The prognosis of malignant pleural mesothelioma (MPM) is extremely poor with an increasing incidence rate worldwide (1,2). Multimodal therapies have been adopted for MPM, including surgery, chemotherapy and radiotherapy (3). In particular, the goal of radical surgery is to achieve macroscopic complete resection (MCR) (4) obtained via extrapleural pneumonectomy (EPP) or pleurectomy/decortication (P/D). EPP involves the en bloc resection of the ipsilateral lung and the pleura, diaphragm, and pericardium (5–7). P/D includes the removal of the ipsilateral pleura and preservation of the lung parenchyma (8). Therefore, EPP is a lung-sacrificing surgery, whereas P/D is a lung-sparing surgery.

Certain studies identified benefits of P/D regarding postoperative survival (9,10). Two previous meta-analyses comparing EPP and P/D revealed no statistical difference in long-term survival, although P/D was associated with lower morbidity and mortality rates (11,12). Due to these results, many centers have shifted their surgical approach for MPM from EPP to P/D. Comparing EPP and P/D, little is understood regarding recurrence rates, and few studies have focused on survival and treatment implementation rates following recurrence.

Recently, we adopted P/D due to the reports of its comparable survival and decreased morbidity and mortality rates. The purpose of the present study was to investigate surgical results, including post-recurrence survival (PRS), in patients treated with EPP or P/D for MPM.

A total of 50 patients underwent a surgical intervention for MPM following neoadjuvant chemotherapy. From April 2005 to July 2011, 25 patients underwent EPP. Among them, 20 patients attained MCR via EPP and 5 patients required an exploratory thoracotomy. From August 2011 to December 2017, 25 patients underwent P/D. Among them, 15 patients attained MCR via P/D and 9 patients shifted to EPP to achieve MCR. Only 1 patient required exploratory thoracotomy. MCR was achieved in 44 patients (29 EPP and 15 P/D; Fig. 1), 18 patients (62%) in the EPP group received adjuvant radiotherapy and 8 patients (53%) in the P/D received adjuvant chemotherapy.

Patient demographics, including pre-/intra-/postoperative data, are presented in Tables I–II. All patients, except one, received pemetrexed and cisplatin as neoadjuvant chemotherapy. The median operative time was significantly longer for P/D compared with EPP (P<0.0001). In the P/D group, the histology of all patients except one was the epithelial type. The 90-day mortality rate was 6.9% for the EPP group and 0% for the P/D (P=0.429).

The median follow-up period was 48 months (54 months in the EPP group and 43 months in the P/D group). The median and 5-year OS time of all patients was 22 months [95% confidence interval (CI)=14.2–27.8] and 20%, respectively (Fig. 2A). The median and 5-year PFS time of all patients was 14 months (95% CI, 11.9–16.1) and 13%, respectively (Fig. 2B). OS was significantly different between the EPP and P/D groups for all patients (median OS time, 17 vs. 34 months; 5-year OS, 11 vs. 44%; P=0.019; Fig. 2C), whereas no difference was observed for PFS (median PFS time, 13 vs. 21 months; 5-year PFS, 11 vs. 17%; P=0.373; Fig. 2D). In patients with epithelial histology, no significant difference was identified in OS (median OS time, 28 vs. 34 months; 5-year OS, 20 vs. 47%; P=0.162; Fig. 2E) or PFS (median PFS time, 21 vs. 21 months; 5-year PFS, 20 vs. 18%; P=0.910; Fig. 2F) between the two groups. Univariate analysis, including age, final histology, pathological stage, histopathological effect and surgical procedure, identified epithelial histology [hazard ratio (HR), 0.203; 95% CI, 0.085–0.471; P=0.0003] and P/D (HR=0.388; 95% CI, 0.154–0.855; P=0.018) as significant favorable prognostic factors of OS (Table III). Following multivariate analysis, only epithelial histology (HR, 0.224; 95% CI, 0.086–0.557; P=0.001) remained a significant prognostic factor for OS (Table III).

The recurrence rates following EPP and P/D were 76% (22/29 patients) and 67% (10/15 patients), respectively (Table IV). No significant difference was revealed for the type of tumor recurrence, including locoregional, distant and both, between the groups (P=0.705). Among the 32 patients who developed recurrence, 17 patients (53%) received chemotherapy, including 9 patients (41%) in the EPP group and 8 patients (80%) in the P/D (Table IV). Patients in the P/D group demonstrated a higher likelihood of receiving chemotherapy following recurrence (P=0.046).

The median PRS time was 5 months (95% CI, 1.755–8.245). PRS was significantly longer in the P/D group compared with the EPP group (median PRS time, 3 vs. 20 months; 1.5-year PRS, 5 vs. 54%; P=0.003; Fig. 3A), even among patients with epithelial-type disease (median PRS time, 6 vs. 20 months; 1.5-year PRS, 8 vs. 61%; P=0.012; Fig. 3B). Patients who received chemotherapy following recurrence presented with a significantly longer PRS compared with those who did not receive chemotherapy (median PRS time, 15 vs. 1 months; 1.5-year PRS time, 36 vs. 0%; P=0.0001; Fig. 3C). The same was true for patients with epithelial-type disease (median PRS time, 18 vs. 2 months; 1.5-year PRS, 49 vs. 0%; P=0.001; Fig. 3D). Univariate analysis, including age, final histology, pathological stage, chemotherapy following recurrence, histopathological effect and surgical procedure, indicated that epithelial histology (HR, 0.255; 95% CI, 0.100–0.631; P=0.004), chemotherapy following recurrence (HR, 0.225; 95% CI, 0.092–0.52; P=0.0005), and P/D (HR, 0.238; 95% CI, 0.076-.618; P=0.002) were significant favorable prognostic factors for PRS (Table V). With multivariate analysis, chemotherapy following recurrence (HR, 0.318; 95% CI, 0.101–0.911; P=0.033) was the only significant prognostic factor for PRS (Table V).

In total, 22 patients were assessed for VC and FEV1 (12 EPP and 10 P/D) and 16 patients were assessed for DLCO (7 EPP and 9 P/D). FEV1 (average reduction rate, 41 vs. 19%; P=0.017), VC (average reduction rate, 43 vs. 28%; P=0.036) and DLCO (average reduction rate, 25 vs. 0.01%; P=0.042) following surgery were significantly improved in the P/D group compared with the EPP group (Fig. 4)

The present retrospective study compared surgical outcomes between EPP and P/D. A number of studies have demonstrated that patients with epithelial histology have a significantly longer survival time compared with those with non-epithelial histology (17–19); the current multivariate analysis of OS time revealed the same result. When the study subjects were limited to those with epithelial MPM, no significant difference was identified in OS or PFS time between the two groups. However, PRS was significantly improved in patients who underwent P/D compared with those who underwent EPP. In addition, patients who received chemotherapy following recurrence had a significantly longer PRS time. The majority of patients undergoing P/D could receive chemotherapy.

Several studies have demonstrated that patients who undergo P/D have improved survival compared with those who undergo EPP (9,10,20); however, other studies have reported no difference in survival (21–23). The optimal surgical procedure with resectable MPM remains to be established due to the lack of randomized prospective studies comparing the surgical procedures; however, the majority of studies including the present study have demonstrated that prognosis is comparable between EPP and P/D, which suggests that P/D can be an alternative procedure to EPP.

Locoregional recurrence remains the main problem for MPM and various treatments have been employed, including adjuvant chemotherapy or radiotherapy (24,25). A number of studies have reported that EPP can provide improved local control, i.e free from recurrence in the ipsilateral hemithorax. P/D has a major disadvantage of high local recurrence rates (26,27). However, in the current analysis, the PFS curve between the two groups was similar, suggesting that P/D can also adequately provides local control for MPM. It is essential for EPP or P/D to conduct trimodality therapy and remove as much visible tumor as possible. Even when EPP is performed, positive microscopic margins are almost inevitable (17,28). Recently, two retrospective studies reported that disease-free survival rates were similar between EPP and P/D (29,30). Similar to the present study, these studies also adopted cisplatin/pemetrexed as induction chemotherapy and resected the diaphragm and/or pericardium if required for MCR. Precise surgery for obtaining MCR including trimodality therapy as well as selection of surgical procedures appeared to serve a critical role in acquiring the local control.

Although PFS was similar between the two groups, OS was improved for P/D compared with EPP, which suggests that surgical results could differ following recurrence. Therefore, the current study also assessed PRS and treatments following recurrence, which demonstrated that patients undergoing P/D exhibited a significantly improved PRS compared with those undergoing EPP regardless of the histologic type. In addition, the majority of patients who underwent P/D received chemotherapy following recurrence, unlike the majority of those who underwent EPP. A recent retrospective study reported that second-line treatment following EPP can prolong PRS (31). In the present study, chemotherapy following recurrence improved PRS for patients with MPM regardless of the surgical procedure. To the best of our knowledge, this is the first report to demonstrate that P/D prolongs survival rates following recurrence in patients with MPM.

The current results raise the question of why P/D can improve PRS. P/D is a less invasive surgical procedure compared with EPP. In the present study, the 90-day mortality rate following P/D was 0%, compared with a mortality rate of 6.9% for EPP, although this difference was not statistically significant. By preserving the ipsilateral lung, the postoperative pulmonary function of patients following P/D can be preserved to some extent (32,33). This advantage of P/D appeared to have a favorable effect on PRS in the present study. The current analysis indicated that pulmonary function following P/D was significantly improved compared with that following EPP. Therefore, chemotherapy was more conducive for patients in the P/D group compared with those in the EPP group following recurrence. Due to the high implementation rates of chemotherapy following recurrence, improvements in PRS may be achieved.

The current study had certain limitations. Firstly, the present study had possible patient selection bias. By chance, after August 2011, all but one tumor had epithelial histology and a significant difference was revealed in the histologic types between the two groups. Therefore, the current study also assessed surgical results in patients according to epithelial-type. In addition, 36% patients intended for P/D were converted to EPP during the surgery to attain MCR, which may have led to possible selection bias. Furthermore, the present study was a single-center trial with a small sample size. Further research will be necessary to confirm the current findings.

In conclusion, the present study evaluated the survival efficacy of P/D for MPM. P/D provided a similar PFS and improved PRS by preserving postoperative pulmonary function compared with EPP, leading to improved OS. Therefore, P/D may be an alternative procedure to EPP for patients with resectable MPM if MCR is achieved.