Lymphedema is a debilitating disease commonly believed to result from progressive fluid accumulation. However, fat accumulation and fibrosis feature in the natural course of this disease.1–5 Impaired lymphatic function and lymph stasis may lead to a chronic inflammatory state that stimulates collagen deposition and adipocyte accumulation over time.4
Although the phenomenon of fat accumulation is well described, current staging systems for lymphedema are purely clinical and do not account for this variable. Patients undergoing surgical or nonsurgical interventions are typically assessed by limb volume difference and International Society of Lymphology stage.6 However, the fluid and fat composition of the excess limb volume can vary considerably and may confound results, affect prognosis, and influence surgical decision-making. For example, we frequently observe that patients with the same International Society of Lymphology stage of lymphedema (in particular, stage 2) have markedly different patterns of fluid and fat excess. In addition, the regional patterns of lymph accumulation and fibroadipose deposition have not been previously described. It is also unclear how the regions of fluid and fat excess relate to the location of the major lymphatic collectors of the lower limb.7,8 Magnetic resonance angiography is routinely used in our practice to provide information regarding not only fat and fluid composition but also vascular anatomy for lymph node transplantation, and assess for potential vein stenosis and occult recurrence. The purpose of this study was to review these magnetic resonance angiographs to evaluate the patterns of fluid and fat accumulation in patients with lower extremity lymphedema.
PATIENTS AND METHODS
The Memorial Sloan Kettering Institutional Review Board (16-275) approved this study. Patients who underwent magnetic resonance angiography for primary or secondary lymphedema involving the lower extremity between December 1, 2012, and December 31, 2018, were retrospectively identified from our prospectively maintained database. All magnetic resonance angiographic examinations were performed by the same radiologist. Axial and sagittal high-resolution T1 fat-saturated gradient-echo images of the involved extremities were obtained after administration of a contrast agent. An intravenous gadolinium agent was used, and delayed-phase vascular imaging was performed. Magnetic resonance angiographic sequences were obtained on a 1.5-T magnetic resonance imaging unit (GE Healthcare, Waukesha, Wis.). Patients were scanned in the supine position, with a dedicated peripheral vascular coil. Coronal acquisition was obtained with T1-weighted postcontrast imaging without fat saturation using gadofosveset trisodium contrast agent. Axial and sagittal reformats were then obtained. Patients with lower extremity lymphedema were deidentified by an investigator not involved in the review process.
Assessment of Degree of Fat and Fluid Composition
The grading system proposed in this study was developed as a first attempt to stratify patients in a reasonable and reproducible way to evaluate their fluid and fat composition, as we have observed a wide spectrum of differences in patients with the same International Society of Lymphology stage. It is admittedly subjective; an objective method will clearly be preferable in the future. This is only an initial step toward this end and involved a lengthy process of identifying recurring patterns of fluid and fat accumulation. Working with an imaging software programmer, we initially attempted to quantify the absolute fluid and fat volumes on magnetic resonance angiography. However, because of the presence of significant artifact, the use of absolute pixel brightness thresholds to differentiate between fluid and fat was not feasible. Consequently, we manually delineated the areas of interest for fat and fluid. Without an automated and reliable system, however, this was not practical for clinical use and was abandoned. We subsequently developed and tested different grading systems on the basis of viewing magnetic resonance angiographic sections representing the largest focus of lymphedema. These approaches were assessed for reproducibility, clinical relevance, and ease of use. Although more-involved grading systems were initially tested, lack of reproducibility and time-consuming methodology led to the current system presented in this article. The recurring patterns of reticular formation of fluid accumulation within the fat, compared with the more-pronounced fluid excess, which resembled a contiguous stripe of lymph between fat and fascia, became easily recognizable. Although this grading system was qualitative, it provided a means to stratify patients with different tissue composition who otherwise have the same International Society of Lymphology stage. The following is a brief summary of the four International Society of Lymphology stages: 0 = subclinical lymphedema (impaired lymphatic function without evidence of limb swelling; 1 = mild lymphedema (mild but visible limb swelling that temporarily improves with elevation); 2 = moderate lymphedema (moderate limb swelling with evidence of fibrosis that may or may not be associated with pitting edema that does not improve with elevation); and 3 = elephantiasis (profound limb swelling with severe fibrosis, hyperkeratosis, and often verrucous skin changes).
For fluid and fat grading, coronal images were used to evaluate the largest area of swelling in the affected limb (lower leg, thigh, or both). The location of the highest concentration of fluid and fat was then recorded using coronal, axial, and sagittal sections. Finally, the dominant contributor to the limb volume excess was determined (fluid, fat, or codominant).
Next, the fluid and fat components were graded on a three-point scale. Scoring for fluid accumulation grade was as follows: 0 = no fluid, 1 = honeycombing/reticular pattern of fluid within the subcutaneous fat, and 2 = continuous visible stripe of fluid between the fat and investing muscle fascia. Scoring for fat accumulation was as follows: 0 = no excess fat, 1 = fat accumulation less than twice the width of the widest fat stripe on the unaffected side, and 2 = fat accumulation greater than twice the width of the widest fat stripe on the unaffected side. Representative images of fluid and fat combinations observed on the basis of our magnetic resonance angiographic grading system are included as Figures 1 through 6.
Analysis of Interobserver Agreement
Images were deidentified and reviewed by three independent investigators, including a radiologist (A.K.), a plastic surgeon (J.H.D.), and a general surgery resident (R.V.). Overall percentage agreement and Light’s kappa (κ) with bootstrapped 95 percent confidence intervals were used to assess agreement among the three independent readers across 47 cases analyzed. Cohen’s simple κ with 95 percent CI and percentage agreement were used to assess pairwise agreement among each pair of readers.
Patients’ demographic characteristics, fluid and fat grade, predominance, and location were summarized by a plastic surgeon (J.H.D.). The relationships between available patient characteristics and overall dominance, fat grade, and fluid grade were assessed using Fisher’s exact test and the Kruskal-Wallis test, where appropriate. International Society of Lymphology stages were then compared with results from our magnetic resonance angiographic grading system using Fisher’s exact test. Analysis was performed using SAS 9.4 (SAS Institute, Inc., Cary, N.C.).
Patient characteristics are listed in Table 1. A total of 76 patients were included in our study, of whom 63 (82.9 percent) were female and 13 (17.1 percent) were male. Malignancy accounted for the majority of cases, with gynecologic cancer being the most common cause of lower extremity lymphedema [32 of 76 (42.1 percent)] (Fig. 7). The mean duration of lymphedema was 125 ± 113 months (range, 3 to 348 months). Four patients had lymphedema for more than 30 years; 32.9 percent of patients (25 of 76) had lymphedema for less than 3 years. The average age was 51 years (range, 12 to 73 years), and the mean body mass index was 25.9 kg/m2 (range, 17.0 to 42.0 kg/m2). Of the total 76 patients, eight (10.5 percent) had International Society of Lymphology stage 1, 65 (85.5 percent) had stage 2, and 3 (4 percent) had stage 3 disease. Duration of lymphedema was not significantly associated with International Society of Lymphology stage (r2 = −0.13; p = 0.29).
Table 1. -
|Body mass index, kg/m2
|History of radiation treatment
|Duration of lymphedema, mo
SL, International Society of Lymphology.
Fluid and Fat Distribution
The vast majority of patients in this series had some degree of fat accumulation [68 of 76 (89.5 percent)]. In our cohort, eight of 76 patients (10.5 percent) had a fat grade of 0, 30 of 76 (39.5 percent) had a fat grade of 1, and 38 of 76 (50 percent) had a fat grade of 2. All but two patients [74 of 76 (97.4 percent)] had clinical lymphedema with no evidence of fluid accumulation on magnetic resonance angiography. Two of 76 patients (2.6 percent) had a fluid grade of 0, 37 of 76 patients (48.7 percent) had a fluid grade of 1, and 37 of 76 patients (48.7 percent) had a fluid grade of 2. Fat was the dominant component of excess limb volume in 29 of 76 patients (38.2 percent), fluid was the dominant component in the same number of patients [29 of 76 (38.2 percent)], and fat and fluid contributed equally in 18 of 76 patients (23.7 percent). Fluid grade was positively but weakly associated with fat grade (r2 = 0.25; p = 0.029). Overall fat or fluid dominance was not associated with age or sex. Lymphedema duration was not associated with higher grade of fluid (r2 = −0.03; p = 0.79) or fat (r2 = −0.07; p = 0.52) accumulation.
The most common locations of fluid and fat accumulation are presented in Table 2. The most common location of fluid accumulation was the lower leg [59 of 76 (77.6 percent)]—in particular, the lateral lower leg [34 of 59 (57.6 percent)]. Twelve patients (15.8 percent) had fluid accumulation in the thigh, most commonly the lateral thigh [five of 12 (41.7 percent)]. Only five patients (6.6 percent) had fluid accumulation distributed equally in the lower leg and thigh.
Table 2. -
Distribution of Fluid and Fat Accumulation in the Lower Extremity by Dominant Location and Laterality
| Lower leg
| Lower leg
The most common site of fat accumulation was the lower leg [43 of 76 (56.6 percent)]; the predominant distribution pattern in the lower leg was global [23 of 43 (53.5 percent)]. When fat accumulation was present in the thigh, it was most commonly pronounced medially [13 of 18 (72.2 percent)]. Fifteen patients [15 of 76 patients (19.7 percent)] had fat accumulation distributed equally in the lower leg and thigh.
Association between International Society of Lymphology Stage and Fluid-Fat Grade
We next analyzed the data to determine whether International Society of Lymphology stage was associated with fluid and fat grade. Most of the patients in this study were diagnosed with International Society of Lymphology stage 2 lymphedema [65 of 76 (85.5 percent)]. Fluid grade was not significantly associated with lymphedema stage (Fig. 8). In contrast, fat grade was significantly associated with lymphedema stage (Fig. 9). For example, only 12.5 percent of patients with stage 1 disease had a fat grade of 2, whereas approximately half of patients with stage 2 disease had a fat grade of 2, and 100 percent of patients with stage 3 disease had a fat grade of 2. These findings demonstrate a highly variable mix of fluid and fat grades across stages. This is most striking in patients with stage 2 lymphedema, among whom nearly every combination of fluid and fat grade was represented.
Agreement among the three readers was high (Table 3). Agreement for overall fat or fluid dominance was 91.5 percent (Light’s κ = 0.90; 95 percent CI, 0.78 to 0.98). Concordance for location of both fat and fluid accumulation was 91.5 percent. Interobserver agreement was 95.7 percent (κ = 0.95; 95 percent CI, 0.83 to 1.00) for fluid grade and 87.2 percent (κ = 0.86; 95 percent CI, 0.70 to 0.95) for fat grade. Overall, the interreader agreement for this grading system was high, making this a reasonable starting point for grading the fluid-fat excess in patients with lymphedema.
Table 3. -
Overall Interobserver Agreement among Three Independent Readers
||Light’s κ (95% CI)
The findings in this study demonstrate that fat accumulation is a prominent feature of lymphedema and may be an important factor in patient assessment. Nearly 90 percent of patients had some degree of fat accumulation, and fat was the dominant component of excess limb volume in more than one-third of patients. This demonstrates the significance of the contribution of fat to limb volume and may account for variable outcomes following lymphatic reconstruction or decongestive therapy. Many current outcome studies focus on limb volume and bioimpedance, which have inherent limitations.9 Limb volume does not account for the fluid-fat composition. Bioimpedance is an indirect means of measuring extracellular fluid content, which may be low in a fat-dominant limb even though the patient may have advanced disease.10 Although the grading system proposed in this study is limited by its qualitative nature, it does provide a simple and useful method to stratify patients on the basis of fluid-fat composition. This is a significant confounding variable when studying any medical-, surgical-, or physiotherapy-based intervention and is not specifically accounted for in existing staging systems. An ideal method of grading fluid and fat would be quantitative. We are currently evaluating several techniques for obtaining quantitative data that would be feasible in a clinical setting.
The specific patterns of fat and fluid accumulation are also of interest and do not appear to have any relationship to the location of the dominant ventromedial lymphatic collectors of the lower extremity.8 This may be relevant to cases of lymphatic reconstruction. Typically, lymph node transplants are placed along the ventromedial axis.11–13 Placing lymph nodes laterally, where most of the fluid accumulates, may be worth investigating in the future.
Curiously, the duration of lymphedema was not associated with either fluid or fat grade or International Society of Lymphology stage. The ability to draw conclusions from this may be limited by the distribution of lymphedema stages among the patients in this study. However, patients with a short duration of lymphedema and early disease stage had profound fat accumulation. For example, 25 percent of patients with a severe fat grade of 2 had lymphedema for less than 2 years. This counters the generalized notion that fluid is serially replaced by fat in a linear relationship over time. The lack of a time-dependent association with fat or fluid accumulation suggests that these patients do not all fall along the same spectrum of disease. These variable manifestations of fibrofatty accumulation likely reflect fundamental differences in immune pathophysiology among patients, with some forms of lymphedema that are more aggressive and proliferative than others. In addition, some patients have intractable cellulitis (the cause of which remains unclear), whereas others do not.14 These findings may have implications for how we conduct studies among patients with lymphedema. Instead of approaching patients with lymphedema as a homogeneous population, patients may be specifically stratified into groups—for example, fat dominant, fluid dominant, cellulitis-prone, and no history of infection. The differences observed among patients may explain the variable outcomes following any intervention and may allow us to better direct specific treatments to the most appropriate subgroups of patients.
Understanding a patient’s fluid-fat composition may help determine the most appropriate clinical approach for them. A limb with advanced fibrofatty proliferation may not respond well to decongestive therapy or physiologic procedures, such as vascularized lymph node transplantation and lymphaticovenous anastomosis. In these cases, liposuction, as advocated by Brorson, has been beneficial, provided that there is no significant pitting and the patient is 100 percent compliant with compression.15,16 In contrast, a fluid-dominant limb may respond better to decongestive therapy, vascularized lymph node transplantation, or lymphaticovenous anastomosis. However, as observed in this study, many patients have a combination of fluid and fat excess, and multimodal treatment may be most appropriate for such patients. Ideal timing and patient selection remain topics of debate.17,18
The staging of lymphedema is currently dominated by systems that use physical examination alone and is inadequate in the modern era of imaging and biomarkers. The basic approach to staging affects the level at which we approach the disease and, in turn, the treatment offered to patients with lymphedema. In this study, the majority of patients had International Society of Lymphology stage 2 disease. However, within this group, there was a wide array of fluid and fat compositions. In effect, the use of magnetic resonance angiography has unmasked significant differences in patients with the same stage of lymphedema. For example, Figures 4 and 5 demonstrate two patients with stage 2 lymphedema who have similar limb volumes but completely inverse fluid-fat ratios. Consequently, it is likely that these two patients may have different outcomes following a new medical treatment, surgical intervention, or physiotherapy. In our practices, we routinely review the results of our magnetic resonance angiography grading system with patients so that they can better understand why they are likely or unlikely to respond to treatment.
As imaging techniques continue to improve and further insights into the pathophysiology of lymphedema are made, staging of lymphedema will continue to evolve. The system proposed here provides only a fluid-fat grade and is not a staging system. It is meant to complement and contribute to a more complete staging of a patient’s lymphedema. The consideration of additional components, such as physiologic function19 and, one day, pathophysiologic grade, will help to provide a more complete view of the patient’s overall disease burden. There is currently no single imaging modality or staging system that adequately reflects the status of a patient’s lymphatic system in total. Magnetic resonance angiography serves to fill a gap, providing information not only on fluid-fat composition but also on the status of the venous system, the vascular anatomy (which can inform surgical planning), and the presence or absence of occult cancer. Thus, magnetic resonance angiography has become a key component of our clinical workup and an important part of studies investigating outcomes after lymphatic surgery.20–22 This fluid-fat grading system is intended to be used not in place of existing staging systems but as an additional tool to clarify differences among patients. In summary, the closer we delve into lymphatic imaging, the more clear it becomes that staging of a lymphedema patient using physical examination alone is inadequate.
Knowledge of the local patterns of fluid and fat accumulation and of the overall tissue composition is important to understand the individual nature of a patient’s lymphedema. Current lymphedema staging systems are inadequate, as they do not specifically differentiate patients on the basis of these confounding variables. The fluid-fat grading system proposed in this study is a simple and reliable way of stratifying patients and can serve as a complement to existing clinically based staging systems. This grading system may be useful for evaluating patients and planning treatment and may help to identify the influence of fat and fluid factors on outcomes.
This research was funded by the New York State Department of Health through the Empire Clinical Research Investigator Program (grant no. 49650/38090). This research was also funded in part through National Institutes of Health/National Cancer Institute Cancer Center Support Grant P30 CA008748.
1. Avraham T, Yan A, Zampell JC, et al. Radiation therapy causes loss of dermal lymphatic vessels and interferes with lymphatic function by TGF-beta1-mediated tissue fibrosis. Am J Physiol Cell Physiol. 2010;299:C589–C605.
2. Avraham T, Zampell JC, Yan A, et al. Th2 differentiation is necessary for soft tissue fibrosis and lymphatic dysfunction resulting from lymphedema. FASEB J. 2013;27:1114–1126.
3. Dayan JH, Ly CL, Kataru RP, Mehrara BJ. Lymphedema: Pathogenesis and novel therapies. Annu Rev Med. 2018;69:263–276.
4. Ghanta S, Cuzzone DA, Torrisi JS, et al. Regulation of inflammation and fibrosis by macrophages in lymphedema. Am J Physiol Heart Circ Physiol. 2015;308:H1065–H1077.
5. Mihara M, Hara H, Hayashi Y, et al. Pathological steps of cancer-related lymphedema: Histological changes in the collecting lymphatic vessels after lymphadenectomy. PLoS One 2012;7:e41126.
6. Executive Committee. The diagnosis and treatment of peripheral lymphedema: 2016 consensus document of the International Society of Lymphology. Lymphology 2016;49:170–184.
7. Sappey MPC. Anatomie, Physiologie, Pathologie des Vaisseaux Lymphatiques consideres chez L’homme at les Vertebres. 1874.Paris: A. Delahaye et E. Lacrosnier.
8. Tourani SS, Taylor GI, Ashton MW. Anatomy of the superficial lymphatics of the abdominal wall and the upper thigh and its implications in lymphatic microsurgery. J Plast Reconstr Aesthet Surg. 2013;66:1390–1395.
9. Fu MR, Cleland CM, Guth AA, et al. L-Dex ratio in detecting breast cancer-related lymphedema: Reliability, sensitivity, and specificity. Lymphology 2013;46:85–96.
10. Seward C, Skolny M, Brunelle C, Asdourian M, Salama L, Taghian AG. A comprehensive review of bioimpedance spectroscopy as a diagnostic tool for the detection and measurement of breast cancer-related lymphedema. J Surg Oncol. 2016;114:537–542.
11. Cheng MH, Huang JJ, Huang JJ, et al. A novel approach to the treatment of lower extremity lymphedema by transferring a vascularized submental lymph node flap to the ankle. Gynecol Oncol. 2012;126:93–98.
12. Nelson JA, Mehrara BH, Dayan JH. Vascularized lymph node transfer to the profunda artery perforator pedicle: A reliable proximal recipient vessel option in the medial thigh. Plast Reconstr Surg. 2017;140:366e–367e.
13. Smith ML, Molina BJ, Dayan E, et al. Heterotopic vascularized lymph node transfer to the medial calf without a skin paddle for restoration of lymphatic function: Proof of concept. J Surg Oncol. 2017;115:90–95.
14. Zampell JC, Yan A, Avraham T, et al. Temporal and spatial patterns of endogenous danger signal expression after wound healing and in response to lymphedema. Am J Physiol Cell Physiol. 2011;300:C1107–C1121.
15. Brorson H. Complete reduction of arm lymphedema following breast cancer: A prospective twenty-one years’ study. Plast Reconstr Surg. 2015;136:134–135.
16. Brorson H. Liposuction normalizes lymphedema induced adipose tissue hypertrophy in elephantiasis of the leg: A prospective study with a ten-year follow-up. Plast Reconstr Surg. 2015;136:133–134.
17. Batista BN, Germain M, Faria JC, Becker C. Lymph node flap transfer for patients with secondary lower limb lymphedema. Microsurgery 2017;37:29–33.
18. Gould DJ, Mehrara BJ, Neligan P, Cheng MH, Patel KM. Lymph node transplantation for the treatment of lymphedema. J Surg Oncol. 2018;118:736–742.
19. Mihara M, Hara H, Araki J, et al. Indocyanine green (ICG) lymphography is superior to lymphoscintigraphy for diagnostic imaging of early lymphedema of the upper limbs. PLoS One 2012;7:e38182.
20. Liu N, Wang C, Sun M. Noncontrast three-dimensional magnetic resonance imaging vs lymphoscintigraphy in the evaluation of lymph circulation disorders: A comparative study. J Vasc Surg. 2005;41:69–75.
21. Notohamiprodjo M, Weiss M, Baumeister RG, et al. MR lymphangiography at 3.0 T: Correlation with lymphoscintigraphy. Radiology 2012;264:78–87.
22. White RD, Weir-McCall JR, Budak MJ, Waugh SA, Munnoch DA, Sudarshan TA. Contrast-enhanced magnetic resonance lymphography in the assessment of lower limb lymphoedema. Clin Radiol. 2014;69:e435–e444.