The interpretation and diagnostic utility of postoperative computed tomography (CT) imaging can be complicated by the presence of absorbable biomaterials (1–5). A commonly misdiagnosed biomaterial is oxidized regenerated cellulose (ORC) (1).
ORC is a widely used hemostatic agent in nearly all surgical specialties. ORC induces hemostasis by means of a primary local hemostyptic action and secondary activation of platelets to form a temporary platelet plug (6). Although ORC can be removed once hemostasis is achieved, it can also be left at the surgical site and degraded through normal wound healing (7). In a recent prospective, blinded study investigating the interpretation of CT images of patients treated with ORC, only 2 of 18 patients (11%) were accurately diagnosed as being treated with ORC (8). Interpretation of the remaining 16 patients was inconclusive, or ORC was misdiagnosed as an abscess, a hematoma, a lymphocele or a collection with hydroaeric levels.
Oxidized nonregenerated cellulose (ONC) is an alternate form of oxidized cellulose that was first used clinically in 2011 (9). ONC lacks organized cellulose fibers and has higher material density, which can give it a different appearance from ORC on CT images. Unorganized cellulose fibers of ONC have been shown to provide superior hemostasis relative to ORC (6). As the use of ONC becomes more widespread in surgical procedures, understanding the imaging characteristics of ONC on CT will aid in accurate interpretation and diagnosis.
This image report describes the postoperative noncontrast CT appearance of ONC and compares it with that of ORC when used to treat a femoral artery needle puncture in a rabbit at 0 and 14 days after implantation.
Material and Methods
Oxidized Cellulose Hemostatic Agents
Tabotamp (Ethicon, Inc., Somerville, New Jersey), also known as Surgicel® in North America, is the ORC used in this study. Traumastem (Baxter Healthcare SA, Zurich, Switzerland), also known as Celstat in North America, is the ONC used in this study.
All animal activities were performed according to the United States Animal Welfare Act and The Guide for the Care and Use of Laboratory Animals in an institution accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International following Institutional Animal Care and Use Committee Approval.
A rabbit femoral artery puncture model mimicking peripheral vascular surgery was performed in a 2.8-kg New Zealand White female rabbit (6). A fentanyl patch was placed 1 day before surgery. On the day of surgery, acepromazine (0.1 mg/kg, subcutaneously [SC]), glycopyrrolate (0.01 mg/kg, SC), and ketamine (25 mg/kg, SC) were administered to the animal. Following tracheal intubation, anesthesia was maintained using isoflurane.
The femoral arteries were exposed and isolated. An RB-1 curved tapered needle with 1-cm of 5-0 polyglactin 910 suture was passed through each artery in a proximal-to-distal direction. The needle entrance and exit holes were allowed to bleed freely for 6 seconds and then treated with a single application of 2 layers of 3- × 3-cm oxidized cellulose.
ORC was applied to the right femoral artery and ONC to the left femoral artery according to a randomization procedure using 30 seconds of digital pressure. Once hemostasis was achieved, the incisions were closed in an identical, routine fashion. The focus of this investigation was postoperative imaging, so hemostatic success was not measured.
Following postoperative imaging on the day of surgery (day 0), the animal was recovered and received meloxicam once daily for 3 days (0.2 mg/kg, intramuscularly). On day 14, the same anesthetic protocol was used for reimaging. After imaging, the animal was euthanized with pentobarbital (50 mg/kg, intravenously).
With the rabbit in the dorsal positon, noncontrast CT imaging was performed after surgery and 14 days postoperatively (Figure 1). These time points were selected based on previous research by Pampal et al. (10). A preoperative CT was not conducted, as the interest of this report was the appearance of the implanted material. CT imaging was performed using a multiple-detector-row CereTom™ scanner (NeuroLogica Corporation, Danvers, Massachusetts).
No intravenous contrast was administered so that the retained oxidized cellulose material can be identified without obscuration (ie, beam hardening or scatter) by the intimately adjacent intravascular iodine contrast (11). Contiguous 160-mm (day 14) and 260-mm (day 0) step-and-shoot data acquisition of the pelvis was conducted from the iliac crests to the feet with 140-kVp, 4-mAs, and a field of view radius of 126.656-mm. The images were acquired to minimize subject radiation dose to as low as reasonably achievable while maintaining diagnostic imaging quality. Coronal images were generated with 0.625-mm thickness.
On day 0, the oxidized cellulose hemostatic agents are seen medial to the femoral arteries with symmetrical beam attenuation and with the same attenuation as observed in the thigh muscles (Figure 2). The appearance of ORC in the right groin is more heterogeneous than that of ONC in the left groin. Surgical sutures are seen in both groins. There is minimal fat stranding and foci of air in the regions of interest, but these findings are to be expected in the immediate postoperative period with or without the use of hemostatic agents.
On day 14, neither of the oxidized cellulose hemostatic agents is definitively identifiable (Figure 3). No femoral pseudoanuerysms are appreciated in either groin. In comparison with the immediate postoperative images, there are expected changes of normal wound healing, including some skin thickening and reduction in local fat stranding. There are no foci of air along either femoral artery on the 14-day postoperative image. There are again focal hyperattenuating surgical sutures in both groins. No adverse tissue reactions are seen.
Based on this investigation, ONC has a more homogeneous appearance compared with ORC in noncontrast CT imaging immediately following surgery that is not identifiable on postoperative CT images on day 14. The difference between the oxidized cellulose on day 0 is likely because of the different density structures. The different presentations of oxidized cellulose over time, however, are not uncommon and represent their 2-phase degradation process. In the first phase, polyuronic acid is solubilized within 18 hours, whereas in the second phase, fibrous material is phagocytized and hydrolyzed by local macrophage within 4 weeks (6).
During the first phase, ORC can be misdiagnosed as a bowel leakage (12), obstructive mass (13), hematoma (14), mediastinitis (15) or, most frequently, abscess (16–19). These misdiagnoses are attributed to the appearance of ORC as a mixed- or low-attenuation mass with focal collections of gas that is absent of air–fluid levels and contrast enhancement in or near the operative site on immediate postoperative CT scans (1), which is similar to their appearance in this investigation. Given the comparable appearance to ORC, ONC can be misdiagnosed during the acute postoperative period. As oxidized cellulose is degraded or encapsulated, its appearance may vary. An interesting follow-up study should investigate its appearance more frequently in the postoperative period and in different anatomical locations (eg, intracranial, intrathoracic, or retroperitoneal sites). Further, serial observations in multiple animals or clinical patients should be used to investigate its varying presentations.
During the second phase, oxidized cellulose degradation and wound healing can be delayed depending on the amount of material used, blood flow, and tissue bed. When delayed too long, ORC forms a fibrous-encapsulated sterile granuloma, that is, a gossypiboma or “Surgiceloma.” In this phase, ORC is misdiagnosed as a tumor (20–24). This misdiagnosis is attributed to the appearance of ORC as a heterogeneous soft tissue mass with or without rim enhancement and with variable to no contrast enhancement. In this study, the quantity of ORC and ONC used was insufficient to form a gossypiboma. Because a gossypiboma was not induced, this is a potential direction for further investigation. Similarly, the appearance of ONC in contrast-enhanced CT imaging or less conventional modalities for soft tissue postoperative imaging (eg, plain radiography, ultrasonography, and magnetic resonance imaging [MRI]) can be investigated.
Oxidized cellulose has lower attenuation than a hematoma and a simpler appearance. Abscesses generally have scattered, variably sized, and patterned foci of gas with air–fluid levels and rim enhancement (25). Such distinguishing CT characteristics are suggestive but not confirmatory. Radiographic and sonographic examinations do not aid in the differentiation of oxidized cellulose from the above misdiagnoses (26, 27). Because of superior contrast between tissues, MRI may provide a more definitive diagnosis of oxidized cellulose postoperatively. Oto et al. described ORC to be hypointense in the early postoperative period on T2-weighted images, but variable thereafter (28).
Pampal et al. performed an exhaustive study comparing attenuation of ORC with normal tissue on CT and signal intensity on MRI within the intraperitoneal and retroperitoneal cavities of guinea pigs (10). Based on their CT findings, ORC is consistently hypoattenuating relative to paravertebral muscles and liver parenchyma when placed in the intraperitoneal space and hyperattenuating relative to fat when placed in the retroperitoneal space. In comparison with Pampal et al., this investigation shows that ORC is hyperattenuating to fat and overall isoattenuating to thigh musculature. The difference between the findings can be because of variation in the imaging technique, radiation dose, and anatomic location between the studies.
Although the CT imaging characteristics of ORC have been described over several years, this is the first report of the noncontrast CT imaging characteristics of ONC. This investigation characterized the appearance of ONC in noncontrast CT imaging as being similar to that of ORC. The only distinction between the two was that ONC appeared more homogenous. The homogeneity of ONC is likely because of the unorganized fiber structure or greater material density.