Indirect lymphography for sentinel lymph node detection in dogs with mast cell tumors

Abstract

Objective:

The objective of this report is to describe the technique and diagnostic utility of indirect lymphography (IL) using water-soluble contrast for sentinel lymph node (SLN) mapping in dogs with mast cell tumors.

Animals:

Fifty-three dogs with 59 mast cell tumors were included.

Procedure:

Medical records were retrieved for dogs with a cytological diagnosis of mast cell tumor which also had IL performed for lymph node mapping. Dogs were excluded when surgery had been performed before presentation. Images were reviewed by a Board-certified radiologist for uptake of contrast within the sentinel lymph node.

Results:

Lymphography studies from 34 tumors (57.6%) were diagnostic (clearly identifiable lymphatics and sentinel lymph node). Lymphography studies from 12 tumors (20.3%) were partially diagnostic (identifiable lymphatics, but sentinel lymph node not highlighted). Lymphography studies from 13 tumors (22%) were non-diagnostic. Indirect lymphography studies were interpreted as either diagnostic or partially diagnostic in 77.9% of tumors.

Conclusion:

The results indicate that IL is a simple, available technique to allow for identification of a sentinel lymph node in dogs with mast cell tumors.

Clinical relevance:

Indirect lymphography is a simple and widely accessible technique for SLN mapping in dogs with mast cell tumors, particularly for the general practice environment.

Introduction

Mast cell tumors (MCTs) are the most common skin tumor in the dog, representing ~16 to 21% of all cutaneous tumors (1,2). Staging animals with mast cell disease is a critical element in determining prognosis and effecting loco-regional control in canine MCTs (3–7). Staging tests typically use diagnostic imaging studies such as abdominal ultrasound to evaluate for lymphadenopathy and changes to the spleen and/or liver, fine-needle aspiration (FNA) with cytology of the spleen and liver, and computed tomography for surgical planning and potentially lymph node evaluation. Lymph node status in dogs with MCTs is one of the most important components of staging (8). Multiple studies have confirmed the benefit of identification and extirpation of metastatic lymph nodes with the addition of chemotherapy for dogs with MCTs (8–10,11–13). However, identification of lymph node metastasis is complicated by the fact that lymph node size is a poor indicator of metastatic status, and FNA sampling does not always reflect the actual status of the lymph node (14,15).

The sentinel lymph node (SLN) is defined as the first lymph node to receive drainage from a tumor site (16). Lymphoscintigraphy, indirect computed tomographic lymphography with aqueous contrast, and radiography using lipid soluble iodinated contrast have been described for identification of SLN(s) in various canine tumors (12,17–20). Many of these methods for SLN identification are limited to academic institutions with advanced imaging modalities and nuclear medicine capabilities. Techniques that are more widely applicable to various practice settings allow for staging patients that would not otherwise see a specialist. The use of lipid soluble iodinated contrast and radiography has been described, but it is often cost prohibitive and requires a 24 to 48 h delay from time of injection to optimal identification of the SLN (18).

A pilot study performed at the authors’ institution described indirect radiographic lymphography (IL) using aqueous contrast in clinically normal dogs (21). This study demonstrated that IL was safe and able to delineate lymphatic channels and lymph nodes draining the evaluated site. Another study compared IL and lymphoscintigraphy for SLN identification (22). An SLN was identified in 8 of 8 dogs using lymphoscintigraphy and 7 of 8 dogs using IL. Agreement between the results of lymphoscintigraphy and IL studies showed complete match in 3 dogs, partial match in 4 dogs, and no match in 1 dog (22). This demonstrates that SLN detection can vary by technique, but from this study 7 of 8 dogs had some degree of concordance in the SLN detected. No SLN mapping technique has proven to be 100% accurate, but a more accessible SLN mapping technique with comparable results to scintigraphy is an attractive option for clients and clinicians without access to centers that have nuclear medicine capability.

The effects of surgery on SLN mapping as detected with lymphoscintigraphy have also been investigated by Hlusko et al (23). That study evaluated the effect of surgery on the canine brachium in a simulated tumor model to determine if surgery altered identification of the SLN postoperatively. Conclusions of Hlusko et al (23) were that SLN identification occurred faster postoperatively, and agreement or partial agreement between the results of the preoperative and postoperative lymphoscintigraphy studies was observed in 8 of 8 dogs. Additional studies are required to compare preoperative and postoperative SLN mapping patterns in tumor-bearing dogs. However, the study by Hlusko et al (23) provided preliminary information regarding the effect of surgery on SLN identification.

The aim of this retrospective, descriptive study is to describe the procedure and to establish the value of IL for the identification of the SLN in dogs with spontaneously occurring MCTs using aqueous iodinated contrast. Cytologic and histopathologic status of identified lymph nodes is also reported if available.

Materials and methods

Case selection and medical records review

Computerized medical records from 2018 to 2020 were searched at a single institution to identify dogs that met the inclusion criteria. Dogs were selected if they had a cytological diagnosis of MCT and had IL performed for SLN mapping. Dogs were included if they had de novo disease. Previous surgery at the site was a criterion for exclusion. Data collected from the medical record included breed, age, sex, weight, tumor location, and cytologic results from the tumor and SLN, and the SLN mapping results.

Lymphography technique

Dogs were sedated at the discretion of the attending clinician prior to lymphography. Indirect lymphography was performed in all cases using 3 to 4 mL of aqueous iodinated contrast [Isovue, Iopamidol injection 76%; Bracco Diagnostics, Monroe township, New Jersey, USA), 755 mg/mL or Omnipaque (iohexol); BIPSO GmbH, Singen, Germany, 350 mg/mL] injected peritumorally using a 4-quadrant technique and a 25-gauge needle. The injection protocol was based on previously described techniques (15,18,24). To decrease the risk of mast cell degranulation and subsequent tumor seeding, peri-tumoral injections were placed 0.5 to 1 cm away from the border of the mass (18). The amount of contrast injected was evenly distributed among the 4 injection sites. Following the initial injection, each subsequent injection was done immediately and sequentially. The injection sites were gently massaged for 10 to 15 s. The person performing the injection was not always the same. Lateral projections were performed immediately after injection of contrast material and continued every 1 to 2 min until lymphatic vessels were radiographically identified. Orthogonal views were immediately performed when the SLN was identified on the lateral projection. Most cases were discontinued after 5 min if no lymph node was noted. Due to the retrospective nature of the study, timing of the radiographs was not standardized. Fine-needle aspirates (FNA) were obtained of the SLN after identification. This was done by direct percutaneous aspiration when accessible on palpation or with ultrasound guidance. Ultrasound guidance was used when lymph nodes were not palpable or were intra-abdominally positioned. Because multiple lymph nodes can be present in a lymphocentrum, FNA was used for a single identifiable or palpable lymph node. When more than one tumor was evaluated in a dog, each tumor was individually imaged, and studies were done sequentially. Tumors were imaged simultaneously only if they could be included in the same radiographic study.

Data interpretation

All radiographic images were reviewed by an ACVR Board-certified radiologist and evaluated for uptake of contrast within the SLN. The radiologist was blinded to clinical information. Studies were considered diagnostic when the SLN was identified by tracing afferent lymphatic vessels from the tumor towards a contrast enhanced LN (Figure 1). Studies were considered partially diagnostic when there was contrast movement within lymphatic vessels approximately 50% of the distance from the site of injection towards a known lymphocentrum, but a lymph node was not definitively visualized with contrast. Studies were considered non-diagnostic when no or minimal lymphatic vessels were visible and a SLN was not identified. The data were evaluated using descriptive statistics.

Figure 1.

Lateral radiographic projection demonstrating a diagnostic IL image.

Results

Fifty-three client-owned dogs met the inclusion criteria. Of the 54 dogs, 6 had 2 MCTs each; therefore, 59 tumors were included in the study. Dogs included 19 castrated males, 2 intact males, 30 spayed females, and 2 intact females. Breeds included mixed breed dogs (n = 13); Labrador retrievers (n = 10); Boston terriers (n = 3); boxers (n = 3); bulldogs (n = 3); Chihuahuas (n = 2); American Staffordshire terriers (n = 2); pugs (n = 2); and one of each of the following breeds: Boykin spaniel, border collie, schnauzer, golden retriever, German shepherd dog, rat terrier, Great Dane, Chinese shar-pei, Alaskan malamute, cocker spaniel, mastiff, pit bull, Weimaraner, Yorkshire terrier, poodle, and Rhodesian ridgeback. The median age was 9.5 y (range: 5 to 14 y). The median weight was 29.6 kg (range: 3.75 to 55.5 kg).

Overall, 34 studies (57.6%) were diagnostic, 12 were partially diagnostic (20.3%), and 13 (22.0%) were non-diagnostic. In the diagnostic and partially diagnostic studies, lymph node or channel enhancement occurred immediately in 19 lymphography studies and at a mean of 3.5 min (range: 1 to 18 min) in the remaining studies. Non-diagnostic studies were ended at a mean of 24 min (range: 8 to 90 min) and were always ended because of an overt lack of contrast movement. Pre-contrast radiographs were taken in 23 lymphography studies. In cases in which pre-contrast radiographs were taken, the pre-contrast images were compared to the post-contrast images to identify more subtle contrast uptake in multiple cases. Mild peritumoral swelling, after injection, was noted in some cases, but unexpected reactions or adverse events related to lymphography were not observed.

A secondary goal of this study was to detail the cytological and histopathological findings of the 56 lymph nodes sampled with FNA via palpation (n = 21) or with ultrasound guidance (n = 35) (Table 1). Diagnostic lymphography studies were obtained for 34 tumors in 33 dogs. Metastatic disease was not observed with histopathology or cytology in 18 tumors. Metastatic disease was confirmed with either histopathology or cytology in 12 of those tumors. In 4 tumors, either a LN was not evaluated, or the information was unavailable. Cytology and histopathology were concordant for 4 tumors. Four tumors were suspected to have metastatic disease, but mast cell tumor was only confirmed histopathologically in 1 of these. There were 5 tumors in which 2 or more lymph nodes were evaluated for lymphography and cytology/histopathology.

Table 1.

Tumor, number, location, lymphography study assessment, lymph node identified, tumor grade, and node status for diagnostic and partially diagnostic studies.

Tumor	Location	Diagnostic quality	Presumed SLN	Tumor grade	LN cytology	LN histopathology
1	Left brachium, pectoral	Diagnostic	Left superficial cervical	Low grade	NA	−
2	Right pinna	Diagnostic	Right superficial cervical	Low grade	−	NA
3	Right lateral brachium	Diagnostic	Right superficial cervical	Low grade II	NA	−
4	Right pinna	Diagnostic	Right superficial cervical	Low grade	NA	−
5	Left flank	Diagnostic	Left inguinal	NA	+	NA
6	Right scapular region	Diagnostic	Right superficial cervical	NA	NA	NA
7	Left gluteal region	Diagnostic	Left MILN	SQ, no grade	NA	−
8	Left proximal scapula	Diagnostic	Left superficial cervical, Left axillary	NA	+, +	NA
9	Left gluteal region	Diagnostic	Left MILN	Low grade	−	NA
10	Left scrotal region	Diagnostic	Left inguinal	Low grade II	+	+
11	Right ear	Diagnostic	Right superficial cervical	High grade III	+	+
12	Right cervical/shoulder	Diagnostic	Right superficial cervical	NA	−	NA
13	Right prepuce	Diagnostic	Right inguinal, Right MILN	High grade	NA	+, +
14	Left caudolateral thoracic wall	Diagnostic	Left axillary	Low grade	Increased mast cells	−
15	Left antebrachium	Diagnostic	Left superficial cervical	NA	+	NA
16	Left front dorsal foot	Diagnostic	Left superficial cervical	Low grade II	NA	−
17	Left caudal thigh/perivulvar	Diagnostic	Left inguinal	Low grade I	−	NA
18	Right distal tibia	Diagnostic	Right popliteal, Right inguinal, Right MILN	NA	+, +, +	NA
19	Left caudal stifle	Diagnostic	Left popliteal, Left inguinal	High grade III	−, +	NA, +
20	Right pinna/skull	Diagnostic	Right superficial cervical	NA	−	NA
21	Left ear base	Diagnostic	Left superficial cervical	NA	−	NA
22	Right chest/caudal axilla	Diagnostic	Right axillary, Right superficial cervical	NA	NA, +	NA
23	Left distal lateral tibial	Diagnostic	Popliteal	NA	Increased mast cells	NA
24	Right lateral thigh	Diagnostic	Right inguinal	Low grade II	Increased mast cells	+
25	Left dorsolateral flank	Diagnostic	Left axillary	NA	−	NA
26	Mid perineal region	Diagnostic	Bilateral inguinal	NA	−On left inguinal, increased mast cells on right	NA
27	Left flank/proximal thigh	Diagnostic	Left MILN	NA	NA	NA
28	Med aspect of left proximal tibial	Diagnostic	Right MILN	NA	+	NA
29	External left pinna base	Diagnostic	Left superficial cervical	NA	+	NA
30	Left hip/epaxial region	Diagnostic	Left MILN	NA	NA	NA
31	Left thoracic limb, digits	Diagnostic	Left superficial cervical	Low grade	−	NA
32	Right ear base	Diagnostic	Right superficial cervical	SQ, no grade	−	NA
33	Left shoulder	Diagnostic	Left superficial cervical	NA	−	NA
34	Right side of chest wall	Diagnostic	Right superficial cervical	Low grade II	NA	NA
35	Left proximal antebrachium	Partially diagnostic	Left superficial cervical	SQ, no grade	NA	+
36	Left proximal antebrachium	Partially diagnostic	Left superficial cervical	NA	+	NA
37	Left caudolateral thigh	Partially diagnostic	Left superficial cervical	SQ, no grade	+	+
38	Left proximal antebrachium	Partially diagnostic	Left inguinal	NA	NA	NA
39	Left flank	Partially diagnostic	Left inguinal	NA	NA	NA
40	Left perineal mass, vulva	Partially diagnostic	Left inguinal, MILN	NA	+, +	NA
41	Right lateral thigh	Partially diagnostic	Right MILN	Low grade II	Increased mast cells	+
42	Right caudal thigh	Partially diagnostic	Right MILN, Right inguinal	NA	+ MILN, increased mast cells inguinal	NA
43	Right caudal thigh	Partially diagnostic	Right inguinal, Right MILN	NA	+	+
44	Right caudal thigh	Partially diagnostic	Right inguinal, Right MILN	Low grade	Increased mast cells	−
45	Right distolateral humeral region	Partially diagnostic	Right axillary, Right superficial cervical	NA	−	NA
46	Prepuce	Partially diagnostic	Right inguinal	NA	NA	NA

Grade is based on the Patnaik Grades I, II, III and the Kiupel 2-tier grading system.

Mets — Metastatic disease; SQ — Subcutaneous; LN — Lymph node; MILN — Medial iliac lymph node; NA — Not applicable.

Cytology or histopathology confirmed metastasis in 7 lymph nodes from the partially diagnostic studies and were concordant in confirming metastasis in 4. One lymph node was positive on cytology but not removed for histopathology. Two cytologically confirmed metastatic SLNs were removed and were negative for metastasis on histopathology. Two tumors had more than 1 lymph node evaluated. Thirteen lymphography studies were non-diagnostic.

Twenty-six MCTs from 24 dogs were removed along with the SLN identified by IL. Some owners declined surgery after staging or did not return for recommended mass removal. Seventeen of the 26 tumors (65.4%) had diagnostic lymphography studies and 4 (15.4%) had partially diagnostic studies. The remaining 5 tumors (19.2%) had non-diagnostic studies. Fourteen (53.8%) tumors were low grade on the Kiupel system (Patnaik Grade I or II). Nine of these (64.3%) low-grade tumors did not have metastatic lymph nodes on cytology or histopathology, whereas 3 (21.4%) of the low-grade tumors had lymph node metastasis, and the remaining 2 (14.3%) tumors either had low cellularity on cytology of the lymph node, or the medical record did not show evidence of cytology or histopathology. Three tumors (11.5%) were high grade on the Kiupel system (Patnaik Grade III), and all 3 had lymph node metastasis. Four (15.4%) MCTs were subcutaneous, and therefore were not graded but 2 (50%) of them had metastatic lymph nodes on histopathology. Four of the 5 tumors with non-diagnostic studies were surgically excised. In 2 of these cases, the regional lymph node was not excised. In the remaining 2 cases, the regional lymph node was cytologically normal and not removed.

Discussion

The results of this study demonstrate that IL can be used as a diagnostic tool for SLN mapping in dogs with MCT. Indirect lymphography is particularly useful when more advanced SLN mapping techniques are not an option. Indirect lymphography involves peritumoral injection of water-soluble contrast and serial imaging for identification of the SLN. This study evaluated 53 dogs with 59 MCTs and discovered IL studies diagnostic or partially diagnostic in 77.9% of tumors imaged.

Lymph node status in dogs with MCT is important because treatment of metastatic LNs can improve survival (25–30). Approximately 55 to 96% of undifferentiated or high grade MCTs will metastasize (4,6). The most common metastatic pattern recognized for MCT is dissemination to the draining lymph node, followed by the spleen and liver, with rare involvement of the lungs, other visceral organs, and the bone marrow (1,4,31,32). The behavior of MCTs can be unpredictable, independent of grade, therefore staging is important to determine the extent of the disease. If the tumor is amenable to curative intent surgical excision, and there is no evidence of negative prognostic indicators at the time of diagnosis, then FNA of the SLN may be the only necessary staging test prior to surgery (8). There is often ambiguity related to lymph node palpation, including even identification of the SLN. Ferrari et al (29) reported that histologically detectable metastatic disease can be seen in half of cases with non-palpable and normalsized regional lymph nodes in dogs with MCT. Sentinel lymph node mapping should be considered when working up any MCT (14,15,18). This can sometimes be considered impossible in the first opinion practice environment in which access to lymphoscintigraphy and CT is often not available.

In our study, IL was used for identification of the SLN, in the partially diagnostic and diagnostic imaging studies. Although not the primary aim of the study, cytological and histological evaluation of the LNs extirpated, from the 24 dogs treated surgically, were also evaluated. Of the tumors that were surgically removed, 4 were subcutaneous and 2 metastasized to the SLN. Although this is a small number of subcutaneous tumors, these findings underscore the importance of SLN mapping despite the previously reported less aggressive biologic activity of subcutaneous MCTs (33,34). The low-grade MCTs in this study (Patnaik Grades I and II) had a metastatic rate of 21.4%. A previous paper reported metastasis in 5.8% (3 of 52) of dogs with Patnaik Grade 1 tumors, 16.5% (48 of 291) of dogs with Patnaik Grade 2 tumors, and 14.9% (44 of 295) of dogs with Kiupel low-grade tumors (35). The 3 high-grade (Patnaik Grade III) MCTs in the present study all had metastasis to the SLN. This is similar to the reported metastatic rate for undifferentiated MCTs, which is up to 96% (4,6). In a recent study, Lapsley et al (30) compared the effect of SLN histology versus locoregional lymph node (LRLN) FNA cytology. In that study, the SLN differed anatomically from LRLN in 5 of 18 scans (27.8%). Metastases were detected by histology in 9 of 20 (45%) SLN compared with in 1 of 20 FNA of LRLN (5%). These results did change the stage and adjunctive treatment recommendations in 8 of 20 (40%) tumors. Only 6 of 19 (31.6%) LRLN FNA samples were diagnostic (30). The authors were able to show that in 1/4 of tumors, lymph nodes were consistently identified with indirect computed tomographic lymphangiography, and this differed from the LRLN. They concluded that histopathological examination of SLN altered recommendations in half of the dogs compared with the previous standard of care (30). Our data showed that at least 10/24 (41.7%) of the LNs removed were positive for metastatic disease.

Severe complications with IL have not been observed in small animals, but it is documented in human medicine that allergic reactions or pulmonary embolism may occur (16). As with other previous lymphography studies in veterinary medicine, mild peri-tumoral swelling was noted in our study, but no major adverse events or complications were documented. Ultrasound guidance may be necessary for FNA of the lymph nodes, depending on the location and size of the identified lymph node. Sedation was typically used for the IL study. Based on the authors’ clinical experience, an injection volume < 4 mL can yield inconsistent results. Cases imaged in the earlier period included in this study were injected with 3 mL; however, this was increased as subjectively higher quality studies were performed with 4 mL. A pre-contrast image can be helpful for comparison purposes and imaging should begin immediately after injection. Pre-contrast radiographs were performed in selected cases at the discretion of the clinical team.

In a previously reported lymphography study by Mayer et al (24), radiography showed lymph node enhancement was consistently observed 24 h after injection with iodized oil (IO) in 25 cases. In that study, re-injection of IO was necessary in 4 of 25 cases, in which diagnostically satisfying contrast uptake was observed 24 h after the second injection. In our study, when diagnostic or partially diagnostic lymph node enhancement was noted, this occurred immediately or within a mean of 3.5 min, rather than the previously reported 24 h (24). Non-diagnostic studies were ended at a mean of 24 min. Our data indicated an overlap in the time ranges for diagnostic or partially diagnostic and non-diagnostic studies. The wide range of timing between the studies could be attributed to differences in the massage technique used, since the person performing the technique was not always the same. The exact amount of time to continue a study or to determine when to end a partially or non-diagnostic study is not clear. A prospective study is needed to determine the most appropriate time needed between imaging, as well as how to determine when to end the IL study.

Computed tomography lymphography for SLN mapping has been reported in dogs with cancer (15,16,18,19,22). Computed tomography IL may be more sensitive than radiographic IL, but compassion studies are needed to evaluate this. Computed tomography IL requires advanced cross-sectional imaging capabilities that are not available to most first opinion practices, whereas radiographic IL uses readily available equipment and simple techniques.

Retrospective studies have inherent limitations. Also due to the retrospective nature, the exact timing of all imaging exposures and sequences was not standardized. Another limitation identified in our study was than not every SLN mapped patient was treated surgically. There were various reasons for why this occurred. As such, tumor grade and lymph node status were not determined for every dog; however, the main goal of this report was to detail the findings of IL. A third limitation can be noted for the subjective nature of radiographic interpretation. In addition, an inherent obstacle to SLN mapping studies is that it can be difficult to ensure the opacified node (the SLN) is sampled or removed when multiple nodes are present in the regional lymphatic basin. The corresponding detection rates between SLNs radiographically identified and those discovered intraoperatively with methylene blue was reported by Brissot and Edery (18). Sentinel lymph nodes were successfully identified in 29 out of 30 tumors using the radiographic lymphography technique (96.7%) and agreement between the blue-stained SLNs and those detected by radiographic lymphography was seen in 88.5% (23/26) of cases (16,18). In the authors’ practice, all the LNs of the observed lymphocentrum are extirpated and submitted for histopathology.

In conclusion, the results of this study indicate that the use of water-soluble contrast and radiography can be a simple, readily available technique for identification of SLNs in dogs with MCTs. Indirect lymphography for SLN mapping is a more easily accessible option than CT IL or lymphoscintigraphy. Prospective clinical studies are needed to validate indirect lymphography as a staging tool for dogs with MCT.