Abstract
Background
Breast cancer is the most common malignancy among women globally, with 2.3 million new cases in 2022. Advances in early detection and treatment have improved survival, but long-term complications like breast cancer-related lymphoedema (BCRL) remain underrecognized. Affecting 6–40% of patients following axillary lymph node dissection, BCRL leads to chronic swelling, pain, and reduced quality of life. The risk is heightened in low- and middle-income countries (LMICs), where delayed diagnosis and limited awareness necessitate more aggressive interventions. Despite its profound physical, psychosocial, and economic impact, BCRL remains understudied in LMICs, highlighting the need for region-specific data and preventive strategies.
Methods
This prospective observational study, which was conducted in a tertiary care centre over 18 months (July 2022–December 2023), aimed to identify clinicopathological predictors of the development of early ipsilateral arm lymphoedema in patients undergoing Modified radical mastectomy (MRM). We diagnosed lymphoedema via circumferential measurements (≥ 10% increase in ipsilateral arm) at 1 day pre-surgery and followed up the patients at 1-, 3-, and 6-months post-surgery. Variables included patient demographics, tumour characteristics, and treatment factors.
Results
We recruited 80 patients, and 21 of them (26.25%) developed lymphoedema in 6 month period, with significant predictors identified as BMI (mean 28.42 vs. 26.04 kg/m2, p = 0.008), Left-sided tumours (40% vs. 18% for right-sided, p = 0.03), number of lymph nodes removed (mean 21.14 vs. 13.49, p = 0.003), and positive lymph nodes (mean 3.81 vs. 1.15, p = 0.009), Axillary radiation (55% vs. 16.7%, p < 0.001). Multivariate analysis confirmed BMI (aOR = 1.19, 95%CI:1.00–1.42; p = 0.045), total lymph nodes removed (aOR = 1.08, 95%CI:1.01–1.15; p = 0.027), and axillary radiation (aOR = 4.57, 95%CI:1.37–15.26; p = 0.013) as independent predictors, while positive lymph nodes lost significance when adjusted for confounders (aOR = 1.04, p = 0.718). Age, hormone receptor status, and chemotherapy type showed no significant associations.
Conclusion
The study highlights the critical role of pre-operative risk stratification. These findings underscore the need for personalized surgical and adjuvant strategies to mitigate Lymphoedema risk, enhancing long-term survivorship care in breast cancer patients.
Trial registration
Clinical Trials Registry India—CTRI/2023/08/056411 (Registered prospectively on 11/08/2023).
Keywords: Breast cancer, Lymphoedema, Survivorship, Prospective studies, Axillary dissection
Background
Breast cancer remains the most prevalent malignancy among women globally, with an estimated 2.3 million new cases diagnosed in 2022, as per GLOBOCAN 2022 data, accounting for nearly 12% of all cancer diagnoses worldwide [1]. While advancements in early detection and multimodal therapies—including surgery, radiation, and systemic treatments—have significantly improved 5-year survival rates to over 90% for localized disease, these interventions often carry long-term sequelae that may compromise survivorship outcomes [2]. Among these, lymphoedema stands out as a debilitating and underrecognized complication. Characterized by progressive swelling of the ipsilateral arm due to impaired lymphatic drainage, it affects 6–40% of patients undergoing axillary lymph node dissection (ALND), with variability attributed to surgical techniques, adjuvant therapies, and patient-specific risk factors [3, 4]. In low- and middle-income countries (LMICs), delayed diagnoses leading to advanced disease, often necessitates more aggressive interventions, further amplifying the lymphedema burden. However, region-specific data remain scarce, hindering tailored preventive strategies [5].
The pathophysiology of lymphoedema is multifactorial, rooted in the disruption of lymphatic architecture during surgery and exacerbated by post-treatment inflammation and fibrosis. Axillary clearance, a cornerstone of breast cancer staging, disrupts axillary lymph nodes and collateral pathways, impairing fluid transport and increasing hydraulic resistance within the lymphatic system [6]. Subsequent radiation therapy compounds this damage by inducing perivascular fibrosis and sclerosis, reducing lymphatic contractility by up to 60% [7]. Molecular mechanisms further drive chronicity: transforming growth factor-beta (TGF-β) activation promotes fibroblast proliferation and extracellular matrix (ECM) deposition. In contrast, vascular endothelial growth factor-C (VEGF-C)-mediated lymphangiogenesis often yields dysfunctional vessels incapable of restoring drainage [8, 9]. Concurrent oxidative stress, mediated by reactive oxygen species (ROS), exacerbates endothelial damage, perpetuating a cycle of inflammation and tissue remodelling [10]. These processes culminate in the hallmark features of lymphoedema—non-pitting oedema, hyperkeratosis, and recurrent cellulitis, which diminish quality of life and impose substantial economic burdens.
The ramifications of lymphoedema extend far beyond physical morbidity. Psychosocially, patients report profound anxiety, depression, and body image disturbances, often exacerbated by cultural stigmatization of visible disfigurement [11]. Functionally, reduced range of motion and chronic pain limit activities of daily living, with 30% of patients unable to return to work within two years of diagnosis [12]. Economically, the lifelong costs of compression garments, physiotherapy, and hospitalizations exceed US$1,800 annually per patient in high-income countries. At the same time, out-of-pocket expenses frequently lead to catastrophic health expenditures in resource-limited settings like India, with a recent study reporting a mean total cost of breast cancer treatment alone to be INR/₹ 258,095/US$ 3531, with adequate data for lymphoedema treatment in Indian setting not available [13]. Despite these challenges, lymphoedema remains underprioritized in oncology care pathways, particularly in regions where survival outcomes dominate clinical focus.
Globally, obesity, extensive nodal dissection, and post-operative radiation to the axilla are well-established risk factors for BCRL. However, emerging evidence suggests regional disparities in risk profiles. In South Asia, for instance, higher rates of advanced-stage diagnoses—40% of Indian patients present with stage III/IV disease—necessitate more aggressive surgical interventions, potentially elevating lymphoedema incidence [14]. Cultural factors, including delayed healthcare-seeking behaviour and limited lymphoedema awareness, further complicate early detection and intervention. Such gaps underscore the urgent need for context-specific research to inform risk stratification and guide resource allocation in LMICs.
The primary aim of this study was to identify clinicopathological predictors associated with the development of ipsilateral arm lymphoedema in breast cancer patients undergoing Modified Radical Mastectomy (MRM). The study sought to stratify risk and inform targeted strategies for lymphoedema prevention and early intervention in a resource-limited setting by analysing patient demographics, tumour characteristics, and surgical and adjuvant treatment factors.
Methods
This prospective observational study was conducted at a tertiary care centre over 18 months (July 2022–December 2023) following approval by the Institutional Ethics Committee (AIIMS/IEC/23/212). Consecutive sampling of eligible patients was done, with adherence to inclusion and exclusion criteria rigorously maintained throughout the study. All patients were treated in a specialised breast unit, with standardised pre-operative protocols, workup, standard surgical technique, post operative care and rehabilitation protocols, with no major changes over the period of 18 months.
Eligible participants included in the study were female patients aged ≥ 18 years with histologically confirmed breast carcinoma scheduled for Modified Radical Mastectomy (MRM), while exclusion factors were preexisting ipsilateral arm lymphoedema, pre-operative axillary vein thrombosis confirmed via imaging, metastatic disease at presentation, or a history of recurrent breast malignancy. Consecutive sampling was performed within the specified time frame.
The diagnostic criteria for lymphoedema were a ≥ 10% increase in ipsilateral arm circumference relative to the contralateral arm using circumferential measurements at four anatomical landmarks, a method validated in resource-constrained settings for its cost-effectiveness and reproducibility. All measurements were done by single trained observer. Arm measurements were made at 8AM ± 1 h, all patient were asked not to wear compression garment on day of measurement and were not given instructions with regard to activity before measurement (Fig. 1, ) [15].
Fig. 1.
4 points of arm measurement. F-MP: Midpoint of forearm; C-10: 10 cm below lateral epicondyle of humerus; MP: Midpoint of arm; C-15: 15 cm above lateral epicondyle of humerus
Pre-operative assessment documented clinical and tumour characteristics, including baseline arm circumference measurements. Postoperatively, patients were evaluated at 1, 3, and 6 months to monitor arm circumference changes, with lymphoedema defined as a ≥ 10% increase in circumference compared to pre-operative measurements or the contralateral limb.
The data collected was tabulated in Microsoft Excel (Microsoft Corporation). Qualitative (categorical) variables were expressed in frequency, percentage, and proportions, and quantitative (continuous) variables were expressed by means and standard deviation or median with inter-quartile range. The chi-square test was used to determine the association between the patient, tumour, and treatment factors and the development of lymphoedema. Unpaired t-tests were used to check the hypothesis in normally distributed variables (tested by the Shapiro–Wilk test). The Wilcoxon-Mann–Whitney U test was used to compare variables that were not normally distributed in the two subgroups of the variable lymphoedema. Fisher's exact test was used to explore the association between lymphoedema and variables where more than 20% of the total number of cells had an expected count of less than 5. Regression analysis was done using backward selection and bidirectional selection method. All Analyses were done using SPSS v25 (IBM Corporation). A p-value of < 0.05 was considered statistically significant.
Results
The following table summarizes the regression analysis for the dependent variable using all the predictor variables together in one go. The 'OR (univariable)' column lists the odds ratios for each of the variables with respect to the dependent variable, when these variables are used as single predictors of the dependent variable, without entering the rest of the variables in the model. The 'OR (multivariable)' column lists the odds ratios for all the variables when they are entered in the model together (and are now thus controlling for each other). The first category in each of the categorical variables is the reference category, against which the odds ratios of the rest of the variables is calculated.
The following table is the same as the previous table, except in this table bidirectional stepwise selection is used to select only the most useful variables to include in the final multivariable predictive model for the dependent variable.
This study recruited 86 eligible patients, of whom four were lost to follow-up after surgery, and two patients died before completion of the study duration. The attrition rate in our study was 7% (n = 6). So, data from 80 patients was taken up for the final Analysis (Fig. 2).
Fig. 2.
Flowchart for patient assessment for eligibility, inclusion, follow up, attrition and final analysis
The clinicodemographic characteristics of the patients are detailed in Table 1. Table 2 presents the final histopathological profile obtained postoperatively, providing insight into the spectrum of pathological entities encountered in the cohort. Table 3 outlines the results of univariate Analysis, identifying factors significantly associated with the outcome of interest, while Tables 4 and 5 displays the multivariate analysis, demonstrating the independent predictors after adjusting for potential confounders Tables 6 and 7.
Table 1.
Clinicodemographic profile of patients and treatment received
| Factors | Mean ± SD || Median (IQR) || Min–Max OR N (%) |
|---|---|
| Age (Years) | 47.11 ± 11.74 || 45.00 (38.00–55.00) || 28.00—77.00 |
| Age | |
| < 40 Years | 24 (30.0%) |
| 41- 60 Years | 43 (53.8%) |
| 61–70 Years | 13 (16.2%) |
| Gender | |
| Female | 79 (98.8%) |
| BMI (kg/m2) | 26.67 ± 3.86 || 27.15 (24.00–29.13) || 18.97—35.30 |
| Dominant Limb | |
| Left | 0 (0.0%) |
| Right | 80 (100.0%) |
| Side of Carcinoma | |
| Left | 30 (37.5%) |
| Right | 50 (62.5%) |
| Stage As Per 8th AJCC | |
| IA | |
| IB | |
| IIA | 23 (28.7%) |
| IIB | 16 (20.0%) |
| IIIA | 10 (12.5%) |
| IIIB | 30 (37.5%) |
| IIIC | 1 (1.2%) |
| Chemotherapy Regimen | N (%) |
| Neoadjuvant Therapy | |
| Not Given | 19 (23.8%) |
| Taxane Based | 48 (60.0%) |
| Trastuzumab Based | 13 (16.2%) |
| Chemotherapy Drug Given | |
| Taxane-Based Given | 67 (83.8%) |
| Trastuzumab Based Given | 13 (16.2%) |
Table 2.
Histopathology features
| Tumour factors | Mean ± SD || Median (IQR) || Min–Max OR N (%) |
|---|---|
| Hormone Receptor Status (Positive) | 51 (63.7%) |
| HER-2-Neu Receptor (Positive) | 20 (25.0%) |
| Nottingham Grade | |
| 1 | 4 (5.0%) |
| 2 | 18 (22.5%) |
| 3 | 58 (72.5%) |
| DCIS Component | |
| Present | 16 (20.0%) |
| Not Present | 64 (80.0%) |
| Lymphovascular Invasion | |
| Present | 9 (11.2%) |
| Not Present | 71 (88.8%) |
| Capsular Invasion | |
| Present | 10 (12.5%) |
| Extra Nodal Extension | |
| Present | 11 (13.8%) |
| Not Present | 69 (86.2%) |
| Number of Lymph Nodes Removed | 15.50 ± 9.60 || 14.00 (9.75–20.00) || 0.00—48.00 |
| Positive Lymph Nodes Removed | 1.85 ± 3.48 || 0.00 (0.00–2.00) || 0.00—18.00 |
Table 3.
Univariate analysis
| Parameters | Lymphoedema | p-value | |
|---|---|---|---|
| Present (n = 21) |
Absent (n = 59) |
||
| Age (Years) | 46.14 ± 10.24 | 47.46 ± 12.29 | 0.635 |
| BMI (kg/m2) | 28.42 ± 3.13 | 26.04 ± 3.93 | 0.008 |
| Side of Carcinoma | 0.030 | ||
| Left | 12 (57.1%) | 18 (30.5%) | |
| Right | 9 (42.9%) | 41 (69.5%) | |
| Dominant Limb | 1.000 | ||
| Left | 0 (0.0%) | 0 (0.0%) | |
| Right | 21 (100.0%) | 59 (100.0%) | |
| Chemotherapy Drug Given | 0.496 | ||
| Taxane-Based Given | 19 (90.5%) | 48 (81.4%) | |
| Trastuzumab Based Given | 2 (9.5%) | 11 (18.6%) | |
| Stage As Per 8th AJCC | 0.614 | ||
| IIA | 6 (28.6%) | 17 (28.8%) | |
| IIB | 5 (23.8%) | 11 (18.6%) | |
| IIIA | 1 (4.8%) | 9 (15.3%) | |
| IIIB | 6 (28.6%) | 18 (30.5%) | |
| IIIC | 0 (0.0%) | 1 (1.7%) | |
| Nottingham Grade | 0.536 | ||
| 1 | 0 (0.0%) | 4 (6.8%) | |
| 2 | 6 (28.6%) | 12 (20.3%) | |
| 3 | 15 (71.4%) | 43 (72.9%) | |
| DCIS Component | 0.751 | ||
| Present | 5 (23.8%) | 11 (18.6%) | |
| Not Present | 16 (76.2%) | 48 (81.4%) | |
| Lymphovascular Invasion | 0.433 | ||
| Present | 1 (4.8%) | 8 (13.6%) | |
| Not Present | 20 (95.2%) | 51 (86.4%) | |
| Hormone Receptor Status (Positive) | 14 (66.7%) | 37 (62.7%) | 0.746 |
| HER-2-Neu Receptor (Positive) | 4 (19.0%) | 16 (27.1%) | 0.463 |
| Capsular Invasion | 0.118 | ||
| Present | 5 (23.8%) | 5 (8.5%) | |
| Not Present | 16 (76.2%) | 54 (91.5%) | |
| Number of Lymph Nodes Removed | 21.14 ± 10.33 | 13.49 ± 8.55 | 0.003 |
| Extra Nodal Extension | 0.146 | ||
| Present | 5 (23.8%) | 6 (10.2%) | |
| Not Present | 16 (76.2%) | 53 (89.8%) | |
| Radiation to Axilla (Post-Operative) | 11 (52.4%) | 9 (15.3%) | < 0.001 |
| Positive Lymph Nodes Removed | 3.81 ± 4.92 | 1.15 ± 2.50 | 0.009 |
| Level of Lymph Node Dissection | 0.262 | ||
| ALND | 20 (95.2%) | 59 (100.0%) | |
| ALND + Level 3 | 1 (4.8%) | 0 (0.0%) | |
| Seroma (Post-Operative) (Yes) | 6 (28.6%) | 11 (18.6%) | 0.363 |
| Wound Infection (Post-Operative) (Yes) | 3 (14.3%) | 8 (13.6%) | 1.000 |
| Restriction in Arm Movement (Post-Operative) (Yes) | 12 (57.1%) | 21 (35.6%) | 0.085 |
Table 4.
Multivariate analysis
| Dependent: Lymphoedema | Absent | Present | OR (univariable) | OR (multivariable) | |
|---|---|---|---|---|---|
| BMI (kg/m2) | Mean (SD) | 26.0 (3.9) | 28.4 (3.1) | 1.20 (1.04–1.43, p = 0.020) | 1.19 (1.01–1.44, p = 0.048) |
| Number of Lymph Nodes Removed | Mean (SD) | 13.5 (8.5) | 21.1 (10.3) | 1.09 (1.03–1.16, p = 0.004) | 1.07 (1.01–1.16, p = 0.037) |
| Positive Lymph Nodes Removed | Mean (SD) | 1.2 (2.5) | 3.8 (4.9) | 1.23 (1.06–1.47, p = 0.010) | 1.04 (0.85–1.29, p = 0.718) |
| Radiation to Axilla (Post-Operative) | No | 50 (83.3) | 10 (16.7) | - | - |
| Yes | 9 (45.0) | 11 (55.0) | 6.11 (2.05–19.28, p = 0.001) | 3.87 (0.85–18.38, p = 0.077) | |
MODEL FIT: χ2(4) = 21.67, p = < 0.001 Pseudo-R2 = 0.24
Number in data frame = 80, Number in model = 80, Missing = 0
AIC = 80.4, C-statistic = 0.82, H&L = Chi-square (8) 10.33 (p = 0.243)
Table 5.
Multivariate analysis (bidirectional stepwise selection)
| Dependent: Lymphoedema | Absent | Present | OR (univariable) | OR (multivariable) | |
|---|---|---|---|---|---|
| BMI (kg/m2) | Mean (SD) | 26.0 (3.9) | 28.4 (3.1) | 1.20 (1.03–1.41, p = 0.020) | 1.19 (1.00–1.42, p = 0.045) |
| Number of Lymph Nodes Removed | Mean (SD) | 13.5 (8.5) | 21.1 (10.3) | 1.09 (1.03–1.16, p = 0.004) | 1.08 (1.01–1.15, p = 0.027) |
| Positive Lymph Nodes Removed | Mean (SD) | 1.2 (2.5) | 3.8 (4.9) | 1.23 (1.05–1.44, p = 0.010) | - |
| Radiation to Axilla (Post-Operative) | No | 50 (83.3) | 10 (16.7) | - | - |
| Yes | 9 (45.0) | 11 (55.0) | 6.11 (2.01–18.58, p = 0.001) | 4.57 (1.37–15.26, p = 0.013) | |
MODEL FIT: χ2(3) = 21.54, p = < 0.001 Pseudo-R2 = 0.23
Number in data frame = 80, Number in model = 80, Missing = 0
AIC = 78.6, C-statistic = 0.816, H&L = Chi-square (8) 11.86 (p = 0.158)
Table 6.
Performance of study parameters for predicting lymphedema
| Variable | Total Positives | True Positives | True Negatives | False Positives | False Negatives | ||
|---|---|---|---|---|---|---|---|
| Lymphoedema | Present | Absent | 21 (26.2%) | - | - | - | - |
| BMI (kg/m2) (Cutoff: 27.2 by ROC) | >= 27.2 | < 27.2 | 40 (50.0%) | 16 (20.0%) | 35 (43.8%) | 24 (30.0%) | 5 (6.2%) |
| Number of Lymph Nodes Removed (Cutoff: 16 by ROC) | >= 16 | < 16 | 37 (46.2%) | 16 (20.0%) | 38 (47.5%) | 21 (26.2%) | 5 (6.2%) |
| Positive Lymph Nodes Removed (Cutoff: 3 by ROC) | >= 3 | < 3 | 18 (22.5%) | 10 (12.5%) | 51 (63.7%) | 8 (10.0%) | 11 (13.8%) |
Table 7.
Performance of study parameters for predicting lymphedema
| Variable | AUROC | Sensitivity | Specificity | PPV | NPV | Diagnostic Accuracy |
|---|---|---|---|---|---|---|
| BMI (kg/m2) (Cutoff: 27.2 by ROC) | 0.696 (0.57–0.82) | 76.2% (53–92) | 59.3% (46–72) | 40.0% (25–57) | 87.5% (73–96) | 63.7% (52–74) |
| Number of Lymph Nodes Removed (Cutoff: 16 by ROC) | 0.719 (0.60–0.84) | 76.2% (53–92) | 64.4% (51–76) | 43.2% (27–61) | 88.4% (75–96) | 67.5% (56–78) |
| Positive Lymph Nodes Removed (Cutoff: 3 by ROC) | 0.674 (0.54–0.81) | 47.6% (26–70) | 86.4% (75–94) | 55.6% (31–78) | 82.3% (70–91) | 76.2% (65–85) |
In this prospective study of 80 breast cancer patients undergoing Modified Radical Mastectomy (MRM), 26.25% (n = 21) developed ipsilateral arm lymphoedema within 6 months post-surgery. Key predictors included:
Body Mass Index (BMI): Patients with lymphoedema had significantly higher BMI (mean 28.42 vs. 26.04 kg/m2, p = 0.008). With T test identifying a significant difference between the two groups in terms of BMI (kg/m2) (t = 2.783, p = 0.008). Strength of Association measured by Point-Biserial Correlation = 0.27 (Medium Effect Size). Univariable OR: 1.20 (95% CI: 1.04–1.43, p = 0.02). On multivariate Analysis (Stepwise Model), aOR: 1.19 (95% CI: 1.00–1.42, p = 0.045). Thus, BMI remains an independent predictor after controlling for nodal dissection and radiation. Each 1-unit increase confers 19% higher lymphedema risk.
Side of tumour: Left-sided tumours were associated with a higher lymphoedema incidence (40% for left side vs. 18% for right side. On the chi-square test, there was a significant difference between the various groups in terms of the distribution of lymphoedema (χ2 = 4.688, p = 0.030). The odds ratio and relative risk are 3.04 (CI = 1.09–8.48) and 2.22 (CI = 1.08–4.59), respectively, for the development of lymphoedema on the left side. Laterality was not included in multivariate analysis as it is less likely affected by other confounders.
Lymph node dissection: A greater number of lymph nodes were removed in patients who developed lymphoedema (mean 21.14 vs. 13.49, p = 0.003). Univariable analysis confirmed this association (OR = 1.09, 95% CI: 1.03–1.16, p = 0.004). Crucially, the total number of nodes removed remained a significant independent predictor in both multivariable models. After adjusting for BMI, positive nodes, and radiation, each additional node removed increased risk by 7% (aOR = 1.07, 95% CI: 1.01–1.16, p = 0.037). In the optimized stepwise model (adjusting for BMI and radiation), each additional node removed increased risk by 8% (Adj. OR = 1.08, 95% CI: 1.01–1.15, p = 0.027).
Number of positive lymph nodes: The Number of positive lymph nodes was higher in patients who developed lymphoedema (mean 3.81 vs 1.15). On using the Wilcoxon-Mann–Whitney U Test, there was a significant difference in terms of Positive Lymph Nodes Removed (p = 0.009), strength of Association (Point-Biserial Correlation) = 0.34 (Medium Effect Size). univariable analysis indicated a significant association (OR = 1.23, 95% CI: 1.05–1.44, p = 0.010), this variable was not an independent predictor of lymphoedema risk in the multivariate models. When included in the full multivariable model simultaneously adjusting for BMI, total lymph nodes removed, and post-operative axillary radiation, the association became non-significant (aOR = 1.04, 95% CI: 0.85–1.29, p = 0.718). Furthermore, the number of positive lymph nodes was excluded from the final optimized model derived through bidirectional stepwise selection.
Post-operative radiation: Axillary radiation significantly increased risk (55% vs. 16.7%). Chi-square test showed a significant difference between the various groups in terms of distribution of lymphoedema (χ2 = 11.386, p = < 0.001). Strength of association by Cramer's V test = 0.38 (Moderate Association) and strength of association by Bias Corrected Cramer's V = 0.36 (Moderate Association). Univariable OR: 6.11 (95% CI: 2.01–18.58, p = 0.001). On Multivariate Analysis (Stepwise Model): aOR: 4.57 (95% CI: 1.37–15.26, p = 0.013). which implies radiation remains the strongest independent predictor, with 357% higher risk after adjusting for BMI/nodal dissection.
BMI (kg/m2) demonstrated poor diagnostic performance for predicting lymphoedema with an AUROC of 0.696 (95% CI: 0.57—0.82), which was statistically significant (p = 0.008); at a cutoff ≥ 27.2, sensitivity was 76% and specificity was 59%. The Number of Lymph Nodes Removed showed fair diagnostic performance with an AUROC of 0.719 (95% CI: 0.60—0.83), statistically significant (p = 0.003); at a cutoff ≥ 16, sensitivity was 76% and specificity was 64%. Positive Lymph Nodes Removed demonstrated poor diagnostic performance with an AUROC of 0.674 (95% CI: 0.53—0.81), statistically significant (p = 0.009); at a cutoff ≥ 3, sensitivity was 48% and specificity was 86%. All three parameters demonstrated high negative predictive values (82.3–88.4) (Fig. 3).
Fig. 3.
Area under ROC curve analysis for predictors of Lymphedema (BMI, Total number of Lymph nodes Removed, Number of Positive lymph nodes)
Non-significant factors included age (p = 0.635), chemotherapy type (taxane vs. trastuzumab, p = 0.496), hormone receptor status (p = 0.746), pre-operative T- stage and N- stage, presence of in situ component (p = 0.751), Lymphovascular invasion (p = 0.433), capsular invasion (p = 0.116), extra nodal extension (p = 0.146) and post-operative complications like seroma (p = 0.363) or wound infection (p = 1.000).
Discussion
This prospective observational study sought to address gaps by identifying clinicopathological predictors of early lymphoedema in Indian breast cancer patients undergoing modified radical mastectomy (MRM).
Lymphoedema developed in 21 out of 80 patients (26%). This finding is concordant with data projected by similar studies in the literature, reporting incidence ranging from 10 to 45%. In a recent meta-analysis, in which 57 studies were analysed, the overall estimated incidence of chronic arm oedema following axillary lymph node dissection was documented to be 21.4% [16].
Higher BMI (mean 28.42 vs 26.04 kg/m2; t = 2.783, p = 0.008) and left-sided tumours (40% vs 18% incidence; χ2 = 4.688, p = 0.030; OR = 3.04, CI = 1.09–8.48) emerged as significant clinical predictors of lymphoedema. Crucially, BMI maintained independent predictive status after multivariate adjustment (Adj. OR = 1.19, 95% CI:1.00–1.42, p = 0.045), with each unit increase conferring 19% higher risk. The total number of lymph nodes removed demonstrated robust independent association (mean 21.14 vs 13.49; W = 891.000, p = 0.003), remaining significant in both full (Adj. OR = 1.07, 95% CI:1.01–1.16, p = 0.037) and stepwise models (Adj. OR = 1.08, 95% CI:1.01–1.15, p = 0.027), indicating 7–8% increased risk per additional node excised after controlling for confounders. Post-operative axillary radiation constituted the strongest independent predictor (55% vs 16.7% incidence; χ2 = 11.386, p < 0.001; stepwise Adj. OR = 4.57, 95% CI:1.37–15.26, p = 0.013), translating to 357% elevated risk when adjusted for BMI and nodal dissection. While positive lymph node count was elevated in lymphoedema cases (mean 3.81 vs 1.15; W = 835.500, p = 0.009) and significant univariably (OR = 1.23, 95% CI:1.05–1.44, p = 0.010), it lost independence in multivariate analysis (full model Adj. OR = 1.04, 95% CI:0.85–1.29, p = 0.718) and was excluded from the stepwise model, suggesting its effects are mediated through total nodal excision. Non-significant factors included age (p = 0.635), chemotherapy type (p = 0.496), hormone receptor status (p = 0.746), pre-operative T/N stage, histopathological features (in situ component p = 0.751; LVI p = 0.433; capsular/extra nodal invasion p = 0.116–0.146), and post-operative complications (seroma p = 0.363; infection p = 1.000).
The association of a higher BMI with increased incidence of lymphoedema concurs with similar studies documented in the literature, including meta-analysis of 57 studies by Manirakiza et. al. and prospective studies by Degnim et al., Clark et al., Dominick et al., and Kwan et al., among others [17–21]. In our study, we observed that patients with lymphoedema had a higher BMI (28.42) than those without (26.04) (p = 0.0081), indicating that a higher BMI could increase the risk for lymphoedema. The cause of this effect is complex and multifactorial. A functional link has emerged between lymphatic malfunction and the pathogenesis of obesity. Patients with higher BMI may require more lymphatic channels for higher lymph flow to facilitate drainage. It is likely to result from the capacity of the lymph and circulatory imbalance [22]. So, either it could be that a heavier arm needs more lymph flow to maintain optimum lymph content, or it is because the surgery is more extensive/destructive due to the presence of more adipose tissue in the axilla [16, 23]. Studies have also claimed that obese patients are susceptible to fat necrosis, poor wound healing, and infection, and the separation of deep lymphatic channels by additional subcutaneous fat, thus leading to lymphoedema [24].
In our study, patients with left-sided breast carcinoma demonstrated a significantly higher incidence of ipsilateral arm lymphoedema (40%) compared to those with right-sided carcinoma (18%), with this difference being statistically significant (χ2 = 4.688, p = 0.030). While this association is clearly observed, the underlying mechanisms remain speculative and require further investigation. Potential explanations might include anatomical asymmetries in venous or lymphatic drainage between sides, or the fact that all patients had right-hand dominance—possibly leading to differential limb use patterns that could theoretically influence fluid dynamics. However, these remain hypotheses needing rigorous validation; causation cannot be inferred from the present data. Similar associative findings were reported by Singh et al. [25], though the biological or behavioural basis for this lateralization effect warrants dedicated mechanistic studies.
A significant difference emerged in lymph node removal between groups (mean 21.14 vs. 13.49; W = 891.000, p = 0.003), with univariable analysis showing each additional node increased lymphedema risk by 9% (OR 1.09, 95% CI 1.03–1.16, p = 0.004). Crucially, this association persisted after multivariate adjustment for BMI and radiation (Adj. OR 1.08, 95% CI 1.01–1.15, p = 0.027), suggesting an independent 8% risk increase per node. While this effect may reflect mechanical disruption of lymphatics, alternative explanations like more aggressive disease biology in extensively dissected cases could theoretically contribute – particularly given these patients' higher positive node burden (mean 3.81 vs 1.15). The findings align with Clark et al. and Dominick et al. [18, 19] but contradict Yen et al. and Beaulac et al. [26, 27], highlighting the need for larger studies dissecting causal pathways between surgical extent, tumour biology, and lymphedema development.
A significant difference was observed in positive lymph node counts between groups (mean 3.81 vs. 1.15; W = 835.500, p = 0.009), with univariable analysis indicating a 23% increased lymphedema risk per positive node (OR 1.23, 95% CI 1.05–1.44, p = 0.010). However, this association did not persist in multivariate analysis when adjusted for BMI, total nodes removed, and radiation (Adj. OR 1.04, 95% CI 0.85–1.29, p = 0.718). While this univariable relationship might theoretically reflect greater lymphatic disruption or more aggressive disease biology, causation cannot be established from these data. The effect appears largely explained by its correlation with the extent of nodal dissection (total nodes removed), which remained independently significant. Similar associative patterns were reported by Ahmed et al. and Zou L et al. [28, 29], though the non-independent nature of this variable in our models suggests these findings require careful reinterpretation in light of surgical extent confounders.
Notably, the study found no association between hormone receptor status or chemotherapy type (taxane vs. trastuzumab) and lymphoedema, contrasting with Western cohorts where HER2-targeted therapies correlate with higher incidence [30].
The ROC-derived cutoffs (BMI ≥ 27.2 kg/m2; nodal dissection ≥ 16 nodes; positive nodes ≥ 3) offer clinically actionable thresholds for risk stratification despite suboptimal individual diagnostic performance (AUCs ≤ 0.72). These parameters should be integrated into a combinatorial risk assessment model rather than used in isolation, prioritizing patients with multiple risk factors—particularly those receiving axillary radiation (highest independent risk) or with left-sided tumours. We propose intensified surveillance (e.g. Clinical examination and assessment patient-reported outcomes, quarterly bioimpedance measurements, early physiotherapy referral) for high-risk cohorts defined by:
Major criterion: Axillary radiation
- Minor criteria:
-
oBMI ≥ 27.2 kg/m.2
-
o≥ 16 lymph nodes removed
-
oLeft-sided primary tumour
-
o
The high negative predictive values (NPV 82–88%) support using these cutoffs to identify low-risk patients who may require less intensive monitoring. Future studies should validate these algorithms and incorporate radiation status and laterality to optimize resource allocation in lymphedema prevention programs.
By bridging critical gaps in region-specific evidence, this study provides hope for prioritizing high-risk patients and adopt preventive measures such as pre-operative BMI optimization, tailored radiation planning, early identification and referrals for high-risk cohorts. Future research should explore genetic and biomarker profiles (e.g., miR-21, TGF-β levels) to refine predictive models and enable precision medicine approaches. In resource-constrained settings, integrating BCRL surveillance into national cancer control programs and training community health workers in early symptom recognition could mitigate long-term disability, underscoring the imperative of holistic, patient-centred survivorship care.
Limitations
The study's sample size limits its statistical power for detecting small-effect associations and subgroup analyses, increasing Type II errors for small-effect predictors. While BMI, nodal dissection, and radiation emerged as robust predictors in our cohort, their effect sizes and clinical utility require validation in larger, multi-centre studies before implementation in guidelines. While our findings identified risk factors for early-onset lymphedema, long-term studies are needed to assess predictors of late onset lymphedema. Several confounders exist, like comorbidity granularity (diabetic status recorded but not glycaemic control/neuropathy severity), treatment specifics (chemotherapy timing/agents and radiation dosimetry/fields), and post-operative factors (drain duration, compression compliance, rehabilitation intensity),) may influence risk estimates. Future studies should incorporate detailed treatment and comorbidity data. This study lacks patient-reported outcomes, potentially limiting sensitivity for stage 0–1 lymphedema. Circumferential measurement with 10% cut- off was chosen due to resource-limited settings, balancing specificity (reducing false positives) with feasibility. The association of left-sided tumours remains speculative. Though inter observer reliability was not done which may contribute to systemic error, but similar studies on circumferential methods demonstrate high intra observer reliability [31]. Generalizability is limited by single-centre model and may not reflect community-level LMIC resource limitations affecting treatment decisions.
Conclusion
This study confirms a substantial early onset lymphoedema burden (26%) in Indian patients after Modified Radical Mastectomy. Key findings support a practical risk-stratification approach using readily available clinical factors: elevated BMI, left-sided tumour laterality, extent of nodal dissection, and post-operative axillary radiation. Notably, the novel finding of increased lymphoedema risk with left-sided tumours suggests potential anatomical or biomechanical influences, but needs further exploration before causation can be determined. Future studies should combine circumferential measurements with validated PRO tools to correlate objective findings with symptom burden. We propose prioritizing intensified surveillance for patients possessing multiple risk factors (Axillary radiation, higher BMI, extensive lymph node dissection). This targeted strategy can be crucial for optimizing resource allocation in constrained settings, enabling pre-operative weight management and early intervention in high-risk groups. Validation of these predictors through larger multi-centre studies is essential to establish standardized, evidence-based lymphoedema prevention protocols.
Abbreviations
- AIC
Akaike Information Criterion
- AJCC
American Joint Committee on Cancer
- ALND
Axillary Lymph Node Dissection
- aOR
Adjusted Odds Ratio
- AUC
Area Under the Curve
- AUROC
Area Under the Receiver Operating Characteristic Curve
- BCRL
Breast Cancer-Related Lymphoedema
- BMI
Body Mass Index
- CI
Confidence Interval
- CTRI
Clinical Trials Registry—India
- DCIS
Ductal Carcinoma In Situ
- ECM
Extracellular Matrix
- FOXC2
Forkhead Box Protein C2
- HER2
Human Epidermal Growth Factor Receptor 2
- H&L
Hosmer-Lemeshow test
- IEC
Institutional Ethics Committee
- IQR
Interquartile Range
- LMICs
Low- and Middle-Income Countries
- LVI
Lymphovascular Invasion
- miR-21
MicroRNA-21
- MRM
Modified Radical Mastectomy
- NPV
Negative Predictive Value
- OR
Odds Ratio
- PPV
Positive Predictive Value
- ROC
Receiver Operating Characteristic
- ROS
Reactive Oxygen Species
- SD
Standard Deviation
- TGF-β
Transforming Growth Factor Beta
- VEGF-C
Vascular Endothelial Growth Factor-C
Authors’ contributions
Conceptualization: FH, SB. Data Collection: MO, FH, SR, VSK, KA, NR, AJ, AA. Manuscript Preparation (Drafting): MO. Manuscript Review & Editing: FH, SB. Final Approval of Manuscript: All authors. Accountability for All Aspects of the Work: All authors. Guarantor of the Article: FH. All authors meet the ICMJE criteria for authorship and have read and approved the final manuscript.
Funding
No funding was received to conduct this study.
Data availability
Data supporting research can be accessed/obtained by contacting the first author, Mohd Ozair Khan: [mailto:ozairkhan321@gmail.com](mailto:ozairkhan321@gmail.com).
Declarations
Ethics approval and consent to participate
The study was approved by the Institutional Ethics Committee (All India Institute of Medical Sciences, Rishikesh) (AIIMS/IEC/23/212) (In accordance with the Declaration of Helsinki).
All Patients provided informed consent to participate in the study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229–63. 10.3322/caac.21834. [DOI] [PubMed] [Google Scholar]
- 2.DiSipio T, Rye S, Newman B, Hayes. Incidence of Unilateral Arm Lymphoedema after Breast Cancer: A Systematic Review and Meta-Analysis. Lancet Oncol.2013; 14(6): 500–515. 10.1016/s1470-2045(13)70076-7. [DOI] [PubMed]
- 3.Dunne M, Keenan K. Late and Long-Term Sequelae of Breast Cancer Treatment. Am J Nursing. 2016;116(6):36–45. 10.1097/01.naj.0000484223.07306.45. [DOI] [PubMed] [Google Scholar]
- 4.Rockson SG. Lymphedema. Am J Med. 2001;110(4):288–95. 10.1016/S0002-9343(00)00727-0. [DOI] [PubMed] [Google Scholar]
- 5.Mathur P, Sathishkumar K, Chaturvedi M, Das P, Sudarshan KL, Santhappan S, Nallasamy V, John A, Narasimhan S, Roselind FS. Cancer Statistics, 2020: Report from National Cancer Registry Programme, India. JCO Global Oncology. 2020; 6(6): 1063–75. 10.1200/go.20.00122. [DOI] [PMC free article] [PubMed]
- 6.Suami H, Taylor GI, Pan W-R. The lymphatic territories of the upper limb: anatomical study and clinical implications. Plast Reconstr Surg. 2007;119(6):1813–22. 10.1097/01.prs.0000246516.64780.61. [DOI] [PubMed] [Google Scholar]
- 7.Shaitelman SF, Schlembach PJ, Arzu I, Ballo M, Bloom ES, Buchholz D, et al. Acute and short-term toxic effects of conventionally fractionated vs hypofractionated whole-breast irradiation: a randomized clinical trial. JAMA Oncol. 2015;1(7):931–41. 10.1001/jamaoncol.2015.2666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Itoh F, Watabe T. TGF-β signaling in lymphatic vascular vessel formation and maintenance. Front Physiol. 2022. 10.3389/fphys.2022.1081376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zampell JC, Yan A, Avraham T, Daluvoy S, Weitman ES, Mehrara BJ. HIF-1α: coordinates lymphangiogenesis during wound healing and in response to inflammation. FASEB J. 2012;26(3):1027–39. 10.1096/fj.11-195321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mittal M, Siddiqui MR, Tran K, Reddy SP, Malik AB. Reactive oxygen species in inflammation and tissue injury. Antioxid Redox Signal. 2014;20(7):1126–67. 10.1089/ars.2012.5149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rhoten BA. Hide and Seek: Body Image-Related Issues for Breast Cancer Survivors with Lymphedema. J Womens Health Issues Care. 2015; 4(2). 10.4172/2325-9795.1000180.
- 12.Rockson SG, Tian W, Jiang X, Kuznetsova T, Haddad F, Zampell J, et al. Pilot studies demonstrate the potential benefits of antiinflammatory therapy in human lymphedema. JCI Insight. 2018. 10.1172/jci.insight.123775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wadasadawala T, Mohanty SK, Sen S, Kanala TS, Maiti S, Puchali N, et al. Out-of-pocket payment and financial risk protection for breast cancer treatment: a prospective study from India. The Lancet Regional Health. 2024. 10.1016/j.lansea.2023.100346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mehrotra R, Yadav K. Breast cancer in India: present scenario and the challenges ahead. World J Clin Oncol. 2022;13(3):209–18. 10.5306/wjco.v13.i3.209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Starritt EC, Joseph D, McKinnon JG, Lo SK, De Wilt JH, Thompson JF. Lymphedema after complete axillary node dissection for melanoma. Ann Surg. 2004;240(5):866–74. 10.1097/01.sla.0000143271.32568.2b. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Gillespie TC, Sayegh HE, Brunelle CL, Daniell KM, Taghian AG. Breast cancer-related lymphedema: risk factors, precautionary measures, and treatments. Gland Surg. 2018;7(4):379–403. 10.21037/gs.2017.11.04. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Manirakiza A, Irakoze L, Shui L, Manirakiza S, and Ngendahayo L. Lymphoedema after Breast Cancer Treatment Is Associated with Higher Body Mass Index: A Systematic Review and Meta-Analysis. East African Health Res J. 2019; 3(2): 178–92. 10.24248/EAHRJ-D-19-00009. [DOI] [PMC free article] [PubMed]
- 18.Degnim AC, Miller J, Hoskin TL, Boughey JC, Loprinzi M, Thomsen K, et al. A prospective study of breast lymphedema: frequency, symptoms, and quality of life. Breast Cancer Res Treat. 2012;134(3):915–22. 10.1007/s10549-012-2004-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Clark B, Sitzia J, Harlow W. Incidence and risk of arm oedema following treatment for breast cancer: a three-year follow-up study. QJM. 2005;98(5):343–8. 10.1093/qjmed/hci053. [DOI] [PubMed] [Google Scholar]
- 20.Dominick SA, Madlensky L, Natarajan L, Pierce JP. Risk factors associated with breast cancer-related lymphedema in the WHEL study. J Cancer Surviv. 2013;7(1):115–23. 10.1007/s11764-012-0251-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kwan ML. Risk factors for lymphedema in a prospective breast cancer survivorship study. Arch Surg. 2010;145(11):1055. 10.1001/archsurg.2010.231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Mak SS, Yeo W, Lee YM, Mo KF, Ka Tse KY, Tse SM, et al. Predictors of Lymphedema in Patients with Breast Cancer Undergoing Axillary Lymph Node Dissection in Hong Kong. Nurs Res. 2008;57(6):416–25. 10.1097/nnr.0b013e31818c3de2. [DOI] [PubMed] [Google Scholar]
- 23.Greene AK, Maclellan RA. Obesity-induced upper extremity lymphedema. Plast Reconstr Surg Glob Open. 2013;1(7):e59. 10.1097/gox.0b013e3182a96359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Arngrim N, Simonsen L, Holst JJ, Bülow J. Reduced adipose tissue lymphatic drainage of macromolecules in obese subjects: a possible link between obesity and local tissue inflammation? Int J Obes. 2012;37(5):748–50. 10.1038/ijo.2012.98. [DOI] [PubMed] [Google Scholar]
- 25.Singh AP, Singh A, Arya R, Naik A, Gurjar OP. Prevalence of lymphedema and associated risk factors following multimodal breast cancer treatment. J Med Sci Health. 2023;9(3):310–5. 10.46347/jmsh.v9i3.22.300. [Google Scholar]
- 26.Yen TW, Fan X, Sparapani R, Laud PW, Walker AP, Nattinger AB. A contemporary, population-based study of lymphedema risk factors in older women with breast cancer. Ann Surg Oncol. 2009;16(4):979–88. 10.1245/s10434-009-0347-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Beaulac SM. Lymphedema and quality of life in survivors of early-stage breast cancer. Arch Surg. 2002;137(11):1253. 10.1001/archsurg.137.11.1253. [DOI] [PubMed] [Google Scholar]
- 28.Zou L, Liu FH, Shen PP, Hu Y, Liu XQ, Xu YY, et al. The incidence and risk factors of related lymphedema for breast cancer survivors post-operation: a 2-year follow-up prospective cohort study. Breast Cancer. 2018;25(3):309–14. 10.1007/s12282-018-0830-3. [DOI] [PubMed] [Google Scholar]
- 29.Ahmed RL, Schmitz KH, Prizment AE, Folsom AR. Risk factors for lymphedema in breast cancer survivors, the Iowa women’s health study. Breast Cancer Res Treat. 2011;130(3):981–91. 10.1007/s10549-011-1667-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kim MM, Dawood S, Allen P, Sahin AA, Woodward WA, Smith BD, et al. Hormone receptor status influences the locoregional benefit of trastuzumab in patients with nonmetastatic breast cancer. Cancer. 2012;118(20):4936–43. 10.1002/cncr.27502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bakar Y, Özdemir ÖC, Sevim S, Duygu E, Tuğral A, Sürmeli M. Intra-observer and inter-observer reliability of leg circumference measurement among six observers: a single blinded randomized trial. J Med Life. 2017;10(3):176–181. PMCID: PMC5652265 [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data supporting research can be accessed/obtained by contacting the first author, Mohd Ozair Khan: [mailto:ozairkhan321@gmail.com](mailto:ozairkhan321@gmail.com).