Contralateral lower-limb growth after systemic chemotherapy and surgery for high-grade osteosarcoma in children: a retrospective longitudinal MRI study

Abstract

Background

Osteosarcoma occurs most commonly during adolescence, a period of active skeletal growth. Beyond concerns regarding survival, patients and their families often worry that surgery, perioperative chemotherapy, and the physiological stress of treatment may adversely affect growth. Therefore, the purpose of this study was to evaluate growth patterns in the contralateral limb following treatment for high-grade osteosarcoma.

Methods

Consecutive patients 14 years of age or younger treated for high-grade osteosarcoma of the distal femur or proximal tibia between 2012 and 2020, with a minimum follow-up of 5 years and contralateral limb MRI available, were retrospectively reviewed. Serial lower-extremity MRI was used to measure assess longitudinal and transverse skeletal growth. Growth trajectories were analyzed using linear mixed-effects models. Age- and sex-matched percentile curves were used to assess deviations from expected growth.

Results

Twenty patients (8 boys, 12 girls; mean age 11.7 ± 2.0 years) were analyzed. Continued contralateral femoral and tibial growth was observed in all patients until physeal closure, without premature growth arrest. Mean annual growth rates were 0.64 ± 0.47 cm/year and 0.41 ± 0.14 cm/year, respectively. Transverse skeletal growth remained proportionate. Nine of twenty patients (45%) demonstrated downward crossing of age- and sex-matched growth percentile bands during skeletal maturation.

Conclusions

Pediatric osteosarcoma patients treated with surgery and systemic chemotherapy demonstrated continued growth of the contralateral lower limb without premature physeal closure or disproportionate transverse remodeling. However, downward shifts in growth percentile trajectories were observed, and their clinical significance warrants further investigation.

Level of Evidence III, Retrospective cohort study.

Introduction

Osteosarcoma is among the most common primary malignant bone tumor in children, typically arising in the metaphyseal region of long bones, most frequently in the distal femur and proximal tibia, occurs most commonly during adolescence, a period of active skeletal growth [1–3]. Approximately 42% of tumors occur in the femur (with about 75% located distally) and 19% in the tibia (with about 80% located proximally) [4]. Advances in multimodal treatment, including systemic chemotherapy and limb-salvage surgery, have markedly improved survival rates and functional outcomes [5–7]. However, the surgical, therapeutic, and psychosocial stresses associated with treatment, including perioperative chemotherapy, may adversely affect growth on the normal limbs, when correlating growth charts derived from the Green-Anderson tables [8]. This is an important concern for patients and their families, in addition to survival. Moreover, the current understanding of transverse growth patterns remains limited. Improved understanding of treatment-associated growth alterations is valuable for preoperative education, anticipating the progression of LLD, and refining strategies for growth modulation. Therefore, we hypothesized that children treated with surgery and systemic chemotherapy for high-grade osteosarcoma would maintain continued longitudinal growth of the contralateral lower limb without premature physeal closure or major disturbance of overall skeletal growth trajectories compared with normative reference populations. In addition, we hypothesized that transverse skeletal growth would remain proportionate during skeletal maturation following treatment.

Methods

Study design and patient selection

This retrospective cohort study was conducted at a single tertiary center after approval by the Institutional Review Board. All procedures were performed in accordance with the Declaration of Helsinki, and the requirement for written informed consent was waived because of the retrospective nature of the study and the use of de-identified data.

Consecutive patients treated for high-grade osteosarcoma of the distal femur or proximal tibia between January 1, 2012, and October 31, 2020 were included for analysis (Fig. 1). All patients were treated according to our institutional osteosarcoma protocol, which included neoadjuvant chemotherapy, tumor resection with or without reconstruction, and adjuvant chemotherapy.

Fig. 1.

Patient selection flowchart

Patients were excluded if they were (1) 15 years of age or older at the time of diagnosis; (2) had a follow-up duration of less than 5 years; (3) had undergone prior surgery or surgery during the follow-up period on the contralateral limb for trauma, tumor, or growth-modulating procedures; (4) had congenital deformities affecting the contralateral limb; or (5) had osseous metastatic lesions outside the primarily affected limb.

Data collection and measurement

All demographic and clinical data, including age, sex, tumor location, surgical procedure, chemotherapy regimen, and tumor recurrence, were comprehensively recorded. MRI data were obtained from the institutional picture archiving and communication system (PACS).

For each enrolled patient, lower-limb MRI was performed using 1.5-T or 3.0-T scanners with standardized coronal and axial sequences encompassing the entire lower limbs. MRI was selected as the imaging modality to assess longitudinal and transverse growth of the lower-limb long bones because it avoids the magnification deviations commonly encountered with plain radiography.

For the femur, the whole femoral longitudinal length was defined as the distance from the apex of the femoral head to the most distal point of the lateral femoral condyle on the coronal plane [9, 10]. The distal femoral bicondylar width was defined as the linear distance between the most lateral and most medial margins of the femoral condyles [11]. The femoral diaphyseal isthmus width was defined as the minimum transverse diameter of the femoral diaphyseal cortex on the axial plane [12].

For the tibia, the whole tibial longitudinal length was defined as the distance from the center of the tibial plateau to the center of the talar dome [13]. The proximal tibial plateau width was defined as the maximum mediolateral distance across the articular surface of the tibial plateau on the axial plane [13]. The tibial diaphyseal isthmus width was defined as the minimum transverse diameter of the cortical bone at the diaphyseal level on the axial plane [14]. To assess relative growth trend change, femoral and tibial lengths were compared with age- and sex-matched normative growth percentiles reported by Chen et al. [10]. Patients were categorized into percentile bands (< 10th, 10–25th, 25–50th, 50–75th, 75–90th, > 90th). A downward shift into another percentile band at skeletal maturity was defined as off-track growth, reflecting divergence from normative reference trajectories. This cutoff has not been formally validated as a clinical indicator of growth impairment and was applied solely as an exploratory descriptive metric to characterize relative deviations from age- and sex-matched normative trajectories within this cohort.

Bone measurements were independently performed by two orthopedic surgeons who were blinded to the study objectives and patient identities. Interobserver reliability was assessed using the intraclass correlation coefficient (ICC).

Statistical analysis

All statistical analyses were performed using R software (version 4.3.2; R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were expressed as means ± standard deviations (SD). Longitudinal growth trajectories were analyzed using linear mixed-effects models to account for repeated MRI measurements within individual patients. Patient-specific random intercepts were included to account for within-subject correlation, and a first-order autoregressive covariance structure [AR(1)] was specified to model the correlation between repeated measurements over time. Age was treated as a continuous variable, and no time-centering strategy was applied. Linear, quadratic, and cubic polynomial functions of age were evaluated according to the observed growth patterns and model fit. Model selection was guided by the Akaike information criterion (AIC) and residual diagnostic assessment.

Because of the relatively small sample size and limited repeated observations per patient, random slope models, spline-based models, and higher-order interaction analyses were not performed to avoid model overfitting. Mixed-effects modeling was selected because it accommodates irregular MRI follow-up intervals and missing observations under the missing-at-random assumption. Residual normality and homoscedasticity were assessed by visual inspection of residual plots.

The Wilcoxon signed-rank test was used for paired non-normally distributed continuous variables. Interobserver reliability was assessed using the intraclass correlation coefficient (ICC). A p-value < 0.05 was considered statistically significant.

Results

Patient characteristics

From January 2012 to October 2020, a total of 20 patients (8 boys and 12 girls) were enrolled for analysis (Fig. 1). The mean age at diagnosis was 11.7 ± 2.0 years (range, 8–14 years). The mean follow-up duration was 8.8 ± 2.6 years (range, 5–14 years). All patients had follow-up until skeletal maturity. The primary tumor site was distal femur in 16 patients (80.0%) and proximal tibia in four patients (20.0%). Nineteen of the twenty patients (95.0%) received MAP ± IFO (methotrexate, doxorubicin, and cisplatin, with or without ifosfamide) while one patient (5.0%) received MEPI regimen (methotrexate, etoposide, cisplatin, and ifosfamide) [15]. Surgical procedures included wide resection combined with biological reconstruction using structural allograft (n = 9, 45.0%), recycle autograft (liquid nitrogen freezing (n = 3, 15.0%), or extracorporeal irradiation and reimplantation (ECIR) (n = 8, 40.0%). During follow-up, one patient experienced local recurrence, and five patients developed distant pulmonary metastases. All these patients underwent surgical resection of the recurrent or metastatic lesions, followed by adjuvant chemotherapy. Baseline demographic and treatment characteristics are summarized in Table 1. No patient had premature closure of physis and growth arrest.

Table 1.

Patient demographics

Tumor location	Distal femur (N = 16)	Proximal tibia (N = 4)
Age (years)	11.6 ± 1.7	11.8 ± 2.1
Female %	9 (56.3%)	3 (75.0%)
Initial height (cm)	149.8 ± 12.4	152.4 ± 14.9
Initial BMI (kg/m2)	18.9 ± 3.3	19.1 ± 3.3
Conventional	Osteoblastic: 9 (56.3%)	Osteoblastic: 2 (50.0%)
Osteosarcoma subtype	Chondroblastic: 5 (31.3%)	Chondroblastic: 1 (25.0%)
	Fibroblastic: 1 (6.3%)	Fibroblastic: 1 (25.0%)
	Telangiectatic: 1 (6.3%)	Telangiectatic: 0 (0%)
Treatment Option	Structural allograft: 8 (50.0%)	Structural allograft: 1 (25.0%)
combined with wide resection	ECIR: 5 (31.3%)	ECIR: 3 (75.0%)
	Arthroplasty: 5 (31.3%)	Arthroplasty: 0 (0%)
	Liquid Nitrogen: 3 (18.8%)	Liquid Nitrogen: 0 (0%)
Metastasis	5 (31.3%)	0 (0%)
Local Recurrence	1 (6.3%)	0 (0%)

ECIR: ExtraCorporeal Irradiation and Reimplantation

Mean age was 11.4 ± 2.7 years in males and 11.9 ± 1.6 years in females

Contralateral limb length growth and prediction modeling

Figures 2 and 3 illustrates the predicted longitudinal growth trajectories of the femur and tibia. Depending on the observed growth pattern, the most appropriate time function (linear, quadratic, or cubic) was selected for each mixed-effects model to ensure optimal fit and predictive accuracy. This modeling approach allowed the fitted curves to reflect individual variations in skeletal development. MRI evaluations demonstrated continuous longitudinal growth of both bones until physeal closure, with no evidence of premature growth arrest in any patient before 15 years of age. The mean annual longitudinal growth rates were 0.64 ± 0.47 cm/year for the femur and 0.41 ± 0.14 cm/year for the tibia.

Fig. 2.

Prediction models for femoral and Tibial Growth. A Estimated Whole Femoral Longitudinal Length =  − 9.417 + 8.455 × Age − 0.445 × (Age)² + 0.008 × (Age)³ B Estimated Distal Femur Bicondylar Width = 4.894 + 0.376 × Age − 0.010 × (Age)² C Estimated Femoral Diaphyseal Isthmus Width = 0.75 + 0.14 × Age − 0.003 × (Age)² D Estimated Whole Tibial Longitudinal Length =  − 58.98 + 18.403 × Age − 1.204 × (Age)² + 0.026 × (Age)³ E Estimated Proximal Tibial Plateau Width =  − 4.327 + 2.296 × Age − 0.151 × (Age)² + 0.003 × (Age)³ F Estimated Tibial Diaphyseal Isthmus Width = 1.618 + 0.036 × Age

Fig. 3.

presents a 9-year-old girl diagnosed with chondroblastic osteosarcoma (A), who underwent wide resection and knee arthrodesis. At age 11 follow-up, the limb-length discrepancy measured approximately 4 cm (B), which progressed to 6 cm by age 13 (C). By 18 years of age, the discrepancy remained stable at around 6 cm (D). This case clearly illustrates that most femoral and tibial longitudinal growth occurs before approximately 14 years of age. A Preoperative X ray. B Two-year postoperative lower-limb triple-film radiographs obtained when the patient was 11 years old. C Four-year postoperative lower-limb triple-film radiographs obtained when the patient was 13 years old. D Nine-year postoperative lower-limb triple-film radiographs obtained when the patient was 18 years old

Transversely, both the distal femoral condyle and tibial plateau showed gradual widening throughout skeletal maturation. The distal femoral bicondylar width increased at an average rate of 0.8 ± 0.5 mm/year, whereas the femoral diaphyseal isthmus width expanded at 0.4 ± 0.3 mm/year. Similarly, the tibial plateau widened at 1.1 ± 0.4 mm/year, and the tibial diaphyseal isthmus width increased at 0.4 ± 0.3 mm/year.

The isthmus width-to-length ratio of both the femur and tibia remained stable throughout follow-up, with no significant changes observed (P = 0.48 and P = 0.28, respectively), indicating proportionate longitudinal and transverse skeletal maturation.

Assessment of growth percentile concordance showed that 9 of 20 patients (45%) demonstrated a downward shift into a lower percentile band at skeletal maturity, meeting the definition of off-track growth. Ten patients (50%) remained within their original percentile band throughout follow-up, and one patient (5%) initially classified in the 10–25th crossed just above the 25-percentile threshold after follow-up.

Reliability analysis

Interobserver reliability for all MRI-based measurements was excellent, with ICC values of 0.93 for femoral and 0.92 for tibial parameters.

Discussion

In this study, we found that longitudinal growth of the contralateral femur and tibia continued until skeletal maturity in children treated for high-grade osteosarcoma. Nine of twenty (45%) patients did not maintain their expected age-matched growth percentile trajectories. Using MRI-based longitudinal assessment and mixed-effects growth modeling, we observed continuous physeal growth without evidence of premature growth arrest. Femoral and tibial longitudinal growth in healthy children begins to plateau after approximately 14 years of age, which is consistent with our findings and closely mirrors the growth trajectories observed in this cohort [10].

According to the Green-Anderson growth tables, although longitudinal growth may vary depending on factors such as race and genetic background, the distal femur contributes approximately 0.9 cm/year, and the proximal tibia approximately 0.6 cm/year to lower-limb growth during childhood [8]. Patients in the present cohort reported averaging 0.64 cm/year for the whole femur and 0.41 cm/year for the whole tibia. 45% (9 of 20 patients) demonstrated downward crossing of growth percentile bands during skeletal maturation compared with age- and sex-matched reference data [8, 16]. Previous experimental and clinical studies have suggested that chemotherapy and other treatment-related factors may influence skeletal growth through effects on growth plate chondrocyte activity, endochondral ossification, endocrine regulation, nutritional status, and overall treatment burden [17–21]. In addition to local effects on the growth plate, systemic chemotherapy may also influence overall endocrine function, thereby indirectly affecting skeletal growth. Chemotherapeutic agents commonly used in osteosarcoma treatment, such as methotrexate, cisplatin, and ifosfamide, have been reported to cause suppression of the hypothalamic–pituitary–gonadal axis and alterations in growth hormone and thyroid hormone levels in pediatric cancer patients [22–24]. Su et al. pointed out that methotrexate, a drug commonly used in osteosarcoma treatment, may lead to increased osteoclast formation and bone resorption, as well as decreased bone formation, thereby impairing the function of the growth plate [25]. Hoshi et al. observed that in young osteosarcoma patients treated with systemic chemotherapy, skeletal growth in terms of final height was not significantly impaired in the cohort as a whole, although patients diagnosed at ≤ 10 years of age tended to show a higher risk of short stature [26]. Including current study, given the absence of a matched control cohort, these findings should be interpreted cautiously.

The natural growth patterns of transverse dimensions, including bicondylar width and isthmus thickness, observed in this study population are valuable because the existing literature provides limited information on transverse skeletal development, particularly with respect to femoral and tibial width. These data may provide preliminary reference information for future studies evaluating treatment planning and skeletal development in pediatric osteosarcoma patients, thereby optimizing treatment strategies for this patient population.

Importantly, continued longitudinal growth and preserved transverse-to-longitudinal growth relationships were observed in the contralateral lower limb throughout skeletal maturation in this cohort, without evidence of premature physeal closure. Although downward crossing of growth percentile bands was observed in some patients, the clinical significance of these longitudinal growth changes remains uncertain. Given the established survival benefits of multimodal therapy in high-grade osteosarcoma, and the minimal effect on growth from perioperative treatment, patients and their families should still follow the standard treatment protocol without undue concern about potential growth disturbance. Optimizing nutritional and psychological support to help patients through the stressful treatment process is an area worthy of further investigation.

Limitations

This study has several limitations. First, it was a retrospective, single-center study with a relatively small sample size from a single ethnic population, which may limit the generalizability of the findings. Second, no matched healthy control cohort was available; reference percentile curves served only as a descriptive framework, and downward crossing of percentile bands should not be interpreted as conclusive evidence of pathological growth impairment. Third, the observed longitudinal growth changes are likely multifactorial and cannot be attributed solely to chemotherapy, as surgery, systemic illness, nutritional status, recurrence, physical activity, and overall treatment burden may also influence skeletal growth. Fourth, nutritional status, endocrine profiles, bone age, and growth-related biomarkers were not systematically available for analysis. Finally, because of the limited sample size, we did not perform sex-stratified analyses or higher-order interaction analyses. Future prospective, multicenter studies with larger cohorts and systematic longitudinal assessments are needed to determine the clinical significance and reproducibility of these findings.

Conclusion

Children treated for high-grade osteosarcoma of the distal femur or proximal tibia demonstrated continued longitudinal and proportional transverse growth of the contralateral lower limb until skeletal maturity, without premature physeal closure or growth arrest. 45% of patients showed downward crossing of age- and sex-matched growth percentile bands, suggesting that subtle treatment-associated alterations in growth trajectory may occur. These findings support continued growth surveillance during follow-up and warrant validation in larger prospective studies.

Abbreviations

AIC

Akaike information criterion

AR(1)

First-order autoregressive covariance structure

ECIR

Extracorporeal irradiation and reimplantation

ICC

Intraclass correlation coefficient

IFO

Ifosfamide

LLD

Limb-length discrepancy

MAP

Methotrexate, doxorubicin, and cisplatin

MEPI

Methotrexate, etoposide, cisplatin, and ifosfamide

MRI

Magnetic resonance imaging

PACS

Picture archiving and communication system

SD

Standard deviation

Author contributions

YHT and KCK collected the data and drafted the manuscript. KLC conceived the study design, contributed to data interpretation and substaintially revised the manuscript. PHW, CMC, and YKL prepared Table 1 and Fig. 1. SWT and PKW performed data analysis and prepared Fig. 2 and Fig. 3. CFC and WMC supervised the project, administered the study, and performed validation. All authors reviewed and approved the final version of the manuscript and agree to be accountable for all aspects of the work.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to patient privacy restrictions but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Institutional Review Board of Taipei Veterans General Hospital (IRB number: 2025–02-020AC) and was conducted in accordance with the Declaration of Helsinki. We confirmed that all methods were carried out in accordance with relevant guidelines and regulations. The requirement for informed consent was waived by the Institutional Review Board due to the retrospective nature of the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.