Long-term outcome and prognostic factors after chest wall resection and reconstruction

Abstract

OBJECTIVES

Our study aimed to evaluate factors influencing perioperative and long-term outcomes of patients undergoing curative-intent chest wall resection and reconstruction.

METHODS

A retrospective single-centre analysis was conducted on all patients undergoing curative-intent chest wall resection and reconstruction from 2010 to 2023. Perioperative outcome was analysed for the entire cohort. Overall survival and disease-free survival were analysed using the Kaplan–Meier method and log-rank test and multivariable Cox proportional hazards regression models with a particular focus on patients with lung cancer and chest wall involvement.

RESULTS

A total of 143 consecutive patients (median age 62 years, 44.1% females) were included, and 75% of patients received perioperative systemic therapy or radiation. Rib resection alone was performed in 69.9%, additional sternal resection in 16.1%, spine resection in 11.9% and clavicle resection in 2.1%. Additional resections included the lung (n = 75), diaphragm (n = 6), pericardium (n = 2), subclavian vein (n = 2), pulmonary artery (n = 1) and multiple structures (n = 22). Reconstruction was performed using synthetic protheses (n = 89), metallic (n = 6) or combinations of materials (n = 32). Median tumour diameter was 10 cm; 88% were malignant. Local recurrence rate was 9.5%. Median disease-free survival (86 events) was 36 months, and median overall survival (62 events) was 80 months. The 5-year disease-free survival and overall survival were 54.1% and 74.1%, respectively. In patients with lung cancer, overall survival was significantly affected by age (P = 0.028), histology (P < 0.001), resection size >10 cm (P = 0.018), postoperative performance status (P < 0.001) and postoperative complications (P < 0.001) in multivariable analysis and disease-free survival by postoperative performance status (P < 0.001).

CONCLUSIONS

Postoperative performance status is correlated with overall survival after chest wall resection.

Chest wall tumours encompass benign and malignant primary neoplasms, metastatic tumours and direct invasion from adjacent organs [1].

GRAPHICAL ABSTRACT

INTRODUCTION

Chest wall tumours encompass benign and malignant primary neoplasms, metastatic tumours and direct invasion from adjacent organs [1]. The incidence of chest wall tumours is relatively low, accounting for only 2–5% of all thoracic tumours [2, 3]. Malignant chest wall neoplasms comprise a heterogeneous spectrum and can be categorized into primary and secondary malignancies [4–6]. Primary chest wall tumours are often sarcomas, representing 15–20% of all sarcomas and classified according to the fourth edition of World Health Organization (WHO) Classification [7]. Clinical management of these tumours is stage dependent. Due to the low incidence of chest wall sarcoma and the paucity of high-level clinical evidence, no chest wall-specific TNM staging-system has been established [8]. Lung cancer invading the chest wall is one of the most frequent indications for chest wall resection.

These resections and reconstructions are advanced procedures, and meticulous surgical technique is a critical factor in determining favourable perioperative and long-term outcomes. Frequent interdisciplinary collaboration with plastic surgeons, orthopedic surgeons or neurosurgeons is imperative for complex procedures [9]. Except for Ewing sarcomas, no definitive guidelines for the multimodal management of primary chest wall tumours have been established to date [5, 10, 11]. Consequently, several contentious issues regarding surgical and multimodal therapy remain unresolved.

To address this knowledge gap, we conducted a comprehensive review of 13 years of experience in chest wall resection and reconstruction for primary and secondary chest wall tumours, with a particular focus on the impact of surgical techniques on perioperative outcomes. In addition, we aimed to identify further prognostic factors for overall survival (OS) and disease-free survival (DFS) with a special focus on patients with lung cancer involving the chest wall.

MATERIALS AND METHODS

Ethical statement

This study was approved by the ethics committee of the Medical University of Vienna (1988/2024) and was performed in accordance with the declaration of Helsinki and the WMA Declaration of Taipei. Due to the retrospective study design, written informed consent was waived. The STROBE guidelines were followed in this study [12, 13].

Study population and data collection

We retrospectively reviewed a prospectively maintained database of all chest wall resections and reconstruction performed for chest wall tumours from 2010 to 2023 at the Thoracic Surgery Department of Vienna. Tumours were classified as primary, metastatic chest wall tumours and infiltrative lung cancer. Patients undergoing partial resection (<1 complete rib) and those who had previously undergone surgery at another centre were excluded.

Collected data are presented in Supplementary Table S1.

For patients who received additional treatment and/or follow-up outside of our centre, information regarding their status was obtained through telephone follow-up. Follow-up concluded on 13 November 2023.

Clinical management

Treatment decisions were based on a dedicated multidisciplinary tumour board. The type of resection was guided by tumour histology and extension. The type of reconstruction was at the discretion of the surgeon and based on intraoperative findings. Chest wall reconstruction included synthetic materials (Vicryl mesh, Gore-Tex, Polypropylene or Permacol mesh) or metallic materials (Titanium systems, 3D-printed, Nuss-system or Synthes/Pioneer sternal cable system) alone or in combination with biological flaps. Sternal reconstructions required dedicated materials. Biological flaps included muscle flaps, pedicled or free myocutaneous flaps. Muscles utilized included latissimus-dorsi, pectoralis, diaphragm, rectus-abdominis muscle and transverse myocutaneous gracilis flap. Follow-up was performed within the Department of Thoracic Surgery or cooperating departments.

Statistical analysis

Categorical variables are presented as number and percentage; continuous variables are presented as median and range. Clinicopathological characteristics in univariable settings were statistically analysed by χ² test and Kruskal–Wallis rank-sum-tests (for the association between a categorical and a continuous variable), Pearson-correlation coefficients (for two continuous variables) or χ² tests (for two categorical variables). Multivariable linear regression models were utilized for the prediction of continuous variables, and logistic regression was used to predict categorical parameters. Model selection was based on maximizing either the adjusted R-squared value (continuous target variables) or the area under the receiver operating characteristic curve (categorical target variables). This approach circumvents the introduction of threshold values during model selection (as is the case for step-wise regression methods) and allows for directly choosing the best-performing model out of the complete set of all models with every possible parameter combinations.

OS was defined as time from the date of surgery to death with surviving patients being censored or the date of the last follow-up. DFS was defined as the interval between the date of surgery and the date of local or distant recurrence. The inverse Kaplan–Meier method was used to estimate follow-up information. OS and DFS were analysed using the Kaplan–Meier-method and log-rank-test in univariable settings. Parameters associated with OS or DFS in univariable models (P-value < 0.3) were included in a multivariable Cox proportional hazards regression model to access their independent effects.

The investigated independent variables were preoperative American Society of Anesthesiologists (ASA) classification, age, gender, pulmonary and cardiovascular comorbidities, previous tumours, type of resection, surgical approach, additionally resected organs, preoperative therapies, type of reconstruction, tumour type and histology, maximum diameter of resection, and resection. All statistical analyses were performed using R version 4.2.1 (R Foundation, Vienna, Austria).

RESULTS

Patients and clinicopathological characteristics

A total of 143 patients were operated for chest wall tumours from January 2010 to November 2023 (Table 1); 55.9% were males and 44.1% females. Median age was 62 years (range 24–85). Preoperative status was ASA III in 44.1% of the cases, ASA II in 41.3%, ASA I in 12.6% and ASA IV in 2.1%. Most patients had cardiovascular comorbidities (65.7%), and no previous tumours (60.3%). Forty-two percent of patients also had pulmonary comorbidities. Seventy-five (75%) of patients received perioperative systemic therapy or radiation, of whom 51.7% underwent neoadjuvant therapy.

Table 1:

Summary of patient characteristics

	Overall (N = 143)
General parameters
Gender
Male	80 (55.9%)
Female	63 (44.1%)
Age (years)
Median (range)	62.0 (24.0, 85.0)
Preoperative status
ASA
ASA I	18 (12.6%)
ASA II	59 (41.3%)
ASA III	63 (44.1%)
ASA IV	3 (2.1%)
ASA V	0 (0.0%)
Pulmonary comorbidities
None	83 (58.0%)
Any	60 (42.0%)
Cardiovascular comorbidities
None	49 (34.3%)
Any	94 (65.7%)
Previous tumours
NA	2
None	85 (60.3%)
Any	56 (39.7%)
Type of tumour
Primary lung cancer with chest wall infiltration	79 (55.2%)
Primary chest wall tumour	38 (26.6%)
Metastatic chest wall tumour	23 (16.1%)
Other malignancies	3 (2.1%)
Histology (if lung cancer)
NA	64
Adenocarcinoma	42 (53.2%)
Neuroendocrine tumours	3 (3.8%)
Squamous cell carcinoma	34 (43.0%)
Preoperative therapy
None	69 (48.3%)
Neoadjuvant treatment	74 (51.7%)
Surgical parameters
Resection approach
Posterolateral thoracotomy	66 (46.2%)
Anterolateral thoracotomy	23 (16.1%)
Sternotomy	3 (2.1%)
Sternotomy with lateral extension	8 (5.6%)
Posterolateral thoracotomy + spine approach	12 (8.4%)
Thoracoscopy	3 (2.1%)
Combinate: open surgery + VATS	10 (7.0%)
Hemiclamshell	14 (9.8%)
Clamshell	4 (2.8%)
Resected sections
Not necessary (other)	2 (1.4%)
Ribs	100 (69.9%)
Sternum (and ribs)	23 (16.1%)
Spine (and ribs)	17 (11.9%)
Other	1 (0.7%)
Additionally resected organs
None	36 (25.2%)
Any	107 (74.8%)
Maximum diameter of resection (cm)
Median (range)	10.0 (2.0, 29.0)
Postoperative parameters
Resection status
NA	1
R0	122 (85.9%)
R1	20 (14.1%)
R2	0 (0.0%)
Hospital stay (days)
NA	3
Median (range)	8,5 (2.0–180.0)
Complications
NA	3
No	104 (74.3%)
Yes	36 (25.7%)
Recurrence
No	77 (53.8%)
Yes	66 (46.2%)
Location of recurrence
Local	14 (9.5 %)
Regional	15 (10.1 %)
Distant	40 (27.0%)
Redo-operation
NA	4
No	105 (75.5%)
One	30 (21.6%)
More than one	4 (2.9%)
Neuropathic pain
NA	2
No	122 (86.4%)
Yes	19 (13.5%)
VAC therapy
No	135 (94.4%)
Yes	8 (5.6%)

Perioperative results

Median tumour diameter was 10 cm (range 2–29). Eighty-eight percent were malignant: primary lung cancer with chest wall infiltration (55.2%) of which 53.2% were adenocarcinoma (AC) and 43% squamous cell carcinoma (SCC); primary chest wall tumour (26.6%) and metastatic chest wall tumour (16.1%). Surgical approaches were posterolateral thoracotomy (46.2%), sternotomy (21%), anterolateral thoracotomy (16.1%), hemiclamshell (9.8%), posterolateral thoracotomy plus spine approach (8.4%), thoracoscopy combined with open surgery (7%), sternotomy with lateral extension (5.6%), clamshell (2.8%) and thoracoscopy (2.1%). Rib resection only was performed in 69.9% of cases, additional sternal resection in 16.1%, spine in 11.9% and clavicle/scapula in 6 cases (0.7%). Additional resections were lung (n = 75), diaphragm (n = 6), pericardium (n = 2), subclavian vein (n = 2), pulmonary artery (n = 1) and multiple structures (n = 22). Synthetic (n = 89), metallic (n = 6) and material combinations (n = 32) were used for reconstruction. Biological flaps were used in 24 patients (15 muscle flap, 6 pedicled myocutaneous flaps, 3 free myocutaneous flaps). R0 resection was achieved in 85.9%.

Median duration of surgery was 180 min (range 20—700 min). In a multivariable linear regression model, patients undergoing neoadjuvant treatment had significantly longer operations than those who did not (P = 0.016). The type of surgical procedures and their association with the type of resected sections and operation time are shown in Fig. 1.

Figure 1:

Surgical approach and its association with resected sections and operation time

Complications occurred in 36 patients (25.7%) including surgical wound infection in 9 (6.1%), empyema in 5 (3.4%), bleeding in 4 (2.7%), pneumonia or pneumonitis in 2 (1.4%), exacerbation of underlying pulmonary disease in 1 (0.7%) and other complications such as flap necrosis and chest wall instability in 11 cases (7.4%). Multiple complications were noted in four cases (2.7%). Age (P = 0.007), sternal resection (P = 0.0039) and duration of surgery (P = 0.012) were independent predictors of complications in a multivariable logistic regression model (Fig. 2). Interestingly cardiovascular comorbidities were associated with reduced perioperative complications (P = 0.008).

Figure 2:

Multivariable analysis of factors associated with perioperative complications

Median length of hospital stay was 8.5 days (range 2–18). In a multivariable linear regression model, ASA score (P = 0.015), cardiovascular comorbidities (P = 0.038), surgical approach (P = 0.043), resection diameter (P = 0.005), type of reconstruction (P = 0.049) and postoperative complications (P = 0.034) had an impact on the duration of hospital stay (Fig. 3).

Figure 3:

Multivariable analysis of factors associated with hospital stay

Neuropathic pain occurred in 19 patients (13.5%). In a multivariable logistic regression model, posterolateral thoracotomy was more likely to cause neuropathic pain compared to an anterolateral approach (P = 0.035) as well as longer duration of surgery (P = 0.011) (Fig. 4).

Figure 4:

Multivariable analysis of factors associated with neuropathic pain

Long-term outcome

Median DFS (86 events) and OS (62 events) was 36 and 80 months, respectively. Five-year DFS and OS were 54.1% and 74.1%, respectively.

The results of univariable analyses for OS and DFS in all groups of patients are shown in Supplementary Figs S1 and S2. OS was significantly associated with the following factors: age >62 years (P = 0.0044), SCC (log-rank P = 0.0031), larger than median resection diameter (log-rank P = 0.022), postoperative complications (log-rank P = 0.025), limited functional status (log-rank P < 0.0001) and tumour recurrence (log-rank P < 0.0001). Specifically, patients older than 62 years had a 25% OS compared to 50% in those aged 62 years or younger. OS in patients with SCC was reduced to 50%, whereas those with AC had an OS close to 75%. Tumours > 10 cm reduced OS to 30%, while ≤10 cm associated with a 50% 5-year OS. Severely limited functional status reduced OS to 50%, compared to 80% in patients with preserved functional status.

DFS was significantly influenced by the maximum resection diameter (log-rank P = 0.025), resection margins (log-rank P = 0.045), vacuum-assisted closure (VAC) therapy (log-rank P= 0.044) and performance status (PS) (log-rank P < 0.0001).

In order to homogenize the long-term results regarding OS, multivariable analyses were performed for patients with lung cancer involving the chest wall and for all other patients separately. A multivariate Cox proportional hazards regression model analysis was fitted for all patients with non-missing metadata for all the covariates of the model (Fig. 5).

Figure 5:

Multivariable analyses of prognostic factors for overall survival

OS in lung cancer patients (N = 65) is significantly affected by the following parameters:

• Older patients tend to have shorter survival than younger ones (P = 0.028).

• Patients with SCC fare worse than patients with AC (P < 0.001),

• Resection of >10 cm has a worse prognosis (P = 0.018).

• Post-operative complications shorten survival times (P < 0.001).

• Patients who retain a ‘fully limited’ functional status after surgery have worse survival (P < 0.001).

• Interestingly, patients who stay longer in hospital after their operation (more than 8.5 days) have a better prognosis (P = 0.008).

• Patients who experience recurrence have shorter survival times (P < 0.001).

The concordance of the model is 88.13%, which outperforms the usual prognostic models (concordance in the range of 60–80%) for survival data.

Due to the fact that chest wall resection might be necessitated by other pathologies than lung cancer, we performed a multivariable Cox proportional hazards regression analysis without the lung cancer histology variable. The concordance of this model was 79.59%, and the following significant prognostic associations were identified:

• Older patients tend to have shorter survival compared to younger ones (P = 0.009).

• Patients with primary chest wall tumours have better outcomes than those with primary lung cancer involving chest wall infiltration (P = 0.011).

• Resection >10 cm has a worse prognosis (P = 0.005).

• Patients with a ‘fully limited’ functional status after surgery have worse survival outcomes (P < 0.001).

• Patients who experience recurrence have shorter survival times (P < 0.001).

As with OS, including lung cancer, histology as a model parameter limits analysis to lung cancer patients. To address this, we conducted a separate multivariable analysis (Supplementary Fig. S3). In this analysis of DFS in lung cancer patients, only the following factor significantly influences DFS:

• Patients who retain a ‘fully limited’ functional status after surgery have worse DFS (P < 0.001).

We also performed a multivariable analysis excluding lung cancer histology. In this model, the only variable with a significant prognostic effect was the retained functional status (P < 0.001). Model performance (62.71% concordance) is somewhat diminished, but still in the expected range (60–80%) for survival data.

DISCUSSION

Our study identified several important predictors of survival in patients who underwent chest wall resection and reconstruction. Five-year DFS and OS were 54.1% and 74.1%, respectively, like the results obtained by others reports based on sarcoma chest wall resection and reconstruction [14–20].

Tumour resection margin status and local recurrence were associated with diminished survival rates in some reports [18, 21]. Consequently, the current ESMO guidelines recommend en-bloc resection with an R0 margin for local control in soft tissue and visceral sarcomas, though they do not specify the minimum distance required for the resection margin. In desmoid tumours, medical therapy is now playing a central role in modern treatment options [22].

The maximum resection diameter has been also discussed extensively in existing literature. We demonstrated that a larger resection diameter of >10 cm was associated with decreased OS and DFS in univariable analyses. When planning such interventions, factors such as tumour histology and grade, the need for neoadjuvant treatments and the patient’s clinical status should also be considered. In our multivariable analyses, patients with lung cancer and chest wall infiltration had poorer survival outcomes compared to those with primary chest wall tumours. Additionally, factors such as extended resection (>10 cm), postoperative complications, limited PS, VAC therapy and recurrence significantly impacted OS. These findings are consistent with other reports [23]. Furthermore, fully impaired PS was the factor most strongly associated with adverse long-term outcomes, markedly reducing both OS and DFS, in all uni- and multivariable analyses conducted. However, the cohort with fully impaired PS was small. Given these factors, a centralized and multidisciplinary approach is crucial. Additionally, a more thorough assessment of PS both before and after surgery would be valuable for better understanding these results.

Chest wall resection surgery requires adequate preoperative pulmonary function testing. Several studies have demonstrated that outcomes following lung resection are influenced by the predicted postoperative DLCO, particularly in patients who have undergone induction therapy [19, 24]. However, in our study, pulmonary comorbidities did not influence the OS and DFS outcomes in any groups, including the subgroup of lung cancer patients. This aspect may require further investigation by separately evaluating the different preoperative and postoperative parameters of the respiratory function tests similarly to the studies by Jones et al. and Liu et al. [24, 25].

The technical aspects of resection and particularly reconstruction of the chest wall are critical factors determining the postoperative outcome. The observed association between wide resection margins and improved survival rates in sarcoma has led to the adoption of more extensive resections, which, in turn, require more complex reconstruction techniques [26–29]. Additionally, it is crucial to restore both the anatomical structure and functional integrity of the chest wall (biomimesis) [26, 30]. The aim is to avoid flail chest, thus preventing respiratory complications and allowing faster patient recovery and shorter hospital stay [31].

Over the past decade, the range of available options for reconstruction, including dynamic prostheses, has increased exponentially in number and variety [32–35]. Nevertheless, in the literature, there is still no consensus defining a standardized approach for optimal chest wall reconstruction [26, 29], and recommendations for reconstruction techniques are given based on the individual experiences [9, 20]. Our centre has introduced more flexible, stronger and safer prostheses leading to fewer postoperative complications related to prosthesis failure. An in-depth study might be considered in the future to compare long-term outcomes between static and dynamic chest wall prostheses. In our study, no statistically significant differences in perioperative outcome were found based on the reconstruction technique used. Also, no significant differences in survival or postoperative complications were observed with the use of different prostheses, in line with the results of other studies [26]. Despite the substantial proportion of extended resections involving pulmonary resection in our study (50.7%), the overall incidence of pulmonary complications (8.2%), including pneumonia, pneumonitis, empyema, haemothorax and exacerbation of underlying pulmonary disease, was consistent with rates reported in previous research [26, 29, 36]. To prevent complications, daily physiotherapy and early patient mobilization are crucial. Effective pain management, particularly using epidural catheters or intercostal nerve blocks, is essential to support these interventions and is a key component of postoperative care [18].

Given these considerations an individualized patient-tailored approach seems more promising than the search for a single ideal method [26]. This approach should consider the unique characteristics and underlying conditions of each patient, the benefits and limitations of each type of prosthesis, and the resources available at each institution. Additionally, achieving local control at each site through surgery, chemotherapy and/or radiotherapy is a critical factor to consider prior to planning a complete resection of chest wall tumour in concordance with current ESMO guidelines [10].

Strengths and limitations

Strengths of our study include an extensive follow-up of up to 13 years and a focus exclusively on patients with chest wall resection and reconstruction. To our knowledge, our report represents one of the largest and most comprehensive analyses of survival outcomes in patients who required chest wall resection and reconstruction due to tumour excision.

Despite our efforts to mitigate bias and the influence of confounding variables, retrospective studies inherently remain susceptible to selection and reporting bias. In this context, our study reflects the experience of a single high-volume centre. However, the limited sample size and in particular the limited number of events for OS must be considered when interpreting the results. The single-centre design limits the generalizability of the results. Granular functional follow-up data were not available in a systematic way due to the retrospective design. Additionally, the inclusion of both primary and secondary chest wall tumours introduces heterogeneity in our results.

CONCLUSION

A comprehensive preoperative evaluation and multimodal approach is essential for achieving optimal functional outcomes and survival rates. R1 was shown as a significant negative prognostic factor for survival. Additionally, age, lung cancer histology, maximum resection diameter, postoperative complications, PS and recurrence are independent prognostic factors impacting OS. Specifically, in lung cancer patients, perioperative mortality is low and postoperative complications, which are a prognostic factors for OS and DFS, are of low severity. The type of chest wall reconstruction had no significant impact on survival or postoperative complications.

Current knowledge is based on retrospective analyses of limited series only; therefore, multicentre studies are needed to define the optimal management approach.

Glossary

ABBREVIATONS

AC

Adenocarcinoma

ASA

American Society of Anesthesiologists

DFS

Disease-free-survival

IQR

Interquartile range

LOS

Length of hospital stay

OS

Overall survival

PS

Performance status

SCC

Squamous cell carcinoma

VAC

Vacuum-assisted closure

WHO

World Health Organization

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

FUNDING

No funding was available for this project.

 

Conflict of interest: None declared.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request.

Author contributions

Alba M. Fernandez Gonzalez: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Validation; Visualization; Writing—original draft; Writing—review & editing. Jose R. Matilla: Conceptualization; Data curation; Investigation; Methodology; Project administration; Resources; Writing—review & editing. Orsolya Anna Pipek: Formal analysis; Visualization; Writing—original draft. Laura Gonzalez Sanchez: Data curation; Investigation; Methodology; Writing—review & editing. Merjem Begic: Writing—review & editing. Zsolt Megyesfalvi: Writing—review & editing. Balazs Döme: Writing—review & editing. Clemens Aigner: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Supervision; Validation; Writing—original draft; Writing—review & editing

Reviewer information

Interdisciplinary CardioVascular and Thoracic Surgery thanks Erik R. de Loos, Pascal Alexandre THOMAS, Yuji Shiraishi, Larry R Kaiser and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.