Systemic Therapy Use and Outcomes After Relapse from Preoperative Radiation and Extrapleural Pneumonectomy for Malignant Pleural Mesothelioma

This article reports outcomes of patients after relapse from preoperative radiation followed by extrapleural pneumonectomy and chemotherapy.

Abstract

Background.

Multimodality therapy with preoperative radiation (RT) followed by extrapleural pneumonectomy (EP) for patients with operable malignant pleural mesothelioma (MPM) has demonstrated encouraging results. At relapse, there are few data on the tolerance and efficacy of systemic therapies after prior multimodality therapy.

Materials and Methods.

We conducted a retrospective analysis of patients with relapsed MPM after RT and EPP ± adjuvant chemotherapy to determine overall survival (OS; date of relapse to death) and the proportion of patients that received systemic therapy and associated response rate (RR). OS was estimated using Kaplan–Meier method and potential prognostic variables were examined.

Results.

Fifty‐three patients were included (2008–2016). Median OS was 4.8 months (median follow‐up 4.4 months, range 0.03–34.8). Eastern Cooperative Oncology Group (ECOG) performance status (PS) ≥2, disease‐free interval (DFI) <1 year, and hemoglobin ≤110 g/L at recurrence were associated with worse prognosis. Thirty‐six percent of patients received any systemic therapy, whereas it was omitted in 62% because of poor PS. RR was 15% (0 complete responses, 15% partial responses) in 13 individuals with response‐evaluable disease. Therapy was discontinued because of toxicity (6/15) or disease progression (5/15), and median number of cycles was four.

Conclusion.

Patients with relapsed MPM following RT and EPP, especially those with ECOG PS ≥2, DFI <1 year, and hemoglobin ≤110 g/L at recurrence, have poor prognosis and low RR to first‐line systemic therapy. Earlier detection and novel diagnostic markers of relapse as well as potential neoadjuvant or adjuvant systemic therapy should be investigated in future studies.

Implications for Practice.

The results of this study have reinforced the importance of careful selection of appropriate candidates for this combined‐modality approach and favor prompt detection of recurrence with early and regular postoperative imaging and biopsy of suspected relapsed disease along with rapid initiation of systemic therapy even in patients with very low burden of disease. Furthermore, with the emergence of new systemic agents targeting different histological subtypes of malignant pleural mesothelioma, histological sampling of recurrence could inform therapeutic decisions in the future.

Introduction

Despite modern therapeutic modalities, the prognosis of patients diagnosed with malignant pleural mesothelioma (MPM) remains poor [1]. Even today, in part due to the rare nature of this disease, the optimal management of these patients remains controversial. In the context of resectable disease, various protocols using a combination of chemotherapy, radiation, extrapleural pneumonectomy (EPP), or extended pleurectomy decortication are currently used [2], [3]. The benefit of these multimodality regimens, however, has never been shown in a phase III, randomized, controlled trial. One feasibility study, randomizing patients to EPP and postoperative radical hemithoracic radiotherapy versus observation after initial chemotherapy, showed worse outcomes in those treated with surgical resection [4]. Conversely, single‐arm trials have demonstrated promising results with longer reported survival rates in patients treated with combined modality regimens than might be expected with nonsurgical interventions [2], [3].

In our institution, hemithoracic intensity‐modulated radiation therapy followed by EPP ± adjuvant chemotherapy is employed in selected individuals. Prolonged disease‐free interval (DFI) and promising overall survival (OS) have been shown in patients treated with this regimen [5], [6]. However, we have subjectively noticed that at the time of relapse, a proportion of patients do not proceed to subsequent lines of therapy or appear to have poor response to standard first‐line systemic therapy.

The aim of this retrospective study is therefore to examine outcomes of patients after relapse from preoperative radiation followed by EPP ± adjuvant chemotherapy by determining OS (defined as date of relapse to death) in this population, deliverability of subsequent palliative systemic therapy, and associated response rates.

Materials and Methods

We conducted a single‐center, retrospective, post hoc analysis of patients with recurrent disease after preoperative hemithoracic radiation followed by EPP and adjuvant chemotherapy in patients with ypN2‐3 disease on final pathologic review who were fit for treatment within 24 weeks of surgery (cisplatin/carboplatin and raltitrexed/pemetrexed for three cycles) who were originally enrolled on the Surgery for Mesothelioma After Radiation Therapy (SMART) clinical trial [5], [6]. Patients were identified from a database of all individuals having undergone this procedure at our institution from 2008 to 2016 [5], [6]. All were subsequently followed for relapse clinically, and thoracic and abdominal computed tomography (CT) scans were repeated at 3, 6, 12, 18, and 24 months, and then annually with additional investigations at the clinician's discretion. Recurrence was determined from radiologic imaging and confirmed pathologically as needed. All patients with diagnosis of relapsed disease were included in our study. A small proportion of patients received preoperative chemotherapy and given that it was unclear at the time of diagnosis whether these patients had resectable disease and therefore represented a different clinical subgroup, they were excluded from our study. Finally, individuals treated with systemic therapy after relapse outside our institution were only included in OS analyses.

Patient demographics, laboratory values, and disease characteristics at relapse were collected retrospectively from medical records. Date of radiation and surgical interventions as well as information regarding palliative therapies including all lines of systemic treatments and use of surgery or radiation were abstracted from charts. When evaluable, response to first‐line systemic therapy was measured. An independent thoracic radiologist reviewed imaging, and response to treatment was reported as per RECIST 1.1 criteria and modified RECIST criteria for mesothelioma when measurable disease involved the pleura [7], [8]. Overall response rate (ORR) was defined as the sum of complete responses and partial responses.

The primary endpoint of this trial was OS as defined by time from relapse to death; secondary endpoints were deliverability of palliative systemic therapy, reported as the proportion of patients treated with any systemic agent after relapse, and associated response rate. Descriptive statistics were used to report demographic, laboratory, and clinical data and were presented as medians and ranges for continuous factors and frequencies for categorical factors. Disease‐free interval was defined from start of radiation to relapse. OS for all patients was estimated using Kaplan–Meier method. Exploratory univariable analyses (log‐rank tests) were used to evaluate potential prognostic variables [9], [10], [11]. All statistical analyses were conducted with R‐3.2.2. Our study was approved by the University Health Network ethics review board.

Results

Patient Demographics

Fifty‐three patients were included in this study; 85% were male with a median age of 66 years, 62% were past or current smokers, 45% had an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0–1, and 64% had epithelioid histology (pathological diagnosis from resected disease; Fig. 1; Table 1). Twenty‐nine patients had nodal involvement and therefore an indication for adjuvant chemotherapy, but only 10 patients received it as per protocol. Chemotherapy was omitted in others because of poor PS (n = 15), patient (n = 1) or physician (n = 1) preference, and unknown reason (n = 1). Treatment status was unknown in the remaining patient (n = 1). Thirty‐four percent of patients (n = 18) had pathological confirmation of relapse with either cytological or surgical biopsies (n = 12 and 6, respectively; Table 1). Sites of disease are summarized in Figure 2. Disease was limited to intrathoracic metastases in 26 patients (49%), whereas others had distant metastases with or without thoracic involvement (n = 27, 51%). The most common sites of distant spread were the abdomen (n = 22), followed by liver (n = 3), bone (n = 2), and spleen (n = 1). Isolated local recurrences in the ipsilateral thorax were limited (n = 4, 7.5%).

Figure 1.

Consort diagram.Abbreviations: chemo, chemotherapy; EPP, extrapleural pneumonectomy; ORR, objective response rate; OS, overall survival; RT, radiation; Tx, systemic therapy.

Table 1. Patient characteristics at relapse (n = 53).

^aOmitted due to poor PS (n = 15), patient preference (n = 1), physician preference (n = 1), and unknown reason (n = 1).

Abbreviations: ECOG PS, Eastern Cooperative Oncology Group performance status; Hb, hemoglobin; LDH, lactate dehydrogenase; Plt, platelets; WBC, white blood cells.

Figure 2.

Pattern of recurrent disease (n = 53). Abdomen alone: intra‐abdominal nodes and/or peritoneal; bilateral: bilateral chest wall and/or lung; contralateral: contralateral pleura and/or lung; ipsilateral: ipsilateral chest wall; nodes: intrathoracic nodes alone; other: extrathoracic metastases.

Systemic Therapy

Of the 53 patients, 42 were assessed at our institution for palliative treatments after relapse (Fig. 1). Of these, only 15 patients received systemic therapy (36%). Poor performance status was the predominant reason for not proceeding with systemic therapy (n = 26; Table 2). In an exploratory analysis, we found that 26% (11/42, range 64–162 days) of patients relapsed <180 days from surgery, of whom only 27% (3/11) received systemic therapy; 73% (8/11) did not because of poor PS. In those diagnosed with relapse ≥180 days from surgery (74%, 31/42, range 181–2,156 days), 39% (12/31) were treated with systemic therapy, 58% (18/31) were not because of poor PS, and 3% (1/31) were not because treatment was not recommended in the context of indolent disease.

Table 2. Systemic therapy administration (n = 42).

^aThree patients treated with prior adjuvant chemotherapy as per protocol.

^bIndolent disease followed with observation.

Abbreviations: PD, progressive disease; PS, performance status; Tx, systemic therapy.

Systemic therapy delivery is summarized in Table 2. In treated patients, platinum and antifolate doublet (n = 14) or investigational agent (n = 1) was administered as first‐line systemic therapy. The median number of cycles of first‐line systemic therapy was 4 (range 1–11). Recommended ideal doses were administered in cycle 1 for all but one patient; however, six patients required dose delays at cycle 2 and one patient required dose reduction because of side effects. Treatment was discontinued because of chemotherapy‐related toxicity in six patients and because of clinical or radiological progression of disease in five patients. Only four patients completed the planned number of treatment cycles as recommended by their oncologist. Finally, the majority of treated patients underwent a single line of systemic therapy (n = 8). Notably, two patients were treated with ≥3 lines. Median time between diagnosis of relapsed disease and start of systemic therapy was 45 days (range 9–403; mean 102.7 days).

Efficacy

After a median follow‐up of 4.4 months (range 0.03–34.8), median OS was 4.8 months with a 6‐month OS rate of 41% (Fig. 3). Univariable analysis of patient‐, laboratory‐, and disease‐related factors at relapse showed that ECOG performance score of ≥2, DFI of <1 year since local therapy, and hemoglobin concentration ≤ 110 g/L were associated with statistically significant shorter OS (2.8 vs. 10.7 months, p = .001; 3.0 vs. 8.7 months, p = .039; 2.1 vs. 9.6, p = .0037, respectively) whereas age, sex, smoking status, surgical histology, laboratory values at time of recurrence, and sites of relapse were not (Table 3; Fig. 4).

Figure 3.

Overall survival in patients with recurrent malignant pleural mesothelioma after preoperative radiation and extrapleural pneumonectomy.

Table 3. Univariable analysis for overall survival (n = 53).

Bolded p values are statistically significant.

Abbreviations: DFI, disease‐free interval; ECOG PS, Eastern Cooperative Oncology Group performance status; Hb, hemoglobin; LDH, lactate dehydrogenase; Plt, platelets; WBC, white blood cells.

Figure 4.

Overall survival in patients with recurrent malignant pleural mesothelioma after preoperative radiation and extrapleural pneumonectomy according to patient characteristics at relapse. (A): ECOG performance status, (B): hemoglobin (g/L), and (C): disease free interval.Abbreviations: ECOG, Eastern Cooperative Oncology Group; Hb, hemoglobin; Rel, relapse; y, year.

Of the 15 patients treated with systemic therapy, 13 had disease evaluable for response as reviewed by an independent thoracic radiologist. ORR to first‐line systemic therapy was 15% (0 complete responses, 2 partial responses). Forty‐six percent of patients had stable (n = 6) and 39% had progressive disease (n = 5). Of note, three patients had previously received adjuvant platinum and antifolate chemotherapy 10, 13, and 38 months prior. Two of these patients had stable disease on subsequent rechallenge with platinum doublet chemotherapy (Table 2).

Pathology at Recurrence

Eighteen patients had pathological diagnosis of recurrence, six with surgical or core biopsies allowing detailed review of recurrent histological subtype. Interestingly, two patients with epithelioid histology on original surgical specimens had sarcomatoid and small cell features on biopsy of relapsed disease. Two individuals with baseline biphasic mesothelioma had relapse with sarcomatoid component, whereas others showed no such change in subtype (n = 2; epithelioid and biphasic at baseline and relapse).

Discussion

There are limited data to inform the management of individuals with recurrent MPM following EPP. Our study aimed to evaluate survival after relapse as well as deliverability and efficacy of systemic therapy in this population based on the subjective observation that some patients did not receive systemic therapy or had poor response. Our series included 53 patients and reported poor prognosis with a median OS of 4.8 months from the time of recurrence following hemithoracic radiation and EPP. In those followed after relapse (n = 42), subsequent palliative chemotherapy was administered in only 36% of patients with an ORR of 15%.

In patients who are not candidates for resection, median OS is approximately 12–16 months in clinical trials of first‐line platinum doublet chemotherapy and 18 months with the addition of bevacizumab [12], [13], [14]. OS otherwise ranges from 9 to 12 months in retrospective series of patients treated with platinum‐based doublets in a real‐world setting with chemotherapy delivered in 62% of patients, surgery ± radiation in 7.7%, and combination of surgery and chemotherapy ± radiation in 19.5% [15], [16], [17], [18], [19], [20].

Few studies have analyzed outcome after recurrence following multimodality therapy. Kostron et al. published a cohort of 106 patients with recurrent MPM after treatment with induction chemotherapy followed by EPP ± adjuvant radiation [21]. They reported a median postrecurrence survival (PRS) of 7 months (95% confidence interval 5–9 months). Subsequent treatments after relapse included chemotherapy alone (45%, n = 48), local radiation (9%, n = 9), repeat surgery (9%, n = 10), or a combination of modalities (10%, n = 11). The use of any palliative chemotherapy and local relapse only were associated with a longer PRS. Baldini and colleagues also examined the pattern of failure in 25 patients with relapsed MPM after trimodality treatment with EPP, adjuvant chemotherapy, and hemithoracic radiation and reported a median time from relapse to death of 3 months (range 0–24 months) [22]. Administration of subsequent therapy, however, was not detailed.

Our median postrecurrence OS of 4.8 months was more in line with Baldini's report. These poor outcomes could partly be due to a higher rate of patients with resectable isolated local recurrences with more favorable prognosis (31% vs. 7.5% in our study) in Kostron's study. In fact, none of the patients in our study had relapsed disease amenable to second surgical resection, and only a minority were treated with additional radiation. Method of surveillance for recurrent disease may account for this differing rate of isolated local recurrences, as Kostron et al. used both CT and positron emission tomography (PET)‐CT scans whereas we used CT alone. PET‐CT has been shown to diagnose relapse more accurately and thus may have led to earlier administration of subsequent therapy prior to clinical deterioration [23], [24].

Our reported poor outcomes are also likely related to the low rate of systemic therapy delivery and dose intensity after relapse. In the study by Kostron et al., 71% of patients received any form of chemotherapy after relapse, whereas only 36% of patients were treated with at least one cycle of systemic therapy in our study. Sharkey et al. reported survival outcomes after radical surgery according to timing of chemotherapy and found that only an estimated 15% of patients were unfit for systemic therapy throughout the course of their disease [25]. Retrospective series examining the proportion of patients with unresectable advanced MPM who are candidates for systemic therapy reported rates in the range of 40%–90% [20], [26], [27]. In a population‐based study, Enewold and colleagues reported that 62% of patients diagnosed in the U.S. in 2011 were treated with some form of systemic therapy [20]. In our study, poor PS was the reason for omission in all but one patient. One concern regarding multimodality regimens is that the recovery time following completion of treatment could limit delivery of subsequent systemic therapy because of poor functional reserve following EPP. In our study, patients having relapsed within 180 days from surgery were less likely to receive chemotherapy (27%) than those with recurrent disease ≥180 days from surgery (39%), which is consistent with this hypothesis.

Single‐arm and retrospective studies have reported that platinum doublet chemotherapy improves survival when administered in the perioperative setting, although there are no randomized studies to confirm this. Current guidelines recommend a multimodality approach for operable disease, incorporating other antineoplastic treatments (chemotherapy or radiation) to surgical resection [28]. However, timing of systemic therapy administration (neoadjuvant, intraoperative, or adjuvant) remains an area of debate. Sharkey et al. showed that delivery of chemotherapy, whether perioperative or delayed until disease progression, did not affect progression‐free survival or OS, except in patients with nonepithelioid histology or nodal involvement who were negatively impacted by delayed treatment [25]. It raises the question if systemic therapy should be routinely administered perioperatively to try and reduce risk of recurrence, particularly given our finding of poor prognosis at the time of relapse. The SMART trial only recommended adjuvant chemotherapy for patients with nodal involvement, but in a high proportion of patients, their poor PS in the postoperative setting precluded timely administration [5], [6]. In our study, prior adjuvant chemotherapy was not significantly associated with OS (Table 3). Future approaches may include the evaluation of checkpoint inhibitors, which may have greater efficacy and tolerability especially after radiation therapy. There are prospective trials accruing to address the role of checkpoint inhibitors, vaccine therapy, or targeted therapies in the perioperative setting in mesothelioma (NCT02707666, NCT03228537, NCT02004028, NCT01265433).

ORR to first‐line systemic therapy was lower than reported rates in both clinical trials and retrospective series [12], [13], [14], [15], [16], [17], [18], [19]. Such low response rates and poor OS led us to question whether aggressive multimodality therapy favors the emergence of more resistant clonal populations at relapse. Little has been reported about transformation to more aggressive histology in patients with relapsed MPM after trimodality therapy or its prognostic significance. Kostron et al. examined tissue samples of recurrent disease in 27 patients and found a change from epithelioid to biphasic histology in 3 cases and biphasic to epithelioid in 1 case. Surprisingly, in our series, initial surgical histology was not associated with OS. However, two cases showed transformation to more aggressive histological subtypes with sarcomatoid and small cell histology on surgical biopsy of relapsed disease. The former is associated with an aggressive disease course, resistance to chemotherapy, and poor prognosis, and the latter is a rare variant with limited clinical data described in the literature [29], [30], [31]. These findings are limited by small numbers but are relevant with respect to selection of subsequent systemic therapy given emerging treatments and clinical trials restricted to histological subtype such as antimesothelin agents in epithelioid mesothelioma, or ADI‐PEG 20 in patients with biphasic or sarcomatoid subtypes (NCT02709512), as well as the lower response rates to chemotherapy in nonepithelioid subtypes [30].

There are limitations to this study: It is a single‐center retrospective study, and not all patients were included in analyses regarding systemic therapy because of treatment delivery at other institutions. Further, the follow‐up intervals and imaging of patients from the time of relapse until death was not prospectively mandated as part of the original SMART study and assessments were according to according to local practices; this includes assessment of palliative systemic therapy response. In addition, the study population consists of a limited number of subjects. This is, however, to be expected in a rare disease where only a small proportion of patients are suitable candidates for aggressive local interventions. Furthermore, we do not know what proportion of patients with inoperable disease received first‐line palliative chemotherapy in the real‐world setting at our institution as a comparator to place our results in context. Another limitation is the short follow‐up time from relapse to death. However, at time of analysis, only 8 of 53 patients were alive; our reported median follow‐up of 4.4 months reflects the low rate of outlying survivors and poor prognosis in others. Finally, we did not include a control group of other patients with MPM treated at our institution and limited our comparison to other published series.

Conclusion

Our findings provide data to highlight the poor outcomes of patients diagnosed with relapsed MPM after preoperative radiation and EPP ± adjuvant chemotherapy, reinforcing the need for careful selection of appropriate candidates for this approach. Results of this study have informed our practice; we favor prompt detection of recurrence with early postoperative imaging and biopsy of suspected relapsed disease along with rapid initiation of systemic therapy even in patients with very low burden of disease. Furthermore, with the emergence of new systemic agents targeting different histological subtypes of MPM, histological sampling of recurrence could inform therapeutic decisions in the future. Finally, future studies should investigate the benefits of systemic therapy prior to or following local multimodality to reduce recurrence and evaluate novel diagnostic markers of relapse to identify recurrence as early as possible.

Author Contributions

Conception/design: Sara V. Soldera, Melania Pintilie, Marc de Perrot, John Cho, Penelope A. Bradbury

Provision of study material or patients: Sara V. Soldera, Marc de Perrot, John Cho

Collection and/or assembly of data: Sara V. Soldera, John Kavanagh, Natasha B. Leighl, Marc de Perrot, John Cho, Andrew Hope, Ronald Feld, Penelope A. Bradbury

Data analysis and interpretation: Sara V. Soldera, Melania Pintilie, Penelope A. Bradbury

Manuscript writing: Sara V. Soldera, John Kavanagh, Melania Pintilie, Natasha B. Leighl, Marc de Perrot, John Cho, Andrew Hope, Ronald Feld, Penelope A. Bradbury

Final approval of manuscript: Sara V. Soldera, John Kavanagh, Melania Pintilie, Natasha B. Leighl, Marc de Perrot, John Cho, Andrew Hope, Ronald Feld, Penelope A. Bradbury

Disclosures

The authors indicated no financial relationships.