Fax: (617) 632-4247.
Pneumonectomy after chemoradiation
The Dana-Farber Cancer Institute/Brigham and Women's Hospital experience
Version of Record online: 18 JAN 2008
Copyright © 2008 American Cancer Society
Volume 112, Issue 5, pages 1106–1113, March 2008
How to Cite
Allen, A. M., Mentzer, S. J., Yeap, B. Y., Soto, R., Baldini, E. H., Rabin, M. S., Sugarbaker, D. J. and Bueno, R. (2008), Pneumonectomy after chemoradiation. Cancer, 112: 1106–1113. doi: 10.1002/cncr.23283
- Issue online: 19 FEB 2008
- Version of Record online: 18 JAN 2008
- Manuscript Accepted: 24 SEP 2007
- Manuscript Revised: 20 SEP 2007
- Manuscript Received: 16 MAY 2007
- nonsmall cell lung cancer;
- trimodality therapy
The current study was conducted to examine the outcomes of pneumonectomy after induction chemoradiotherapy in patients with locally advanced nonsmall cell lung cancer (NSCLC).
All patients undergoing pneumonectomy after induction therapy at the Brigham and Women's Hospital were retrospectively evaluated for 30-day and 100-day mortality and treatment-related complications with Institutional Review Board approval. Multivariate and univariate analyses for clinical factors correlating with toxicity and/or survival were calculated.
Between 1995 and 2005, 73 patients underwent pneumonectomy for NSCLC after induction therapy. All patients received radiation (median dose of 54 gray [Gy]) and 69 patients (95%) received concurrent chemotherapy. The median age was 62 years and 43 patients (59%) were male; Thirty-seven patients (51%) had American Joint Committee on Cancer stage IIIA NSCLC, 27 (37%) had stage IIIB, 6 had stage IIB, and 4 had stage IV NSCLC because of a resected solitary brain metastasis. A majority (44; 60%) of patients received the combination of carboplatin and paclitaxel, whereas 15 (21%) received the combination of cisplatin and etoposide. Forty-five patients (62%) underwent left pneumonectomy. With a median follow-up of 28 months, the 1-year and 2-year overall survival rates were 70% and 49%, respectively. The 30-day and 100-day mortality rates were 6% and 10%, respectively. Only 4 of 73 patients (6%) died of acute respiratory distress syndrome. The rate of nonfatal treatment-related morbidity was 11%. On univariate analysis, right-sided pneumonectomy was associated with a higher risk of treatment-related mortality (P = .099).
With an acceptable mortality rate, a single-institutional series demonstrated that trimodality therapy including pneumonectomy can be safely accomplished in patients with advanced NSCLC. Cancer 2008. © 2008 American Cancer Society.
The treatment of locally advanced nonsmall cell lung cancer (NSCLC) depends on the stage of the cancer, the cardiorespiratory state of the patient, and locally available expertise. Many of the paradigms that have been developed in the modern era usually require a combination of chemotherapy, radiotherapy (RT), and surgical resection. The challenge in treatment of locally aggressive NSCLC is in the delivery of a locally aggressive therapy to eliminate large volumes of disease within the chest, while simultaneously addressing the risk of distant metastasis. In addition, we must ensure that the toxicity of the therapy does not outweigh the benefits of the treatment. In 1995, Albain et al.1 published a phase 2 trial that attempted to address the optimal treatment for a locally advanced NSCLC. The Southwest Oncology Group (SWOG) investigators utilized a preoperative regimen of concurrent chemotherapy and RT followed by surgical resection. The published results were encouraging, with a 3-year survival rate of 25%. In addition, the overall treatment-related mortality was acceptable at 10%, with pneumonectomy patients having a higher mortality rate (15%), which is consistent with the known increased mortality observed in patients undergoing pneumonectomy compared with lesser resections.2, 3
The results of this trial prompted a North American Intergroup trial evaluating chemoradiation followed by surgery versus chemoradiotherapy alone for patients with stage III NSCLC. This latter trial demonstrated a statically significant reduction in local failure (22% vs 10%; P = .002) and improvement in progression-free survival (PFS) (22.4% vs 11.1%; P = .017) in patients receiving the trimodality approach without a difference in overall survival.4 The lack of survival benefit was attributed to higher treatment-related mortality in the surgical group, specifically, a 26% mortality rate among the 54 patients who underwent pneumonectomy after induction chemoradiotherapy. These data led many oncologists and surgeons to conclude that pneumonectomy after chemoradiation should be abandoned as a treatment strategy because of the high mortality rate.
We were surprised by the high mortality reported in the Intergroup study. In our long-term experience with extrapleural pneumonectomy for mesothelioma, we have reported mortality in the range of 2% to 4%.5 The importance of a high-volume tertiary care center in the performance of complex surgical procedures with lower morbidity and mortality has become recognized in many fields.6 We therefore sought to determine whether the results from the Intergroup trial were consistent with the experience at our center.
MATERIALS AND METHODS
The records of patients who underwent pneumonectomy at the Brigham and Women's Hospital (BWH) between 1995 and 2005 were reviewed to identify all patients who had received neoadjuvant therapy that included RT before surgery. Clinical parameters and outcome data were obtained. Seventy-three consecutive patients who underwent preoperative therapy followed by pneumonectomy were identified for this study. Patients were staged using TNM staging according to the 5th edition of the American Joint Committee on Cancer (AJCC).
Patients were chosen for this analysis if they received definitive RT (defined as >4 weeks) followed by pneumonectomy at BWH. Patients with previous lung resections, prior RT, recurrent cancer, or other uncontrolled malignancies were excluded. All patients had stage III NSCLC with 2 exceptions. A limited number of patients (n = 4) had stage IV NSCLC with solitary resected brain metastasis. In addition, 6 patients with tumors adjacent or potentially invading the mediastinum that were believed to require preoperative therapy to render their disease resectable and found at the time of surgery to have T3N0, stage II NSCLC were also included.
RT was performed at many institutions with a wide variety of techniques. A majority of patients underwent computed tomography (CT) simulation with 3-dimensional conformal treatment planning. In general, the field arrangement was anteroposterior/posteroanterior fields followed by an off-cord oblique cone-down after 46 gray (Gy) with the final cone down to 54 Gy. Beam energies varied from 6 to 15 megavolts. The majority of patients had elective lymph node regions included in the primary treatment fields; these fields did not include the supraclavicular fossa, unless macroscopically involved. The dose was calculated without heterogeneity corrections in all but the most recent patients.
Follow-up and Statistical Analysis
Overall survival was measured from the time of surgery to the date of death and was estimated by the method of Kaplan and Meier. Patients who had not died were censored at their date of last follow-up. The log-rank test was used to compare overall survival between patient groups. A Fisher exact test was used to determine the association of patient and treatment characteristics with treatment-related mortality. Factors that were found to be significant at the P = .200 level in the univariate analysis of each outcome were analyzed jointly as independent variables in a multivariate model. The relative risk of death associated with a patient group was based on the hazards ratio (HR) estimated by the proportional hazards model in the multivariate analysis. All P values were based on 2-sided hypothesis tests. Computations were performed using SAS 9.1 software (SAS Institute, Cary, NC).
From 1995 to 2005, 73 patients underwent pneumonectomy at BWH after preoperative therapy. The median age of the patients was 62 years (range, 37–78 years), and 43 patients were male (59%). Thirty-seven patients (51%) (37 of 73) had stage IIIA NSCLC and 27 (37%) had stage IIIB NSCLC. Six patients (8%) had stage IIB disease and 4 patients had stage IV because of a solitary resected brain metastasis (Table 1). Forty-one percent of patients had pretreatment forced expiratory volume in 1 second (FEV1) >60% of predicted.
|Characteristic||No. of patients|
|Sex||Male: 43 (59%)|
|Female: 30 (41%)|
|Median age (range), y||62 (37–78)|
|Histology||Squamous cell: 38 (52%)|
|Adenocarcinoma: 33 (45%)|
|Nonsmall cell unspecified: 2 (3%)|
|AJCC stage||IIIA: 37 (51%)|
|IIIB: 27 (37%)|
|IIB: 6 (8%)|
|IV: 3 (4%)|
|Race||White: 69 (95%)|
|Hispanic: 3 (4%)|
|Other: 1 (1%)|
|Laterality||Left: 45 (62%)|
|Right: 28 (38%)|
|Pneumonectomy||Simple: 45 (62%)|
|Complex: 28 (38%)|
|Chemotherapy||Carboplatin/paclitaxel: 44 (60%)|
|Cisplatin/etoposide: 15 (21%) or nontaxane|
|Other regimens: 10 (14%)|
|No chemotherapy: 4 (5%)|
|Radiotherapy dose, Gy||Median, 54 (range, 40–66)|
All patients received preoperative RT. The median dose was 54 Gy (range, 40–66 Gy). Sixty-nine patients (95%) received preoperative chemotherapy as well. The median number of cycles of chemotherapy was 3 (range, 1–8). The majority of the patients received the combination of carboplatin and paclitaxel (44; 60%) whereas 15 patients (21%) received the combination of cisplatin and etoposide. The median time from the completion of RT to surgery was 6 weeks (range, 2–12 weeks). Forty-five patients (62%) underwent left-sided pneumonectomy and 28 (38%) patients underwent a right pneumonectomy. Complex pneumonectomy, defined as intrapericardial pneumonectomy or extended pneumonectomy, was performed in 28 patients (38%), whereas 45 patients (62%) underwent standard pneumonectomy. The treatment details are summarized in Table 1.
The median follow-up of all 25 living patients was 28 months (range, 6–108 months). The overall median survival was 23 months (95% confidence interval [95% CI], 17–30 months). The 1-year and 2-year survival rates were 70% and 49%, respectively (Fig. 1). Univariate analysis for survival revealed that none of the clinical variables analyzed reached statistical significance, although there was a trend toward improved survival for patients who received RT doses >50 Gy (P = .079). This trend was also noted on multivariate analysis (P = .076; HR of 1.8) (Table 2) (Fig. 2).
|Median, months (95% CI)||1-Year||2-Year||3-Year||P|
|All patients||23.0 (17.2–30.0)||70%||49%||29%|
|Squamous cell||23.0 (17.2–35.4)||70%||49%||31%|
|Platinum plus taxane||27.8 (17.0–30.2)||71%||53%||28%|
|Platinum plus etoposide (or other nontaxane)||25.9 (14.4–60.1)||80%||51%||44%|
|Type of surgery||.45|
|Complex pneumonectomy||19.1 (8.8–30.2)||65%||41%||29%|
|Preoperative XRT dose, cGy||.08|
|FEV1 % predicted||.29|
Four patients died within the first 30 days after surgery, for a 30-day mortality rate of 6%. One patient died of a combination of pneumonia and cardiac failure. Three additional patients died of acute respiratory distress syndrome (ARDS). Furthermore, 3 other patients died after 30 days but within 100 days after surgery for an overall treatment-related mortality rate of 10%. One patient developed a fatal pulmonary embolism, 1 patient had ARDS, and 1 patient died from pneumonia after complications of a bronchopleural fistula. On univariate analysis, only a right-sided pneumonectomy was found to be a borderline significant predictor of treatment-related mortality (P = .099). In particular, neither complex pneumonectomies, RT dose, histology, nor type of chemotherapy were found to be correlated with treatment-related mortality (Table 3). Aside from the 7 patients who died from treatment-related mortality, an additional 8 patients (11%) had complications possibly related to their treatment. These included brochopleural fistula (1 patient), atrial fibrillation (1 patient), pneumonia (2 patients), respiratory failure (1 patient), wound infection (1 patient), wound dehiscence (1 patient), and vocal cord paralysis (1 patient). All these complications were successfully treated and did not appear to affect the patient's survival. In all of these patients, the cause of death was directly related to disease progression.
|Squamous cell||38 (52%)||8%|
|Platinum plus taxane||44 (59%)||7%|
|Platinum plus etoposide (or other nontaxane)||15 (27%)||7%|
|Type of surgery||1.00|
|Complex pneumonectomy||28 (38%)||11%|
|Preoperative XRT dose, cGy||.69|
|FEV1 % predicted||.25|
The appropriate management of stage III NSCLC continues to be debated. Surgery as a frontline treatment has largely been abandoned for 2 major reasons: inadequate local control and insufficient treatment of systemic disease. It has been previously reported that when surgery is used as an upfront treatment modality in patients with stage IIIA NSCLC, up to 30% of patients will have incomplete resection leaving behind macroscopic disease.7 The 5-year survival rate for patients with incomplete resections is < 5%.7–10 However, even in patients who can undergo complete resections, the risk of systemic disease is significant and the overall survival without adjuvant therapy at 5 years is only 10%.8 Furthermore, it is very difficult to administer adjuvant therapy after major surgery, with patient compliance limiting completion of adjuvant therapy.11 For these reasons, many clinicians prefer neoadjuvant therapy.
Another option for treating locally advanced NSCLC is definitive chemoradiotherapy. Randomized trials have shown an improvement in survival of concurrent chemoradiotherapy over sequential therapy.12–14 Although sequential therapy leads to a median survival of 14 months and a 3-year survival rate of 15% to 20%, these trials have shown that concurrent chemoradiation leads to a 17-month median survival and 25% to 30% 3-year survival rate. Further improvements in outcome have been observed with the addition of consolidation chemotherapy, with the best results demonstrating a 26-month median survival and 3-year survival rate of 37%.4 However, when one examines the patterns of failure throughout these studies, an interesting trend emerges. From the initial sequential studies through the recently published SWOG 9504 described above, the limitation continues to be the locoregional failure rate, which has remained consistent between 47% and 53% and has not altered survival.4, 12, 14, 15 This observation suggests that, although systemic therapy advances can lead to improvements in survival via reduction in distant metastasis, further survival gains may be achieved through enhanced local control. The addition of surgical resection to chemoradiation is a potential option for improving local control.
One of the first major studies to address the addition of surgery to chemoradiation, often called trimodality therapy, for locally advanced NSCLC was SWOG 8805. Patients enrolled in this phase 2 study for stage III NSCLC received 45 Gy and concurrent cisplatin and etoposide followed by surgical resection. This therapy was found to be tolerable, with a 10% treatment-related mortality, and led to an encouraging median survival of 13 months for patients with stage IIIB disease and 17 months for those with stage IIIA disease. The overall 3-year survival rate was 26%.1 On the surface, these numbers appear to be very comparable to those achieved with chemoradiotherapy without surgery, as presented above. However, local failure in the patients who had surgery was only 10% and the combined local failure rate (local + local and distant) decreased to 38%, an improvement compared with the studies of chemoradiation alone. In addition to these promising results, an intergroup trial testing chemoradiation followed by surgery for pancoast tumors (all N0 or N1) was completed that demonstrated a 55% 2-year survival rate with only 2.7% treatment-related mortality.16 On the basis of these encouraging results, a North American Intergroup study was mounted to compare chemoradiotherapy with trimodality therapy in patients with stage III, N2-positive NSCLC.
From 1994 to 2001, 429 patients were enrolled in a phase 3 randomized trial comparing chemoradiotherapy with 61 Gy versus 45 Gy of RT combined with chemotherapy followed by surgical resection. The median survival for both treatment arms was 23.6 months for the trimodality arm and 22.2 months for the chemoradiation arm (P not significant). However, the PFS and local failure rate was statistically better for the surgical arm compared with chemoradiation alone (median PFS of 12.5 months vs 10.5 months [P = .01]; local failure only 10% vs 22% [P = .02]).4 The overall survival of the surgical arm was limited by early treatment-induced mortality. Sixteen patients (7.9%) had treatment-related mortality in the surgical group compared with 4 patients in the chemoradiotherapy arm (2.1%). This was particularly striking in the subgroup of patients who underwent pneumonectomy. Fourteen of 54 patients (26% of the pneumonectomy subgroup) had treatment-related mortality after pneumonectomy. Nearly all of these (11 of 14 patients; 79%) were because of respiratory causes or ARDS. These results led many clinicians to conclude that pneumonectomy after induction chemoradiation therapy should be abandoned or severely limited. The high mortality rate in the Intergroup trial prompted a subsequent trial comparing trimodality therapy, as was done in Intergroup 0139, to induction chemotherapy alone (ie, without RT) followed by surgery in an attempt to reduce the toxicity of chemoradiation.
In our retrospective review, we sought to determine whether the results from the Intergroup study were consistent with our own experience. As we have shown, the overall treatment-related mortality rate from our series was 6% at 30 days (standard benchmark of operative mortality) and 10% at 100 days. This mortality is 2.5–fold to 4-fold lower than the 26% mortality reported from the Intergroup trial and consistent or slightly higher than other reports of pneumonectomy alone or chemotherapy followed by pneumonectomy (Table 4).2, 3, 17, 18 The results from this study of a 6% 30-day mortality rate are comparable to our own published data showing a 30-day mortality rate of 4% after pneumonectomy without induction therapy.19 In addition, only 4 patients in the current series developed fatal pulmonary complications compared with the 11 patients who died as a result of ARDS (21%) in the Intergroup study. This is significant because one would expect that if RT increased toxicity from surgical resection it would be because of decreased pulmonary reserve. We did find that right pneumonectomy was a higher risk procedure than left pneumonectomy, as previously reported.
|Study (No. of patients)||Chemotherapy||Radiotherapy dose, Gy||Treatment mortality rate|
|Albain 19951 (41)||Cisplatin/etoposide||45 Gy||15% (6/41)|
|Albain 19951 (54)||Cisplatin/etoposide||45 Gy||26% (14/54)|
|Bedini 20032 (11)||Daily cisplatin||RT to 50–60 Gy||45% (5/11)|
|Bernard 20013 (24)||Details not given||Details not given||Overall 7% CT/RT rate not defined|
|Cyjon 20024 (8)||Cisplatin daily||45 Gy||38% (3/8)|
|Daly 20065 (30)||Cisplatin/etoposide||59.4 Gy||13% (4/30)|
|Pezzetta 20056 (19)||Cisplatin/docetaxel||1.6 Gy twice daily for 5 d every 2 wk||0% (0/19)|
|Sonett 20047 (29)||Varied platinum-based chemotherapies||Once daily >59 Gy||0% (0/29)|
|Takeda 20068 (19)||Cisplatin-based chemotherapy||RT, median 41.5 Gy||11% (2/19)|
|Vora 20009 (9)||Cisplatin/etoposide||59.4 Gy||0% (0/9)|
|Current study (73)||Cisplatin-based chemotherapy||RT once daily, median 54 Gy||10% (7/73)|
An obvious caveat to the results we have presented is the inherent selection bias in our retrospective series. Although we included all patients who underwent preoperative RT with or without chemotherapy followed by pneumonectomy over a 10-year time period, we have not reported on those patients who either developed progressive disease during induction therapy or those who developed toxicity during induction therapy and therefore did not go on to receive surgical therapy. In addition, those patients treated with preoperative chemotherapy (without RT) followed by surgery are herein excluded. However, patients with progressive disease and toxicity were also excluded from the surgical arm of the Intergroup trial and other efforts to investigate trimodality therapy.
It has been well documented for many complex surgical procedures including pancreaticoduodenectomy, esophagectomy, and liver resection that the volume of both the surgeon and the hospital has a significant impact on mortality.6 However, there has been some controversy regarding whether the same data exist for pneumonectomy. In the study by Begg et al.6 no difference was observed in mortality based on hospital volume for pneumonectomies. However, other reports have identified both the hospital volume as well as the specialization of the surgeon in only noncardiac thoracic surgery as important prognostic factors,20–22 as well as the existence of specialized teams and postoperative monitoring and treatment units.23 Our institution and its support staff have specific expertise in pneumonectomies because of our high volume of extrapleural pneumonectomies for mesothelioma (90 performed in 2006) and have published improvements in the surgical techniques of these patients.5, 24, 25 Whether this experience in mesothelioma translates to an improved outcome in the management of pneumonectomy patients after chemoradiation in NSCLC is unknown.
To our knowledge, the current study is the largest series reported to date of patients undergoing neoadjuvant therapy followed by pneumonectomy. However, other smaller series from selected institutions have also demonstrated an acceptable mortality of pneumonectomy after chemoradiotherapy. The University of Maryland group reported a series of patients who received high-dose RT (>59 Gy) with concurrent chemotherapy followed by surgery. They reported no postoperative deaths in the 11 patients who underwent pneumonectomy.26 Another reported small series from the M. D. Anderson Cancer Center of patients receiving chemoradiotherapy before surgery noted a major pulmonary event in 18% of patients with no mortality.27 Finally, Choi et al.28 reported a series of 30 of patients treated with 60 Gy of radiation and concurrent chemotherapy followed by pneumonectomy at the New England Medical Center and Boston Medical Center with a 13% mortality. The above reports are all single-institution, nonprospective series with surgery performed by dedicated thoracic surgeons. However, it is very likely that there is an inherent selection bias that is present in all of these series, including our own.
Another interesting note that arose from the results of the current study was a suggestion of improved survival with higher doses of preoperative RT. In the Intergroup study, a subgroup analysis indicated that, in patients in the trimodality arm, those patients who achieved lymph node downstaging to pT0 had a remarkable 5-year survival rate of 41%.4 Similar findings by our own group as well as others have indicated that lymph node downstaging after induction chemoradiation is the most significant predictor of overall survival.29 In the current study, not all patients had N2-positive lymph nodes documented at the time of diagnosis, such that a lymph node downstaging rate could not be accurately assessed. However, the hypothesis that intensified induction chemoradiation is safe and may improve lymph node downstaging before surgery is an intriguing concept that should invite further study.
In conclusion, the results of the current single-institution experience with a dedicated thoracic surgical team and systematized approach to patient care have shown that preoperative chemoradiotherapy followed by pneumonectomy results in an acceptable mortality rate in a larger cohort than that of the Intergroup study. Trimodality therapy, including pneumonectomy, should remain an option for the treatment of patients with stage III NSCLC. We also suggest that this approach should be reserved for specialized high-volume centers of general thoracic surgery in clinical practice and in future trials.
- 4Phase III comparison of concurrent chemotherapy plus radiotherapy (CT/RT) and CT/RT followed by surgical resection for stage IIIA(pN2) non-small cell lung cancer (NSCLC): initial results from intergroup trial 0139 (RTOG 93–09). Proc Am Soc Clin Oncol. 2003; 22. Abstract 2497., , , et al.
- 12Phase III comparison of sequential vs concurrent chemo radiation for pts. with unresected stage III non-small cell lung cancer (NSCLC): initial report of Radiation Therapy Oncology Group (RTOG) 94–10. Proc of Am Soc Clin Oncol (ASCO), New Orleans, LA. J Clin Oncol. 2000; 484a. Abstract 1891., , , et al.
- 14Randomized phase III trial of sequential chemoradiotherapy compared with concurrent chemoradiotherapy in locally advanced non- small-cell lung cancer: Groupe Lyon-Saint-Etienne d'Oncologie Thoracique-Groupe Francais de Pneumo-Cancerologie NPC 95–01 Study. J Clin Oncol. 2005; 23: 5910–5917., , , et al.