Postoperative Assessment of Surgical Clip Position in 16 Dogs With Cancer: A Pilot Study

Introduction

The radiation treatment volume in an animal with gross tumor is the palpable or visible disease that can be identified on physical examination and imaging studies. However, many veterinary cancer cases are presented for radiation therapy postoperatively and have only microscopic disease. Radiation fields based only on a surgical scar may be inaccurate and result in ultimate treatment failure.^1–3 Tumors may recur at the edge or just outside the irradiated field. Metallic markers (e.g., hemoclips, metal sutures, or staples) placed at the time of surgery allow radiographic visualization and determination of the full breadth and depth of the surgical site, and they help localize areas that may contain residual microscopic neoplasia. The metallic markers, in conjunction with the surgical scar, delineate the surgical volume and thereby provide the radiation target volume.

Inaccuracies are encountered when lumpectomy scars have been used alone for planning a radiation field in women with breast carcinoma.¹ A study involving human breast cancer patients identified the tumor bed based on radiographic examination and visualization of hemoclip position.² The study showed an overall potential for topographical errors of 52.8% (14/27 patients) if the fields were based on the surgical scar and hemoclips were not used to help design the radiation treatment fields.² In another study, the clinically established fields were inadequate in 34 of 50 (68%) patients. The relative positions of the surgical scar and clips in some patients were widely disparate, such that fields based on a scar alone would be inadequate.³

Computed tomography (CT)-based, computer-generated treatment plans are now utilized at veterinary radiation facilities to more accurately deliver radiation doses to tumors while sparing surrounding normal tissues.⁴ While CT scans are routinely performed to image animals with macroscopic disease, the use of CT imaging in the postoperative setting to define the radiation field has not been evaluated.

Placement of surgical clips is a relatively inexpensive and easy method for postoperatively identifying the tumor bed in an animal. Five radiopaque clips, placed at the margins of the tumor bed (e.g., medial, lateral, proximal [cranial], distal [caudal], and deep), represent the minimum number of clips needed to define the surgical bed in conjunction with visualization of the surgical incision.⁵ In some cases, hemoclips may be used during surgery for hemostasis, and a larger number of hemoclips may be inserted. The placement of metallic clips in the surgical bed to delineate the extent of the resected tumor and involved adjacent tissue may provide useful information for the radiation oncologist. Visualization of surgical clips is best accomplished by obtaining diagnostic-quality, routine radiographs that define the radiation treatment field (i.e. simulation films). Simulation films can be compared to portal films (i.e., films made using the radiation source itself to expose the film) that delineate the actual radiation treatment field, and an assessment can be made as to whether the treatment field should be adjusted. In addition to accurate treatment planning, immobilization and positioning of the animal during image acquisition and during daily radiation therapy are also critical. In veterinary radiation oncology, only limited evaluation of the methods to define the radiation target volume, portal imaging, and animal positioning have occurred.^6–8

The use of metallic clips raises two concerns about the presence of metallic objects in radiation treatment fields: (1) generation of electrons and backscatter off the implant, resulting in increased radiation dose deposition in the immediate vicinity of the metal; and (2) shadowing distal to the metal, thereby reducing delivery of the radiation dose beyond the marker. Clinically, the presence of hemoclips or metal staples in the radiation field does not appear to create problems, and recent investigations in dose alterations have focused on more substantial implants (such as orthopedic prostheses), which may cause more interference with dose delivery.^9–13

This pilot study investigated the use of hemoclips and surgical staples placed at the time of tumor resection as a means of subsequently visualizing the margins of tumor bed. The hypothesis examined in this study was that a shift in marker position could be expected as the surgical wound contracted; therefore, information obtained by imaging animals at two different points in time would differ. The goals of the study were to characterize the causes and extent of surgical clip movement postoperatively and to develop preliminary recommendations for the use of metallic markers in animal radiation cases.

Materials and Methods

Tumor-bearing dogs that underwent resection of their tumor and had either hemoclips or metallic surgical staples placed at the time of surgery were consecutively entered into the study. All dogs had radiographs taken immediately postoperatively and had follow-up radiographs ≥2 weeks after surgery. In surgery, hemoclips or staples were inserted after tumor resection to delineate the extent (peripheral aspects) of the tumor bed. The number of hemoclips or staples that were placed was left to the discretion of the individual surgeon. Hemoclips were not routinely used for other purposes, except to provide intraoperative hemostasis. When hemoclips were placed for hemostasis during tumor resection, a larger number of hemoclips ultimately delineated the tumor bed.

Orthogonal radiographs were made immediately postoperatively and were then repeated at the time of suture removal or after wound healing. Animals were positioned for radiographs in a routine fashion, based on the anatomical site to be imaged. Positioning devices were not used to immobilize the animals during radiography. The animals were under general anesthesia for the immediate postoperative radiographs, but they were not anesthetized for the radiographs taken at the 2+-week recheck. All radiographs were reviewed by one radiologist (Samii). For each dog, the two sets of films were evaluated for animal positioning, surgical clip number, and surgical clip position. Surgical clip position was assessed in relation to the other clips, to normal tissue, and to osseous landmarks. Measurements of the separation of the surgical clips were taken from both sets of radiographs for comparison. Measurements included the separation distance in both the craniocaudal and dorsoven-tral directions or equivalent directions for each set of films for a total of four measurements per animal at both time points (immediately postoperatively and at wound healing). In one case, only one measurement was taken per radiograph, as only two surgical clips were identified on the films. The change in separation of the surgical clips was recorded as a positive (+) number in centimeters if there was increased separation of the surgical clips at the time of the second set of films, and a negative (−) number in centimeters if there was a decrease in the separation of the hemoclips on the second set of films. The percentage change in separation of the hemoclips between the two sets of films was calculated by dividing the change in separation distance by the overall separation distance of the surgical clips for each direction on the films.

Results

A total of 16 tumor-bearing dogs at the Veterinary Medical Teaching Hospital, University of California, Davis were entered into the study [see Table]. Hemoclips were used as the surgical markers in 14 dogs, and surgical staples were employed in two dogs. Tumor types included mast cell tumors (n=7), fibrosarcomas (n=2), and one each of an apocrine gland adenocarcinoma, synovial cell sarcoma, mammary gland carcinoma, hemangiopericytoma, salivary gland adenocarcinoma, thyroid carcinoma, and undifferentiated sarcoma. Tumor locations included the extremities (n=5), the cervical area (n=4), thoracic wall (n=3), abdominal wall (n=2), and head (n=2).

The number of surgical clips placed at surgery ranged from two to 78 (median eight; mean 14). The largest number of hemoclips (n=78) inserted was in a dog (case no. 14) that underwent surgical resection of a thyroid carcinoma, and most of the clips were placed for hemostasis. All dogs except one (case no. 2) had a minimum of five metallic markers placed to delineate the surgical bed. In one dog (case no. 5), wound complications after incomplete resection of a grade II mast cell tumor required surgical debridement, and the number of hemoclips at the time of wound healing was less than the number seen immediately postoperatively. In another dog (case no. 3), wound dehiscence occurred 5 days postoperatively, and there was delayed primary wound closure. At the time of the second set of films, one less hemoclip was identified.

On the immediate postoperative films, the distance (i.e., separation) between the surgical clips ranged from 1.3 to 28.3 cm (median 6.5 cm; mean 7.8 cm), based on a total of 62 measurements. The difference in separation of the hemoclips (i.e., shift) was obtained in two orthogonal directions (e.g., cranial-caudal and dorsoventral, see Table) for 15 dogs and in only one direction for the dog (case no. 2) with two surgical staples. The difference in separation of the surgical clips between the two sets of films ranged from -1.8 to +3.2 cm (median 0.25 cm; mean 0.3 cm), based on a total of 62 measurements. The relative percentage change in separation of the hemoclips based on the same 62 measurements ranged from −24.1% to +55.6% (median +4%; mean +5.3%). In a small (n=3) subset of dogs, a large shift (range of 2.4 to 3.2 cm) in surgical clip position was identified radiographically.

Shifts in surgical clip location were considered to arise from a difference in positioning of the dog (nine dogs), an artifactual shift (two dogs), or an absolute shift from migration of the surgical clips (three dogs). The cause of the shift was undetermined in two dogs. An example of a positioning shift was a dog with a dorsal cervical grade II mast cell tumor [Figure 1A], in which the hemoclips were all located dorsal to the cervical vertebrae in the immediate postoperative film. In the repeat radiograph [Figure 1B], there was a ventral shift in the hemoclips, with one clip overlying the cervical spinal cord as a result of positioning. An example of an artifactual shift included a dog in which an enlarged urinary bladder on the second set of radiographs caused abdominal distention and a large shift in the hemoclips [Figures 2A, 2B]. An absolute shift in surgical clip position was identified in a dog (case no. 3) that had clustering of five hemoclips at the most cranial extent of the surgical field at the time of the second set of radiographs [Figures 3A, 3B]. In two dogs (case nos. 5, 16), there was increased soft-tissue swelling associated with the surgical site immediately postoperatively that was not identified on the recheck radiographs after wound healing, so a subsequent decrease in separation of the surgical clips was found on the recheck radiographs in four of the eight measurements made for the two dogs.

Seven dogs had skin staples that were visualized on the first set of radiographs. In three dogs, the skin staples made radiographic assessment of hemoclip position more difficult, because they were superimposed on the hemoclips. For example, in a dog with a mast cell tumor of the caudal thigh, the postoperative films [Figure 4A] revealed numerous skin staples used in the skin flap that were placed for wound closure and for anchoring a drain. In the second set of films [Figure 4B], the hemoclips were easier to identify. In four dogs, no difficulty was encountered in determining hemoclip location as opposed to skin staples. An additional finding was that in some of the dogs, the skin staples used to close the incision were found to be in a more lateral position than the surgical clips on the immediate postoperative radiograph. The skin staples did not appear to provide an accurate delineation of the location of the underlying tumor bed in these dogs.

Discussion

Surgical clips may be of benefit in cancer patients for identifying the postoperative surgical bed for the sake of establishing radiation treatment fields. The original intent of the study was to determine if a shift in clip position occurred as the surgical wound healed, which could impair accurate delineation of the tumor bed. Although it was possible to identify an absolute shift in hemoclip position in a small number of cases in this study, a substantial component of the shift in hemoclip position appeared to arise from positioning changes that occurred between the two sets of radiographs. Further statistical analysis of the measurements acquired in this study was not performed, as it was thought that the analysis would not provide useful information because of the confounding factor of animal positioning.

Dogs were under general anesthesia for the postoperative radiographs but were not anesthetized for the recheck films, and this may have contributed to the difficulty in accurately repositioning the animals. In addition, for surgical sites involving skin or subcutaneous tissues, tissue looseness played a role in hemoclip position when compared to clips inserted in more fixed, deeper structures. A substantial shift in the position of hemoclips relative to normal tissues underscores the importance of exactly reproducing the position of the animals for each radiation treatment. Such positioning may be aided by the use of repositioning devices.⁶ Repositioning devices such as vacuum lock mattresses and bite block positioners are used to position animals for radiation therapy on a regular basis, and it is recommended that the positioning device(s) be dedicated for use on one animal for the duration of a specific animal’s radiation treatment course.

Radiographic assessment of surgical clip number and position immediately postoperatively is worthwhile, as surgical clips can shift as the site heals, or the number may change with wound dehiscence or debridement [Figures 5A, 5B]. Loss or shifts in the clips can alter the presumed extent of the tumor bed. Soft-tissue swelling in the immediate postoperative setting may alter the apparent position of surgical clips relative to radiographs obtained after resolution of the swelling. Although postoperative swelling could theoretically result in a greater separation of the clips compared to their separation after wound healing, the mean percentage change in separation was actually 5.3% higher at the time of the second set of films in this study.

Orthogonal radiographs (as opposed to one view) were important in visualizing surgical clip position, in locating the tumor site, and in defining the location of potentially normal tissues that would be irradiated. A more thorough assessment of the surgical site, delineation of the tissue planes that were disrupted, and the potential location of residual tumor would require cross-sectional diagnostic imaging such as with CT or magnetic resonance (MR) imaging.

An additional finding in the study reported here was that immediately following surgery, skin staples were sometimes found in a more lateral position than the surgical clips. This is important to consider, in that setting up a radiation treatment field based on a surgical scar alone may not accurately predict the location of the underlying tumor bed. A surgical approach may not directly overly the tumor, and the result can be a more lateral location of the surgical scar. When establishing radiation treatment fields, the authors concluded that using a combination of hemoclip position in conjunction with identification of the surgical scar (marked with wire during a simulation or portal film) would more accurately delineate the location of any residual microscopic neoplasia than would location of the surgical incision alone.

Based on this pilot study, a number of recommendations can be made that may enhance the utility of surgical clips during radiation treatment planning. Immediate postoperative radiographs should be made to assess hemoclip positioning. These films can then be compared with radiographs taken for establishing radiation therapy fields, because some clips may shift independent of positioning. Precision in repositioning animals between radiographic examinations is paramount and may be aided by the use of general anesthesia in conjunction with repositioning devices.

Conclusion

Radiation treatment fields in animals are typically defined based on a specific distance or margin (e.g., 3 cm) from the surgical scar. Because of positioning problems in this study, it was not clearly shown that placement and radiographic visualization of hemoclips in conjunction with the scar location are of potential benefit in designing radiation treatment fields. Future studies are needed to identify the dimensions of the tumor bed based on the location of hemoclips as compared to a set distance around a surgical scar. Additional studies exploring the utility of hemoclips in radiation treatment planning are also necessary to address variations based on anatomical site and to further define sites where surgical clip migration may be an issue.