Abstract
Background
Breast cancer-related lymphedema (BCRL) can substantially impair health-related quality of life. Patient-reported outcome measures (PROMs) may complement objective limb measurements by capturing symptoms, functional limitations, and psychosocial consequences; however, comparative evidence on BCRL-specific PROMs and validated Chinese versions remains limited. This study aimed to cross-culturally adapt four BCRL-specific PROMs and compare their classical and Rasch measurement properties, known-groups validity, and exploratory discrimination relative to a circumference-based reference classification.
Methods
The Upper Limb Lymphedema 27 (ULL-27), Lymphedema Quality of Life Questionnaire for Upper Limbs (LYMQOL-ARM), Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema (Lymph-ICF-UL), and Lymphedema Life Impact Scale Version 2 (LLIS) were translated and culturally adapted according to international standards. In total, 204 Chinese women treated for breast cancer participated: 104 completed ULL-27 and LYMQOL-ARM, and 100 completed Lymph-ICF-UL and LLIS. Content validity, internal consistency, construct validity, known-groups validity across BCRL stages, and Rasch measurement properties were evaluated. Exploratory receiver operating characteristic (ROC) analyses assessed discrimination relative to the study reference classification. PROM scores were also compared between axillary lymph node dissection (ALND) and sentinel lymph node biopsy (SLNB) recipients who did not meet the circumference-based BCRL criterion.
Results
The four PROMs were generally comprehensible and culturally relevant. All subscales demonstrated adequate weighted likelihood estimation reliability (0.730–0.921) and high internal consistency (Cronbach’s α >0.90). Of 93 items, 79 (84.9%) showed acceptable fit, and all response-category thresholds were ordered. Evidence generally supported essential unidimensionality, although local dependence, floor effects, and selected measurement-invariance concerns were identified. LLIS and Lymph-ICF-UL scores differed most consistently across BCRL stages. The largest area under the curve (AUC) estimates were observed for Lymph-ICF-UL Symptoms [0.734; 95% confidence interval (CI): 0.633–0.834], LLIS Physical (0.725; 95% CI: 0.625–0.825), and LLIS total score (0.680; 95% CI: 0.573–0.787); no pairwise AUC difference remained significant after Holm adjustment. Among circumference-negative participants, LLIS total, Physical, and Psychosocial scores and Lymph-ICF-UL Symptoms and Mental Function scores were significantly higher after ALND than after SLNB.
Conclusions
The Chinese versions of the four PROMs showed generally supportive but nonuniform measurement properties. LLIS and Lymph-ICF-UL performed comparatively well in distinguishing clinically defined BCRL stages, while LLIS Physical and Lymph-ICF-UL Symptoms showed favorable lower-end targeting and the largest exploratory AUC estimates. These PROMs may complement clinical evaluation and objective limb assessment by capturing patient-reported burden not fully reflected by circumference measurements. Because the ROC analyses used an imperfect reference classification and the study was cross-sectional, diagnostic and predictive validity cannot be established. Larger, prospectively powered longitudinal studies incorporating more sensitive objective assessments are required.
Keywords: Breast cancer-related lymphedema (BCRL), patient-reported outcome measures (PROMs), psychometric properties, Rasch analysis, surveillance
Highlight box.
Key findings
• The four culturally adapted Chinese patient-reported outcome measures (PROMs) for breast cancer-related lymphedema (BCRL) demonstrated generally acceptable but nonuniform measurement properties. Lymphedema Life Impact Scale Version 2 (LLIS) and selected Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema (Lymph-ICF-UL), domains showed relatively favorable stage-related and exploratory discriminatory findings.
What is known and what is new?
• PROMs capture symptoms and functional and psychosocial burden that objective limb measurements may not fully reflect, but validated Chinese BCRL-specific instruments and comparative psychometric evidence remain limited.
• This study cross-culturally adapted four PROMs and comparatively evaluated their classical and Rasch measurement properties, known-groups validity, and exploratory discrimination, identifying instrument-specific strengths and limitations.
What is the implication, and what should change now?
• BCRL-specific PROMs should complement clinical evaluation and objective limb assessment. Instrument selection should be guided by the intended construct and measurement context, with LLIS and selected Lymph-ICF-UL domains showing particular promise for assessing clinical burden.
Introduction
Breast cancer is the most commonly diagnosed malignancy among women worldwide, and its incidence continues to rise (1). Advances in early detection and treatment have substantially improved survival (2); consequently, greater attention is being directed toward the long-term health-related quality of life (HRQoL) of breast cancer survivors. Breast cancer-related lymphedema (BCRL) is a potentially debilitating complication, affecting approximately 20% of patients after axillary lymph node dissection (ALND) and 6% after sentinel lymph node biopsy (SLNB) (3). Notably, upper-limb discomfort and functional limitations are also frequently reported after SLNB (4). Through physical symptoms such as limb swelling and pain, recurrent infections, prolonged treatment requirements, and psychosocial distress, BCRL can profoundly impair HRQoL (5).
Improving quality of life is a central goal of lymphedema management (6). Sackey et al. (7). demonstrated that patient-perceived lymphedema burden was more strongly associated with long-term deterioration in HRQoL after breast cancer surgery than objective measurements. This finding highlights the essential role of patient-reported outcome measures (PROMs) in capturing the symptoms, functional limitations, and psychosocial consequences experienced by patients, including aspects of disease burden that may not be fully represented by objective limb measurements.
Early recognition and intervention are clinically important because intervention at an earlier stage may lead to more favorable outcomes (8-11). However, some patients experience upper-limb symptoms and functional concerns before changes meet conventional objective thresholds (12,13). The limited ability of circumference-based measurements alone to characterize symptom burden and treatment-related consequences during this period represents an important clinical gap. By directly capturing the patient experience, PROMs may provide complementary information that helps identify concerns warranting closer clinical evaluation, repeated surveillance, or more sensitive objective assessment. Serial PROM assessment may also be useful for monitoring changes in symptoms and treatment response over time. Nevertheless, whether PROM findings can predict subsequent objectively detectable BCRL requires confirmation in prospective longitudinal studies.
Several validated lymphedema-specific PROMs are available internationally, including the Upper Limb Lymphedema 27 (ULL-27), Lymphedema Quality of Life Questionnaire for Upper Limbs (LYMQOL-ARM), Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema (Lymph-ICF-UL), and Lymphedema Life Impact Scale Version 2 (LLIS) (6). Previous systematic reviews have primarily examined their psychometric properties (14-16). However, direct comparative evidence regarding their measurement performance, ability to distinguish clinically defined BCRL stages, and potential to complement conventional surveillance remains limited. Heterogeneity in instrument selection across clinical and research settings also impedes comparison and synthesis of patient-reported data. A head-to-head evaluation is therefore needed to clarify the relative measurement strengths of these PROMs and identify domains that may be particularly informative within BCRL surveillance and management.
To address these gaps, we cross-culturally adapted four widely used lymphedema-specific PROMs for Chinese patients with breast cancer and conducted a comprehensive comparative evaluation of their classical and Rasch measurement properties. We further examined their known-groups validity across clinically defined BCRL stages and explored their discriminatory performance relative to the study reference classification. Through this work, we aimed to establish a measurement basis for selecting appropriate Chinese BCRL-specific PROMs and for further evaluating their potential integration into longitudinal surveillance and management, where they may systematically capture symptoms and consequences not fully represented by circumference measurements and inform further assessment.
Methods
PROMs of upper limb lymphedema
The lymphedema-specific QoL scales used included the ULL-27 (17), LYMQOL-ARM (18,19), Lymph-ICF-UL (20,21), and LLIS (22). The reliability and validity of these original scales have been tested previously. The components and score calculation of the four PROMs are presented in Table S1. Higher scores indicate lower QoL. Any reference to LLIS refers to version 2 unless stated otherwise in this article.
The validated Chinese version of the 36-item Short-Form Health Survey Questionnaire (SF-36) was used to measure general QoL (23). SF-36 scores comprise a physical component summary (PCS) and a mental component summary (MCS). Higher scores indicate better QoL.
Study design
This study comprised two phases: standardized translation and cultural adaptation (24,25) of four BCRL-specific PROMs, followed by a comparative psychometric evaluation based on classical test theory (CTT) (26) and Rasch measurement principles (27). The study design and interpretation were informed by current Consensus-based Standards for the selection of health Measurement Instruments (COSMIN) guidance (28,29).
This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Fudan University Shanghai Cancer Center Institutional Review Board (SCCIRB) (No. 2302269-16). Informed consent was obtained from all individual participants included in the study.
Linguistic validation
The scales were translated into Chinese through a rigorous process (24,25). Two native Chinese speakers, a linguist and a clinical medical professional proficient in English, performed independent forward translations. The two forward-translated versions were subsequently merged, and any discrepancies were discussed to reach a consensus. Subsequently, two native English speakers bilingual in Chinese blindly translated the synthesized forward translation back into English to identify misunderstandings or inaccuracies. The revised version was reviewed and approved by a committee of clinicians and translators. A pilot test with ten women was conducted to assess the comprehensibility and cultural relevance of the questionnaires. Based on feedback from the pilot test, the expert committee re-evaluated the text and finalized the Chinese versions (Appendix 1).
Participants
Consecutive patients attending the lymphedema outpatient clinic at the Fudan University Shanghai Cancer Center were included. The eligibility criteria were as follows: (I) clinical diagnosis of lymphedema, (II) age 18 years or older, (III) previous breast cancer diagnosis and surgical treatment, and (IV) native Chinese speakers. The exclusion criteria were as follows: (I) upper limb edema unrelated to breast cancer treatment, (II) acute infection, (III) undergoing radiotherapy, (IV) metastasis, and (V) cognitive impairment. Participants in the comparison group met the same criteria, except that they were not diagnosed with lymphedema. To facilitate the psychometric validation process, each participant completed the SF-36 along with two of the four Chinese lymphedema-specific PROMs: ULL-27 with the LYMQOL-ARM or Lymph-ICF-UL with the LLIS.
We prospectively targeted at least 100 participants for each questionnaire set in accordance with the commonly used COSMIN sample-size criterion for studies of measurement properties. A total of 104 participants completed the ULL-27 and LYMQOL-ARM, and 100 completed the LLIS and Lymph-ICF-UL. No separate a priori power calculation was performed for the ROC analyses, which were considered exploratory.
Diagnosis and staging of lymphedema
The circumferences of both upper limbs were measured at seven specific levels: thumb web space, wrist, 10 cm above the wrist, 10 cm below the elbow, elbow, 10 cm above the elbow, and axilla. Lymphedema was diagnosed when the circumference of the affected arm was at least 2 cm larger than that of the unaffected arm at any level. Lymphedema staging was determined based on a combination of circumference measurements and physician assessment following the guidelines of the International Society of Lymphology (30). Clinical staging assessments were performed by trained nurses and clinicians who were blinded to PROM scores at the time of assessment.
Content validity
To assess content validity, face-to-face interviews were conducted with 15 patients diagnosed with BCRL for each scale to determine whether the items on the scales accurately and comprehensively captured the challenges and impact of lymphedema on their lives. The interviews also aimed to identify redundant items or important problems missing from the scale.
Rasch analysis
Partial credit models were fitted separately to each subscale using the TAM and eRm packages in R. Model performance was evaluated in terms of weighted likelihood estimation (WLE) reliability, item fit, unidimensionality, response-category functioning, local dependence, targeting, and differential item functioning (DIF).
WLE reliability of ≥0.70 was considered adequate for group-level comparisons, and Infit and Outfit mean-square values of 0.5–1.5 indicated acceptable item fit. Unidimensionality was assessed using principal component analysis of Rasch residuals, Smith’s paired t-test, and the Martin–Löf likelihood-ratio test. Response categories were evaluated for threshold ordering, and local dependence was examined using adjusted Yen’s Q3 statistics. Targeting was assessed from person–item distributions and floor and ceiling effects. DIF by age, education, and surgical side relative to arm dominance was examined using Andersen likelihood-ratio and item-level Wald tests, with Holm adjustment where estimable. Some DIF statistics were nonestimable because of limited subgroup sizes and sparse response patterns, and DIF was not assessed for Lymph-ICF-UL because of its large number of response categories.
CTT indices
Internal consistency of subscales in the four PROMs was assessed using Cronbach’s α coefficients. Values >0.9 were categorized as excellent, values >0.7 as good, and values >0.6 as acceptable.
Construct validity was examined using prespecified correlations between PROM scores and conceptually relevant SF-36 domains and between conceptually related domains of the BCRL-specific instruments. The direction and magnitude of expected correlations were specified according to the constructs measured.
Known-groups validity was evaluated by comparing PROM subscale and total scores across BCRL stages using the Kruskal-Wallis test. Associations between PROM scores and BCRL stage were also assessed using Spearman’s rank correlation coefficients.
ROC curves were used to assess the discrimination of PROMs relative to the objective classification. AUCs were reported with 95% CIs. Sensitivity, specificity, and data-derived cutoffs were calculated where applicable. Pairwise differences in AUCs were assessed using DeLong tests for unpaired ROC curves. P values for pairwise comparisons were adjusted using the Holm method.
Exploratory analysis in objective-negative group
As an exploratory, hypothesis-generating analysis, PROM scores were compared between ALND and SLNB recipients who did not meet the circumference-based BCRL criterion. ALND was treated as a surgical exposure associated with greater risk of BCRL morbidity. The purpose of this analysis was to explore whether PROMs captured symptoms or consequences among circumference-negative patients that were not reflected by the circumference criterion. Findings were interpreted in consideration of both lymphatic and nonlymphatic effects of axillary surgery.
Statistical analysis
Continuous variables were summarized using means and 95% CIs or distribution-appropriate statistics. Categorical variables were summarized using frequencies and percentages.
Questionnaire completion rates were 100% for all four PROMs. No item responses were inadvertently omitted in the collected questionnaires. “Not applicable” responses (only from Lymph-ICF) were treated as structurally missing item responses. They were retained as distinct codes in the source dataset and recoded as missing (NA) for scoring and Rasch analyses.
IBM SPSS Statistics 25 software (IBM Corp., Armonk, NY, USA) and R (version 4.2.1) were used for all statistical analyses. Statistical significance was set at P<0.05.
Results
Sociodemographic and clinical features
A total of 204 female breast cancer patients were included. Of these, 104 patients (61 diagnosed with lymphedema and 43 without lymphedema) completed LYMQOL-ARM and ULL-27, and 100 patients (59 with lymphedema and 41 without lymphedema) completed Lymph-ICF-UL and LLIS. The demographic and clinical characteristics of the patients are shown in Table 1. Most patients with BCRL underwent ALND, comprising 93.4% of patients in the LYMQOL-ARM and ULL-27 groups and 86.4% of patients in the Lymph-ICF-UL and LLIS groups. Most patients completed the questionnaires >1 month postoperatively, with 93.3% in the LYMQOL-ARM and ULL-27 groups and 96% in the Lymph-ICF-UL and LLIS groups, thereby minimizing the potential interference of surgical wounds on their responses to BCRL.
Table 1. Demographic and medical information of participants (n=204).
| Variable | ULL-27 & LYMQOL-ARM | LLIS & Lymph-ICF | P value | ||||
|---|---|---|---|---|---|---|---|
| Lymphedema (N=61) | Non-lymphedema (N=43) | Overall (N=104) | Lymphedema (N=59) | Non-lymphedema (N=41) | Overall (N=100) | ||
| Age (years) | 52.9±12.0 | 48.8±10.8 | 51.2±11.6 | 53.2±11.5 | 50.5±10.7 | 52.1±11.2 | 0.57 |
| BMI (kg/m2) | 23.8±3.13 | 23.5±3.72 | 23.7±3.37 | 23.6±2.64 | 23.5±2.93 | 23.5±2.75 | 0.75 |
| Education | 0.45 | ||||||
| Middle school or below | 17 (27.9) | 11 (25.6) | 28 (26.9) | 25 (42.4) | 9 (22.0) | 34 (34.0) | |
| High school | 13 (21.3) | 3 (7.0) | 16 (15.4) | 9 (15.3) | 7 (17.1) | 16 (16.0) | |
| Diplomas | 9 (14.8) | 10 (23.3) | 19 (18.3) | 10 (16.9) | 11 (26.8) | 21 (21.0) | |
| Bachelor or higher | 22 (36.1) | 19 (44.2) | 41 (39.4) | 15 (25.4) | 14 (34.1) | 29 (29.0) | |
| Marital status | 0.54 | ||||||
| Single | 2 (3.3) | 3 (7.0) | 5 (4.8) | 2 (3.4) | 0 (0) | 2 (2.0) | |
| Married | 52 (85.2) | 33 (76.7) | 85 (81.7) | 50 (84.7) | 35 (85.4) | 85 (85.0) | |
| Divorced | 4 (6.6) | 5 (11.6) | 9 (8.7) | 2 (3.4) | 4 (9.8) | 6 (6.0) | |
| Widow | 3 (4.9) | 2 (4.7) | 5 (4.8) | 5 (8.5) | 2 (4.9) | 7 (7.0) | |
| Occupation | 0.88 | ||||||
| Employed | 21 (34.4) | 22 (51.2) | 43 (41.3) | 20 (33.9) | 19 (46.3) | 39 (39.0) | |
| Housewife | 12 (19.7) | 7 (16.3) | 19 (18.3) | 14 (23.7) | 7 (17.1) | 21 (21.0) | |
| Retired | 28 (45.9) | 14 (32.6) | 42 (40.4) | 25 (42.4) | 15 (36.6) | 40 (40.0) | |
| Time since surgery | 0.37 | ||||||
| ≤1 months | 5 (8.2) | 2 (4.7) | 7 (6.7) | 4 (6.8) | 0 (0) | 4 (4.0) | |
| 1–6 months | 29 (47.5) | 35 (81.4) | 64 (61.5) | 23 (39.0) | 31 (75.6) | 54 (54.0) | |
| 0.5–2 years | 19 (31.1) | 3 (7.0) | 22 (21.2) | 21 (35.6) | 10 (24.4) | 31 (31.0) | |
| ≥2 years | 8 (13.1) | 3 (7.0) | 11 (10.6) | 11 (18.6) | 0 (0) | 11 (11.0) | |
| Axillary surgery | 0.97 | ||||||
| SLNB | 4 (6.6) | 17 (39.5) | 21 (20.2) | 8 (13.6) | 12 (29.3) | 20 (20.0) | |
| ALND | 57 (93.4) | 26 (60.5) | 83 (79.8) | 51 (86.4) | 29 (70.7) | 77 (80.0) | |
| Receiving chemotherapy | 0.26 | ||||||
| No | 44 (72.1) | 20 (46.5) | 64 (61.5) | 44 (74.6) | 25 (61.0) | 69 (69.0) | |
| Yes | 17 (27.9) | 23 (53.5) | 40 (38.5) | 15 (25.4) | 16 (39.0) | 31 (31.0) | |
| Affected extremity | 0.91 | ||||||
| Bilateral | 2 (3.3) | 0 (0) | 2 (1.9) | 1 (1.7) | 1 (2.4) | 2 (2.0) | |
| Right | 34 (55.7) | 15 (34.9) | 49 (47.1) | 26 (44.1) | 24 (58.5) | 50 (50.0) | |
| Left | 25 (41.0) | 28 (65.1) | 53 (51.0) | 32 (54.2) | 16 (39.0) | 48 (48.0) | |
| BCRL stage | 0.97 | ||||||
| 0 | 0 (0) | 43 (100) | 43 (41.3) | 0 (0) | 41 (100) | 41 (41.0) | |
| 1 | 37 (60.7) | 0 (0) | 37 (35.6) | 37 (62.7) | 0 (0) | 37 (37.0) | |
| 2 | 24 (39.3) | 0 (0) | 24 (23.1) | 22 (37.3) | 0 (0) | 22 (22.0) | |
Continuous variables are expressed as mean ± SD and compared using an unpaired, 2-tailed t-test. Categorical variables were compared using the χ2 test. ALND, axillary lymph node dissection; BCRL, breast cancer-related lymphedema; BMI, body mass index; LLIS, Lymphedema Life Impact Scale Version 2; Lymph-ICF, Lymphedema Functioning, Disability and Health Questionnaire; LYMQOL-ARM, Lymphedema Quality of Life Questionnaire for Upper Limbs; SLNB, sentinel lymph node biopsy; ULL-27, Upper Limb Lymphedema 27.
Content validity
During pretesting, minor revisions were made to improve cultural relevance while preserving the original concepts. In the Lymph-ICF-UL, “Work in the garden” was adapted to “Arrange flowers and plants at home”, and “Sunbathe” to “Engage in outdoor activities in the sun”. Patient interviews generally supported the comprehensibility and relevance of the four PROMs but identified potential content limitations: LYMQOL-ARM did not explicitly address swelling or tightness, whereas “difficulties walking” in Lymph-ICF-UL and ULL-27 was considered less relevant to BCRL.
Rasch analysis results
All 15 partial credit models converged successfully, with WLE reliability ranging from 0.730 to 0.921. Fourteen subscales had first residual-component eigenvalues below 3.0; Smith’s test supported unidimensionality in 8 of 14 evaluable subscales, and all 12 evaluable Martin–Löf tests were nonsignificant. Overall, these findings supported essential unidimensionality for most subscales, although the ULL-27 Psychological subscale remained uncertain and some tests were nonestimable.
Of the 93 items, 79 (84.9%) showed acceptable fit, 10 showed potential overfit, and 4 showed potential underfit. All response-category thresholds were ordered. Local dependence was identified in several subscales, particularly ULL-27 Physical Function (Table 2 and Tables S2-S4).
Table 2. Item functioning of LLIS.
| Item | Subscale | Item fit | Threshold ordering | Response dependency | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Outfit MNSQ | Outfit Z | Infit MNSQ | Infit Z | Thresholds | Ordered | Pairs with adjusted Q3 >0.20/No. of item pairs | Flagged pairs | ||||||
| Item_1 | Item_2 | Mean Q3 | Q3 | Adjusted Q3 | |||||||||
| 1 | Physical | 1.411 | 2.39 | 1.284 | 1.8 | −0.391; 1.948; 3.998 | True | 3/15 | llis_2 | llis_3 | −0.166 | 0.231 | 0.397 |
| 2 | 0.449 | −4.7 | 0.439 | −4.74 | −0.972; 1.331; 3.18; 4.917 | True | llis_3 | llis_4 | −0.166 | 0.095 | 0.261 | ||
| 3 | 0.785 | −1.41 | 0.736 | −1.8 | −1.053; 1.422; 2.804; 3.729 | True | llis_5 | llis_6 | −0.166 | 0.075 | 0.241 | ||
| 4 | 0.925 | −0.32 | 0.832 | −1.03 | −0.048; 1.584; 2.875; 3.439 | True | |||||||
| 5 | 0.593 | −2.85 | 0.614 | −2.86 | −0.616; 1.237; 2.821; 3.737 | True | |||||||
| 6 | 1.11 | 0.74 | 1.1 | 0.71 | −1.023; 0.257; 2.268; 4.24 | True | |||||||
| 7 | Psychosocial | 0.545 | −2.67 | 0.456 | −3.97 | 0.314; 1.812; 3.014; 4.194 | True | 3/15 | llis_7 | llis_8 | −0.116 | 0.331 | 0.447 |
| 8 | 0.339 | −3.53 | 0.337 | −4.94 | 0.729; 1.868; 3.282; 3.685 | True | llis_8 | llis_10 | −0.116 | 0.214 | 0.33 | ||
| 9 | 0.816 | −0.59 | 0.976 | −0.06 | 1.046; 2.132; 2.912; 4.183 | True | llis_7 | llis_10 | −0.116 | 0.088 | 0.204 | ||
| 10 | 0.617 | −1.75 | 0.539 | −2.76 | 0.506; 1.859; 2.544; 2.769 | True | |||||||
| 11 | 0.826 | −0.95 | 0.806 | −1.17 | −0.1; 1.343; 2.614; 3.536 | True | |||||||
| 12 | 2.329 | 4.89 | 1.891 | 4.42 | −1.395; 0.521; 1.198; 1.796 | True | |||||||
| 13 | Function | 0.822 | −0.73 | 0.852 | −0.73 | 0.968; 2.621; 3.509; 5.276 | True | 2/10 | llis_13 | llis_16 | −0.213 | 0.164 | 0.377 |
| 14 | 0.666 | −2.11 | 0.619 | −2.36 | −0.398; 2.007; 3.416; 4.359 | True | llis_14 | llis_15 | −0.213 | 0.161 | 0.373 | ||
| 15 | 0.763 | −1.39 | 0.736 | −1.61 | 0.034; 1.644; 2.899; 4.507 | True | |||||||
| 16 | 0.73 | −1.41 | 0.767 | −1.32 | 0.535; 2.191; 3.399; 4.765 | True | |||||||
| 17 | 1.3 | 1.41 | 1.074 | 0.46 | 0.406; 1.993; 2.714; 5.537 | True | |||||||
LLIS, Lymphedema Life Impact Scale Version 2; MNSQ, mean-square fit statistic.
As for targeting, the alignment between item difficulty and the sample was generally satisfactory at the level of mean locations. However, examination of the complete score distribution identified less satisfactory coverage at the lower end of the latent trait. Consistent with this, ceiling effects were low across all subscales (0–3.85%), but floor effects were more prominent and ranged from 3.85% to 43.00%. The largest floor effects were observed for Lymph-ICF-UL Mental function (43.00%), ULL-27 Social dimension (33.65%), LLIS Function (31.00%), LYMQOL-ARM Appearance (26.92%), and LYMQOL-ARM Function (25.00%). By contrast, lower floor effects were found for ULL-27 Physical function (3.85%), Lymph-ICF-UL Symptoms (11.00%), Lymph-ICF-UL Mobility activities (11.00%), and LLIS Physical (13.00%) (Table 3).
Table 3. Reliability, unidimensionality and targeting of the four PROMs.
| PROM | Item | Subscale | WLE reliability | Residual PCA first eigenvalue | Smith t result (95% CI) | Martin Loef result | Person mean (SD) | Target difference | Threshold range | Below threshold range% | Above threshold range% | Floor % | Ceiling % |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ULL-27 | 1-5 | Physical function | 0.921 | 2.758 | 12.5% (7.5–20.2) | χ2=156.56, df=895, P>0.99 | 0.01 (2.02) | 0.008 | −2.12 to 5.12 | 13 | 0 | 4 | 0 |
| 16-22 | Psychological dimension | 0.859 | 3.425 | 17.3% (11.2–25.7) | Not estimated | −0.08 (2.47) | −0.08 | −2.77 to 5.55 | 13 | 2 | 13 | 2 | |
| 23-27 | Social dimension | 0.777 | 1.762 | 1.9% (0.5–6.7) | Not estimated | 0.29 (2.50) | 0.286 | −0.77 to 5.76 | 41 | 2 | 34 | 1 | |
| LYMQOL-ARM | 1(a-h)-3 | Function | 0.730 | 1.312 | Not estimated | Not estimated | 0.20 (2.52) | 0.197 | −1.31 to 4.92 | 39 | 5 | 25 | 4 |
| 4-8 | Appearance | 0.776 | 2.116 | 4.8% (2.1–10.8) | χ2=58.82, df=53, P=0.27 | 0.14 (2.58) | 0.141 | −1.70 to 4.87 | 27 | 6 | 27 | 4 | |
| 9-14 | Symptom | 0.848 | 1.609 | 5.8% (2.7–12.0) | χ2=55.05, df=80, P=0.99 | 0.05 (2.51) | 0.053 | −1.50 to 4.96 | 31 | 3 | 13 | 2 | |
| 15-20 | Mood | 0.828 | 2.154 | 5.8% (2.7–12.0) | χ2=62.68, df=80, P=0.92 | 0.10 (2.56) | 0.099 | −2.37 to 5.25 | 17.31 | 3 | 17 | 2 | |
| LLIS | 1-6 | Physical | 0.831 | 1.746 | 4.0% (1.6–9.8) | χ2=69.20, df=131, P>0.99 | 0.08 (1.98) | 0.082 | −1.05 to 4.92 | 28 | 2 | 13 | 0 |
| 7-12 | Psychosocial | 0.757 | 2.166 | 13.0% (7.8–21.0) | χ2=87.97, df=143, P>0.99 | 0.14 (1.90) | 0.141 | −1.40 to 4.19 | 20 | 2 | 20 | 2 | |
| 13-17 | Function | 0.738 | 1.826 | 4.0% (1.6–9.8) | χ2=88.22, df=95, P=0.68 | 0.32 (2.07) | 0.322 | −0.40 to 5.54 | 43 | 1 | 31 | 1 | |
| Lymph-ICF-UL | 1-7 | Symptoms | 0.802 | 2.011 | 29.0% (21.0–38.5) | χ2=215.29, df=1169, P>0.99 | 0.01 (1.43) | 0.013 | −0.81 to 3.39 | 21 | 0 | 11 | 0 |
| 8-11 | Mental function | 0.801 | 1.860 | 1.0% (0.2–5.4) | χ2=88.42, df=399, P>0.99 | 0.37 (2.47) | 0.368 | −0.61 to 5.57 | 43 | 3 | 43 | 3 | |
| 12-15 | Household activities | 0.847 | 1.910 | 15.1% (9.2–23.7) | χ2=124.42, df=399, P>0.99 | −0.03 (2.15) | −0.03 | −1.80 to 4.92 | 20 | 1 | 19 | 0 | |
| 16-23 | Mobility activities | 0.736 | 2.190 | 9.0% (4.8–16.2) | χ2=312.14, df=1599, P>0.99 | 0.18 (0.89) | 0.184 | −0.44 to 2.63 | 15 | 0 | 11 | 0 | |
| 24-29 | Life and social activities | 0.794 | 2.230 | 10.4% (5.8–18.1) | χ2=283.888, df=899, P>0.99 | −0.17 (1.54) | −0.17 | −1.22 to 3.24 | 20 | 2 | 18 | 2 |
CI, confidence interval; LLIS, Lymphedema Life Impact Scale Version 2; Lymph-ICF-UL, Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema; LYMQOL-ARM, Lymphedema Quality of Life Questionnaire for Upper Limbs; PCA, principal component analysis; PROM, patient-reported outcome measures; SD, standard deviation; ULL-27, Upper Limb Lymphedema 27; WLE, weighted likelihood estimation.
No item-level DIF remained significant after Holm adjustment, although global DIF was detected for LLIS Psychosocial by surgical side relative to arm dominance and LYMQOL-ARM Mood by education. For the remaining evaluable subscale-group combinations, the global LR tests did not provide statistically significant evidence of DIF (Table 4).
Table 4. Differential item functioning of LLIS, ULL-27 and LYMQOL-ARM.
| PROM subscale | Grouping variable | Groups, n | Items evaluated | Items not estimable | Andersen LR χ2 [df] | LR P | No. significant Wald tests (raw) | No. significant Wald tests (Holm-adjusted) |
|---|---|---|---|---|---|---|---|---|
| LLIS Physical | Age | Old, 54; young, 46 | 1, 5, 6 | 2, 3, 4 | 4.834 [10] | 0.90 | 0 | 0 |
| Education | ≤ high school, 50; > diplomas, 50 | 1, 6 | 2, 3, 4, 5 | 1.150 [6] | 0.98 | 0 | 0 | |
| Surgery side | Dominant, 51; non-dominant, 49 | 1, 3, 5 | 2, 4, 6 | 14.353 [10] | 0.16 | 2 | 0 | |
| LLIS Psychosocial | Age | Old, 54; young, 46 | 9, 10, 11, 12 | 7, 8 | 14.423 [15] | 0.49 | 0 | 0 |
| Education | ≤ high school, 50; > diplomas, 50 | 9, 12 | 7,8,10,11 | 9.604 [7] | 0.21 | 0 | 0 | |
| Surgery side | Dominant, 51; non-dominant, 49 | 7, 10, 11, 12 | 8,9 | 32.894 [15] | 0.005 | 0 | 0 | |
| LLIS Function | Age | Old, 54; young, 46 | 14, 15 | 13,16,17 | 3.686 [7] | 0.82 | 0 | 0 |
| Education | ≤ high school, 50; > diplomas, 50 | 14, 15 | 13,16,17 | 8.074 [7] | 0.33 | 0 | 0 | |
| Surgery side | Dominant, 51; non-dominant, 49 | None | 13-17 | / | / | |||
| ULL-27 Physical function | Age | Old, 54; young, 50 | 1, 2, 14 | 3–13, 15 | 6.356 [11] | 0.85 | 0 | 0 |
| Education | ≤ high school, 44; > diplomas, 60 | 1–5, 9, 11, 14 | 6–8, 10, 12, 13, 15 | 39.443 [31] | 0.14 | 2 | 0 | |
| Surgery side | Dominant, 55; non-dominant, 49 | 1–7, 9, 11, 14 | 8, 10, 12, 13, 15 | 38.858 [39] | 0.48 | 3 | 0 | |
| ULL-27 Psychological dimension | Age | Old, 54; young, 50 | 21, 22 | 16–20 | 2.775 [7] | 0.91 | 0 | 0 |
| Education | ≤ high school, 44; > diplomas, 60 | None | 16-22 | / | / | / | / | |
| Surgery side | Dominant, 55; non-dominant, 49 | 21, 22 | 16–20 | 5.744 [7] | 0.57 | 3 | 0 | |
| ULL-27 Social dimension | Age | Old, 54; young, 50 | 23, 24, 27 | 25, 26 | 4.415 [11] | 0.96 | / | / |
| Education | ≤ high school, 44; > diplomas, 60 | 23, 25–27 | 24 | 9.797 [15] | 0.83 | / | / | |
| Surgery side | Dominant, 55; non-dominant, 49 | 23, 24, 27 | 25, 26 | 9.471 [11] | 0.58 | / | / | |
| LYMQOL-ARM Function | Age | Old, 54; young, 50 | 1–3 | None | 7.585 [8] | 0.48 | / | / |
| Education | ≤ high school, 44; >diplomas, 60 | 1–3 | None | 7.288 [8] | 0.51 | 0 | 0 | |
| Surgery side | Dominant, 55; non-dominant, 49 | 1–3 | None | 10.758 [8] | 0.22 | 0 | 0 | |
| LYMQOL-ARM Appearance | Age | Old, 54; young, 50 | 4–8 | None | 13.832 [14] | 0.46 | / | / |
| Education | ≤ high school, 44; > diplomas, 60 | 4–8 | None | 15.538 [14] | 0.34 | 7 | 6 | |
| Surgery side | Dominant, 55; non-dominant, 49 | 4–8 | None | 16.819 [14] | 0.27 | 0 | 0 | |
| LYMQOL-ARM Symptom | Age | Old, 54; young, 50 | 9–14 | None | 16.823 [17] | 0.47 | 0 | 0 |
| Education | ≤ high school, 44; > diplomas, 60 | 9–14 | None | 11.114 [17] | 0.85 | 0 | 0 | |
| Surgery side | Dominant, 55; non-dominant, 49 | 9–14 | None | 11.147 [17] | 0.85 | 1 | 0 | |
| LYMQOL-ARM Mood | Age | Old, 54; young, 50 | 15–20 | None | 12.417 [17] | 0.77 | 0 | 0 |
| Education | ≤ high school, 44; > diplomas, 60 | 15–20 | None | 38.942 [17] | 0.002 | 0 | 0 | |
| Surgery side | Dominant, 55; non-dominant, 49 | 15–20 | None | 10.883 [17] | 0.86 | 0 | 0 |
“/” indicates that the corresponding item- or threshold-level Wald statistics could not be estimated because of limited subgroup sample sizes or sparse response patterns. LLIS, Lymphedema Life Impact Scale Version 2; LYMQOL-ARM, Lymphedema Quality of Life Questionnaire for Upper Limbs; LR, likelihood ratio; PROM, patient-reported outcome measure; ULL-27, Upper Limb Lymphedema 27.
CTT indices
Cronbach’s α exceeded 0.90 for all subscales and total scores, indicating high internal consistency (Table S5).
All convergent-validity hypotheses were confirmed for ULL-27 and LLIS, compared with 75% for LYMQOL-ARM and 60% for Lymph-ICF-UL. The proportions of divergent-validity hypotheses confirmed were 83.3%, 66.7%, 62.5%, and 40.0%, respectively (Table 5).
Table 5. Spearman’s rank correlation coefficient between the SF-36 and the Chinese versions of the four BCRL PROMs to determine convergent and divergent validity.
| Subscales | SF36-PCSa | SF36-MCSb |
|---|---|---|
| Lymphedema patients (n=61) | ||
| ULL-27 | ||
| Physical function†a | −0.497**‡ | −0.384**§ |
| Psychological dimension†b | −0.432** | −0.605**‡ |
| Social dimension†b | −0.344**§ | −0.449**‡ |
| LYMQOL-ARM | ||
| Functiona | −0.376** | −0.517** |
| Appearance†b | −0.371**§ | −0.474**‡ |
| Symptom†a | −0.532**‡ | −0.353**§ |
| Mood†b | −0.498** | −0.554**‡ |
| Lymphedema patients (n=59) | ||
| LLIS | ||
| Physical†a | −0.500**‡ | −0.288*§ |
| Psychosocialb | −0.600** | −0.435**‡ |
| Function†a | −0.633**‡ | −0.398**§ |
| Lymph-ICF-UL | ||
| Symptomsa | −0.449**‡ | −0.252 |
| Mental function†b | −0.377**§ | −0.498**‡ |
| Household activitiesa | −0.368** | −0.577** |
| Mobility activitiesa | −0.387** | −0.405** |
| Life and social activities†b | −0.391**§ | −0.572**‡ |
Hypothesis testing: For convergent validity, it was hypothesized that the correlation coefficient between the subscales of theoretically similar aspects of the BCRL PROMs and SF-36 would be at least moderate (correlation coefficient <–0.4). For divergent validity, it was assumed that the correlation coefficient between the subscales of different aspects of BCRL PROMs and SF-36 would be weak (correlation coefficient ≥–0.4) and weaker than that of similar aspects. Specifically, Subscales of the same superscript lowercase letter (a,b) represent similar concepts. Thus, the number of hypotheses used to assess the content validity of each scale was thrice the number of subscales. Convergent or divergent validity was considered excellent if ≥90% of the hypotheses were supported, good if between 75% and 90% of the hypotheses were confirmed, and moderate if between 40% and 74% of the hypotheses were met. The lower scores of BCRL PROMs and higher scores of SF-36 indicate poorer quality of life. Spearman’s rank correlation coefficient between the SF-36 and the Chinese versions of the four BCRL PROMs was calculated. †, accepted hypotheses for divergent validity that the correlation coefficient between the subscales of different concepts of BCRL PROMs and SF-36 was weaker than that of similar concepts. ‡, accepted hypotheses for expected moderate correlations assessing convergent validity. §, accepted hypotheses for expected weak correlations assessing divergent validity. *, P<0.05; **, P<0.01. BCRL, breast cancer-related lymphedema; Lymph-ICF-UL, Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema; LYMQOL-ARM, Lymphedema Quality of Life Questionnaire for Upper Limbs; MCS, Mental Component Summar; PCS, Physical Component Summary; PROMs, patient-reported outcome measures; SF-36, 36-item Short-Form Health Survey Questionnaire; ULL-27, Upper Limb Lymphedema 27.
All LLIS and Lymph-ICF-UL subscale and total scores differed significantly across BCRL stages (all P<0.05), whereas stage-related differences were less consistent for ULL-27 and LYMQOL-ARM (Table S6 and Figure 1). Correlations with BCRL stage exceeded 0.40 for LLIS Physical, LLIS total, and Lymph-ICF-UL Symptoms. Correlations among PROM scores were generally stronger than their correlations with clinically determined BCRL stage, indicating that PROMs may provide unique insights into disease conditions (Table S7 and Figure S1).
Figure 1.
Scores of the subscales and global scale of LLIS across different stages. LLIS, Lymphedema Life Impact Scale Version 2.
Discrimination relative to objective classification
The largest AUC estimates were observed for the Lymph-ICF-UL Symptoms subscale (0.734; 95% CI: 0.633–0.834), LLIS Physical subscale (0.725; 95% CI: 0.625–0.825), and LLIS total score (0.680; 95% CI: 0.573–0.787). The ULL-27 Physical Function subscale had an AUC of 0.654 (95% CI: 0.548–0.759). Other AUC estimates were generally poor to modest. Many estimates had relatively wide CIs, indicating uncertainty in the discriminatory performance of the corresponding domains. No pairwise difference in AUC remained statistically significant after Holm adjustment (Figure 2, Table S8).
Figure 2.
Sensitivity and specificity of the subscales and global scale of each BCRL PROM. AUC, area under the curve; BCRL, breast cancer-related lymphedema; LLIS, Lymphedema Life Impact Scale Version 2; Lymph-ICF-UL, Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema; LYMQOL-ARM, Lymphedema Quality of Life Questionnaire for Upper Limbs; PROM, patient-reported outcome measure; ULL-27, Upper Limb Lymphedema 27.
Sensitivity is crucial for clinicians to detect early-stage patients (31). The Lymph-ICF-UL and LYMQOL-ARM total scores had the highest sensitivity (79.61% and 75.00%, respectively), while the specificity was relatively low at only 43.90% and 55.81%, respectively. LLIS demonstrated a relatively high sensitivity and specificity (64.41% and 65.85%, respectively). ULL-27 performed poorly on the total score and subscales. The relatively higher AUCs for physical and symptom domains may partly reflect their conceptual proximity to the physical manifestations represented in the study reference classification (Tables S8,S9). These sample-derived cutoffs are exploratory.
Exploratory analysis in objective-negative group
Among participants who did not meet the circumference-based BCRL criterion, LLIS demonstrated significantly higher global scores in the ALND group compared to the SLNB group (P<0.05), with significantly higher scores also observed for the Physical and Psychological subscales. The Symptoms and Mental Function subscales of Lymph-ICF-UL also showed significant differences between groups (Figure 3). Furthermore, LLIS scores of ALND group in objectively diagnosed non-lymphedema patients approached or exceeded cutoffs based on objective assessment (Table 6). However, these cross-sectional findings should be interpreted cautiously, as they may reflect greater general postoperative morbidity associated with more extensive axillary surgery rather than true subclinical lymphedema.
Figure 3.
Lymphedema-specific PROM scores of SLNB and ALND groups in objectively diagnosed non-lymphedema patients. ALND, axillary lymph node dissection; LLIS, Lymphedema Life Impact Scale Version 2; Lymph-ICF-UL, Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema; LYMQOL-ARM, Lymphedema Quality of Life Questionnaire for Upper Limbs; PROM, patient-reported outcome measure; SLNB, sentinel lymph node biopsy; ULL-27, Upper Limb Lymphedema 27.
Table 6. LLIS scores of ALND group in objectively diagnosed non-lymphedema patients and cutoff.
| Subscale or total | Mean ± SD | Cutoff based on objective diagnosis |
|---|---|---|
| Physical | 19.97±16.86 | ≥20.83 |
| Psychosocial | 20.40±21.06 | ≥16.67 |
| Function | 18.45±22.02 | ≥10 |
| Total score | 18.58±17.80 | ≥18.06 |
ALND, axillary lymph node dissection; LLIS, Lymphedema Life Impact Scale Version 2; SD, standard deviation.
Discussion
This study culturally adapted four BCRL-specific PROMs and compared their classical and Rasch measurement properties in Chinese patients. All subscales achieved WLE reliability adequate for group-level comparisons, most items showed acceptable fit, and all estimated response-category thresholds were ordered. LLIS and Lymph-ICF-UL scores differed most consistently across clinical stages, while LLIS Physical and Lymph-ICF-UL Symptoms showed the largest exploratory AUC estimates. At the same time, the Rasch analyses identified limitations involving dimensionality, local independence, lower-end targeting, and measurement invariance in selected subscales. The results therefore provide generally supportive but nonuniform measurement evidence for the Chinese versions of these instruments.
Content validity comparison revealed LLIS’s advantages over the other PROMs. While ULL-27 (17) and LYMQOL-ARM (18) have longer histories of application, and Lymph-ICF-UL (20) was specifically designed for BCRL, LLIS integrates critical domains often neglected by others—notably self-management strategies and infection risk (22). During cultural validation, LLIS showed higher patient relevance. Several items in LYMQOL-ARM (e.g., “Ride a bike”, “Drive a car”) were deemed culturally irrelevant for many Chinese patients, and Lymph-ICF-UL and ULL-27 contained non-specific items (e.g., “difficulties walking”). Moreover, the ULL-27 Psychological subscale showed a comparatively higher proportion of flagged item pairs (33%) in response dependency, together with inconclusive evidence regarding its unidimensionality. These converging findings suggest possible content overlap or residual multidimensionality within this subscale and may limit the precision and interpretability of its scores.
The principal clinical implication of this study is that BCRL-specific PROMs can systematically capture consequences that are not fully represented by circumference measurements. Circumference primarily reflects differences in limb size, whereas PROMs provide structured information on heaviness, discomfort, functional limitations, emotional effects, and social participation. Incorporating this information into routine follow-up may provide a more comprehensive understanding of treatment-related burden. New, persistent, or worsening patient-reported concerns may inform further assessment, including focused clinical examination, additional objective assessment and evaluation of alternative treatment-related causes.
The Rasch and ROC findings should be interpreted together. LLIS Physical and Lymph-ICF-UL Symptoms combined comparatively favorable lower-end targeting with the largest AUC estimates, suggesting that these domains may be particularly useful for capturing mild physical and symptom burden in the present sample. Their relatively higher AUCs may also partly reflect conceptual proximity to the physical manifestations represented in the study classification. However, the modest AUC estimates, wide CIs, limited sample size, and absence of significant pairwise differences after Holm adjustment warrant cautious interpretation of the ROC findings.
The observation that LLIS and selected Lymph-ICF-UL subscales detected significant differences between ALND and SLNB patients who were objectively classified as non-lymphedema is intriguing but requires cautious interpretation. While one hypothesis is that these PROMs detect early symptom burden in patients with subclinical lymphedema that has not yet progressed to circumference changes >2 cm (7,32), alternative explanations must be considered. The elevated scores in ALND patients may reflect greater general postoperative morbidity (nerve injury, reduced shoulder range of motion, chest wall discomfort, psychosocial impact) associated with more extensive surgery, rather than specifically representing early lymphatic dysfunction. Without longitudinal follow-up data demonstrating that elevated PROM scores in objectively non-lymphedema patients predict subsequent development of clinically detectable BCRL, we cannot establish predictive validity for early disease detection. Prospective longitudinal studies are essential to determine whether baseline PROM scores in at-risk patients can identify those who will progress to overt lymphedema, which would establish their utility for risk stratification and targeted early intervention.
With further validation, BCRL-specific PROMs could become an important component of an integrated surveillance and management pathway. Serial PROM assessment could be combined with clinical examination and objective limb measurements to provide a multidimensional view of patient status. Electronic administration may facilitate integration into telehealth and longitudinal monitoring. Clinically meaningful or worsening scores could guide further objective tests and clinical assessment, indicating enhanced patient education or even necessary early intervention. PROM findings may also assist in monitoring changes after interventions and evaluating whether treatment addresses the outcomes that matter to patients (33). This proposed pathway represents a future application framework that requires prospective clinical validation.
The study has several additional limitations. First, direct cross-national DIF analysis could not be performed because harmonized individual-level item-response data from culturally diverse populations were unavailable. Moreover, limited subgroup sample sizes and sparse response patterns precluded the estimation of some item- or threshold-level Wald statistics. Second, no definitive reference standard is available for early or subclinical BCRL, and no separate a priori power calculation was conducted for the ROC analyses. Accordingly, the reported AUCs, sensitivities, specificities, and data-derived cutoffs reflect discrimination relative to the study reference classification rather than true diagnostic accuracy. Moreover, the cross-sectional design precluded assessment of predictive validity. Consequently, the precision and clinical interpretation of the ROC findings and the exploratory analyses in objective-negative group are limited. These findings require confirmation in larger, prospectively powered studies incorporating more sensitive objective assessments and longitudinal follow-up.
Conclusions
The Chinese ULL-27, LYMQOL-ARM, LLIS, and Lymph-ICF-UL demonstrated generally acceptable measurement properties for group-level assessment, although dimensionality, local independence, targeting, and invariance were not uniformly supported. LLIS and selected Lymph-ICF-UL domains showed relatively favorable stage-related and exploratory discriminatory findings. By systematically capturing symptoms, functional limitations, and psychosocial consequences not fully represented by circumference measurements, these PROMs may contribute clinically meaningful information to BCRL surveillance and management.
Supplementary
The article’s supplementary files as
Acknowledgments
We acknowledge the staff members of lymphedema outpatient clinic at Fudan University Shanghai Cancer Center for the assistance in questionnaire collection. Part of this work was presented as a poster at the 2023 San Antonio Breast Cancer Symposium, and the corresponding abstract was subsequently published in the 2024 meeting supplement of Cancer Research (Volume 84, Issue 9 Supplement).
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Fudan University Shanghai Cancer Center Institutional Review Board (SCCIRB) (No. 2302269-16). Informed consent was obtained from all individual participants included in the study.
Footnotes
Funding: This work was supported by the Program of Shanghai Academic Research Leader (No. LJRC2202), the Shanghai Committee of Science and Technology (No. 22DX1900500), and Science and Technology Talent and Platform Program of the Department of Science and Technology of Yunnan Province (No. 202305AF150153).
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0286/coif). The authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0286/dss
References
- 1. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49.. 10.3322/caac.21660 [DOI] [PubMed] [Google Scholar]
- 2. Tao X, Li T, Gandomkar Z, et al. Incidence, mortality, survival, and disease burden of breast cancer in China compared to other developed countries. Asia Pac J Clin Oncol 2023;19:645-54.. 10.1111/ajco.13958 [DOI] [PubMed] [Google Scholar]
- 3. DiSipio T, Rye S, Newman B, et al. Incidence of unilateral arm lymphoedema after breast cancer: a systematic review and meta-analysis. Lancet Oncol 2013;14:500-15.. 10.1016/S1470-2045(13)70076-7 [DOI] [PubMed] [Google Scholar]
- 4. Reimer T, Stachs A, Veselinovic K, et al. Patient-reported outcomes for the Intergroup Sentinel Mamma study (INSEMA): A randomised trial with persistent impact of axillary surgery on arm and breast symptoms in patients with early breast cancer. EClinicalMedicine 2023;55:101756.. 10.1016/j.eclinm.2022.101756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Jørgensen MG, Toyserkani NM, Hansen FG, et al. The impact of lymphedema on health-related quality of life up to 10 years after breast cancer treatment. NPJ Breast Cancer 2021;7:70.. 10.1038/s41523-021-00276-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Coriddi M, Dayan J, Sobti N, et al. Systematic Review of Patient-Reported Outcomes following Surgical Treatment of Lymphedema. Cancers (Basel) 2020;12:565.. 10.3390/cancers12030565 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Sackey H, Johansson H, Sandelin K, et al. Self-perceived, but not objective lymphoedema is associated with decreased long-term health-related quality of life after breast cancer surgery. Eur J Surg Oncol 2015;41:577-84.. 10.1016/j.ejso.2014.12.006 [DOI] [PubMed] [Google Scholar]
- 8. Whitworth P, Vicini F, Valente SA, et al. Reducing rates of chronic breast cancer-related lymphedema with screening and early intervention: an update of recent data. J Cancer Surviv 2024;18:344-51.. 10.1007/s11764-022-01242-8 [DOI] [PubMed] [Google Scholar]
- 9. Rafn BS, Christensen J, Larsen A, et al. Prospective Surveillance for Breast Cancer-Related Arm Lymphedema: A Systematic Review and Meta-Analysis. J Clin Oncol 2022;40:1009-26.. 10.1200/JCO.21.01681 [DOI] [PubMed] [Google Scholar]
- 10. Torres Lacomba M, Yuste Sánchez MJ, Zapico Goñi A, et al. Effectiveness of early physiotherapy to prevent lymphoedema after surgery for breast cancer: randomised, single blinded, clinical trial. BMJ 2010;340:b5396.. 10.1136/bmj.b5396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Bucci LK, Brunelle CL, Bernstein MC, et al. Subclinical Lymphedema After Treatment for Breast Cancer: Risk of Progression and Considerations for Early Intervention. Ann Surg Oncol 2021;28:8624-33.. 10.1245/s10434-021-10173-0 [DOI] [PubMed] [Google Scholar]
- 12. Lucci A, McCall LM, Beitsch PD, et al. Surgical complications associated with sentinel lymph node dissection (SLND) plus axillary lymph node dissection compared with SLND alone in the American College of Surgeons Oncology Group Trial Z0011. J Clin Oncol 2007;25:3657-63.. 10.1200/JCO.2006.07.4062 [DOI] [PubMed] [Google Scholar]
- 13. Basch E, Iasonos A, McDonough T, et al. Patient versus clinician symptom reporting using the National Cancer Institute Common Terminology Criteria for Adverse Events: results of a questionnaire-based study. Lancet Oncol 2006;7:903-9.. 10.1016/S1470-2045(06)70910-X [DOI] [PubMed] [Google Scholar]
- 14. Meilani E, Zanudin A, Nordin NAM. Psychometric properties of quality of life questionnaires for patients with breast cancer-related lymphedema: A protocol for a systematic review. Medicine (Baltimore) 2020;99:e23897.. 10.1097/MD.0000000000023897 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Pusic AL, Cemal Y, Albornoz C, et al. Quality of life among breast cancer patients with lymphedema: a systematic review of patient-reported outcome instruments and outcomes. J Cancer Surviv 2013;7:83-92.. 10.1007/s11764-012-0247-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Paramanandam VS, Lee MJ, Kilbreath SL, et al. Self-reported questionnaires for lymphoedema: a systematic review of measurement properties using COSMIN framework. Acta Oncol 2021;60:379-91.. 10.1080/0284186X.2020.1862422 [DOI] [PubMed] [Google Scholar]
- 17. Launois R, Megnigbeto AC, Le Lay K, et al. A specific quality of life scale in upper limb lymphedema: The ULL-27 questionnaire. Value in Health 2021;4:407-8. [Google Scholar]
- 18. Keeley V, Veigas D, Crooks S, et al. The development of a condition-specific quality of life measure for lymphoedema (lymqol). Eur J Lymphol 2004;12:36. [Google Scholar]
- 19. Keeley V, Crooks S, Locke J, et al. A quality of life measure for limb lymphoedema (lymqol). J Lymphoedema 2010;5:26-37. [Google Scholar]
- 20. Devoogdt N, Van Kampen M, Geraerts I, et al. Lymphoedema Functioning, Disability and Health questionnaire (Lymph-ICF): reliability and validity. Phys Ther 2011;91:944-57.. 10.2522/ptj.20100087 [DOI] [PubMed] [Google Scholar]
- 21. De Vrieze T, Vos L, Gebruers N, et al. Revision of the Lymphedema Functioning, Disability and Health Questionnaire for Upper Limb Lymphedema (Lymph-ICF-UL): Reliability and Validity. Lymphat Res Biol 2019;17:347-55.. 10.1089/lrb.2018.0025 [DOI] [PubMed] [Google Scholar]
- 22. Weiss J, Daniel T. VALIDATION OF THE LYMPHEDEMA LIFE IMPACT SCALE (LLIS): A CONDITION-SPECIFIC MEASUREMENT TOOL FOR PERSONS WITH LYMPHEDEMA. Lymphology 2015;48:128-38. [PubMed] [Google Scholar]
- 23. Li L, Wang HM, Shen Y. Chinese SF-36 Health Survey: translation, cultural adaptation, validation, and normalisation. J Epidemiol Community Health 2003;57:259-63.. 10.1136/jech.57.4.259 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Beaton DE, Bombardier C, Guillemin F, et al. Guidelines for the process of cross-cultural adaptation of self-report measures. Spine (Phila Pa 1976) 2000;25:3186-91.. 10.1097/00007632-200012150-00014 [DOI] [PubMed] [Google Scholar]
- 25. Wild D, Grove A, Martin M, et al. Principles of Good Practice for the Translation and Cultural Adaptation Process for Patient-Reported Outcomes (PRO) Measures: report of the ISPOR Task Force for Translation and Cultural Adaptation. Value Health 2005;8:94-104.. 10.1111/j.1524-4733.2005.04054.x [DOI] [PubMed] [Google Scholar]
- 26. Sousa VD, Rojjanasrirat W. Translation, adaptation and validation of instruments or scales for use in cross-cultural health care research: a clear and user-friendly guideline. J Eval Clin Pract 2011;17:268-74.. 10.1111/j.1365-2753.2010.01434.x [DOI] [PubMed] [Google Scholar]
- 27. Masters GN. A rasch model for partial credit scoring. Psychometrika 1982;47:149-74. [Google Scholar]
- 28.Mokkink LB, Prinsen C, Patrick DL, et al. COSMIN study design checklist for patient-reported outcome measurement instruments. COSMIN; 2019.
- 29. Mokkink LB, Elsman EBM, Terwee CB. COSMIN guideline for systematic reviews of patient-reported outcome measures version 2.0. Qual Life Res 2024;33:2929-39.. 10.1007/s11136-024-03761-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Executive Committee of the International Society of Lymphology . The diagnosis and treatment of peripheral lymphedema: 2020 Consensus Document of the International Society of Lymphology. Lymphology 2020;53:3-19. [PubMed] [Google Scholar]
- 31. Soran A, Ozmen T, McGuire KP, et al. The importance of detection of subclinical lymphedema for the prevention of breast cancer-related clinical lymphedema after axillary lymph node dissection; a prospective observational study. Lymphat Res Biol 2014;12:289-94.. 10.1089/lrb.2014.0035 [DOI] [PubMed] [Google Scholar]
- 32. Terada M, Yoshimura A, Sawaki M, et al. Patient-reported outcomes and objective assessments with arm measurement and bioimpedance analysis for lymphedema among breast cancer survivors. Breast Cancer Res Treat 2020;179:91-100.. 10.1007/s10549-019-05443-1 [DOI] [PubMed] [Google Scholar]
- 33. Terwee CB, Bot SD, de Boer MR, et al. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol 2007;60:34-42.. 10.1016/j.jclinepi.2006.03.012 [DOI] [PubMed] [Google Scholar]