s of IR imaging. The colors of the ROIs in D correspond to their colored graphs in E. The three arrows on the time axis identify the times when images A, B, and C, respectively, where captured. Interruptions related with storing each trial of 100 images are shown by dotted lines.

Download figure to PowerPoint

Data analysis

On-line visual analysis of the images was followed by an off-line evaluation. Successive trials of baseline IR images acquired with an interval of at least 30 s between trials were stored (Axil Ultima workstation, Axil Computer, Inc., Santa Clara, CA) and analyzed (ENVI software, Research Systems, Inc., Boulder, CO). To investigate the relationship between changes in thermal gradients and the progress of reperfusion, a region-of-interest (ROI) was established as an outline of the exposed renal parenchyma (Figure 2D, yellow line). A separate analysis was performed examining extremes of temperature with two ROIs established in the kidney areas with the lowest and highest IR signals (6 × 6 pixel ROIs each) followed for the entire observation period (Figure 2D, blue and red squares). A temperature profile was obtained from each ROI positioned at the same location on each IR image of the trial.

Infrared signals from each ROI were used to determine the rewarming time (RT): the delay period between start of reperfusion and the saturation point (when temperature at the ROI of the whole kidney reached the maximum value during the observation). The RT was established as the time for a given ROI to reach its saturation point.

The RT was compared with the cold ischemic time of the kidney for each patient. In addition, the relationship between the RT and post transplant allograft function was analyzed. Renal allograft function was determined clinically by standard clinical chemistry evaluation for serum creatinine and blood urea nitrogen daily for the first 7 days post transplant. The slope derivative of a quadratic best-fit curve (Microsoft Excel, Microsoft Corporation, Bellevue, WA, USA) was determined for each patient. Slope derivatives were compared with RT. Correlations were evaluated using anova with statistical significance set at p < 0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Satisfactory measurements were obtained from all individuals. The measurements were feasible despite occasional requirements to interrupt the optical path to make technical adjustments to the kidney. Graft and patient survival was 100% and all patients had return of graft function. No patients required post transplant dialysis, thus no patient met United Network for Organ Sharing criteria for delayed graft function. As anticipated, however, the rate of return of renal function varied considerably between the living donor and cadaveric cases. Cold ischemic time ranged from 35 min to 29 h and was bimodally distributed. Living donor cases had a cold ischemic time of 46 min (35–60 min) and cadaveric donor cases had a cold ischemic time of 21 h (15–29 h).

Renal RT for the entire kidney ROI was closely correlated with cold ischemic time (Figure 3). This was true when considering the correlation with all kidneys considered taking into account the differences between live donor and cadaveric donor kidneys (p < 0.001, R2 = 0.81), and when evaluating the variation within the cadaveric kidneys alone (p < 0.001, R2 = 0.92). High and low ROIs were also correlated with cold ischemic time (p < 0.05, not shown) and varied greatly in their rate of return indicating that renal reperfusion was regionally heterogeneous. Although areas of cyanosis were typically identified as areas with slow reperfusion, the measured rate of reperfusion varied even within visually indistinguishable regions. Thus, RT as determined by IR imaging correlated well with established clinical perimeters of allograft ischemia, and could distinguish fine differences beyond areas visually identified as poorly perfused.

image

Figure 3. Rewarming time as measured by intraoperative infrared imaging correlates with ischemic time and donor type following renal transplantation. Cadaveric donor cases (▪) and liver donor cases (□) are shown. The solid line describes data from cadaveric cases only, while the dashed regression line describes data from all cases. R2 data are shown for each regression line.

Download figure to PowerPoint

Infrared RT also was correlated with the subsequent return of renal function (Figure 4). There was a negative correlation between RT for the entire kidney ROI and both the regression slope of creatinine (p = 0.02, R2 = 0.38), and the regression slope of BUN (p = 0.07, R2 = 0.26). The data were most tightly correlated when the entire kidney was evaluated, and correlations were dependent on the donor source (cadaveric vs. live donor). Single extreme ROI measurements from the best and worst perfused areas trended toward significant relationships with postoperative function but were not significantly correlated with overall allograft function (not shown).

image

Figure 4. Rewarming time as measured by intraoperative infrared imaging correlates with the negative slopes of (A) serum creatinine and (B) serum blood urea nitrogen over the first 7 days post transplant. Cadaveric donor cases (▪) and liver donor cases (□) are shown. Regression lines describe data for all cases.

Download figure to PowerPoint

The temperature of the renal parenchyma immediately before reperfusion was relatively homogeneous in the population studied (12 °C ± 1 °C). Thus, no correlation between starting temperature and RT, or subsequent return of function was seen (not shown).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Renal allograft reperfusion is increasingly recognized as a significant event both physiologically and immunologically. Indeed, many concepts shaping our current understanding of immune activation are increasingly wed to factors liberated as a result of ischemic allograft injury and factors unique to the cadaveric donor condition (16–21). As a result, immune therapies have been proposed to minimize the effects of ischemia either by specifically countering the events at the time of reperfusion, or by altering the exposure to drugs, like calcineurin inhibitors, with known adverse effects on renal perfusion (22–25). However, these therapies have been applied based on risk assessment of delayed allograft function, specifically targeting the most extreme instances: those leading to post transplant dialysis.

In this pilot study, we have introduced a new technology for the quantitation of renal allograft perfusion, and demonstrated that it provides clinically relevant information at the time of surgery that is associated with ischemic time, donor type, and the subsequent return of allograft function. Although additional study will be required to establish clear parameters identifying degrees of reperfusion injury and guidelines for data interpretation, this initial study is encouraging. We have found that overall parenchymal perfusion as measured by IR imaging is a better predictor of subsequent allograft recovery than regional evaluations using the same technology. As creatinine clearance is an aggregate parameter of total renal function, this relationship is not unexpected. Given the redundant nature of the renal parenchyma, and the compensatory nature of whole organ function, it is not surprising that measurements of single areas are not as correlative as whole organ measurements. Nevertheless, studies evaluating therapeutic interventions (e.g. infusion of calcium channel blockers or other vasodilating agents) may benefit from an ability to isolate poorly perfused areas in response to therapies or maneuvers (e.g. after reconstruction of a branch vessel). In addition, the ability to isolate the ureter could be envisioned as a means of determining ureteral viability before creation of a ureteroneocystostomy.

It is unlikely that precise measurement of reperfusion status will be appropriate for standard post transplant assessment. Rather, this technique may best be applied in a research setting where precise flow measurements are important. This could include trials evaluating agents meant to decrease reperfusion injury, or in which preservation injury needed quantification to stratify outcome measurements. One potential application where such precise flow evaluation could be envisioned to be of routine assistance is in the evaluation of adult organs transplanted into infants, or other cases where flow disparities could be anticipated (26). Although the most encouraging use of this technology has been in whole organ perfusion, this method could also be used for evaluation of regional defects following arterial or venous reconstruction. As the surgical procedures in this pilot evaluation were relatively uncomplicated, the parenchymal warming was rather homogeneous. Nevertheless, this technique could also be helpful in assessing renal temperature during complicated implantations, providing objective data prompting topical cooling measures. Finally, we were able to focus the camera on areas that have typically not been amenable to large vessel flow evaluation including the ureter. Clearly, we cannot make assessments of ureteral viability in such a small clinical study, but normal flow characteristics of ureteral perfusion should be established to evaluate ureters compromised by procurement injury.

Several other methods have been developed to measure renal perfusion in the operating room. These include Doppler ultrasound or laser, and thermodilution methods. However, we feel that the described method has many potential advantages over these technologies. Invasive methods or methods requiring direct contact with the parenchyma clearly cause tissue trauma or compression that affect local perfusion (27, unpublished observations). Additionally, most methods provide regional (ultrasound) or aggregate (thermo-dilutional) information, but not both (28). Those methods limited to regional assessment require sequential probe positioning and thus are not capable of simultaneous or retrospective analysis of data from unanticipated sites of interest (28). Laser Doppler Flow measurement is also increasingly difficult to perform with thick or encapsulated tissues such as the kidney (29). Techniques requiring circumferential placement of probes around major vessels are subject to probe dislocation and further interfere with blood flow itself (28,30). While the radioactive microsphere distribution technique is regarded as the gold standard for the measurement, this technique is highly invasive and requires postmortem tissue sampling for radioactive measurements. It is not applicable for blood flow measurements in the clinical setting. Finally, most of the imaging techniques used in previous studies are not capable of collecting data in real time during the performance of clinical surgery. Thus, changes in perfusion resulting from therapeutic manipulation cannot be assessed.

Based on these pilot data, we suggest that IR imaging be considered as a means of objectively assessing allograft reperfusion. This method should be adaptable to other transplanted organs including the heart and liver. Additional clinical study is warranted to determine the role of this technology in clinical investigative practice.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

The authors gratefully acknowledge the superb clinical support of the Warren G. Magnuson Clinical Center operating room nursing staff.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References
  • 1
    Cole E, Naimark D, Aprile M et al. An analysis of predictors of long-term cadaveric renal allograft survival. Clin Transplant 1995; 9: 2828.
  • 2
    Giral-Classe M, Hourmant M, Cantarovich D et al. Delayed graft function of more than six days strongly decreases long-term survival of transplanted kidneys. Kidney Int 1998; 54: 972978.
  • 3
    Gjertson DW. Impact of delayed graft function and acute rejection on kidney graft survival. Clin Transpl 2000: 467480.
  • 4
    Geddes CC, Woo YM, Jardine AG. The impact of delayed graft function on the long-term outcome of renal transplantation. J Nephrol 2002; 15: 1721.
  • 5
    Troppmann C, Gillingham KJ, Benedetti E et al. Delayed graft function, acute rejection, and outcome after cadaver renal transplantation. The multivariate analysis. Transplantation 1995; 59: 962968.
  • 6
    Govani MV, Kwon O, Batiuk TD, Milgrom ML, Filo RS. Creatinine reduction ratio and 24-hour creatinine excretion on posttransplant day two. simple and objective tools to define graft function. J Am Soc Nephrol 2002; 13: 16451649.
  • 7
    Lundell A, Persson NH, Kallen R, Ekberg H. Impaired renal artery blood flow at transplantation is correlated to delayed onset of graft function. Transpl Int 1996; 9: 5761.
  • 8
    Van Weel V, Van Bockel JH, Van Wissen R, Van Baalen JM. Intraoperative renal duplex sonography: a valuable method for evaluating renal artery reconstructions. Eur J Vasc Endovasc Surg 2000; 20: 268272.
  • 9
    Isiklar I, Aktas A, Akgun S, Karakayali H. From donor to recipient. Doppler US, power US and scintigraphy of kidney perfusion before and after transplantation. Acta Radiol 2000; 41: 285287.
  • 10
    Bhatt KA, Karamanoukian HL, Bergsland J, D'Ancona G, Stephan R. Intraoperative graft verification in renal transplants. Vasc Endovascular Surg 2002; 36: 9396.
  • 11
    Shevelev IA, Tsicalov EN, Budko KP, Sharaev GA, Gorbach AM. Dynamic infrared functional mapping of the cerebral cortex. Biomed Sci 1990; 1: 571577.
  • 12
    Gorbach AM. Infrared imaging of brain function. Adv Exp Med Biol 1993; 333: 95123.
  • 13
    Gorbach A, Heiss J, Kufta C, Oldfield E. Intraoperative Localization of Brain Tumors and Surrounding Functional Cortex with Infrared Imaging. New Orleans, The American Association of Neurological Surgeons, 1998.
  • 14
    Gorbach A, Heiss J, Pluta R, Oldfield E. Intraoperative Infrared Assessment of Cerebral Microcirculation. Sydney, Australia, Seventh World Congress for Microcirculation; 2001 August 19–22, 2001.
  • 15
    Watson JC, Gorbach AM, Pluta RM, Rak R, Heiss JD, Oldfield EH. Real-time detection of vascular occlusion and reperfusion of the brain during surgery by using infrared imaging. J Neurosurg 2002; 96: 918923.
  • 16
    Janeway CA Jr. The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol Today 1992; 13: 1116.
  • 17
    Fearon DT. Seeking wisdom in innate immunity. Nature 1997; 388: 323324.
  • 18
    Gallucci S, Matzinger P. Danger signals: SOS to the immune system. Curr Opin Immunol 2001; 13: 114119.
  • 19
    Fuggle SV, Sanderson JB, Gray DW, Richardson A, Morris PJ. Variation in expression of endothelial adhesion molecules in pretransplant and transplanted kidneys – correlation with intragraft events. Transplantation 1993; 55: 117123.
  • 20
    Bryan CF, Luger AM, Martinez J et al. Cold ischemia time: an independent predictor of increased HLA class I antibody production after rejection of a primary cadaveric renal allograft. Transplantation 2001; 71: 875879.
  • 21
    Hoffmann SC, Kampen RL, Amur S et al. Molecular and immunohistochemical characterization of the onset and resolution of human renal allograft ischemia-reperfusion injury. Transplantation 2002; 74: 916923.
  • 22
    Tosa Y, Kollias N, Lee WP, May JW Jr. Reduction of ischemia-reperfusion injury by monoclonal antibody to intercellular adhesion molecule-1. Transplant Proc 1996; 28: 12101211.
  • 23
    Hong JC, Kahan BD. A calcineurin antagonist-free induction strategy for immunosuppression in cadaveric kidney transplant recipients at risk for delayed graft function. Transplantation 2001; 71: 13201328.
  • 24
    Hammer C, Thein E. Visualization of the effect of polyclonal antithymocyte globulins on adhesion of leukocytes. Transplant Proc 2002; 34: 24862487.
  • 25
    Gonwa TA, Mai ML, Smith LB, Levy MF, Goldstein RM, Klintmalm GB. Immunosuppression for delayed or slow graft function in primary cadaveric renal transplantation. use of low dose tacrolimus therapy with post-operative administration of anti-CD25 monoclonal antibody. Clin Transplant 2002; 16: 144149.
  • 26
    Salvatierra O Jr, Singh T, Shifrin R et al. Successful transplantation of adult-sized kidneys into infants requires maintenance of high aortic blood flow. Transplantation 1998; 66: 819823.
  • 27
    Adams T, Spielman WS, Holmes KR, Heisey SR, Chen MM. Proposed methods for the measurement of regional renal blood flow using heat transfer analysis. Ann Biomed Eng 1985; 13: 237258.
  • 28
    Kraus T, Mehrabi A, Angelescu M, Golling M, Allenberg JR, Klar E. Characterization of renal parenchymal perfusion during experimental infrarenal aortic clamping and declamping with enhanced thermodiffusion electrodes. Ann Vasc Surg 2001; 15: 447456.
  • 29
    Hammad FT, Davis G, Zhang XY, Wheatley AM. Intra- and post-operative assessment of renal cortical perfusion by laser Doppler flowmetry in renal transplantation in the rat. Eur Surg Res 2000; 32: 284288.
  • 30
    Martin GT, Bowman HF. Validation of real-time continuous perfusion measurement. Med Biol Eng Comput 2000; 38: 319325.