Abstract
Background
Pulmonary metastasectomy remains the standard of care for resectable metastatic osteosarcoma. However, controversy persists regarding minimally invasive surgery versus thoracotomy for occult nodule detection, the optimal use and duration of preoperative systemic therapy and prognostic factors associated with long-term survival. This study aimed to evaluate surgical approaches, assess the use and duration of preoperative systemic therapy, and identify prognostic factors associated with long-term survival in this population.
Methods
At a single center, we retrospectively reviewed 159 patients who underwent 229 pulmonary metastasectomies from 2000 to 2022. Eligible patients had pathologically confirmed osteosarcoma, resectable pulmonary metastatic disease, surgery intended to remove all known lesions and postoperative follow-up. The primary endpoint was post-metastasis survival (PMS), defined from metastatic diagnosis to death or the last follow-up. Computed tomography (CT) was scheduled every 4–8 weeks during preoperative surveillance and at prespecified postoperative intervals. Surgical approaches included conventional thoracotomy, mini-thoracotomy, and video-assisted thoracoscopic surgery (VATS). Preoperative strategies comprised upfront surgery or preoperative systemic therapy with first- or second-line chemotherapy, tyrosine kinase inhibitors (TKIs), or their combination. Oncological and perioperative outcomes were analyzed with a median follow-up of 79.1 months.
Results
The median age was 17.0 years, 106 (66.7%) patients were male, and 185 of 229 procedures (80.8%) followed preoperative systemic therapy. The 3- and 5-year PMS were 52.2% and 36.0%, respectively. Despite advances in CT resolution, palpation-assisted mini-thoracotomy identified more occult lesions than VATS (43.4% vs. 6.8%, P=0.004), with comparable rates of surgical complications. However, no significant difference in survival was observed between mini-thoracotomy and VATS. Remarkably, response to preoperative systemic therapy was the strongest independent predictor of survival, with a 5-year survival rate of 41.1% in responders versus 0.0% in non-responders (hazard ratio =0.22, P<0.001). Responders also exhibited a trend toward higher immune diversity (P=0.055) and increased CD8+ T cell infiltration (P=0.098) in metastasectomy specimens. Using tumor response modeling, we identified a 2–3-month period as the optimal duration for preoperative surveillance.
Conclusions
In this single-center retrospective cohort, preoperative surveillance and response-adapted strategy, combined with palpation-assisted mini-thoracotomy, was associated with favorable long-term outcomes of metastasectomy in patients with resectable metastatic osteosarcoma. However, given the observational design and modest sample size, these findings are exploratory and require validation in larger multicenter prospective studies.
Keywords: Osteosarcoma, pulmonary metastasis, metastasectomy, preoperative surveillance (POS), systemic therapy
Highlight box.
Key findings
• Palpation-assisted mini-thoracotomy detected more pathologically confirmed occult nodules than video-assisted thoracoscopic surgery, but no survival difference between approaches was demonstrated.
• Response to preoperative systemic therapy was independently associated with post-metastasis survival and exploratory modeling suggested a 2–3-month observation window.
What is known and what is new?
• Complete resection is central to selected patients with pulmonary metastatic osteosarcoma, but the balance between minimally invasive recovery and manual occult-nodule detection remains unsettled.
• This cohort evaluates surgical approaches, response-adapted surveillance, pathological response, and exploratory immune correlates within one longitudinal dataset.
What is the implication, and what should change now?
• Surgical approaches should be individualized by disease burden, treatment response and the feasibility of complete resection. The proposed preoperative surveillance interval is exploratory and should be validated in more prospective studies.
Introduction
Osteosarcoma is a highly metastatic malignant tumor originating in the bone, with an incidence of approximately 4.1 per million in children and 7.9 per million in adolescents (1). Approximately 15% of patients present with pulmonary metastases at the time of diagnosis (2), and about 40% of those with initially localized disease subsequently develop pulmonary metastases. Pulmonary metastasis remains the predominant cause of treatment failure and death in osteosarcoma, with a dismal 5-year survival of 20% (2). Therefore, improving patient selection and optimizing the timing and extent of pulmonary metastasectomy remain critical unmet needs in metastatic osteosarcoma.
Approximately 60% of initial pulmonary metastases of osteosarcoma are considered resectable (3-6), and complete resection of all known disease can produce durable remission or, in some instances, cure (7,8). Yet, to date, two clinically important uncertainties remain. First, the optimal surgical techniques for metastasectomy are still debated (2,9,10). Conventional open posterolateral thoracotomy offers unrestricted bimanual palpation—crucial for detecting occult osteoid-rich nodules that elude high-resolution computed tomography (CT)—but at the expense of larger incisions, greater pain, and longer recovery (7,11,12). Conversely, video-assisted thoracoscopic surgery (VATS) minimizes the surgical trauma but is generally reserved for patients with a limited number of metastases, as it has been reportedly associated with intraoperative tumor dissemination (13), and precludes full manual exploration, leading to missed small but palpable nodules, an omission that has been potentially linked to late relapse in several series (11,12). The oncological tradeoff between minimally invasive recovery and occult nodule detection and its long-term survival consequence remains unresolved (2,9,13). Second, compared with upfront surgery, no consensus defines how long systemic therapy should be observed before metastasectomy or how radiographic response should guide surgical selection (2,9,14). Although empirical evidence suggests that chemotherapy (6,15-17) and targeted agents such as tyrosine kinase inhibitors (TKIs) might reduce the tumor burden (18,19), their preoperative benefits, risks, and optimal monitoring interval are uncertain. Moreover, prognostic factors associated with long-term survival after pulmonary metastasectomy have varied across previous osteosarcoma cohorts, limiting reliable risk stratification and patient selection for surgery.
Against this background, we hypothesized that surgical approaches would influence occult nodule detection and that the radiographic response during preoperative systemic therapy would stratify the prognosis. Given the persistently poor survival after pulmonary metastasis, resolving these uncertainties is important for optimizing surgical selection and timing. We, therefore, retrospectively reviewed patients with metastatic osteosarcoma who underwent pulmonary metastasectomy at our institution between December 2000 and December 2022, with a median follow-up of 79.1 months. We compared perioperative and oncological outcomes across surgical approaches, modeled the optimal duration of preoperative surveillance (POS), and explored pathological and immune correlates in available specimens. By integrating surgical approaches, response-adapted surveillance and tumor immune profiles within a single long-term cohort, this study aimed to inform patient selection, operative strategy and the timing of metastasectomy. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0589/rc).
Methods
Study design and patient collection
This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Ruijin Hospital in 2024 (Approval No. 2024-LLS-252). The requirement for informed consent was waived by the ethics committee because the study involved only retrospective analysis of deidentified clinical data. Since the first case of resectable metastatic osteosarcoma with a modern multidisciplinary discussion paradigm was established in 2000 (Figure S1) in our institution, we conducted a single-center retrospective cohort study of osteosarcoma patients with pulmonary metastasis undergoing metastasectomy at the Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, between December 1, 2000, and December 31, 2022. The inclusion criteria were as follows: (I) pathologically confirmed diagnosis of osteosarcoma; (II) diagnosed as metastatic osteosarcoma, which was considered resectable and eligible for surgery; (III) undergoing metastasectomy, including wedge resection, segmentectomy, or lobectomy, with the intent of removing all known pulmonary lesions; (IV) complete medical information with postoperative follow-up data.
The sample size was determined by the eligible cases available during this prespecified period. We collected patients’ clinicopathological characteristics, surgical information, imaging measurements, pathology findings and long-term survival outcomes from institutional electronic medical records and maintained clinical databases using standardized data collection procedures. Pulmonary comorbidity was defined as a documented chronic or active pulmonary condition present before metastasectomy. Prior local recurrence was recurrence at the primary tumor site before the procedure. Bilateral pulmonary disease was defined radiologically as suspected metastatic nodules in both lungs.
POS and postoperative follow-up
The preoperative therapeutic strategy involved either upfront surgery or POS with systemic therapy. The follow-up assessments generally included chest CT every 4–8 weeks during the POS period, every 2–3 months during the first two postoperative years, every 6 months for years 3–5, and every 6–12 months thereafter (14). Follow-up information was obtained through outpatient visits, institutional electronic medical records, and telephone contact when necessary. Survival status was updated until November 30, 2025. The primary survival endpoint was assessed using post-metastasis survival (PMS), defined as the interval from the date of first diagnosis of metastatic disease to death from any cause. Patients who remained alive were censored at the date of last confirmed contact. Tumor burden was indexed by the sum of the longest diameters (SLD) of target lesions. Progressive disease (PD) was defined according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (20). Patients showing PD to systemic therapy during POS were defined as POS non-responders, whereas those without PD were defined as POS responders. Response measurements were obtained from serial CT assessments documented in the study records. Postoperative pulmonary recurrence was defined radiologically as a newly developed or enlarging pulmonary lesion considered metastatic on serial CT assessment. Pathologically confirmed recurrence was defined as metastatic osteosarcoma verified by biopsy or surgical resection. Because pathological confirmation was not available for all recurrent lesions, recurrence analyses were based primarily on radiological assessment, with pathological confirmation recorded when available. For patients with relapsed lesions after metastasectomy that were still considered resectable, re-resection was usually recommended to achieve further tumor-free survival at the physician’s discretion. To explore the optimal length of the POS with systemic therapy, we performed a patient-specific analysis of serial tumor measurements. Individual SLD data were smoothed with a locally estimated scatterplot smoothing (LOESS) model. For the POS responders, the data corresponding to the global minimum of the fitted curve were recorded as the center of the POS window. For POS non-responders, the center was defined as the first time point at which the smoothed curve exceeded 120% of baseline.
Preoperative evaluation and surgical approach
The preoperative evaluation of pulmonary metastasis was mainly based on chest CT imaging (21). All patients undergoing metastasectomy had at least 1 radiologically diagnosed target nodule for resection. A radiologically suspected nodule was a lesion considered potentially metastatic by the radiology and thoracic surgery team. Indeterminate lesions selected for exploration were included in the preoperative nodule count. A pathologically confirmed nodule required metastatic osteosarcoma in a resected specimen. All metastasectomy surgeries were exclusively performed by 2 experienced thoracic surgeons (X.Y. and H.L.). Surgical approaches include conventional thoracotomy, VATS, and mini-thoracotomy.
Whereas conventional thoracotomy required a large incision (Figure S2A,S2B), mini-thoracotomy was performed through a 6–8 cm incision made at the fifth intercostal space (ICS) along the anterior axillary line (Figure S2C,S2D). To foster a magnified visual exploration and to aid the unilateral manual palpation of the total lung, a thoracoscope was introduced through an additional auxiliary port placed at the seventh ICS along the mid-axillary line, which was then used for chest tube placement (Figure S2E). During the procedure, the surgeon conducted unilateral manual pulmonary palpation through the incision, to explore for potential additional nodules (Figure S2F), aiming to resect all suspected nodules, including but not limited to the target nodules and indeterminate nodules according to the preoperative radiological evaluation. All resected lesions were sent for pathological examination to determine the number of postoperative pathologically confirmed metastases. Postoperative complications were evaluated based on the Clavien-Dindo classification (22). Persistent air leak was defined as an air leak lasting longer than 5 days after the surgery (23). Complications with Clavien-Dindo grade III and IV or causing postoperative hospital stays longer than 5 days were considered severe complications.
Pathological assessment and immune infiltration analysis
Tissue preparation, hematoxylin-eosin (HE) staining and pathological evaluation of tumor samples were performed in the department of pathology at the researchers’ hospital. The non-viable tumor cell fraction was evaluated using paraffin-embedded tissues in 4-µm slices and determined as the percentage of non-viable tumor area within total tumor bed area (24). In 37 patients in whom fresh tumor specimens were available, RNA- sequencing (~50 million reads) of metastasectomy samples was performed by Primbio Genes Co., Ltd. (Wuhan, China). The pre-processing and downstream data analysis were performed at the Center for High-Performance Computing at Shanghai Jiao Tong University. Briefly, RNA-sequencing reads were aligned with STAR (v.2.7.8) and gene expression matrix was quantified as transcripts per million (TPM) with RSEM (v.1.1.16). To estimate the abundance of immune and stromal cell populations, the MCP-counter algorithm was applied to the log2-transformed TPM expression matrix (25). MCP-counter computes cell population-specific abundance scores based on predefined gene signatures. The following cell types were evaluated: T cells, CD8+ T cells, cytotoxic lymphocytes, B lineage, NK cells, monocytic lineage, myeloid dendritic cells, neutrophils, endothelial cells, and fibroblasts. Intratumoral immune diversity was then quantified using the Shannon index (26). RNA-sequencing and immune-profile analyses were restricted to the 37 patients with available and technically adequate tumor specimens and sequencing data, and no missing immune-profile values were imputed.
Statistical analysis
Continuous variables were presented as medians [interquartile range (IQR)]. For two-group comparisons, we used Wilcoxon rank-sum test. For three-group comparisons, we used a non-parametric statistical comparison with the Kruskal-Wallis test, followed by Dunn’s test for post hoc pairwise comparison with the Holm method to control multiplicity. Categorical variables were summarized as counts and percentages, analyzed by χ2 test or Fisher’s exact test where appropriate. The Cochran-Armitage test for linear trend was applied with the DescTools package (27) in R to evaluate trends in the proportion of procedures with additional nodule detection across CT resolution categories.
Analyses of the full unmatched cohort were primary. For a sensitivity comparison of VATS and mini-thoracotomy, propensity scores were estimated using a logistic regression model with the number of preoperative nodules as the matching covariate. We then performed 1:1 nearest-neighbor propensity score matching (PSM) with a caliper of 0.2, employing the MatchIt package (28) in R. Covariate balance was assessed using standardized mean differences (SMDs), with an acceptable threshold of 0.1 for the variable. After matching, the variable exhibited an SMD below 0.1, indicating adequate balance. The matched analysis was considered a sensitivity analysis to assess whether the full cohort findings remained consistent after balancing the number of preoperative nodules, rather than a replacement for the primary unmatched-cohort analysis. To further address potential residual confounding, other clinically relevant covariates were included in the subsequent univariable and multivariable logistic regression model. Subgroup analyses were conducted to evaluate whether the association between surgical approach and outcomes differed by metastasectomy status or by the number of preoperative nodules. Within-subgroup comparisons were performed using the χ2 test or Fisher’s exact test, as appropriate. Interactions were assessed using logistic regression models with an interaction term implemented in R. P values for interaction were reported.
Kaplan-Meier method and log-rank test were used to estimate the survival curve. Univariable and multivariable Cox proportional hazards models were used to estimate the clinicopathological factors associated with patients’ survival. No automated stepwise procedure was used. Competing-risk methods were not applied because death from any cause was considered an event in the definition of PMS. The proportional hazards assumption was assessed for each covariate and for the overall Cox model using scaled Schoenfeld residuals.
All statistical analyses were conducted on a complete-case basis and performed using R (version 4.4.1) (29). Observations with missing data for any variable included in a given analysis were excluded from that analysis, and no multiple imputation was performed. The statistical results were considered statistically significant at a two-sided P value of <0.05. Variables with clinical relevance or P<0.05 in univariate analyses were entered into the multivariable models.
Results
Patient characteristics and surgical approach
A total of 229 metastasectomy surgeries were performed in 159 patients, with a median follow-up of 79.1 months (IQR, 59.6–109.7). The median age was 17.0 years (IQR, 14.0–20.5), with a male-to-female ratio of 106:53. Conventional thoracotomy accounted for 10 (4.4%) procedures, VATS for 44 (19.2%) and mini-thoracotomy for 175 (76.4%) (Table 1). Upfront surgeries were undertaken in 44 (19.2%) cases, whereas POS with systemic therapy strategy was adopted in the remaining 185 (80.8%) cases (Figure 1A). Notably, our surgical approach has evolved from upfront surgery with conventional thoracotomy between 2000 and 2009, to a mixed approach of VATS and mini-thoracotomy between 2010 and 2018, and finally to an almost exclusively mini-thoracotomy approach with POS strategy since 2019 (Figure 1A).
Table 1. Baseline clinical characteristics by surgical approach before propensity score matching.
| Characteristic | Conventional thoracotomy (N=10) | VATS (N=44) | Mini-thoracotomy (N=175) | P value |
|---|---|---|---|---|
| Age, years | 17.5 (17.0–20.0) | 19.0 (15.5–27.0) | 16.0 (14.0–20.0) | 0.073 |
| Sex | 0.321 | |||
| Female | 4 (40.0) | 11 (25.0) | 65 (37.1) | |
| Male | 6 (60.0) | 33 (75.0) | 110 (62.9) | |
| BMI, kg/m2 | 20.7 (19.3–21.8) | 22.1 (18.4–24.5) | 20.4 (17.1–23.6) | 0.283 |
| ASA physical status | 0.216 | |||
| I–II | 6 (60.0) | 36 (81.8) | 128 (82.1) | |
| III–IV | 4 (40.0) | 8 (18.2) | 28 (17.9) | |
| ECOG performance status | 0.038 | |||
| 0–1 | 9 (90.0) | 40 (90.9) | 131 (74.9) | |
| ≥2 | 1 (10.0) | 4 (9.1) | 44 (25.1) | |
| Initial surgical procedure of primary tumor | 0.030 | |||
| Amputation | 5 (50.0) | 7 (16.3) | 26 (15.1) | |
| Limb-salvage | 5 (50.0) | 36 (83.7) | 146 (84.9) | |
| Anatomical region of primary tumor | 0.509 | |||
| Axis or girdle | 1 (10.0) | 1 (2.3) | 5 (2.9) | |
| Distal extremity | 5 (50.0) | 24 (54.5) | 104 (61.2) | |
| Proximal extremity | 4 (40.0) | 19 (43.2) | 61 (35.9) | |
| Longest diameter of primary tumor (cm) | 9.5 (7.8–12.1) | 10.9 (7.2–13.7) | 12.0 (8.9–15.2) | 0.156 |
| Longest diameter of metastatic tumor (cm) | 3.5 (2.9–8.5) | 1.3 (0.8–1.9) | 1.1 (0.7–2.0) | <0.001 |
| Number of preoperative nodules | 1.0 (1.0–3.0) | 1.0 (1.0–2.0) | 2.0 (1.0–3.0) | <0.001 |
| Number of intraoperatively detected nodules | 1.5 (1.0–4.0) | 1.0 (1.0–2.0) | 3.0 (2.0–6.0) | <0.001 |
| Number of pathologically confirmed nodules | 1.0 (1.0–3.0) | 1.0 (1.0–2.0) | 2.0 (1.0–5.0) | <0.001 |
Data are presented as n (%) or median (interquartile range). ASA, the American Society of Anesthesiologists; BMI, body mass index; ECOG, Eastern Cooperative Oncology Group; VATS, video-assisted thoracoscopic surgery.
Figure 1.
Surgical approaches, survival and pulmonary nodules characteristics over the decades. (A) Distribution of surgical approaches and therapeutic strategies from 2000 to 2022. (B) Kaplan-Meier curve of post-metastasis survival for the entire cohort. (C) Kaplan-Meier curves of post-metastasis survival stratified by surgical approach. (D) Longest diameter of nodules, (E) preoperative nodule count, (F) intraoperatively detected nodule count and (G) pathologically confirmed nodule count across different study periods. (H) Proportion of intraoperatively detected additional nodules and pathologically confirmed additional nodules by imaging modality and resolutions. (I) Bar plot showing the proportion of intraoperatively detected additional or missed nodules and intraoperatively detected additional or missed nodules with pathological confirmation by three different surgical approaches. *Due to the COVID-19 pandemic, the number of metastasectomies decreased in 2022; ** The y-axis is displayed on a log2 scale because the data are highly dispersed. VATS, video-assisted thoracoscopic surgery.
The 3-year and 5-year survival rates of our study were 52.2% (95% CI: 45.0–60.6) and 36.0% (95% CI: 29.0–44.6) (Figure 1B), respectively, with a trend of gradual improvement of 5-year PMS from 9.1% between 2000 and 2009 to 42.5% between 2010 and 2018 (P<0.001). Notably, despite the adoption of mini-thoracotomy as the major approach over VATS, our 5-year PMS remained roughly unchanged since 2010 (P=0.594) (Figure S3A). Remarkably, there were no significant differences in PMS between the VATS and mini-thoracotomy approaches (P=0.185) (Figure 1C). Over the three periods, practice shifted from conservative resection of a small number of large metastases to more aggressive early resection of multiple small lesions (Figure 1D-1G). Surprisingly, despite advances in imaging resolution, the intraoperative detection rates of additional lesions did not decrease (P<0.05), with the rate of intraoperatively detected and pathologically confirmed additional nodules being 42.6% and 29.6% with 5–10 mm CT scan slice thickness, 60.7% and 35.7% with 1.2–3 mm slice thickness, and 60.5% and 44.4% with ≤1 mm slice thickness, respectively (Figure 1H). Such a high rate of additional lesion finding was likely due to broader inclusion of surgical candidates with multiple small lung metastases. Compared with VATS, mini-thoracotomy achieved a 63.4% rate of intraoperatively detected additional nodules and a 43.4% rate of pathologically confirmed additional nodules, which was substantially higher than that with VATS, in which 6.8% of intraoperatively detected and 6.8% of pathologically confirmed additional nodules were identified (Figure 1I). Given the observed imbalance in preoperative nodules (P<0.001) between the surgical groups (Table S1), PSM was further implemented to match the number of preoperative nodules, and re-confirmed our findings that mini-thoracotomy yielded significantly more intraoperatively detected [2.0 (IQR, 1.0–3.0), P=0.004] and pathologically confirmed [1.5 (IQR, 1.0–2.0), P=0.037] additional nodules than VATS [1.0 (IQR, 1.0–2.0) and 1.0 (IQR, 1.0–2.0)] (Table S2).
Perioperative parameters and postoperative complications
We then compared the perioperative parameters and postoperative complications among the three surgical approaches of metastasectomy (Table 2, Table S3). No perioperative mortality was observed in this study. Compared with conventional thoracotomy, both VATS and mini-thoracotomy demonstrated a superior surgical recovery with less trauma, with shorter operative time [VATS: 75.0 (IQR, 55.0–122.5) min and mini-thoracotomy: 85.0 (IQR, 60.0–100.0) min vs. conventional thoracotomy: 142.5 (IQR, 100.0–200.0) min, P=0.012], shorter postoperative length of hospital stay [VATS: 4.5 (IQR, 3.5–6.0) days and mini-thoracotomy:5.0 (IQR, 4.0–6.0) days vs. conventional thoracotomy: 8.5 (IQR, 6.0–10.0) days, P<0.001], earlier chest tube removal [VATS: 2.0 (IQR, 2.0–3.0) days and mini-thoracotomy: 4.0 (IQR, 3.0–4.0) days vs. conventional thoracotomy: 5.0 (IQR, 4.0–7.0) days, P<0.001], less drainage volume [VATS: 237.5 (IQR, 72.5–464.0) mL and mini-thoracotomy: 595.0 (IQR, 345.0–900.0) mL vs. conventional thoracotomy: 1,555.0 (IQR, 1,050.0–1,810.0) mL, P<0.001] and less postoperative pain (VATS: 6.8% and mini-thoracotomy: 14.3% vs. conventional thoracotomy: 50.0%, P=0.005). However, mini-thoracotomy still demonstrated a slight disadvantage in postoperative recovery parameters compared with VATS, showing higher intraoperative blood loss (P=0.001), longer hospital stays (P=0.046), longer chest tube duration (P<0.001), and greater drainage volume (P<0.001), while the overall postoperative complication rates were comparable between VATS and mini-thoracotomy (P=0.503). Noteworthily, the postoperative complications were observed in 70.0% of patients after conventional thoracotomy, 18.2% after VATS, and 22.9% after mini-thoracotomy. Compared with conventional thoracotomy, both VATS and mini-thoracotomy were associated with significantly fewer complications (P=0.004), and such difference in complication rate was mainly driven by persistent air leak (conventional thoracotomy: 40.0% vs. VATS: 2.3% and mini-thoracotomy: 6.8%, P=0.008). Other specific complications, such as pleural effusion, pneumonia, and hemorrhage, were not statistically different among the three surgical approaches.
Table 2. Perioperative parameters and postoperative complications by surgical approach.
| Characteristic | Conventional thoracotomy (N=10) | VATS (N=44) | Mini-thoracotomy (N=175) | P value |
|---|---|---|---|---|
| Blood loss (mL) | 125.0 (20.0–200.0) | 20.0 (10.0–50.0) | 50.0 (30.0–100.0) | <0.001 |
| Operative time (min) | 142.5 (100.0–200.0) | 75.0 (55.0–122.5) | 85.0 (60.0–100.0) | 0.012 |
| Intraoperative blood transfusion | <0.001 | |||
| No | 6 (60.0) | 44 (100.0) | 159 (97.5) | |
| Yes | 4 (40.0) | 0 (0.0) | 4 (2.5) | |
| Postoperative hospital stays (d) | 8.5 (6.0–10.0) | 4.5 (3.5–6.0) | 5.0 (4.0–6.0) | <0.001 |
| Chest tube duration (d) | 5.0 (4.0–7.0) | 2.0 (2.0–3.0) | 4.0 (3.0–4.0) | <0.001 |
| Chest tube drainage (mL) | 1,555.0 (1,050.0–1,810.0) | 237.5 (72.5–464.0) | 595.0 (345.0–900.0) | <0.001 |
| Hospitalization cost ($) | 3,444.5 (3,099.5–5,643.6) | 4,711.7 (3,671.9–6,725.0) | 4,822.4 (4,094.5–6,269.8) | 0.193 |
| Postoperative supplemental analgesia | 0.005 | |||
| No | 5 (50.0) | 41 (93.2) | 150 (85.7) | |
| Yes | 5 (50.0) | 3 (6.8) | 25 (14.3) | |
| Postoperative pulmonary complications | 0.004 | |||
| No | 3 (30.0) | 36 (81.8) | 135 (77.1) | |
| Yes | 7 (70.0) | 8 (18.2) | 40 (22.9) | |
| Pleural effusion | 0.217 | |||
| No | 6 (60.0) | 38 (86.4) | 149 (85.1) | |
| Clavien-Dindo I–II | 4 (40.0) | 6 (13.6) | 24 (13.7) | |
| Clavien-Dindo III–IV | 0 (0.0) | 0 (0.0) | 2 (1.1) | |
| Pneumonia | >0.999 | |||
| No | 10 (100.0) | 42 (95.5) | 165 (94.3) | |
| Clavien-Dindo I–II | 0 (0.0) | 2 (4.5) | 9 (5.1) | |
| Clavien-Dindo III–IV | 0 (0.0) | 0 (0.0) | 1 (0.6) | |
| Hemorrhage | >0.999 | |||
| No | 10 (100.0) | 44 (100.0) | 172 (98.3) | |
| Clavien-Dindo I–II | 0 (0.0) | 0 (0.0) | 2 (1.1) | |
| Clavien-Dindo III–IV | 0 (0.0) | 0 (0.0) | 1 (0.6) | |
| Persistent air leak | 0.008 | |||
| No | 6 (60.0) | 43 (97.7) | 163 (93.1) | |
| Clavien-Dindo I–II | 4 (40.0) | 1 (2.3) | 10 (5.7) | |
| Clavien-Dindo III–IV | 0 (0.0) | 0 (0.0) | 2 (1.1) | |
| Unplanned reoperation | >0.999 | |||
| No | 10 (100.0) | 44 (100.0) | 174 (99.4) | |
| Yes | 0 (0.0) | 0 (0.0) | 1 (0.6) | |
Data are presented as n (%) or median (interquartile range). VATS, video-assisted thoracoscopic surgery.
Prognosis and determinants of an increase in pathologically confirmed nodules
Due to the detection of occult nodules as the main advantage of mini-thoracotomy over VATS, we were then interested in whether resecting or missing such additional nodules might have an impact on the patients’ oncological outcome. Notably, patients with additional pathologically confirmed nodules (Figure 2A) failed to demonstrate a worse PMS than those without an increase [hazard ratio (HR) =1.27; 95% CI: 0.85–1.89; P=0.230]. Similarly, no difference in the pulmonary recurrence rates after metastasectomy was observed between mini-thoracotomy and VATS, regardless of whether the preoperative nodules were ≥3 or not (P=0.988). However, among patients with 3 or more metastases, we found a non-significant trend toward more pulmonary local recurrence in the VATS group (100.0%) compared with the mini-thoracotomy group (77.4%) (P=0.993) (Figure S3B).
Figure 2.
Prognostic impact and risk factors for the increase in pathologically confirmed nodules. (A) Kaplan-Meier curves of post-metastasis survival according to whether or not intraoperative detected and pathologically confirmed additional nodule was present. (B) Multivariable logistic regression model identifying factors independently associated with the detection of pathologically confirmed additional nodules. (C) Detection rate of pathologically confirmed additional nodules in first metastasectomy versus repeated metastasectomy. CI, confidence interval; OR, odds ratio; TKIs, tyrosine kinase inhibitors; VATS, video-assisted thoracoscopic surgery.
Using univariate logistic regression analysis, we found that metastasis-to-surgery interval ≥6 months (P=0.029), non-responder to POS systemic therapy (P=0.013), addition of TKIs into therapy (P<0.05), preoperative pulmonary comorbidities (P<0.001), prior locally recurrent diseases (P<0.001), preoperative nodules ≥3 (P<0.001), cavitation changes (P=0.017), bilateral pulmonary metastasis (P<0.001), and the use of MOTA (P<0.001) were associated with a higher likelihood of incidental nodule detection (Table S4). In multivariate analysis (Figure 2B), prior locally recurrent diseases [odds ratio (OR) =4.75, P=0.013], preoperative nodules ≥3 (OR =4.90, P<0.001), and the use of VATS (OR =0.03, P=0.004) were the independent predictors of additional nodule discovery. Consistent with our experience, the proportion with increased nodules was significantly higher with mini-thoracotomy than VATS in the first metastasectomy (45.6% vs. 5.3%, P<0.001), but not in the repeated metastasectomy (35.9% vs. 16.7%, P=0.369) (Figure 2C).
The effect of POS on surgical outcome
Next, we explored the factors associated with postoperative survival outcome using univariate Cox regression analysis. We identified shorter disease-free interval (DFI) (P=0.029), preoperative pulmonary comorbidities (P<0.001), prior extrapulmonary diseases (P<0.05), poor systemic therapy response (P<0.001), incomplete resection (P=0.042), an increase of ≥3 in pathologically confirmed nodules (P=0.029), and larger tumor size (P=0.003) were associated with worse PMS (Table S5). Further multivariate analysis suggested that response to systemic therapy (HR =0.22, P<0.001) and preoperative pulmonary comorbidity (HR =3.65, P<0.001) remained the only independent variables to predict the survival outcome (Figure 3A). Remarkably, POS responders to systemic therapy had substantially better survival than that of POS non-responders (HR =0.22; 95% CI: 0.12–0.43; P<0.001), with a 3-year survival rate of 59.8% versus 10.0% and a 5-year rate of 41.1% versus 0.0% (Figure 3B), highlighting the role of POS as a pivotal strategy to optimize the selection of the surgical candidate. The proportional hazards assumption was not violated for any individual covariate or for the overall multivariable Cox model based on scaled Schoenfeld residuals (all P>0.05; global P=0.939; Table S6). To inform the optimal duration of POS before metastasectomy, we modeled the patient-derived SLD changes with LOESS. Interestingly, for the POS responders of systemic therapy, the tumor size reached an inflection point of 2.98 months at its nadir with an average of 15.33% reduction, followed by a progressive rebound (Figure 3C). In contrast, tumors in the POS non-responder demonstrated a gradual increase in tumor size, where the SLD reached +20% from baseline at approximately 1.54 months (Figure 3D). Therefore, we defined a period of 2–3 months as the optimal length of POS, enabling sufficient time for the detection of tumor progression while allowing the maximum degree of tumor regression for metastasectomy. Furthermore, by comparing the tumor microenvironment between POS responders and POS non-responders, we noticed a trend toward higher intratumoral immune diversity (P=0.055) and greater CD8 T cell infiltration (P=0.098) in patients showing tumor response during POS (Figure 3E).
Figure 3.
Prognostic significance and clinical outcomes of preoperative systemic therapy. (A) Multivariable Cox proportional hazards model identifying factors independently associated with post-metastasis survival. (B) Kaplan-Meier curves of post-metastasis survival stratified by response to preoperative systemic therapy. Spider plot demonstrating the change in SLD during the preoperative surveillance between POS responders (C) and non-responders (D) with a LOESS fitting curve. (E) The immune infiltration assessment between POS responders and non-responders (n=37). (F) Pathologic non-viable tumor cell fraction of the surgical specimens across different preoperative therapeutic regimens. (G) Best change in SLD across therapeutic regimens assessed by preoperative radiological assessment. (H) Proportion of postoperative complications by therapeutic regimens. (I) Proportion of postoperative complications by the preoperative administration of TKIs. CI, confidence interval; HR, hazard ratio; LOESS, locally estimated scatterplot smoothing; POS, preoperative surveillance; SLD, sum of the longest diameters; TKIs, tyrosine kinase inhibitors.
Additionally, we compared the tumor pathological response according to different POS regimens. We observed that the highest tumor non-viable fraction was observed in patients with chemotherapy combined with TKIs during POS, whereas that of the TKI-only regimens was the lowest (P=0.022) (Figure 3F). However, no significant difference was found among different chemotherapy regimens (Figure S3C). In terms of preoperative tumor size downstaging, the TKI plus chemotherapy group also demonstrated the highest degree of tumor shrinkage compared to the chemotherapy-only group (P=0.001) (Figure 3G, Figure S4). As a representative case (Figure S5), a 20-year-old male with osteosarcoma developed a metastasis in the left lower lobe with malignant pleural effusion. After POS of 2 months with antiangiogenic TKI plus gemcitabine-docetaxel chemotherapy, the tumor was dramatically downstaged with surgical resection, followed by a PMS of at least 6 years after his first metastasectomy. Furthermore, there were no significant differences in the postoperative complications among various POS regimens (P=0.159) (Figure 3H). However, a borderline significant trend toward a higher complication rate was observed with the addition of TKI preoperatively compared with conventional chemotherapy (P=0.087) (Figure 3I).
Discussion
To our knowledge, our data represent one of the largest single-center series including 159 patients undergoing multimodal therapy including pulmonary resection and systemic therapy for pulmonary metastatic osteosarcoma, with a long-term follow-up of 79.1 months (IQR, 59.6–109.7). The 3-year and 5-year PMS rates in our cohort were 52.2% and 36.0%, respectively, which are generally comparable to previously reported series of resectable metastatic osteosarcoma. The early MD Anderson experience (n=137, 1980–2000) reported a 5-year survival of 29.0% after metastasectomy (30), while the St. Jude Children’s Research Hospital cohort of 23 patients with lung-only metastases demonstrated an estimated 5-year survival of 40.0% after surgical remission (31). In the subsequent MD Anderson cohort (2001–2011), patients with pulmonary metastasectomy showed 5-year survival of 35–40% (5). The International Registry of Lung Metastases (IRLM), encompassing over 5,000 patients, reported a 5-year survival of about 34% for patients with a median of 2–3 pulmonary metastases (1,226/4,518, 27.1%) (32). Our survival was broadly consistent with the previous reports. However, direct comparison across cohorts should be made cautiously because surgical eligibility, metastatic burden, systemic therapy, repeat metastasectomy practices and follow-up periods differed substantially. Our cohort reflects the clinical outcomes of a 22-year evolution in our surgical strategy—from the early stage of operating mainly on isolated or symptomatic metastases to the more proactive approach toward resection of all known multiple small lesions (with a maximum of 56 metastatic lesions resected). The absence of further survival improvement after 2010 may therefore reflect both a plateau in the efficacy of local surgical control and the increasing inclusion of patients with more extensive metastatic disease, rather than the effect of any single operative strategy.
One of the key findings of our study is the potential advantage of a POS strategy for pulmonary metastasectomy in osteosarcoma. Currently, the optimal timing of surgical resection remains considerably variable across the literature. Common strategies include upfront metastasectomy after diagnosis (33), hybrid strategies with immediate surgery for one side and delayed resection for the contralateral side (34), salvage resections after chemotherapy failure or intolerance (35), and POS of systemic response (36,37)—which aligns with our approach. Although there is no well-established consensus, our findings suggest the POS with systemic therapies may provide clinically useful information for surgical selection due to several reasons. First, it allows for a median preoperative tumor shrinkage of 11.73% in the therapeutic responders and improves resectability with potentially less surgical morbidity; Second, it facilitates the selection of the optimal surgical candidates who are most likely to benefit from metastasectomy since the non-responders had a high rate of early relapse within the first year post-surgery, with a very poor 5-year survival rate of 0.0%. Third, a trend toward higher immune diversity and CD8 infiltration in POS responders suggests the potential immune priming effect of preoperative systemic therapy before metastasectomy. Finally, our data suggest that preoperative tumor change and pathological tumor non-viable rate differ across therapeutic regimens. Such a strategy could provide in vivo chemosensitivity data while testing the patient’s tolerability to the therapy before surgery to further tailor the post-operative regimen. Notably, by modeling the longitudinal tumor size change, we identified 2–3 months as a potentially informative surveillance window. This duration balances the likelihood of achieving maximal tumor downstaging in the responders and avoiding continued tumor progression in non-responders to select the potential surgical candidates.
Current guidance recommends that patients with pulmonary metastases from osteosarcoma be evaluated by a specialist bone sarcoma multidisciplinary team and considered for pulmonary metastasectomy when complete, parenchyma-sparing removal of all lesions identified on high-resolution CT is feasible. Although thoracotomy has traditionally been preferred because it permits manual palpation of the lung, no clear advantage over thoracoscopic resection has been established (38). Our findings both support and refine this framework. As chest imaging resolution improved over the two decades, the proportion of intraoperatively discovered nodules by manual palpation did not decline in our cohort. The 44.4% additional metastasis detection rate with thin-section (≤1 mm) CT imaging is similar to a previously reported miss rate of 12–46% with conventional 5–8 mm slice thickness (39). Over the study period, increasingly sensitive imaging was accompanied by broader surgical indications and more frequent resection of patients with multiple, smaller lesions. Consequently, improved high-resolution radiological detection may have been offset by a shift toward patients with a greater number of subcentimeter metastases. Manual palpation may also detect firm lesions that remain difficult to distinguish from vessels, scars, or other small pulmonary abnormalities on CT. This observation challenges the popular notion that advances in imaging technology could justify replacing open or palpation-assisted approaches with VATS as the major surgical approach. In our practice, prior local recurrence and a preoperative nodule count ≥3 were associated with a higher likelihood of additional findings of metastatic nodules, which indicates that palpation-based procedures are more indicated in patients with these risk factors. Notably, we found that neither the total number of metastasectomies nor intraoperative identification of occult metastases was associated with a worse outcome for patients with metastasectomies, suggesting that surgical remission of all known metastatic lesions might confer satisfactory oncological results, regardless of whether these lesions were pre-planned or from incidental discovery.
Paradoxically, our data suggest no significant survival benefit of mini-thoracotomy to explore additional nodules over VATS without palpating additional nodules. Similar findings have been reported in multi-institutional and cooperative-group cohorts, particularly among patients with solitary or limited pulmonary disease, in whom survival outcomes did not differ significantly between thoracoscopic and open approaches despite the greater capacity for manual lung palpation with thoracotomy (10,40). Several explanations are possible. First, long-term survival in metastatic osteosarcoma may be driven more strongly by systemic tumor biology and treatment response than by the removal of a small number of additional occult nodules. Second, VATS was preferentially used in patients with fewer and smaller lesions, whereas mini-thoracotomy was more commonly selected for greater metastatic burden; this confounding by indication may have obscured or exaggerated differences between approaches. Third, subsequent systemic therapy, repeat metastasectomy, and treatment of later recurrence may have attenuated any survival effect attributable to the initial surgical approach. Finally, the modest VATS sample size limited the ability to detect small but clinically meaningful survival differences. Accordingly, our results support a selective rather than universal role for palpation-assisted mini-thoracotomy. It may be particularly useful in patients with multiple preoperative nodules, prior local recurrence, or a high probability of radiologically occult disease. However, the absence of a demonstrated survival advantage means that the choice between VATS and mini-thoracotomy should also consider metastatic burden, feasibility of complete resection, the need for repeat surgery, and perioperative recovery.
However, this study has several limitations that warrant consideration. First, its retrospective, single-center design and 22-year recruitment period may have introduced selection bias and temporal heterogeneity, as surgical approaches, systemic therapies, imaging techniques, referral patterns, and patient-selection criteria evolved over time. Second, the VATS and mini-thoracotomy groups differed in disease burden and clinical indications. Treatment allocation was non-random, and matching based on a single measured covariate could not eliminate residual confounding. The modest sample size also limited subgroup and interaction analyses, while repeated procedures in some patients may have introduced within-patient dependence in procedure-level comparisons. Third, the complete case analysis and the requirement for postoperative follow-up may have introduced additional selection bias. Radiological assessments were affected by changes in imaging protocols and incomplete standardization of the retrospective process. Finally, the findings were exploratory and lacked external validation. These limitations highlight the need for multi-center prospective studies with standardized surgical selection criteria, radiological assessment and external validation to confirm our findings.
Conclusions
In this single-center retrospective cohort, palpation-assisted mini-thoracotomy increased the detection of pathologically confirmed occult pulmonary nodules but did not demonstrate a survival advantage over VATS. Response to preoperative systemic therapy was independently associated with PMS, and exploratory modeling generated a candidate 2–3-month surveillance window. These associations should inform individualized multidisciplinary assessment rather than universal treatment rules and require prospective multi-center validation.
Supplementary
The article’s supplementary files as
Acknowledgments
We thank all patients and their families for their trust, cooperation and long-term follow-up, and acknowledge the multidisciplinary teams involved in patient care and data collection, which made this study possible. We are grateful to Professor Robin L. Jones from The Royal Marsden NHS Foundation Trust for his insightful comments and suggestions during manuscript preparation. We also thank Dr. Hecheng Li for his contribution to the surgical care of the patients included in this study.
Parts of this work were previously presented as conference abstracts at the ESMO Asia Congress 2023 and ESMO Congress 2025.
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 Ethics Committee of Ruijin Hospital in 2024 (Approval No. 2024-LLS-252). The requirement for informed consent was waived by the ethics committee because the study involved only retrospective analysis of deidentified clinical data.
Footnotes
Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0589/rc
Funding: This work was supported by the National Natural Science Foundation of China (Nos. 82141116 and 82473015, to W.Z.) and Chinese Anti-Cancer Association (No. CETSDHRCORP252-3-008, to Q.B.).
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0589/coif). The authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0589/dss
References
- 1.Siegel RL, Kratzer TB, Giaquinto AN, et al. Cancer statistics, 2025. CA Cancer J Clin 2025;75:10-45. 10.3322/caac.21871 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Meltzer PS, Helman LJ. New Horizons in the Treatment of Osteosarcoma. N Engl J Med 2021;385:2066-76. 10.1056/NEJMra2103423 [DOI] [PubMed] [Google Scholar]
- 3.Buddingh EP, Anninga JK, Versteegh MI, et al. Prognostic factors in pulmonary metastasized high-grade osteosarcoma. Pediatr Blood Cancer 2010;54:216-21. 10.1002/pbc.22293 [DOI] [PubMed] [Google Scholar]
- 4.Briccoli A, Rocca M, Salone M, et al. High grade osteosarcoma of the extremities metastatic to the lung: long-term results in 323 patients treated combining surgery and chemotherapy, 1985-2005. Surg Oncol 2010;19:193-9. 10.1016/j.suronc.2009.05.002 [DOI] [PubMed] [Google Scholar]
- 5.Slade AD, Warneke CL, Hughes DP, et al. Effect of concurrent metastatic disease on survival in children and adolescents undergoing lung resection for metastatic osteosarcoma. J Pediatr Surg 2015;50:157-60; discussion 160. 10.1016/j.jpedsurg.2014.10.038 [DOI] [PubMed] [Google Scholar]
- 6.Leary SE, Wozniak AW, Billups CA, et al. Survival of pediatric patients after relapsed osteosarcoma: the St. Jude Children’s Research Hospital experience. Cancer 2013;119:2645-53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kager L, Zoubek A, Pötschger U, et al. Primary metastatic osteosarcoma: presentation and outcome of patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols. J Clin Oncol 2003;21:2011-8. 10.1200/JCO.2003.08.132 [DOI] [PubMed] [Google Scholar]
- 8.Boam T, Rogoyski BG, Jawaid W, et al. Do Children With Osteosarcoma Benefit From Pulmonary Metastasectomy? A Systematic Review of Published Studies and “Real World” Outcomes. Ann Surg 2024;280:235-40. [DOI] [PubMed] [Google Scholar]
- 9.Ritter J, Bielack SS. Osteosarcoma. Ann Oncol 2010;21 Suppl 7:vii320-5. 10.1093/annonc/mdq276 [DOI] [PubMed] [Google Scholar]
- 10.Lautz TB, Farooqui Z, Jenkins T, et al. Thoracoscopy vs thoracotomy for the management of metastatic osteosarcoma: A Pediatric Surgical Oncology Research Collaborative Study. Int J Cancer 2021;148:1164-71. 10.1002/ijc.33264 [DOI] [PubMed] [Google Scholar]
- 11.Marulli G, Mammana M, Comacchio G, et al. Survival and prognostic factors following pulmonary metastasectomy for sarcoma. J Thorac Dis 2017;9:S1305-15. 10.21037/jtd.2017.03.177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Teke ME, Saif A, Sarvestani AL, et al. A Phase III, Randomized, Controlled Trial Comparing Open Versus Thoracoscopic Management of Pulmonary Metastases in Patients with Osteosarcoma. Ann Surg Oncol 2022;29:7961-3. 10.1245/s10434-022-12488-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Carrle D, Bielack S. Osteosarcoma lung metastases detection and principles of multimodal therapy. Cancer Treat Res 2009;152:165-84. 10.1007/978-1-4419-0284-9_8 [DOI] [PubMed] [Google Scholar]
- 14.Strauss SJ, Frezza AM, Abecassis N, et al. Bone sarcomas: ESMO-EURACAN-GENTURIS-ERN PaedCan Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol 2021;32:1520-36. 10.1016/j.annonc.2021.08.1995 [DOI] [PubMed] [Google Scholar]
- 15.Kempf-Bielack B, Bielack SS, Jürgens H, et al. Osteosarcoma relapse after combined modality therapy: an analysis of unselected patients in the Cooperative Osteosarcoma Study Group (COSS). J Clin Oncol 2005;23:559-68. 10.1200/JCO.2005.04.063 [DOI] [PubMed] [Google Scholar]
- 16.Ferrari S, Briccoli A, Mercuri M, et al. Postrelapse survival in osteosarcoma of the extremities: prognostic factors for long-term survival. J Clin Oncol 2003;21:710-5. 10.1200/JCO.2003.03.141 [DOI] [PubMed] [Google Scholar]
- 17.Bielack SS, Kempf-Bielack B, Branscheid D, et al. Second and subsequent recurrences of osteosarcoma: presentation, treatment, and outcomes of 249 consecutive cooperative osteosarcoma study group patients. J Clin Oncol 2009;27:557-65. 10.1200/JCO.2008.16.2305 [DOI] [PubMed] [Google Scholar]
- 18.Davis LE, Bolejack V, Ryan CW, et al. Randomized Double-Blind Phase II Study of Regorafenib in Patients With Metastatic Osteosarcoma. J Clin Oncol 2019;37:1424-31. 10.1200/JCO.18.02374 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Italiano A, Mir O, Mathoulin-Pelissier S, et al. Cabozantinib in patients with advanced Ewing sarcoma or osteosarcoma (CABONE): a multicentre, single-arm, phase 2 trial. Lancet Oncol 2020;21:446-55. 10.1016/S1470-2045(19)30825-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228-47. 10.1016/j.ejca.2008.10.026 [DOI] [PubMed] [Google Scholar]
- 21.Saifuddin A, Baig MS, Dalal P, et al. The diagnosis of pulmonary metastases on chest computed tomography in primary bone sarcoma and musculoskeletal soft tissue sarcoma. Br J Radiol 2021;94:20210088. 10.1259/bjr.20210088 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. 10.1097/01.sla.0000133083.54934.ae [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dugan KC, Laxmanan B, Murgu S, et al. Management of Persistent Air Leaks. Chest 2017;152:417-23. 10.1016/j.chest.2017.02.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Roy PG, Ganguly S, Sasi A, et al. Determinants of tumor necrosis and its impact on outcome in patients with Localized osteosarcoma uniformly treated with a response adapted regimen without high dose Methotrexate- A retrospective institutional analysis. J Bone Oncol 2024;49:100651. 10.1016/j.jbo.2024.100651 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Becht E, Giraldo NA, Lacroix L, et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol 2016;17:218. 10.1186/s13059-016-1070-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Jia Q, Wu W, Wang Y, et al. Local mutational diversity drives intratumoral immune heterogeneity in non-small cell lung cancer. Nat Commun 2018;9:5361. 10.1038/s41467-018-07767-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Signorell A. DescTools: Tools for Descriptive Statistics. 2014. doi: 10.32614/cran.package.desctools. [DOI]
- 28.Ho DE, Imai K, King G, et al. MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. J Stat Softw 2011;42:1-28. [Google Scholar]
- 29.R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. 2025. Available online: https://www.r-project.org/
- 30.Harting MT, Blakely ML, Jaffe N, et al. Long-term survival after aggressive resection of pulmonary metastases among children and adolescents with osteosarcoma. J Pediatr Surg 2006;41:194-9. 10.1016/j.jpedsurg.2005.10.089 [DOI] [PubMed] [Google Scholar]
- 31.Daw NC, Billups CA, Rodriguez-Galindo C, et al. Metastatic osteosarcoma. Cancer 2006;106:403-12. 10.1002/cncr.21626 [DOI] [PubMed] [Google Scholar]
- 32.Pastorino U, Buyse M, Friedel G, et al. Long-term results of lung metastasectomy: prognostic analyses based on 5206 cases. J Thorac Cardiovasc Surg 1997;113:37-49. 10.1016/s0022-5223(97)70397-0 [DOI] [PubMed] [Google Scholar]
- 33.Briccoli A, Rocca M, Salone M, et al. Resection of recurrent pulmonary metastases in patients with osteosarcoma. Cancer 2005;104:1721-5. 10.1002/cncr.21369 [DOI] [PubMed] [Google Scholar]
- 34.Arndt CA, Koshkina NV, Inwards CY, et al. Inhaled granulocyte-macrophage colony stimulating factor for first pulmonary recurrence of osteosarcoma: effects on disease-free survival and immunomodulation. a report from the Children’s Oncology Group. Clin Cancer Res 2010;16:4024-30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Huerta CT, Rodriguez C, Sainathan S, et al. Pulmonary metastasectomy in pediatric patients: a narrative review. Video-assist Thorac Surg 2024;9:23. [Google Scholar]
- 36.Gossot D, Radu C, Girard P, et al. Resection of pulmonary metastases from sarcoma: can some patients benefit from a less invasive approach? Ann Thorac Surg 2009;87:238-43. 10.1016/j.athoracsur.2008.09.036 [DOI] [PubMed] [Google Scholar]
- 37.Blackmon SH, Shah N, Roth JA, et al. Resection of pulmonary and extrapulmonary sarcomatous metastases is associated with long-term survival. Ann Thorac Surg 2009;88:877-84; discussion 884-5. 10.1016/j.athoracsur.2009.04.144 [DOI] [PubMed] [Google Scholar]
- 38.Gerrand C, Amary F, Anwar HA, et al. UK guidelines for the management of bone sarcomas. Br J Cancer 2025;132:32-48. 10.1038/s41416-024-02868-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Detterbeck FC, Grodzki T, Gleeson F, et al. Imaging requirements in the practice of pulmonary metastasectomy. J Thorac Oncol 2010;5:S134-9. 10.1097/JTO.0b013e3181dcf64d [DOI] [PubMed] [Google Scholar]
- 40.Mettmann VL, Baumhoer D, Bielack SS, et al. Solitary pulmonary metastases at first recurrence of osteosarcoma: Presentation, treatment, and survival of 219 patients of the Cooperative Osteosarcoma Study Group. Cancer Med 2023;12:18219-34. 10.1002/cam4.6409 [DOI] [PMC free article] [PubMed] [Google Scholar]