Post-Recurrence Treatment Patterns and Outcomes After Curative-Intent Multimodality Strategy in Patients with Pleural Mesothelioma: A Retrospective Cohort Study

1. Introduction

Pleural mesothelioma (PM) is an aggressive malignancy, primarily associated with asbestos exposure. Although the use of asbestos has been heavily restricted in many countries, PM remains a significant public health concern due to its long latency period and the absence of a global asbestos ban [1].

Multimodality therapy consisting of surgery, chemotherapy, and more recently, immunotherapy is considered the standard of care for selected patients with clinically resectable PM after multidisciplinary evaluation. Among surgical options, lung-sparing approaches (LSS) such as pleurectomy/decortication (P/D) and extended pleurectomy/decortication (EPD), have been associated with improved quality of life and potentially prolonged survival compared to the radical extrapleural pneumonectomy (EPP) [2]. However, the MARS2 trial has introduced uncertainty regarding the role of surgery within multimodality treatment for PM, reporting differences in survival and quality-of-life outcomes between surgery-based and non-surgical approaches [3,4].

Overall, the prognosis for patients with PM remains poor, with a median overall survival (OS) ranging from 9 to 17 months, even following aggressive treatment [5,6,7]. A key factor limiting long-term survival is the high recurrence rate, reported in up to 75% despite curative multimodality therapy [8]. The high recurrence rate poses a challenge for clinicians, as there is no consensus on post-recurrence treatment and management in PM.

Recognizing the lack of standardized approaches for recurrent pleural mesothelioma, we conducted a retrospective cohort analysis of patients previously treated with surgery at a specialized institution, focusing on long-term survival and post-recurrence therapeutic strategies. We aimed to identify clinical factors and therapeutic approaches associated with improved outcomes following disease recurrence.

2. Materials and Methods

2.1. Study Design

This study is a retrospective, non-interventional, single-center real-world analysis. Patients included in this analysis were diagnosed with PM and treated within a curative-intent multimodality treatment pathway including surgery. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the city of Vienna (EK_14_030_VK) on 10 April 2014 [9]. The requirement for informed consent was waived for this retrospective analysis.

2.2. Patients and Data

We reviewed electronic medical records of all patients with PM treated at a referral center in Vienna, Austria, between 2001 and 2024; data cutoff was February 2025. Patients were eligible if they underwent surgery as part of a curative-intent multimodality treatment pathway. We excluded patients managed without surgery and those undergoing surgery with palliative intent. Tumor stage was reported according to pathological staging. Clinical TNM stage was not systematically available; therefore, the timing and extent of possible preoperative suspicion of stage IV disease could not be reconstructed.

2.3. Treatment

Treatment decisions were made within multidisciplinary institutional practice and were generally aligned with contemporaneous guideline-based standards applicable during the study period [10]. In general, curative-intent treatment was based on histological confirmation of PM, usually obtained by pleural biopsy, and multidisciplinary evaluation. The treatment modalities following initial diagnosis were divided into three categories: (1) surgical intervention only, (2) surgical intervention and chemotherapy, (3) surgical intervention, chemotherapy and radiotherapy. The surgery-only subgroup consisted of two patients in whom additional therapy could not be delivered before death without evidence of disease progression. These patients were retained in the curative-intent cohort because the initial treatment strategy was curative-intent surgical management, but they did not enter the post-recurrence analyses.

In selected cases, preoperative pleural biopsy was inconclusive or the preoperative working diagnosis was another pleural disease, particularly pleural empyema. In these patients, PM was diagnosed only after surgical histopathological assessment, and systemic therapy was therefore delivered postoperatively as adjuvant treatment when clinically feasible. In patients with a preoperative diagnosis or strong clinical suspicion of PM, systemic therapy was administered as induction/neoadjuvant treatment before surgery. Surgery was performed with the goal of macroscopic complete resection when feasible.

Two surgical approaches were distinguished: extrapleural pneumonectomy (EPP) and lung-sparing surgery (LSS; pleurectomy/decortication). Diagnostic and surgical approaches were individualized based on clinical presentation and disease extent, with LSS preferred whenever feasible. EPP was reserved for cases with extensive involvement of both the pleura and lung parenchyma.

Adjuvant radiotherapy consisted of hemithoracic pleural and mediastinal radiotherapy with locoregional consolidation intent and was delivered as intensity-modulated radiotherapy (IMRT) at a dose of 54-60 Gy in 2 Gy daily fractions. Radiotherapy was administered when disease distribution and postoperative anatomy permitted treatment of the ipsilateral pleural and mediastinal regions without excessive expected toxicity to the preserved lung. Patients who did not respond to neoadjuvant chemotherapy or were deemed unsuitable for surgery received subsequent radiotherapy when feasible, followed by surveillance or best supportive care only (BSC).

2.4. Follow-Up and Recurrence Assessment

After completion of curative-intent treatment, patients underwent routine postoperative surveillance according to institutional practice. Follow-up generally included clinical assessment and contrast-enhanced CT imaging at approximately 3-month intervals during the early postoperative period. During longer-term follow-up, intervals were extended, commonly to 6-9 months, depending on clinical course, prior findings, and physician assessment. PET-CT was not used as a fixed alternating surveillance modality but was performed when recurrence was suspected or when CT findings required further clarification. Recurrence was primarily diagnosed radiologically based on cross-sectional imaging. Biopsy confirmation was obtained when imaging findings were equivocal or when histological confirmation was considered necessary for treatment decision-making.

Post-recurrence treatment was categorized as BSC, local therapy with or without systemic therapy (surgery and/or radiotherapy, alone or combined with systemic therapy), or systemic therapy only.

At recurrence, treatment decisions were made within multidisciplinary institutional practice rather than according to a fixed allocation protocol. Selection of local therapy, systemic therapy, or BSC was individualized based on performance status, comorbidity, pulmonary reserve, recurrence extent and localization, prior treatment exposure, disease-free interval, feasibility of local treatment, expected treatment tolerance, patient preference, and availability of systemic treatment options during the respective treatment era.

2.5. Statistical Analysis of Study Endpoints

Overall survival (OS) was defined as the interval from the date of surgery to the date of death or last follow-up. Disease-free survival (DFS) was defined as the interval from surgery to first disease progression or death. Post-recurrence progression-free survival (prPFS) was defined as the interval from first to second disease progression or death. Post-recurrence survival (PRS) was defined as the interval from first disease progression to death or last follow-up. Median follow-up from surgery and from first recurrence was estimated using the reverse Kaplan-Meier method. DFS grouping was not included as a covariate in OS models calculated from surgery to avoid immortal time bias. Post-recurrence treatment was analyzed using post-recurrence endpoints, because treatment at recurrence is not a baseline exposure.

Survival was estimated using the Kaplan-Meier method. Subgroups were compared using the log-rank test. Exploratory analyses were performed using Cox proportional hazards regression. Parameters included age, sex, histological subtype, laterality, treatment modality, Charlson comorbidity index (CCI), type of surgery, disease stage, and DFS group (less than 12 months vs 12 months or more). Variables that were significant in univariable analyses were included in multivariable models. To explore potential temporal bias, calendar year was evaluated in exploratory analyses. Calendar year of surgery was used for OS and DFS, whereas calendar year of first recurrence was used for PRS and prPFS analyses. Treatment era was additionally evaluated using a dichotomized cutoff (pre-2021 vs. ≥2021). To assess temporal changes in surgical practice, logistic regression was used to evaluate the association between calendar year and surgical approach (LSS vs. EPP). A multivariable Cox model including calendar year of surgery and surgical approach was used to evaluate whether the association between surgical approach and OS was attenuated after adjustment for calendar time.

A p-value below 0.05 was considered statistically significant. No adjustment for multiple testing was applied, and p-values should be interpreted as exploratory. Continuous variables are reported as mean (SD) or median (IQR), as appropriate. Categorical variables are reported as counts (percentages) and were compared using the chi-square test. All analyses were performed using R (version 4.4.1) [11].

To provide clinical context for the curative-intent cohort, we performed an exploratory post hoc comparator analysis in the overall institutional cohort. OS was calculated from initial diagnosis to death or last follow-up, because surgery-based time origins were not applicable to patients managed without surgery. Patients were grouped by initial treatment strategy: curative-intent surgery-based pathway, management without surgery, or palliative-intent surgery. This analysis was interpreted descriptively because treatment allocation was non-randomized.

3. Results

3.1. Patient and Treatment Characteristics

Of 172 patients with PM treated at our institution, 87 were excluded because they did not undergo surgery and 18 were excluded because surgery was performed with palliative intent. The final cohort included 67 patients treated within a curative-intent multimodality treatment pathway including surgery (Figure 1). Among the 85 patients who underwent surgery in the overall institutional cohort, 66 (77.6%) achieved macroscopic complete resection (MCR). MCR status was not used as an exclusion criterion. Among the 67 patients in the curative-intent surgical cohort, postoperative complications occurred in 38 patients (56.7%; Supplementary Table S1). The 30-day and 90-day postoperative mortality rates were 1.5% (1/67) and 3.0% (2/67), respectively. By the data cutoff, 58 patients had developed recurrence and were included in post-recurrence analyses. Of the remaining 9 patients, 5 were alive without progression and 4 had died without evidence of disease progression.

Patient characteristics (mean age 62 years, 30% female) are summarized in Table 1. The predominant histological subtype was epithelioid (89.6%). Figure 2 illustrates the treatment trajectories of individual patients, including the initial treatment approach and therapies administered at recurrence. At diagnosis, most patients received bi- or trimodality therapy. LSS was performed in 56 patients (83.6%), and the remaining 11 patients (16.4%) underwent EPP. At data cutoff, 58 patients (86.6%) experienced PM recurrence.

In an exploratory post hoc comparator analysis of the overall institutional cohort (n=172), OS from initial diagnosis differed by initial treatment strategy. Patients treated within the curative-intent surgery-based pathway (n = 67) had a median OS of 28.2 months (95% CI, 21.23-35.6), compared with 11.1 months (95% CI, 8.52-14.6) in patients managed without surgery (n = 87). Patients undergoing palliative-intent surgery (n = 18), who were described separately, had a median OS of 17.4 months (95% CI, 7.84-34.9). In the overall three-group comparison, OS differed significantly by initial treatment strategy (log-rank p < 0.001).

3.2. Overall Survival and Disease-Free Survival

The median follow-up from surgery was 107 months (95% CI, 68-not estimable). The median OS after surgery was 19.8 months (95% CI, 16.1-28.2) (Figure 3). The 1-year, 2-year, and 5-year OS rates were 72%, 45%, and 24%, respectively. Time-point estimates for OS, DFS, PRS, and prPFS are summarized graphically in Figure 4, with detailed numerical estimates provided in Supplementary Table S2.

In univariable analysis of prognostic factors, sex and type of surgery were significantly associated with OS. Table 2 summarizes the prognostic factor analyses reported for the study endpoints. Patients undergoing LSS had longer median OS (25 months) compared with those undergoing EPP (11 months; HR 0.36, 95% CI 0.17-0.74, p=0.006). In a multivariable analysis, LSS remained significantly associated with longer OS.

The median DFS was 9.7 months (95% CI, 7.4-14.0). The 1-year, 2-year, and 5-year DFS rates were 42%, 21%, and 7%, respectively. In univariable analysis, CCI group was significantly associated with DFS; patients with CCI 0-5 had longer DFS than those with CCI >5 (HR 0.55, 95% CI 0.31-0.97, p=0.041).

3.3. Post-Recurrence Survival (PRS) and Post-Recurrence Progression-Free Survival (prPFS)

Of the 58 patients who experienced disease recurrence, 21 (36%) received local +/- systemic therapy, 26 (45%) received systemic therapy alone, and 11 (19%) received BSC. Among these patients, the median potential follow-up from first recurrence was 63.1 months (95% CI, 63.0-not estimable).

Among the 21 patients receiving local +/- systemic therapy at first recurrence, treatments were heterogeneous and included radiotherapy alone (n=8), surgery alone (n=5), surgery plus chemotherapy (n=3), radiotherapy plus chemotherapy (n=3), trimodality treatment with surgery, chemotherapy, and radiotherapy (n=1), and multimodality treatment including surgery, radiotherapy, chemotherapy, and immunotherapy (n=1). Systemic treatment regimens at first recurrence included carboplatin/pemetrexed (n = 14), cisplatin/pemetrexed (n = 5), pemetrexed alone (n = 3), vinorelbine (n = 2), nivolumab/ipilimumab (n = 4), carboplatin/pemetrexed/bevacizumab (n = 1), raltitrexed/oxaliplatin (n = 1), and unknown regimens (n = 4). Eleven patients experienced first disease progression in or after 2021; among these, four received nivolumab/ipilimumab at first recurrence. The small number of patients treated with nivolumab/ipilimumab precluded robust subgroup analyses or reliable assessment of immunotherapy-specific outcomes. At second recurrence (n=29), seven patients (24.1%) underwent further local treatment (four radiotherapy, three surgery), 17 (58.6%) received systemic therapy alone, and 5 (17.3%) received BSC.

Median PRS was 11.7 months (95% CI, 6.6-18.5). Median PRS was 1.6 months for patients managed with BSC (95% CI, 0.9-not estimable), 21.7 months for patients receiving local +/- systemic therapy (95% CI, 13.7-47.8), and 10.8 months for patients receiving systemic therapy alone (95% CI, 6.6-20.9). In multivariable analysis, local +/- systemic therapy was associated with longer PRS compared with BSC (HR 0.05, 95% CI 0.02-0.15, p<0.001), whereas the comparison with systemic therapy alone showed borderline significance (HR 0.51, 95% CI 0.26-1.00, p=0.051). LSS was associated with longer PRS compared with EPP (HR 0.28, 95% CI 0.12-0.68, p=0.005).

The 1-year and 2-year prPFS rates were 34% and 7.3%, respectively; the 5-year prPFS estimate was not estimable. Post-recurrence progression-free survival also differed across first-recurrence treatment groups; these estimates are summarized in Table 3. LSS was associated with longer prPFS than EPP (HR 0.37, 95% CI 0.16-0.84, p=0.018). Patients with DFS ≥ 12 months had longer PRS (HR 0.40, 95% CI, 0.20-0.78, p=0.007) and longer prPFS (HR 0.51, 95% CI 0.27-0.96, p=0.038). In exploratory analyses of temporal trends, calendar year of first recurrence was not significantly associated with PRS (HR 0.96 per year, 95% CI, 0.90-1.01; p = 0.126) or prPFS (HR 0.97 per year, 95% CI, 0.92-1.02; p = 0.258). In contrast, later calendar year of surgery was associated with improved OS from surgery (HR 0.94 per year, 95% CI, 0.89-0.99; p = 0.031), whereas only a non-significant trend was observed for DFS (HR 0.95 per year, 95% CI, 0.91-1.00; p = 0.061). Surgical approach changed over time, with LSS being more frequently performed in later years (OR 1.21 per year, 95% CI, 1.04-1.47; p < 0.05). In multivariable Cox regression including both calendar year of surgery and surgical approach, the association between calendar year and OS was attenuated and no longer statistically significant (HR 0.96 per year, 95% CI, 0.91-1.02; p = 0.184), whereas LSS versus EPP remained associated with improved OS (HR 0.38, 95% CI, 0.17-0.84; p = 0.017). Stratified analyses by treatment era (pre-2021 vs. ≥2021) did not demonstrate significant differences in survival outcomes.