An Imaging Protocol To Discriminate Specialized Conduction Tissue During Congenital Heart Surgery

Abstract

Objectives:

We performed preclinical validation of intraoperative fiber-optic confocal microscopy (FCM) and assessed its safety and efficacy in an ovine model of the pediatric heart.

Methods:

Intraoperative imaging was performed using an FCM system (Cellvizio, Mauna Kea Technology, Paris, France) with specialized imaging miniprobe (GastroFlex UHD, Mauna Kea Technologies). Before imaging, we applied an extracellular fluorophore, sodium fluorescein, to fluorescently label extracellular space. We imaged arrested hearts of ovine (1–6 months) under cardiopulmonary bypass. Image sequences (1–10 s duration) were acquired from regions of the sinoatrial and atrioventricular node, as well as subepicardial and subendocardial working myocardium from atria and ventricle. The surgical process was evaluated for integration of the imaging protocol during the operative procedure. In addition, fluorescein cardiotoxicity studies (n=3 animals) were conducted by comparing electrocardiogram (PR and QRS intervals) and ejection fraction at baseline and after topical application of fluorescein at 1:10, 1:100 and 1:1000 dilutions on a beating ovine heart.

Results:

Our studies suggest that intraoperative FCM can be used to identify regions associated with specialized conducting tissue in ovine hearts in situ. The imaging protocol was integrated with conventional open heart surgical procedures with minimal changes to the operative process. Application of fluorescein in varying concentrations did not affect the normalized PR interval, QRS interval and ejection fraction.

Conclusion:

These preclinical validation studies demonstrated both safety and efficacy of the proposed intraoperative imaging approach. The studies constitute an important step towards first-in-human clinical trials.

Graphical Abstract

Central picture legend: Utilization of confocal microscopy can help with tissue discrimination inside the heart.

ECCentral Message: Fiber-optics confocal microscopy (FCM) can be used intraoperatively to visualize andidentify specialized conduction tissues.

Introduction

The disposition of the cardiac conduction system in the normal heart is well documented by morphological studies¹. Accurate intraoperative localization of specialized conducting tissue in a pediatric population of congenitally deformed hearts presents a unique challenge. In congenitally deformed hearts, the disposition of these vital structures cannot be accurately identified, and the surgeon must approximate the location of these anatomical sites.^2,3 If these tissues are damaged during surgical repair, the cardiac conduction system can be disrupted and the patient may experience temporary or permanent heart block.^4–6 Table 1 lists the incidence rates of conduction abnormalities after surgical repair of ventricular septal defects,^5,7 atrioventricular canal defects,^5,6,8 tetralogy of Fallot,^4,6,8 sinus venosus atrial septal defects^9–11 and transposition of the great arteries.^12,13 Conduction abnormalities such as dysfunction of sinoatrial (SA) and atrioventricular conduction pathways require chronic cardiac rhythm management using permanent pacemaker implantation, creating a high economic and societal burden.^14,15

Table 1:

Incidence rate of conduction abnormalities and permanent pacemaker implantation after surgical repair of various congenital heart defects leading to sinoatrial and atrioventricular conduction abnormalities.

Congenital defect	Conduction abnormality type	Conduction abnormality %	Permanent pacemaker implantation %
Ventricular septal defects (VSD)
Perimenbranous VSD	Atrioventricular	-	1.17
Atrial switch with VSD	Atrioventricular	-	6.55
Arterial switch with VSD	Atrioventricular	-	3.45
Coarctation with VSD	Atrioventricular	-	1.75
Atrioventricular canal defect	Atrioventricular	-	1–2.35,6,8
Tetralogy of Fallot	Atrioventricular	34	0.9–1.46,8
Sinus venosus atrial septal defect	Sinoatrial	6–359,10,11	4.610
Transposition of the great arteries	Sinoatrial	11–6412,13	3.8–7.512,13

Fiber-optics confocal microscopy (FCM) was developed as an alternative to traditional biopsies used in standard endoscopic evaluations. The technology produces optical sections, allowing clinicians to evaluate tissue microstructure at the point of care through real time in vivo visualization at depths of up to 100 μm below the tissue surface. Fiber-optic imaging systems have been approved for clinical use in urogenital,¹⁶ gastrointestinal,¹⁷ and pulmonary¹⁸ applications. In our previous studies, we introduced approaches based on FCM and extracellular fluorophore delivery for imaging of the living heart.^19–21 These approaches allowed us to identify atrial working myocardium and tissue regions associated with the conduction system based on microstructural arrangements in these tissue types. We hypothesize that surgeons can use these approaches to more accurately localize specialized conducting tissue and prevent injury to these vital tissues during surgical repair. Identifying specific tissue types during surgery may reduce post-surgical complications resulting from trauma to the cardiac conduction system.⁶ Though the imaging system and dye delivery methods were suitable for use in a rodent model, validation protocols are required to demonstrate safety and efficacy for clinical use.²²

The aim of this study was to evaluate our previously developed approaches in the operative setting. Our research group has been moving towards clinical translation with rodent studies²¹, ex vivo human studies²⁰, expert human analysis²² and here with preclinical validation studies for safety and efficacy. This approach takes advantage of an US Food and Drug Administration (FDA) approved imaging system in an off-label application that may have a profound effect on the comorbidities associated with pediatric congenital heart surgery. In this paper, we describe the steps that we took to perform preclinical validation and obtain regulatory approval prior to performing first-in-human studies.

Methods

Sheep Anesthesia and Surgical Protocol:

The experimental studies were IACUC approved and conducted at the Animal Resources at Children’s Hospital (ARCH) facility (Boston, MA). An experimental ovine animal was induced with Ketamine (Fort Dodge Animal Health, Overland Park, KS, USA) at a dose of 10 mg/kg IV and Midazolam (Hospira, Lake Forest, IL, USA) at a dose of 0.2–0.5 mg/kg IV and maintained with 1–4% isoflurane (Patterson Veterinary Supplies Inc., Devens, MA, USA) with intubation and mechanical ventilation. Blood from a donor animal was used for priming the cardiopulmonary bypass circuit and for transfusions to replace blood loss during surgery. The experimental ovine was placed in supine position and sternotomy was performed. Bicaval cannulation was setup with 14–18 Fr cannula (Medtronic, Minneapolis, MN, USA). The right femoral artery was then cannulated with 8–12 Fr cannula (Medtronic, Minneapolis, MN, USA) based on the size of the animal. The animal was systemically heparinized (300 units/kg) (Fresenius Kabi, Canton, MA, USA) and the cardiopulmonary bypass was started once Activated Clotting Time (ACT) was >720 s. The heart was arrested using del Nido cardioplegia solution (20 mL/kg) (Central Admixture Pharmacy Services, Inc., Woburn, MA) delivered via the aortic root after cross clamp application. The superior vena cava (SVC) and inferior vena cava (IVC) were snared for total bypass. FCM imaging was first done at the SA node and the working myocardium. Right atriotomy was performed for imaging the atrioventricular node (AV) region. After imaging, 6–0 Prolene marking sutures (Ethicon Inc., Somerville, NJ) were placed on the FCM imaged sites in the SA and AV node regions. The right atrium was closed using 5–0 Prolene (Ethicon Inc., New Jersey, USA) and the animal was taken off bypass. After the heart regained normal beating for approximately 10 mins, the animal was euthanized by intravenous administration of Fatal Plus (Vortech Pharmaceuticals Ltd., Dearborn, MI, USA).

FCM Imaging and OR Integration:

FCM imaging was conducted on arrested hearts of 1–6 month old ovine on cardiopulmonary bypass. Sodium fluorescein (Fluorescite® 10%, Alcon Laboratories Inc., Fort Worth, TX) was delivered either topically (animals: 16) or systemically via cardioplegia (animals: 6). For topical application, 100 mg/mL fluorescein was diluted 1:1000 in saline. The dilute solution was applied via syringe at a volume of 1–2 mL on the tissue surface 2–3 s before imaging. For systemic dye delivery, 100 mg/mL fluorescein was diluted 1:1000 in cardioplegia and delivered by the catheter inserted in the aortic root perfusing the entire heart via the coronary arteries. The Cellvizio 100 series system (Mauna Kea Technologies, Paris, France) with GastroFlex UHD miniprobe was used for imaging. The system consists of (1) a movable tower that houses the laser scanning unit, control-computer and display, and (2) detachable fiber-optics imaging probes. The tower was placed in the animal OR on the opposite side from the surgeon. The foot pedal was connected and placed by the surgeon’s side. A sterile GastroFlex UHD miniprobe with a tip diameter of 2.6 mm was positioned by hand for imaging on the outer heart surface or using forceps inside the heart. Probes were used for up to 20 procedures with reprocessing after each procedure per the manufacturer’s instructions. The calibration step was performed automatically by the system once the probe was attached into the laser source. The laser source was controlled directly by the surgeon using the foot pedal or by an assistant using the keyboard. The probe acquired image sequences at a frame rate of 9 Hz with a lateral (xy directions) resolution of 1.8 μm, field of view diameter of 240 μm, and at a depth of 55–65 μm from the probe lens-tip. FCM imaging was performed in regions of the SA and AV node, as well as working myocardium of the atria and ventricle. The additional time required for the dye delivery and imaging was recorded.

Conventional Confocal Microscopy:

After euthanasia of the animals, the heart was excised and fixed via antegrade perfusion with a syringe and immersion in Tyrode-like solution containing 4% paraformaldehyde. Immunofluorescent labeling was performed on working myocardium and nodal tissue regions. Dissected tissues were labeled with Anti-Hyperpolarization-Activated Cyclic Nucleotide-Gated Potassium Channel 4 (HCN4) antibody (Alomone, APC-052, Jerusalem, Israel) and an antibody conjugated to Alexa Fluor 633 (Thermo Fisher, A-21070, Waltham, MA), wheat germ agglutinin (WGA) conjugated with Alexa Fluor 488 (Thermo Fisher, W11261, Waltham, MA), and 4’,6-Diamidino-2-Phenylindole, Dihydrochloride (DAPI) (Thermo Fisher, D1306, Waltham, MA). Three-dimensional imaging of working myocardium and nodal regions was performed using a scanning confocal microscope (TCS SP8, Leica Microsystems, Wetzlar, Germany). Imaged regions included the endocardial face of right ventricular working myocardium and the endocardial face of the AV nodal regions adjacent to the suture marker.

Assessment of Acute Fluorescein Cardiotoxicity:

Fluorescein is an FDA-approved dye commonly administered intravenously and utilized for ophthalmic vascular imaging. Acute cardiotoxicity of fluorescein was evaluated by applying different concentrations of fluorescein on the beating ovine heart in situ. Four solutions were prepared using saline and 100 mg/mL fluorescein. Saline was used for baseline recordings. The remaining three solutions were 1:1000 (0.1 mg/mL), 1:100 (1 mg/mL) and 1:10 (10 mg/mL) dilutions of stock fluorescein dye in saline. Three animals 1–6 months old and weighed 10 to 35 kg were selected for the study. A median sternotomy was performed on anesthetized animals followed by incision of the pericardium to access the heart. From lowest to highest fluorescein concentration, each solution was sequentially dispensed via syringe at a volume of 5 mL on the beating heart surface. The solutions were dispensed in a manner to ensure uniformity and complete coverage of the heart surface. Immediately following topical application of each solution, echocardiographic imaging of the left ventricle and electrocardiogram (ECG) recordings were performed for 5–10 min. Echocardiographic imaging was done using a diagnostic ultrasound system (Model iE33, Philips, Andover, MA). ECGs were recorded using a Cardio Companion esophageal electrode system (SurgiVet, Smiths Medical, Norwell, MA). ECG signals were acquired and printed using Advisor Vital Signs Monitor (SurgiVet, Smiths Medical, Norwell, MA). The P wave and QRS complex were identified in the ECG strips. The PR and QRS intervals were measured for 5 beats from the ECG strips for each condition for each animal. Ejection fractions were calculated using the analysis software of the diagnostic ultrasound system. Left ventricular areas during diastole and systole were manually identified and marked on the echocardiographs for the analysis.

Expert Analysis:

FCM videos of 1–10 s duration were acquired during surgery and indexed based on anatomical origin, i.e. working myocardium or nodal. Shorter 1–2 s video clips were generated and randomly selected for evaluation by examiners blinded to the tissue type. Video clips were incorporated into a graphical user interface (MATLAB 2017b) that played the video and prompted examiners for classification.

Prior to classification, examiners were trained using a slide presentation consisting of previously indexed still images and video clips that illustrated microstructural features indicative of each tissue type. Examiners reviewed the set of video clips of working myocardium and nodal tissue and asked to classify each video as either working myocardium or nodal tissue. We defined true positive and true negative outcomes as working myocardium correctly classified as working myocardium and nodal tissue correct classified as nodal tissue, respectively. Sensitivity and specificity of the examiners in discriminating working myocardium and nodal tissue in video clips were determined according to these classifications.

Sample Sizes and Statistics:

Fluorescein cardiotoxicity testing was performed with 3 animals. The test protocols were developed in consultation with the FDA as part of the Investigational New Drug (IND) application. The PR intervals, QRS intervals and ejection fractions for each animal were normalized to their respective mean baseline by the following equation:

$$X_{norm}=\frac{X_{condition}}{{\overline{X}}_{baseline}}$$ (1)

where X_norm was the normalized value of the selected parameter X, X_condition was the recorded parameter value and ${\overline{X}}_{baseline}$ was the mean of the recorded values at baseline condition for the respective animal. Normalized ejection fraction values were calculated in percentage, where equation 1 was multiplied by 100. For FCM expert analysis study, an a-priori power analysis (G*Power 3.1.9.2, University of Duesseldorf) was performed.²³ A minimum sample size of 81 per group was estimated to achieve statistical power of 0.8 and detect an effect across the 2 groups with medium effect size of 0.4 and significance criteria of 0.05.

Results

FCM Imaging and OR Integration

A total of 22 female juvenile sheep (age: 0.9–7.7 months, weight: 11.1–37 kg) were used for FCM imaging using either topical (n=16) or systemic (n=6) application of fluorescein. Imaging in atrial and ventricular regions showed regular striated tissue microstructures while those in the SA and AV node regions displayed irregular and reticulated microstructures (Figure 1; Videos 1–6 in Supplementary Material). Similar microstructural features of nodal and non-nodal tissues were present in our previous studies with heart tissues from rodent, neonatal lamb and humans using FCM.^20–22 Figure 2 shows example images from 3D conventional confocal microscopy of a nodal region and a working myocardium region. Both regions stained positive for WGA and DAPI. In contrast to the working myocardium, WGA images from the nodal region exhibited reticulations and are positive for HCN4.

Figure 1.

Video frames from fiber-optic confocal microscopy of living arrested ovine heart. Images were acquired by the surgeon using the imaging probe after topical administration of 1:1000 dilution of fluorescein for visualization of working myocardium (A, C) and nodal region (B, D). Fluorescein remains in the extracellular space and allows visualization of the cell boundaries to highlight the tissue microstructure. The working myocardium (A,C) presents striated tissue microstructure compared to reticulated microstructure in nodal regions (B,D). Scale: 20 μm.

Figure 2.

Confocal images of immunohistochemically labeled tissue from an atrioventricular node region (A-D) and working myocardium (E-H). Tissues were fluorescently labeled with Wheat Germ Agglutinin (WGA, in yellow) (A, E), Anti-Hyperpolarization-Activated Cyclic Nucleotide-Gated Potassium Channel 4 (HCN4, in green) (B, F) and 4’,6-Diamidino-2-Phenylindole, Dihydrochloride (DAPI, in blue) (C, G). WGA identifies the extracellular space. HCN4 is a marker of nodal and pacemaker cells, but not of working myocardium. DAPI marks cell nuclei. (D) and (H) are composite images showing the labels. Note that HCN4 labelled positive only for cells in the nodal region (B) that presented a reticulated tissue microstructure (A) in contrast to the working myocardium that was HCN4 negative (F) and has a striated tissue microstructure (E). Scale: 50 μm.

A distinct difference in contrast was observed during FCM imaging of the SA node region between the two dye delivery methods. Images from topical dye delivery revealed clearly distinguishable tissue microstructure in all regions of the heart. In contrast, images from systemic dye delivery displayed inferior contrast with indistinguishable tissue microstructure in the SA node region.

The addition of the movable tower, foot pedal and probe into the animal OR was viewed favorably by the surgical team. The surgeon spent 4.9±0.7 s (n=4) to load the syringe with dye and apply 2–3 mL of fluorescein and 1.9±1.1 s (n=7) to place the syringe back on the instrument table. Direct administration of dye via the cardioplegia solution did not add to the procedure time. It took the surgeon 2.3±0.6 s (n=7) to pick up the FCM probe for imaging and 2.7±0.7 s (n=3) to place it down after imaging. FCM video clips of 1–10 s durations were acquired during imaging.

Assessment of Acute Fluorescein Cardiotoxicity

Table 2 summarizes the electrocardiographic and echocardiographic values and parameters of sheep from literature^24–30 and those recorded in our studies. The electrocardiographic and echocardiographic parameters vary with breed, age and geographic location, as well as from animal to animal. The recorded parameters from the 3 animals were normalized to their respective mean baseline values in each animal to account for variations observed in the animals used in the study. The normalized ejection fractions for the different fluorescein dilutions are presented in Figure 3A. There were no clinically meaningful differences in mean normalized ejection fraction following application of each dye dilution condition.

Table 2:

Electrocardiographic and echocardiographic parameters of ovine models

	Electrocardiogram		Echocardiogram
	PR interval (s)	QRS complex (s)	Ejection Fraction %
	0.045 ± 0.00824	0.095 ± 0.00524	59.4 ± 4.928
Literature	0.077 ± 0.01525	0.042 ± 0.00925	51.4 ± 12.929
	0.105 ± 0.01926	0.057 ± 0.01326	41.2 ± 6.730
	0.180 ± 0.01627	0.060 ± 0.01027
Dye study: Baseline	0.190 ± 0.033	0.074 ± 0.010	69.3 ± 11.0
Dye study: 1:1000 dilution	0.185 ± 0.077	0.077 ± 0.019	66.0 ± 7.8
Dye study: 1:100 dilution	0.175 ± 0.033	0.076 ± 0.013	69.7 ± 9.0
Dye study: 1:10 dilution	0.190 ± 0.030	0.075 ± 0.018	71.0 ± 10.7

Figure 3.

Fluorescein cardiotoxicity study results. (A) Normalized ejection fraction vs fluorescein dilution. Ejection fractions were normalized to respective mean baseline values (saline solution). The plot presents the values (dots) and the median (bars) of the normalized ejection fraction in 3 animals (n = 3). The effect of fluorescein dilution on normalized ejection fraction appeared to be insignificant. (B) Box and whisker plots of the normalized PR (in blue) and QRS complex (in orange) intervals vs fluorescein dilution. The time intervals were normalized to respective mean interval baseline values (saline). The horizontal line within the box indicates the median, boundaries of the box indicate the 25^th and 75^th percentile, and the whiskers indicate the maximum and minimum values of non-outliers. The “x” marked in the box indicate the mean. The extra dots are the outliers. The presented values are from 15 beats (n = 15) recorded from 3 animals (5 beats per animal). The effect of fluorescein dilution on normalized PR and QRS complex intervals appeared to be insignificant.

Similarly, for the ECG data, PR and QRS intervals for different fluorescein dilutions were normalized to baseline recordings. The mean normalized intervals for all 3 fluorescein dilutions were close to baseline (Supplementary Material Figures 1–4) suggesting that fluorescein application has no effect on cardiac conduction. The normalized results are presented in Figure 3B.

Expert Analysis

A random set of video clips of 81 working myocardium and 81 nodal regions were evaluated by 8 examiners blinded to the tissue type. Examiners evaluated the images after being trained to identify characteristic features in the microstructural arrangement of each tissue type. The examiners achieved a high sensitivity and specificity of 93.5±3.6% and 96.8±3.9%, respectively (Table 3). Examiners incorrectly classified approximately 5 working myocardium videos as nodal tissue (false negative) and approximately 3 nodal region videos as working myocardium (false positive) out of the total 162 videos.

Table 3:

Examiner tissue classification

Examiner #	Specificity %	Sensitivity %
1	100.0	96.3
2	96.3	97.5
3	98.8	91.4
4	93.8	96.3
5	100.0	90.1
6	96.3	92.6
7	100.0	87.7
8	88.9	96.3
Mean ± SD	93.5 ± 3.6	96.8 ± 3.9

Discussion

We did not observe any complications resulting from the cardiac bypass procedures used in this study during surgery or FCM imaging. Direct systemic administration of fluorescein did not add any time to the total procedure time. However, systemic administration of fluorescein produced weak contrast in the SA node region. We did not observe this in our previous studies with small animals.²⁰ Topical dye delivery resulted in superior contrast compared to systemic dye delivery, though it added a short dye application time. While topical dye delivery added ~7 s per application, the net time added to the procedure was < 40 s (~2 mL dye applied each time up to a total of 10 mL). There was an additional probe handling time of < 30 s (5 x probe handling time). Furthermore, systemic administration of dye delivered more dye to the heart than topical application. In systemic application, fluorescein was delivered to the whole heart and ranged from 20 mg to 70 mg based on the amount of cardioplegia infused. In topical application, up to 10 mL of 1:1000 diluted fluorescein was applied, delivering less than 1 mg of fluorescein to the heart. In both cases, the delivered fluorescein was within the tested and reported range of non-toxicity. Operationally, systemic delivery did allow continuous imaging and eliminated handling of the fluorescein loaded syringe. As neither dye delivery method significantly augmented procedure time and delivered fluorescein was within non-toxic limits, image quality was the primary factor in selecting the preferred dye delivery method. High imaging contrast was vital for effective identification of nodal and non-nodal regions. Hence, topical dye delivery was the superior method.

Possible adverse reactions to fluorescein included vasoplegia, allergic reaction, extravasation, and local tissue damage.^31–33 With the exception of vasoplegia, the listed reactions would have caused ventricular dysfunction measurable by reduction in ejection fraction. We did not observe any significant deviations in ejection fraction from baseline on application of fluorescein dilutions ranging from 1:10 to 1:1000 (Figure 2). Furthermore, no significant differences were observed in the ECGs recorded before and after application of different fluorescein dilutions (Figure 3). Our results indicate fluorescein to be non-cardiotoxic and safe to use on the heart.

The OR is a demanding environment. During complex pediatric cardiac cases, surgical team members are focused on their individual areas of expertise and are required to communicate as needed to ensure the best patient outcome possible. Introduction of another piece of equipment into this environment requires careful consideration. With the addition of each piece of equipment, the OR becomes more crowded and less efficient, and potential hazards multiply. The imaging system was positioned to minimize interference with the overall procedure (Supplementary Material Figure 5). We found that the imaging probes were easy to handle and the imaging screen could be placed out of the way and still be easily visualized by the surgeon.

The coordination of imaging probe placement and visualization of the screen was performed without difficulty. Future versions of the probe may include a finger switch similar to a Bovie pen for capturing images. Our results indicate that the equipment can be integrated into the OR with few challenges. However, we note that these studies were performed in an animal OR and the same results may not be found in the human OR.

The laser of the Cellvizio 100 system emits at a wavelength of 488 nm with a maximum power of 15 mW. This was deemed to be a negligible risk as it was approximately 80% less than the maximum allowable retinal exposure as per the standard IEC 60825–1.³⁴ Thus, we suggest that the fluorescein-based FCM system is a safe intraoperative cardiac imaging modality.

Evaluation of FCM video clips by human examiners indicates that samples from atrial, ventricular, and nodal regions can be distinguished with high sensitivity and specificity. We previously showed that human examiners were able to achieve a higher sensitivity and specificity when reviewing still images from a rodent model in a laboratory environment.²² Compared to still images, video clips appeared to have a relatively higher level of noise. Motion, film formation or debris, and image artifacts all contributed to a higher level of noise. In addition, some clips did not focus on the tissue region of interest the entire video, leading to possible misclassification.

A technical limitation of the applied FCM system and GastroFlex UHD miniprobe is that imaging was limited to a tissue depth of 55–65 μm. Probes with increased working distance have been developed. However, FCM imaging will not reach tissues at a depth larger than approx. 150 μm and cannot cover the full thickness of the SA or AV node in larger mammals. Thus, the presented images (Fig. 1) present either superficial regions of the nodes or node-associated structures. In our prior work using conventional confocal microscopy in rodents, we showed that surficial tissues with an irregular and reticulated microstructure label positive for HCN4, which is an established marker for nodal tissue. A similar finding was presented here for ovine cardiac tissue from a region presumed to be nodal during FCM imaging (Fig. 2).

An operational limitation was the requirement for the surgeon to operate the FCM system and record FCM videos. With the current system setup, the surgeon had to position and hold the probe perpendicular to the tissue surface of the subject while looking at the Cellvizio display in front. This required the surgeon to acclimatize as he/she had to intermittently turn to look down at the subject and forward at the Cellvizio display. A possible solution is to provide a remote monitor located near the subject. Probe maneuverability was challenging in deeper tissue regions, eg AV node and His bundle regions, and forceps had to be utilized to maintain the position and orientation. Furthermore, we note that our ground truth for human expert analysis was based on expertise of the surgeon in identifying nodal regions. In our prior work using a rodent model, we established that an irregular and reticulated microstructure corresponds to nodal tissues using molecular markers.²⁰

Conclusion

We performed a preclinical validation of an FCM system on ovine hearts in situ, demonstrating safety and effectiveness of the imaging protocol. The FCM system was successfully integrated into an animal OR suite. Imaging did not add a significant amount of time to the procedure. Expert examiners achieved high sensitivity and specificity discriminating between working myocardium and tissue from nodal regions. The results of this study provide support for advancing to first-in-man studies using real-time FCM for the intraoperative discrimination of cardiac tissue to reduce the incidence of conduction disturbances during surgical repair of congenital heart disease.

Perspective Statement: A significant barrier in pediatric reconstructive heart surgery is the accurate intraoperative localization of conducting tissues in patients with complex congenitalheart disease. Damage to these tissues can lead to sinus node dysfunction andatrioventricular block. Here, we validate the safety and efficacy of intraoperative fiberoptics confocal microscopy in an ovine model of the pediatric heart.

Abbreviations:

ACT

activated clotting time

AV

atrioventricular

ECG

electrocardiogram

FCM

fiber-optics confocal microscopy

FDA

Food and Drug Administration

IVC

inferior vena cava

OR

operating room

SA

sinoatrial

SVC

superior vena cava