Multimodal Microvascular Mapping for Head and Neck, Skull Base Research and Education: An Anatomical Donor Study

Abstract

Objective  This study was aimed to develop a method combining computed tomography (CT) and fluorescence imaging, allowing identification of microvasculature in anatomical donors and facilitating translational research and education.

Methods  We investigated homogeneity and radiopacity of 30 different mixtures including radiopaque substances povidone–iodine (Betadine), barium sulfate (BaSO ₄ ), and bismuth subsalicylate (Pepto-Bismol) varying in suspension and dilution with agar, latex, or gelatin. Three candidate mixtures were selected for testing the extent of perfusion in renal vasculature to establish methodology. From these candidate mixtures, two were selected for mixture with fluorescein and infusion into cadavers based on their ability to perfuse renal vasculature. The extent to which these two candidate mixtures combined with fluorescein were able to perfuse vasculature in a cadaver head was used to determine which mixture was superior.

Results  BaSO ₄ and bismuth subsalicylate–based mixtures demonstrated superior opacity in vials. In terms of solidifying agents, gelatin-based mixtures demonstrated increased friability and lower melting points compared with the other agents, so only latex and agar-based mixtures were used moving forward past the vial stage. Combinations of BaSO ₄ and latex and BaSO ₄ and 3% agar were found to perfuse kidneys superiorly to the mixture containing bismuth subsalicylate. Finally, in cadaver heads, the mixture containing BaSO ₄ , agar, and fluorescein was found to perfuse the smallest vasculature.

Conclusion  A final combination of BaSO ₄ , 3% agar, and fluorescein proves to be a powerful and novel combination enabling CT imaging, fluorescence imaging, and dissection of vasculature. This paves the way for future translational research and education.

Introduction

The need for expanded capabilities with local, regional, and free-tissue flaps in reconstruction has given rise to many methods for preclinical research and education. The need for more of these methods is seen in every subfield of otolaryngology. Anatomical donors are an important component of existing otolaryngology research and training; via gross dissection, they allow the development of new surgical techniques and educational methods for learners at all levels. In addition to gross dissection, intravascular injection with an agent such as latex or gelatin for enhanced identification of vasculature and/or with a contrast medium for computed tomography (CT) imaging has been used in research and surgical training.

Many solutions have been explored for injection in anatomical donors, including silicone, latex, paraffin oil, and agar. ^{1 2 3 4} Colored casting solutions including latex and agar allow for guided anatomical dissection of the vascular structures. Barium sulfate (BaSO ₄ ), suspended in either agar or gelatin for perfusion, is the most common medium used for contrast. It provides an accessible and chemically cooperative option for developing a multimodal vascular and microvascular imaging method.

Anatomical studies using both casting solutions and contrast media are typically done with a two-stage process using two different groups of specimens. ^{5 6 7 8} In the first group, vessels are injected with latex or another colored casting solution for gross anatomical dissection, and in the second group, BaSO ₄ or another contrast medium is injected and CT images are acquired. Alternatively, BaSO ₄ is commonly suspended in colored latex, which stabilizes the radiopaque contrast and allows for radiographic imaging and gross anatomical dissection in the same specimen. ⁹ This has been done in large joint vasculature, ^{1 10} single muscle vasculature, ¹¹ flap vasculature, ¹² and enteric vasculature. ¹³

Fluorescence imaging, however, has not yet been integrated with these methodologies, though it is commonly used in head and neck and neurosurgeries. ^{14 15 16 17} Its significance is underscored by recent findings suggesting that incorporating fluorescence angiography into surgeries to evaluate the perfusion of free flaps might reduce the risk of flap necrosis by identifying areas of poor perfusion. ^{18 19 20} While fluorescence is one of the key modalities used for vascular imaging in the operating room or clinical setting, it is rarely used in anatomical dissection studies, where it could be used to identify and preserve vasculature via visualization of deeper vascular structures through tissue. A common fluorophore used in fluorescence imaging is fluorescein, which has a peak excitation at 494 nm (blue visible light) and emits fluorescence at a wavelength of approximately 520 nm which allows maximum visualization with a blue light filter. This is available in both the research and the clinical setting.

Despite the availability of both fluorescence imaging and radiopaque contrast, there have been no studies combining these techniques to study vasculature and microvasculature via both imaging and dissection in anatomical donors. In this study, we sought to develop an economical, nontoxic, reproducible, injectable technique that (1) highlights vasculature for gross anatomical dissection, (2) is radiopaque for CT imaging, and (3) contains an active fluorophore for fluorescent accentuation of vascular, in particular microvascular anatomy. Our final suspension of BaSO ₄ , agar, and fluorescein demonstrates visualization of complex vessel arborization in the head, neck, and skull base, through adequate perfusion of the arterial tree.

Materials and Methods

BaSO ₄ powder, ReagentPlus, 99% was obtained from Sigma-Aldrich Co. LLC (St. Louis, Missouri, United States), CAS Number 7727–43–7. Bismuth subsalicylate solution (C ₇ H ₅ BiO ₄ ) in the form of Pepto-Bismol Max Strength liquid 1,050 mg/30 mL, was obtained from Procter & Gamble Co. (Cincinnati, Ohio, United States), CAS Number 14882–18–9. Povidone-iodine solution (C ₆ H ₉ NO•I ₃ ) was obtained in the form of Betadine 10% (1% available iodine) from Purdue Products L.P. (Stamford, Connecticut, United States), CAS Number 25655–41–8. Agar ([C ₁₂ H ₁₈ O ₉ ] _n ) was obtained from Fisher Scientific, Inc. (Waltham, Massachusetts, United States), CAS Number 9002–18–0. Fluorescein solution (C ₂₀ H ₁₂ O ₅ ) was obtained in the form of Fluorescite 10% from Alcon Laboratories, Inc. (Geneva, Switzerland), CAS Number 2321–07–5. The pump used for infusion was a Preston Varistaltic Power Pump (115 VAC, 60 Hz, 1.5 Amp) by Manostat Corporation, Cat No. 72–360–000, Ser. No. 221. The anatomical donors were imaged with a General Electric LightSpeed 64 slice CT Scanner (General Electric Co., Boston, Massachusetts, United States) and were endoscopically visualized with a Tele Pack X portable all-in-one fluorescein blue filter system from Karl Storz Endoscopy-America, Inc. (Southbridge, Massachusetts, United States).

Trial Mixtures

To highlight vessels for dissection and radiographic identification, 30 different combinations created with one of three radiopaque chemicals with different suspension agents were assessed at varying concentrations in vials with plain film X-rays. The complete pipeline for selecting a final mixture is shown in Fig. 1 . Radiopaque chemicals that were trialed in the initial vial stage included povidone–iodine, BaSO ₄ , and bismuth subsalicylate. BaSO ₄ was suspended in dilutions of latex, a mixture of dH ₂ O and agar of varying concentrations, or a mixture of dH ₂ O and gelatin (300 Bloom) of varying concentrations. Bismuth subsalicylate and povidone–iodine, likewise, were each diluted in varying concentrations of agar or gelatin. Latex was only used with BaSO ₄ because BaSO ₄ was in powdered form. All mixtures were placed in 10 mL plastic vials and X-ray imaged. Concentrations of agar ranged from 0.5 to 5% (w/v), and concentrations of gelatin ranged from 10 to 30% (w/v).

Fig. 1.

A flow diagram depicting the process of selecting a final radiopaque suspension beginning with three candidate radiopaque substances and suspension agents. Agar and gelatin were trialed with all three radiopaque agents, and latex was trialed with barium sulfate, as barium was the only powdered radiopaque agent. From thirty initial mixtures examined in vials via X-ray, three were selected for renal perfusion. Based on renal perfusion, two mixtures were selected and combined with fluorescein for injection into anatomical donors, yielding our final mixture based on assessment of perfusion and fluorescence.

Next, vascular perfusion was assessed. Three suspensions were chosen from the vial stage based on homogeneity, radiopacity, friability, and melting point. Each suspension was tested by infusing approximately 10 mL into kidneys from anatomical donors, chosen as surrogates for the vasculature of the nasal cavity due to the density of this organ's capillary bed. Renal artery perfusions were analyzed by plain film X-ray to compare the filling capabilities of each mixture, and suspensions were optimized for use in blood vessels based on the results.

Final Mixture Syntheses

Based on the extent of perfusion in the renal artery infusions, two final mixtures, barium latex and barium with 3% agar, were selected to be injected into intact heads of anatomical donors. The processes by which these mixtures were constructed are shown in Fig. 2 . Quantities of ingredients were as follows. For the final barium latex mixture used in cadavers, 62.5 g BaSO _4(s) was combined with 30 mL of dH ₂ O and 210 mL of red latex. For the barium agar mixture, 250 mL dH ₂ O was combined with 62.5 g of BaSO _4(s) and 8 g of agar. To each, a synthetic fluorescent tracer, fluorescein (100 mg/mL), was added to give final BaSO ₄ concentrations of 0.25 mg/mL. The integrity of the final suspensions was confirmed by X-ray and digital photography of injected vessels illuminated at 494 nm.

Fig. 2.

( A ) A flow diagram depicting the process of constructing the latex and barium sulfate (BaSO ₄ )-based candidate mixture used for injection into an anatomical donor. ( B ) A flow diagram depicting the process of constructing the final agar and BaSO ₄ -based mixture selected based on perfusion and vessel highlighting in anatomical donors.

The final method used to ensure proper mixing and suspension of the barium latex mixture ( Fig. 2A ) was as follows: first, BaSO _4(s) was vortexed into the dH ₂ O; this suspension was stirred at 40°C for 5 minutes while slowly adding latex, prewarmed to 40°C; and finally, the fluorescein was added.

For the final 3% agar mixture ( Fig. 2B ), dH ₂ O was stirred at 25°C, and then BaSO _4(s) was slowly added. This was stirred for 5 minutes at 800 rpm. Once the liquid was uniform in consistency, the agar was added and stirred for another 2 minutes. The beaker was then placed in a microwave where it was heated on medium power until it began to boil. Once bubbling, the beaker was removed and the mixture thoroughly stirred with a ceramic spatula. To reduce microbubble formation in the final product, the suspension was left at room temperature until it measured 40°C in the center. The beaker was returned to the microwave and heated until boiling again. Once removed, the mixture was immediately stirred and the beaker continuously tapped on a flat surface for 2 minutes to encourage degassing. Finally, the fluorescein was stirred in. The suspension was then immediately injected.

Injections

Three traditionally embalmed and two lightly embalmed human anatomical donors were used. When a donor passes away, we try to be in receipt as soon as possible. The bodies are not refrigerated. Three gallons of distilled water are mixed with 24 oz of a preinjection chemical (Primol, Hydrol Chemical Company, Index: 4) which is injected into the inferior direction of right carotid. Superior to the injection needle, the carotid is tied off. A drain tube is inserted into the jugular vein, then the limbs and joints are moved and limbs massaged to promote flow. For traditionally embalmed anatomical donors, five gallons of Cornell Embalming Solution (Hydrol Chemical Company. Index: 22) are then injected. For lightly embalmed anatomical donors, 8 ounces of formalin are mixed in with 1 gallon of distilled water which is injected instead. Intermittent drainage is then stilled until the embalming solution first appears out the drain tube. When the embalming solution is visible, the drain tube is removed and the jugular vein is tied off in both directions. Massage and movement of the joints are then continued until the limbs become firm. After the 5 gallons have entered the donor, the carotid is tied off in both directions. Following embalming, the carotid triangle was dissected bilaterally to reveal the common carotid arteries, and a stylet was inserted through a 1-cm longitudinal incision along the anterior aspect of the left common carotid, approximately 4-cm inferior to the carotid sinus. Plastic tubing of minimal length was connected from a 250 mL beaker to the stylet, passing through a Preston Varistaltic Power Pump. The contralateral common carotid was then completely transected as inferior as possible.

Prior to injection, 150 mL of 80°C dH ₂ O was injected through the stylet to soften the tissue, clear debris and clots, warm the equipment and vessels, and delay setting of the suspension in the agar-based mixture. Once patency of the cerebral arterial circle (of Willis) was confirmed via visualization of drainage of dH ₂ O from the contralateral common carotid, the latex- or agar-based mixture was injected into the left common carotid artery under low pounds per square inch via the varistaltic pump.

Once the infusion began, the transected contralateral carotid artery was monitored for leakage. Upon observing several milliliters of drainage, this vessel was clamped with a large hemostat, encouraging backflow. This occurred at roughly 50 mL of total injected compound. Another 150 to 200 mL was added to ensure complete filling of the head and neck vasculature. If at any time fluid began leaking from the stylet or the pump began stalling, the infusion was terminated. The carotid arteries were then clamped bilaterally and the nasal cavity was immediately examined with endoscopic equipment under white and 494-nm light. Once images were acquired, the anatomical donor was stored in a dark room to prevent photobleaching of the fluorescein. Within 12 to 24 hours, each anatomical donor was imaged by CT at Boston Medical Center, rendered with OsiriX v.6.0. Dissection of the common carotid artery and its branches was then performed.

Results

Preliminary Vial Testing/Renal Imaging

Thirty combinations were trialed prior to injection. Of the radiopaque substances, the undiluted povidone–iodine 10% demonstrated lower levels of attenuation for CT imaging than bismuth subsalicylate– or BaSO ₄ -based compounds ( Fig. 3A ). It was thus discarded as an option before renal artery infusions. In addition, the gelatin mixtures demonstrated fragility, friability, and a low-melting point (35°C). Therefore, the following three mixtures were selected for vascular infusion: (1) undiluted bismuth subsalicylate solidified with 1% agar, (2) 2.5-mg BaSO ₄ in 10 mL latex, and (3) 2.5-mg BaSO ₄ in 10 mL dH ₂ O solidified with 3% agar. Of these three, BaSO ₄ with latex and BaSO ₄ with 3% agar mixtures demonstrated the best perfusion of the kidney vasculature, as shown in Fig. 3B . During the renal artery perfusion stage, larger volumes had to be used for the mixtures which resulted in unique issues not seen in the vial stage, such as settling of solids and inadequate perfusion. Therefore, mixing procedures were optimized as described in the methods, including stirring and heating protocols for the final candidate mixtures, before injection into cadaver heads.

Fig. 3.

( A ) X-ray images of experimental combinations of radiopaque suspensions in nonradiopaque plastic vials including 1% Barium sulfate (BaSO ₄ ) in 1% agar, 1% BaSO ₄ in 15% gelatin, 2.5% BaSO ₄ in latex, 2.5% BaSO ₄ in 3% agar, povidone–Iodine undiluted, Pepto-Bismol Max Strength, 5% BaSO ₄ in latex. ( B ) Renal injection trial with a preinjection kidney on the left and a postinjection kidney on the right with 2.5% BaSO ₄ in dH ₂ O solidified with 3% agar (w/v) to determine extent of arterial perfusion.

Injections

Initially, we explored several methods for injection: leading with 50 to 100 mL of 10% formaldehyde, 50 to 100 mL 80°C dH ₂ O, or injecting the given suspension directly. Between preinjecting with water or formaldehyde versus no preinjection, leading with warmed dH ₂ O yielded the most favorable results. Additionally, we explored numerous methods for maintaining arterial access for the duration of the infusion, such as using cyanoacrylates to secure a large-bore needle or heat-shrink rubber tubing to fasten a transected common carotid around intravenous tubing. Ultimately, externally clamping the terminal end of a metal, bulb-tipped stylet provided the most reliable, heat-resistant, and leak proof seal.

Computed Tomography Imaging

Spiral CT imaging of the injected anatomical donors demonstrated adequate perfusion of both internal and external carotid vasculature. The 2.5% BaSO ₄ in latex and fluorescein perfused vessels as small as 0.5 mm in diameter ( Fig. 4A ), including most major branches of the maxillary and facial arteries. However, the 2.5% BaSO ₄ in dH ₂ O solidified with 3% agar and fluorescein mixture highlighted even smaller arteries, down to 0.25 mm in diameter ( Fig. 4B ). Vessels perfused by this suspension included numerous distal and minor vessels such as posterior inferior cerebellar, anterior ethmoid, infraorbital, supraorbital, sphenopalatine, middle meningeal, and superior alveolar arteries, as well as many of their minor branches. In addition to providing higher spatial resolution in spiral CT imaging, this solution was able to provide an excellent visualization of the vasculature in fluorescence imaging. For example, Fig. 5 shows the strong correlation between fluorescence ( Fig. 5A ) and CT imaging ( Fig. 5B ) of the superficial temporal artery.

Fig. 4.

( A ) Three-dimensional volume rendering from CT scans demonstrating degree of perfusion after injection with 2.5% barium sulfate (BaSO ₄ ) in latex (w/v) and ( B ) 0.5% BaSO ₄ in dH ₂ O solidified with 3% agar (w/v). Note higher degree of perfusion with the agar-based mixture indicated by branches of the internal carotid, such as ophthalmic and supraorbital arteries. CT, computed tomography.

Fig. 5.

( A ) Frontal branch of the left superficial temporal artery illuminated with 494 nm light postinjection with 2.5% barium sulfate (BaSO ₄ ) in dH ₂ O solidified with 3% agar (w/v) and fluorescein. ( B ) CT three-dimensional volume rendering of the same branch in the same donor. CT, computed tomography.

Discussion

To develop a new way of examining vasculature and microvasculature for preclinical research, clinical research, and education, we examined and compared several mixtures that can be used in anatomical donors to highlight vasculature for concurrent dissection, CT imaging, and fluorescence imaging. Given that BaSO ₄ is accessible and its properties are well studied, it proves to be an ideal material to aid in visualization via CT imaging. A suspension of 2.5% BaSO ₄ in dH ₂ O solidified with 3% agar provides superior perfusion of minor arterial branches compared with a suspension of BaSO ₄ in latex, demonstrating anastomoses and unique anatomic anomalies while serving as an adequate scaffold for the delivery of fluorescein. Fluorescein provides a substance applied clinically for in vivo visualization. Hence, the three components of vascular visualization in anatomical donors allow for in situ surgical simulation (fluorescein), CT investigation (BaSO ₄ ), and dissection confirmation of the vasculature (latex). This model allows for visualization and study of the vasculature to reveal a level of detail that is needed for developing research and education opportunities, in advance of applying learned principles in the clinic and surgical suite. In addition, it provides a proof-of-concept bridge for considering modifications of these methods for in vivo use.

Using this methodology, a surgeon or surgical trainee would be able to rehearse, practice, and/or learn existing surgeries. First, the final mixture would be infused into an anatomical donor and CT imaging would be obtained to inform dissection. During the dissection, to highlight vasculature locations in real time, fluorescence imaging with a blue-light filtered endoscope could be performed. This would help guide the dissection to avoid certain arteries and to identify anatomic landmarks, and takes on additional relevance when considering ways to better understand normal anatomy and its variations. Such methods can also be used when faced with revision cases, simulating the loss of reconstructive options, and to preemptively build better rungs for a reconstructive ladder.

In addition, this technique could be used to illustrate proof-of-concept surgeries before using them in the operating suite. As detailed previously, fluorescence imaging has been shown to improve monitoring of perfusion in flaps. To trial new approaches to local, regional, or free flap reconstruction, a surgeon could plan a new technique using CT imaging and then dissect the anatomical donor using fluorescence for real-time monitoring of flap microvasculature. This would allow assessment of flap viability and arteries to avoid, followed by anatomic confirmation. The depth of corroboration offered by this technique would also be useful in medical device development, where it is paramount to visualize vasculature and other perfusion-capable anatomic spaces, as well as obtain a knowledge of in situ position. In sum, this technique has the ultimate potential to advance patient care.

Limitations

There are a few limitations to this study. First, it is possible that different concentrations of fluorescein could have yielded improved correspondence of fluorescence visualization with the underlying vasculature; only different concentrations of contrast media, and not fluorescein, were optimized in this study. Fine tuning of the optimal concentration of fluorescein is a possible future direction. It is also possible that the degree of perfusion of fluorescence-highlighted vasculature in situ may underestimate the degree of perfusion in vivo. Last, other fluorophores, like indocyanine green, could serve as superior agents to assess vasculature in situ.

Conclusion

The novel suspension developed in this project allows for surgeons and researchers alike to both radiographically image and fluorescently highlight vessels for reliable postmortem vascular mapping. This has potential applications in helping faculty and trainees to learn existing techniques, develop innovative surgical procedures, and potentially expand translational work through using anatomical donors for other innovations such as developing new medical devices reliant on vascular and microvascular maps.

Funding Statement

Funding This study was funded by Boston University Department of Otolaryngology—Head and Neck Surgery, Boston University Department of Anatomy and Neurobiology, and the Julia & Seymour Gross Foundation, Inc.