Abstract
Background: The International Society of Lymphology (ISL) guidelines have established grading criteria for primary lower limb lymphedema (PLEL), but there is a lack of a unified standard for assessing disease severity. Purpose: This study aimed to establish and validate a predictive model for evaluating severe PLEL. Methods and Materials: This retrospective study included 226 patients with unilateral PLEL (2018–2023), divided into non-severe (143) and severe (83) groups according to ISL criteria. Researchers recorded 26 MRI and 15 clinical features. Logistic regression, following feature selection via Lasso regularization, was used to construct a nomogram. Results: Six independent risk factors were identified: parallel line sign, crescent sign, longitudinal range, band sign thickness, fat area, and fat diameter. The model achieved a training set AUC of 0.908 and a validation set AUC of 0.891. Conclusions: The severity risk prediction model based on MRI and clinical practice has good discriminatory power and accuracy, providing a reference for individualized clinical prediction.
Introduction
Primary lower limb lymphedema (PLEL) is a chronic, debilitating condition resulting from congenital lymphatic malformations. Currently, the International Society of Lymphology (ISL) grades severity based on volume differences between affected and healthy limbs. However, volume measurement—via water displacement or manual circumference checks—is labor-intensive, prone to error, and fails to capture underlying structural changes. There is a critical need for a predictive model that integrates imaging and clinical data to accurately categorize disease severity and guide intervention.
Materials and Methods
This retrospective study analyzed 226 patients with unilateral PLEL. The cohort was categorized into "non-severe" (mild/moderate) and "severe" groups. A total of 41 features (26 MRI-based, 15 clinical) were extracted.
Feature selection was performed using one-way ANOVA followed by L1 regularization (Lasso), which effectively isolates the most significant predictors by penalizing less relevant variables. The final nomogram was built using logistic regression and validated through 5-fold cross-validation and Decision Curve Analysis (DCA) to assess clinical net benefit.
Results and Clinical Characteristics
The analysis identified six independent risk factors for severe PLEL. Notably, clinical symptoms like skin redness and poor elasticity were significantly more prevalent in the severe group.
As shown in Table 1, 96.4% of severe patients exhibited involvement across the entire lower limb (longitudinal range), compared to 69.9% in the non-severe group. The model demonstrated high discriminative power, with a training AUC of 0.908 (95% CI: 0.868–0.947) and a validation AUC of 0.891 (95% CI: 0.847–0.935).
Discussion
The model highlights several key biological markers of disease progression. The inclusion of fat area and fat diameter as risk factors aligns with the understanding that lymphedema triggers adipose and fibrous tissue proliferation in the subcutaneous layer.
The "crescent sign" (subfascial effusion) was identified as a critical tipping point, suggesting that as fluid accumulates, it eventually breaches the superficial fascia into the subfascial space. Conversely, the "parallel line sign" (dilated subcutaneous lymphatic capillaries) was more common in non-severe cases, indicating it may be a hallmark of earlier-stage disease. The high prevalence of whole-limb involvement in severe cases suggests that advanced PLEL involves both peripheral and central lymphatic deficiencies.
Conclusion
The developed nomogram offers a robust, non-invasive tool for predicting PLEL severity. While limited by its retrospective nature and focus on primary lymphedema, the model provides a significant advancement for individualized clinical management. Future research should focus on multi-class models and validation in secondary lymphedema populations.
Summary updated