Concordance between preoperative imaging methods in patients with limb lymphedema undergoing supermicrosurgical lymphaticovenular anastomosis

Abstract

Objective

Supermicrosurgical lymphaticovenular anastomosis (LVA) is increasingly being recognized as a first-line treatment of limb lymphedema because it is minimally invasive and highly effective. Lymphoscintigraphy and indocyanine green (ICG) lymphography are the two most commonly performed diagnostic imaging examinations to establish the indication and plan the procedure for patients affected by limb lymphedema. In a small group of patients, the information between these two imaging tools can be discordant, showing different anatomical drainage pathways or the absence of drainage and dermal backflow in one examination and valid drainage pathways in the other. The purpose of this study is to examine the types of possible discrepancies between lymphoscintigraphy of the superficial system and ICG lymphography and to describe the surgical outcomes after LVA for patients presenting with such discrepancies.

Methods

We retrospectively reviewed the data of all patients who underwent LVA for upper or lower limb lymphedema between July 2015 and July 2023. From this series, we identified a group of patients with nonconcordant imaging results from lymphoscintigraphy and ICG lymphography before lymphatic surgery. Nonconcordant findings were described in terms of “pattern discordance” and “pathway discordance.” The surgical outcome was measured by the change in the mean circumference of the limb after surgery. The changes between the preoperative and postoperative limb measures were analyzed using the Student t test. P values < .05 were considered significant.

Results

A total of 28 patients with limb lymphedema exhibited inconsistencies between preoperative lymphoscintigraphy of the superficial system and ICG lymphography. Among these patients, 14 experienced pattern discordance, 13 had pathway discordance, and 1 patient had both. After LVA, we observed a significant reduction in the average circumference of the affected limb in the analyzed group.

Conclusions

The discrepancy in the information between lymphoscintigraphy and ICG lymphography in the preoperative study of patients affected by limb lymphedema is rare but possible. This phenomenon is still not fully explained; however, our results suggest that it does not correlate with the outcome of supermicrosurgical LVAs.

Article Highlights.

• •Type of Research: A single-center, retrospective cohort study
• •Key Findings: A total of 28 patients with limb lymphedema exhibited inconsistencies between preoperative lymphoscintigraphy of the superficial system and indocyanine green lymphography. Among these patients, 14 experienced “pattern discordance,” 13 had “pathway discordance,” and 1 patient had both. Supermicrosurgical lymphaticovenular anastomosis was feasible in all cases, achieving a positive outcome in the series of patients examined.
• •Take Home Message: The discrepancy of information between lymphoscintigraphy and indocyanine green lymphography in the preoperative study of patients affected by limb lymphedema is rare but possible. This phenomenon is still not fully explained; however, our results suggest that it does not correlate with the outcome of supermicrosurgical lymphovenous anastomoses.

Lymphaticovenular anastomosis (LVA) is widely considered as a primary surgical option for limb lymphedema treatment in many centers.¹ To achieve successful outcomes with this established technique, two crucial factors are the accurate determination of patient eligibility and the planning of incision sites using the preoperative instrumental assessment.² In standard cases, our approach involves lymphoscintigraphy of the superficial system and a “rest, stress, late” protocol to establish the indication for LVA.³ Lymphoscintigraphy can assess both the superficial and the deep lymphatic systems and can confirm the diagnosis of lymphedema.^4,5 When focused on the superficial lymphatic system, lymphoscintigraphy can determine whether functional lymphatic vessels suitable for LVA are present in each segment of the limb and identify their course.⁶ However, the spatial resolution of this examination is low; thus, merging the information with indocyanine green (ICG) lymphography and ultrasound is necessary for precise planning of the incision sites. In our experience, when lymphoscintigraphy aimed at the superficial lymphatic system and ICG lymphography reveal the absence or a marked insufficiency of lymphatic vessels, this indicates how the obstruction to lymphatic flow might have led to sclerosis of the lymphatic vessels,^7,8 often rendering LVA not feasible. This rational approach is based on the principle that lymphoscintigraphy aimed at the superficial system and ICG lymphography are comparable in evaluating lymphatic function and visualizing peristaltic lymphatic vessels because they target the same superficial lymphatic system.⁹ Although our center uses comparable methods for conducting these two examinations (two distal intradermal injections of technetium colloid and ICG, respectively), and, in most cases, the same lymphatic pathways with similar tracer distribution patterns are visualized, a few patients can exhibit significant inconsistencies between the two techniques during their preoperative imaging workup. The purpose of this study is to describe the possible discrepancies between lymphoscintigraphy aimed at the superficial system and ICG lymphography, hypothesize their causes, and report the surgical outcomes after LVA of these patients.

Methods

We conducted a retrospective observational study, reviewing the data of all patients who underwent LVA for upper or lower limb lymphedema between July 2015 and July 2023. The study was conducted in accordance with the principles outlined in the Declaration of Helsinki. All participating patients provided written informed consent. Patients without a clear indication for LVA based on the diagnostic workup were not included. From this series, we identified a group of patients with nonconcordant imaging results from lymphoscintigraphy and ICG lymphography before lymphatic surgery and included them in the study group. The variables of interest that we collected included age, height, weight, body mass index, smoking habits, comorbidities, previous surgery, cause and duration of lymphedema, preoperative lymphoscintigraphy findings, clinical preoperative and postoperative photographs, intraoperative presence of lymphatic vessels, number of anastomoses, complications, outcome of surgery, and duration of follow-up. The surgergy outcome was measured by the change in the mean circumference of the limb after surgery. The limb circumference was measured every 4 cm, and the average circumference was calculated before and after surgery. The patients had an average follow-up of 47 months. LVA was performed using a standard supermicrosurgical technique with intima-to-intima coaptation, as described in the literature.^10,11 For each limb, we analyzed the predominant imaging findings in the proximal or distal halves of the limb, relative to the elbow or knee. We considered the findings and any discordance in terms of the pattern or pathways. “Pattern discordance” was defined as a prevalence of diffuse dermal backflow or no transit on lymphoscintigraphy of the superficial system, with ICG lymphography showing linear or ectasic vessels, or vice versa. “Pathway discordance” was defined as a completely different anatomy of the course of functioning lymphatic vessels on lymphoscintigraphy of the superficial system and ICG lymphography.

Before imaging

As we know, lymphedema is an extremely variable condition; thus, we standardized the acquisition protocols for lymphoscintigraphy and ICG lymphography to ensure consistent results. Lymphoscintigraphy was always performed by the same nuclear medicine physician, and ICG lymphography was always conducted by the same surgeon. Both techniques were performed under optimal clinical conditions, following a cycle of complex decongestive therapy and optimized compression garment use. Complex decongestive therapy, used to prepare the patient for preoperative imaging and surgery, followed a well-defined protocol that includes an initial attack phase and a subsequent maintenance phase. The attack phase aimed for total decongestion of the limb through a multicomponent bandaging regimen with low-stretch bandages applied daily for a duration of 7 to 21 days. During the daily change of the bandages, a physiotherapist also performed pressotherapy and lymphatic drainage treatment. On achieving maximum decongestion, usually after a period ranging from 7 to 21 days, depending on the limb involved (upper or lower) and the severity of the edema, the maintenance phase began. This phase started with measuring the limb's circumference at various points, followed by creation of a custom-fitted compressive garment with a flat knit, essential for maintaining the results obtained in the decongestion phase.

Lymphoscintigraphy protocol

Patients were asked to remove the compressive garment before imaging. Topical anesthesia at the injection site was avoided to prevent interference with tracer washout. Two or three intradermal injections were performed using a 25-gauge needle to administer technetium-99m (^99mTc)–labeled colloidal tracer. Aspiration was done to avoid intravascular injection and ensure that a blood vessel was not accidentally punctured. The total activity administered to adult patients was 74 MBq, or 37 MBq per limb and per compartment being investigated, given in single or multiple aliquots, depending on the patient and the anatomical area being studied. Each aliquot consisted of 0.2 mL of injected volume. The manufacturer's guide of the primary commercially available radiopharmaceutical indicated that the effective dose following subcutaneous administration of 110 MBq of ^99mTc nanocolloids in an adult patient was 0.44 mSv, with an absorbed dose of 65 mGy to the lymph nodes and 1320 mGy to the injection site.¹² In the case of the upper limb, intradermal tracer injections were performed in the second and third interdigital spaces (plus or minus the first space) of each hand. For the lower limb, similar injections were given in the second and third interdigital spaces (plus or minus the first space) to study the superficial lymphatic system. In the presence of dermal backflow or lymphatic stasis, one or two subfascial administrations in the retromalleolar or plantar region were used to evaluate the deep lymphatic system. Additional intradermal injections above the lymphatic stop were performed to assess proximal lymphatic drainage if the lymphatic stop persisted. For lymphoscintigraphy, we did not use any classification system for dermal backflow but only noted its presence or absence. A double-head gamma camera with a sizable field of view and a parallel hole collimator with low energy was used to capture images.

The ^99mTc photopeak was set in the center of the energy window (140 keV, 10%). Anterior and posterior planar imaging was obtained with the patient lying supine on the gamma camera bed. The washout rate of the tracer was measured through static acquisition of the injection site immediately after administration to have a preset count.¹³

Immediately after injection, dynamic acquisition focused on the proximal lymphatic drainage basin was performed to track the symmetry of tracer lymphatic transport. Dynamic acquisition focused on the proximal lymphatic drainage basin (ie, groin for a lower limb, axilla for an upper limb) was performed right after the injection to track the symmetry of tracer lymphatic transport. Planar images were acquired either using the whole-body acquisition technique or serial static acquisitions at 20 minutes (for early images) or 90 minutes (for delayed images) after tracer administration.

Another static acquisition at the injection site was performed at the end to calculate the tracer's washout rate. When necessary, the deep lymphatic system of the lower limbs was assessed by taking static images 90 minutes after subfascial tracer administration, ≥48 hours after the study of the superficial system. In case of inadequate tracer migration, delayed static images were obtained for ≤4 hours after injection. Additional information on the lymphatic pathways was obtained through a quick and/or prolonged exercise protocol. For the upper limbs, physical exercises involving the hands (ie, free exercise, handgrip or rubber ball squeezing) were recommended between initial and delayed images. For the lower limbs, ambulation for ≥10 to 20 minutes was done. Planar images were used for both qualitative and semiquantitative analyses. The qualitative evaluation of the initial and delayed images included the description of the tracer lymphatic transport, visualization of lymphatic vessels and their number and caliber, presence of collateral lymphatic vessels and/or dermal backflow, and visualization of lymph nodes in the main draining lymphatic basins. A combination of a qualitative and semiquantitative analysis provided greater diagnostic accuracy compared with a qualitative analysis alone,^14,15 although visual interpretation of the lymphoscintigraphic image pattern was reliable in many cases. A semiquantitative analysis on planar images assessed the tracer washout from the injection site and the uptake ratio between the initial and delayed acquisitions. The transport index was also used. The use of both qualitative and semiquantitative methods helped standardize the lymphoscintigraphy results, especially for the preoperative studies.^16,17

Lymphatic mapping with ICG lymphography and ultrasound

The second step of the preoperative imaging protocol involved lymphatic mapping using ICG lymphography. ICG lymphography was performed the day before surgery, following a well-documented method in the literature that we routinely use to evaluate patients with lymphedema, lymphoceles, or lymphorrea and patients undergoing high-risk surgical procedures that could affect lymphatic drainage.18, 19, 20, 21 During ICG lymphography, we were able to identify the course of the lymphatic vessels and the pattern of dermal backflow. For ICG lymphography, we classified the dermal backflow by dividing it into linear, splash, stardust, and diffuse types. For the primitive forms, we used the classification of proximal DB, distal DB, less enhancement, and no enhancement.22, 23, 24

We marked the course of the two main lymphatic pathways of the limb, which follow the great and small saphenous veins or the cephalic and basilic veins, with dotted lines on the skin.^25,26 We also marked the visible lymphatic pathways in the thigh or arm. Next, we used high-frequency (18-22 Mhz) color Doppler ultrasound to investigate the lymphatic pathways and identify any points of ectasia in the lymphatic vessels and to identify the presence and location of superficial venules suitable for anastomosis near the lymphatic vessels. If we found a reflux-free superficial venule next to the lymphatic vessels, we marked the site for a skin incision for LVA. When ICG lymphography did not show a clear linear or ectasic drainage pathway but only an area of diffuse extravasation, we followed the progression axis of fluorescence with the ultrasound probe (48 MHz) to search for lymphatic vessels or performed another injection of ICG further upstream, according to a previously described system.²⁷

Statistical analysis

The sample is described in its clinical and demographic features using descriptive statistics techniques. Quantitative variables are reported using the minimum, maximum, range, mean, and standard deviation. Qualitative variables are summarized with absolute and percentage frequency tables. The normality of continuous variables was checked using the Shapiro-Wilk test. The pre- and postoperative mean circumferences are reported as the mean ± standard deviation. The changes between the preoperative and postoperative values of the mean circumference of the limbs were analyzed using the Student t test. P values < .05 were considered significant. All statistical analyses were performed with IBM SPSS Statistics software, version 24 (IBM Corp).

Results

A total of 142 patients with complete preoperative and postoperative data available were included in the study. Of these patients, 28 affected by unilateral limb lymphedema (15 upper limb and 13 lower limb) showed inconsistency between lymphoscintigraphy of the superficial lymphatic system and ICG lymphography. Of these 28 patients, 25 had secondary lymphedema and 3 had primary lymphedema. In one case, secondary lymphedema was a result of limb trauma with damage to the lymphatic vessels but no impairment of the lymph nodes. The remaining 24 patients developed lymphedema after surgery and/or radiotherapy for gynecological or breast cancer. Nine of these patients had undergone previous radiotherapy. The average age of the patients was 57.1 ± 14.01 years (range, 20-81 years), with an average body mass index of 25.4 ± 4.6 kg/m² (range, 19-37 kg/m²) and an average disease duration of 5.5 ± 4 years (range, 6 months to 15 years) from initial diagnosis to surgery. Fifteen patients had no significant comorbidities, and one had diagnosed venous insufficiency. Type 2 diabetes mellitus, dyslipidemia, obesity, and hypertension were present in nine patients. Of the 28 patients, 14 had inconsistency in the pattern of lymphatic vessels, 13 had inconsistency in the pathway, and 1 had both pattern and pathway inconsistencies. Specifically, in the upper limb, eight patients had pattern inconsistencies (Fig 1, Fig 2, Fig 3) and seven had pathway inconsistencies (Fig 4). In the lower limb, six patients had pattern inconsistencies, six had pathway inconsistencies, and one had both. The combined discrepancies imply that the lymphoscintigraphy exhibited a mere initial advancement of the anterior pathway of the lower limb, which promptly halted without any further lymphatic progress and that ICG lymphography revealed a highly efficient lymphatic vessel in the lateral region of the leg. All patients underwent supermicrosurgical LVA, with an average of 2.5 anastomoses performed (range, 1-5; Table). In the lower limb, the average circumference was 37.1 ± 2.6 cm before surgery and 36 ± 2.5 cm after surgery. In the upper limb, the average circumference was 24.9 ± 2.2 cm before surgery and 24.2 ± 2.6 cm after surgery. In both groups of patients, the reduction in circumference after surgery was significant (lower limb, P < .002; upper limb, P = .017).

Fig 1.

A typical case of pattern discordance. A, Photograph of an 18-year-old girl affected by lymphedema of the left lower limb, secondary to road accident trauma. The swelling is particularly evident in the leg and ankle. B, Lymphoscintigraphy of the superficial lymphatic system of her lower limbs. Diffuse dermal backflow was present at the level of the left leg (black arrow), and no drainage pathways were evident. At the thigh level, a pathway following the main anterior axis of lymphatic drainage was highlighted (blue arrow), leading to an inguinal lymph node. C, Results of indocyanine green (ICG) lymphography with a well-established, ectasic drainage pathway that started from the lateral surface of the foot, ascended along the leg, splitting into three, and then joining again to reach an extravasation area at the knee. D, Drawings on the skin of the leg of the highlighted lymphatic pathways from ICG lymphography. Dotted lines indicate the lymphatic vessels.

Fig 2.

Intraoperative findings of patient shown in Fig 1. A, Vessels found in an incision made on the leg. The venous vessel had a bifurcation (black arrows), and the lymphatic vessel (white arrow) had ideal characteristics in terms of wall and caliber. B and C, Measurement of the calibers of the vessels chosen for anastomosis. The lymphatic vessel had an outer diameter of 1.1 mm and the vein a diameter of 0.65 mm. D and E, Anastomoses performed (white arrows) in optical and fluorescence imaging. The venous bifurcation made it possible to connect both the distal and the proximal lymphatic stump, obtaining the venous outflow of both the anterograde and the retrograde lymphatic flow.

Fig 3.

Pre- and postoperative photographs 2 years after supermicrosurgical lymphovenous anastomosis (LVA) performed at two incision sites in the leg.

Fig 4.

Photographs showing a typical pathway discrepancy. A, Preoperative photograph of a patient with lymphedema of the left upper limb, secondary to treatment of left breast cancer. B, Photograph of the same patient 2 years after supermicrosurgical LVA performed at three incision sites. C, Preoperative lymphoscintigraphy of the superficial lymphatic system, highlighting the beginning of a lymphatic pathway following the cephalic vein (black arrow) and another following the basilic vein (blue arrow). Lateral (D) and dorsal (E) views of the affected forearm where two dorsal lymphatic pathways were identified using indocyanine green (ICG) lymphography, anatomically discordant with that shown by lymphoscintigraphy. The identified lymphatic vessels were marked on the skin with a blue marker using dotted lines.

Table.

Summary of patient data

Year of surgery	Patients, No.	Average No. of incisions	Outcome	Complications
2015	1	5	Improvement, 1; neutral, 0; worsening, 0	None
2016	1	2	Improvement, 1; neutral, 0; worsening, 0	None
2017	7	3	Improvement, 4; neutral, 3; worsening, 0	None
2018	6	3.33	Improvement, 4; neutral, 1; worsening, 1	None
2019	3	2.5	Improvement, 2; neutral, 0; worsening, 1	None
2020	2	2	Improvement, 2; neutral, 0; worsening, 0	None
2021	4	2.5	Improvement, 4; neutral, 0; worsening, 0	None
2022	2	1.5	Improvement, 2; neutral, 0; worsening, 0	None
2023	2	2	Improvement, 2; neutral, 0; worsening, 0	None

Discussion

To ensure a successful LVA procedure, it is important to identify properly functioning lymphatic vessels that can effectively pump lymph into the venous system. To optimize the surgical outcomes, reduce the need for additional procedures, minimize patient discomfort, and reduce resource costs, it is important to determine the presence of these vessels through appropriate instrumental assessment before surgery.²⁸ Lymphoscintigraphy, which focuses on the surface level lymphatic system, can evaluate the drainage function of the same lymphatic pathways that are targeted during preoperative planning for LVA. Thus, this examination is often useful in determining whether LVA is appropriate for an individual patient. In our approach, if there are no visible functioning lymphatic vessels, LVA is often not feasible with positive results, and alternative procedures such as lymph node transfer or liposuction should be considered. Our perspective aligns with widely accepted literature that does not recommend LVA for advanced-stage lymphedema.^29,30 In contrast, early-stage lymphedema refers to the presence of functioning lymphatic vessels that have not yet become sclerotic due to lymph stasis and obstruction and are suitable for LVA. Although fibrosis of soft tissues and damage to the lymphatic vessels can be suspected in patients with certain clinical features such as severe edema, obesity, previous radiotherapy, previous groin dissection, or prolonged illness,31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42 there is no definitive correlation between the level of sclerosis and a patient’s clinical presentation. Therefore, instrumental assessment is important in determining the appropriate course of action.

Since our first LVA procedure in 2015, we have developed a protocol for determining the indication for LVA using lymphoscintigraphy to evaluate the superficial lymphatic system. Before surgery, we also perform ICG lymphography and ultrasound to map out the lymphatic and venular pathways and identify areas of lymph vessel dilation, which helps us plan the incision sites. This approach has allowed us to identify what we refer to as “pathway discordance,” in which lymphoscintigraphy indicates different courses of lymphatic vessels compared with ICG lymphography. In 2017, we expanded the use of ICG lymphography beyond preoperative planning and started using it to supplement the information obtained from lymphoscintigraphy in determining the need for LVA. We adopted this double approach only for patients who showed a complete absence of lymphatic vessels with diffuse dermal backflow or cessation of lymph transit on lymphoscintigraphy but exhibited only mild swelling during physical examination. This is because severe impairment of lymphatic drainage would typically result in moderate or severe swelling.⁴³ Despite the lymphoscintigraphy findings not favoring an indication for LVA, ICG lymphography often showed functioning lymphatic vessels in either one or both halves of the limb. We referred to this situation as “pattern discordance” because ICG lymphography would show a linear or splash pattern but lymphoscintigraphy of the superficial system would show only a diffuse or no pattern. This is clinically significant because it seemingly contradicts our approach of contraindicating LVA based on lymphoscintigraphy, because the ICG lymphography findings suggest otherwise.

We attempted to speculate on this phenomenon by discussing possible causes with nuclear doctors and considering various hypotheses to explain it. The techniques used for lymphoscintigraphy and lymphography have been standardized and comply with widely accepted guidelines practiced by numerous specialists. Therefore, it can be stated that entrusting the examinations to different operators would not significantly affect the outcomes. Thus, despite minimal technical variability that might exist among different individuals, this could not affect the nature or presence of any discrepancies. Pathway discordance can be explained by the patient's position during lymphoscintigraphy. The patient is required to place the limb against the gamma camera plate, which captures static two-dimensional images. Our analysis revealed that evaluating the precise course and position of the lymph vessels in the most lateral or medial region of the limb is challenging using a standard bidimensional lymphoscintigraphic projection. This is because images are only acquired in the anteroposterior direction. Conversely, ICG lymphography allows for observation from multiple directions by moving the camera and adjusting the limb position. This has been identified as one of the main causes of pathway discrepancy. Introducing a second limb projection during lymphoscintigraphy could improve the understanding of the pathway differences and enable a basic three-dimensional reconstruction of the lymphatics' gross anatomy on the affected limb. The concept of “pattern discordance” can be categorized into two scenarios: (1) lymphoscintigraphy showing a positive indication for LVA and ICG lymphography showing only dermal backflow or no flow; or (2) lymphoscintigraphy indicating the absence of vessels, dermal backflow, or a halt in lymph progression and ICG lymphography revealing the presence of at least one functioning vessel in one half of the limb. In the first situation, we believe that the discordance arises from the limitations of ICG lymphography in visualizing lymphatic vessels situated deeper than 2 cm, because it can only detect superficial dermal backflow or the lack of progression. Therefore, an ultrasound assessment plays a relevant role in this subgroup of patients, because it guides the selection of incision sites for surgery. Our hypothesis is that in a subset of patients, these two examinations exhibit different kinetics of the tracers, resulting in a greater ability for ICG to spread and migrate within the lymphatic vessels. However, a prospective study with a larger number of patients is needed to identify the potential causes for this phenomenon and determine whether nonconcordant imaging findings is associated with patient biometric or anamnestic factors. In our series, nonconcordant imaging findings before surgery did not correlate with the clinical outcome in terms of reducing the limb volume.

Our study has some limitations that should be acknowledged. First, the sample size is small. This limitation can be attributed to the rarity of nonconcordant preoperative imaging findings, which is also the reason this phenomenon has been poorly investigated in previous scientific literature. Second, the study was conducted retrospectively. However, the study strengths include its single-center and single-arm design, which allowed for the enrollment of patients who underwent lymphoscintigraphy performed by the same experienced nuclear doctor (G.T.) and surgery and ICG lymphography by the same experienced surgeon (S.G.), both experts in lymphedema. This ensured a comprehensive standardization and adherence to established protocols.

Conclusions

The discrepancy of information useful for the indication and execution of LVA between lymphoscintigraphy of the superficial lymphatic system and ICG lymphography is rare but possible. Pathway discrepancies can be resolved with the introduction of additional projections. Pattern discrepancies, in which lymphoscintigraphy would contraindicate the intervention, but lymphography shows good lymphatic vessels, still remain unexplained but could be related to differences in tissue washout between the two different tracers occurring in some patients. Further studies are needed to understand whether there are patient-related predictive factors.

Author Contributions

Conception and design: AC, SG

Analysis and interpretation: AC, GT, SG

Data collection: GT, FI, CC

Writing the article: AC, SG

Critical revision of the article: AC, GT, FI, CC, SG

Final approval of the article: AC, GT, FI, CC, SG

Statistical analysis: SG

Obtained funding: Not applicable

Overall responsibility: SG

Disclosures

None.