Pulmonary Resection for Second Lung Cancer After Pneumonectomy ...

Pulmonary Resection for Second Lung Cancer After Pneumonectomy: A Population-Based Study Adil Ayub, MD,* Sadiq S. Rehmani, MD,* Adnan M. Al-Ayoubi, MD, PhD, Wissam Raad, MD, Raja M. Flores, MD, and Faiz Y. Bhora, MD Department of Thoracic Surgery, Icahn School of Medicine, Mount Sinai Health System, New York, New York

Background. Pulmonary resection for a second lung cancer after pneumonectomy is generally considered to be at prohibitive risk. Using a population-based database, we examined treatment patterns and survival in patients who underwent pulmonary resection after pneumonectomy for lung cancer. Methods. We queried the Surveillance, Epidemiology, and End Results (SEER) database (1988–2012) to identify patients who underwent pneumonectomy and subsequently experienced contralateral non-small cell lung cancer (NSCLC). Multivariate logistic regression was performed to identify the factors associated with the receipt of surgical resection. Survival was estimated with the Kaplan-Meier method. Results. Of 13,370 patients who underwent pneumonectomy, 402 (3.0%) experienced subsequent contralateral NSCLC, and 170 (42%) met the selection criteria. Surgical resection was performed in 63 (37.1%) cases (sublobar n [ 56, lobectomy, n [ 7). Patients with stage I/II disease and tumor size 2 cm or smaller were more likely to

undergo surgical procedures. The 1-month and 3-month mortality after resection was 11.1% (sublobar resection 10.7%, lobectomy 14.3%) and 12.7% (sublobar 12.5%, lobectomy 14.3%), respectively. The overall 1-year and 3-year survival after surgical resection was 79% and 54%, respectively. The patients who underwent sublobar resection had higher median overall survival than did those who underwent lobectomy (42 vs 18 months). Similarly, median survival after resection for metachronous tumors was higher than after resection for metastatic cancers (40 vs 28 months). Conclusions. On the basis of our analysis of the SEER database, sublobar resection can be performed in selected patients with small tumors (£2 cm) and early-stage disease (stage I/II). Although perioperative mortality is significant, the favorable 1-year and 3-year survival may justify the role of an additional procedure on the single lung.

P

tumor characteristics (ie, tumor size and disease stage) and outcomes are sparsely reported. In addition, of all patients in whom a post-pneumonectomy lung cancer developed, the characteristics of the subgroup of patients who underwent an additional resection have not been previously described. Therefore, using a populationbased database, we sought to examine patients who experienced additional lung cancer after pneumonectomy, identify factors associated with the receipt of surgical resection, and evaluate survival outcomes after resection.

atients successfully treated for lung cancer remain at risk for the development of a new primary lung tumor or recurrence from previous cancer [1, 2]. Surgical resection can provide a chance of cure; however, additional pulmonary resection is generally considered to be at prohibitive risk after a prior pneumonectomy, mainly because of technical, oncologic, and physiologic reasons. Although evidence from small retrospective series [3–9] has suggested limited additional resection to be a worthwhile procedure in selected cases, resection is generally not offered to these patients even for earlystage resectable disease in the belief that prior pneumonectomy is a relative contraindication to further resection. Furthermore, even though information on physiologic and operative factors has been described before, detailed

Accepted for publication April 17, 2017.

(Ann Thorac Surg 2017;-:-–-) Ó 2017 by The Society of Thoracic Surgeons

Material and Methods Study Population A retrospective cohort was identified by use of the Surveillance, Epidemiology, and End Results (SEER)

*Drs Ayub and Rehmani contributed equally to this work. Presented at the Fifty-third Annual Meeting of The Society of Thoracic Surgeons, Houston, TX, Jan 21–25, 2017. Address correspondence to Dr Bhora, Department of Thoracic Surgery, Ste 2B, 1000 Tenth Ave, New York, NY 10019; email: faiz.bhora@ mountsinai.org.

Ó 2017 by The Society of Thoracic Surgeons Published by Elsevier Inc.

The Supplemental Tables can be viewed in the online version of this article http://dx.doi.org/10.1016/ j.athoracsur.2017.04.043 on http://www.annalsthoracic surgery.org.

0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2017.04.043

2

AYUB ET AL LUNG RESECTION AFTER PNEUMONECTOMY

database, which is sponsored by the National Cancer Institute (www.seer.cancer.gov). We queried 18 SEER registries from 1988 to 2012 (November 2014 submission) and identified patients who underwent pneumonectomy for primary lung cancer (ICD-O-3 site codes C34.0–C34.9) and subsequently experienced a second lung malignancy. Pneumonectomy procedure codes included “site-specific surgery” codes (50, 60, and 70) for cases from 1983 to 1997 and “surgery primary site” codes (40, 50-56, 65, 66, and 70) from 1998 onward. Patients with only complete information on tumor laterality, stage, size, American College of Chest Physicians (ACCP) classification, and survival were included in the study. With regard to second cancer, patients who received ablative therapy (n ¼ 11) or adjuvant radiation therapy (RT) (n ¼ 14) were excluded. In addition, patients with small-cell carcinoma (ICD-O code 8041–45, n ¼ 34) were not included in the study because the primary treatment for such cases is nonsurgical. The patient selection algorithm is described in Figure 1.

Covariate Selection Patient demographics, tumor characteristics, treatment status, and survival were evaluated. Tumor size was Fig 1. Patient selection algorithm. (ACCP ¼ American College of Chest Physicians; NSCLC ¼ non-small cell lung cancer; RT ¼ radiation therapy; SEER ¼ Surveillance, Epidemiology, and End Results.)

Ann Thorac Surg 2017;-:-–-

transformed into a categoric variable with two size groups: 2 cm or smaller, and over 2 cm. Cancers were staged according to the American Joint Commission on Cancer 6th edition [10]. For patients whose cancers were diagnosed after 2004, tumor–node–metastasis (TNM) stage was directly extracted from SEER. For cases diagnosed between 1988 and 2003, TNM stage was manually recoded according to SEER variables (EOD 10 size, EOD 10 nodes, EOD 10 extension). Tumors were further classified as metachronous or metastatic based on the ACCP guidelines [11]. All histologically distinct second tumors were categorized as metachronous. For histologically similar lesions, the second tumors were classified as metachronous if the latency period was at least 4 years with no evidence of systemic metastases. Latency was defined as the interval in number of years between the occurrences of the first and second cancers. To classify cancer-directed surgical procedures, “sitespecific surgery codes” (20 ¼ sublobar resection and 30, 40 ¼ lobectomy) and “surgery of primary site” codes (21, 22 ¼ sublobar resection and 30, 33 ¼ lobectomy) were used.



Statistical Analysis Statistical analysis was performed with the Statistical Package for the Social Sciences Software (SPSS version 17, IBM Corp, Armonk, NY). Unadjusted comparisons between groups were performed wth the c2 test for categoric variables and Student’s t test for continuous variables. Multivariate logistic regression was performed to identify factors associated with the receipt of surgical resection. Unadjusted overall survival (OS) was compared by Kaplan-Meier analysis. A two-sided p value of less than 0.05 was considered statistically significant.

Results Of 13,370 patients who underwent pneumonectomy, 402 (3.0%) experienced subsequent contralateral non-small cell lung cancer (NSCLC), and 170 (42%) met the selection criteria (Fig 1). Sixty-three patients (37.1%) underwent surgical resection alone, 69 (40.6%) received radiation therapy (RT), and 38 (22.4%) received neither resection nor RT. The characteristics of all study patients are described in Table 1. The differences in measured Table 1. Characteristics of Patients in Whom Second Lung Cancer Developed After Pneumonectomy (n ¼ 170) Variable

n (%)

Age at pneumonectomy, mean (SD) Age at contralateral cancer diagnosis, mean (SD) Male sex White race Latency in years, mean (SD) Tumor size in cm, mean (SD) Laterality of contralateral cancer Right Left Stage I II III IV Histology Adenocarcinoma Squamous cell carcinoma Othera Typeb Metachronous Metastatic Treatment Operation alone RT alone No operation, no RT

61.8 67.5 106 151 5.7 2.7

(9.0) (9.6) (62.4%) (88.8%) (4.8) (1.9)

105 (61.8%) 65 (38.2%) 111 10 25 24

(65.3%) (5.9%) (14.7%) (14.1%)

70 (41.2%) 83 (48.8%) 17 (10.0%) 106 (62.4%) 64 (37.6%) 63 (37.1%) 69 (40.6%) 38 (22.4%)

a

Includes large cell carcinoma, carcinoid tumor, neuroendocrine tumor, adenosquamous carcinoma, NSCLC NOS, and mixed malignant b Defined according to American College of Chest tumors. Physicians guidelines [10]. NOS ¼ not otherwise specified; cancer; RT ¼ radiation therapy.

NSCLS ¼ non-small cell lung

3

covariates and outcomes between patients with complete and incomplete information are shown in Supplemental Tables S1 and S2. Among patients who underwent surgical resection (n ¼ 63), 56 received sublobar resection and 7 underwent lobectomy (right middle lobe ¼ 1, right upper lobe ¼ 2, right lower lobe ¼ 1, left upper lobe ¼ 1, left lower lobe ¼ 1, right unspecified ¼ 1). The mean age at diagnosis for these patients was 65.1 years, and the average latency period was 3.48 years. Adenocarcinoma was the most common histologic subtype (58.7%), the mean tumor size was 1.75 cm, and most patients had stage I/II disease (92.1%). The tumors in the surgical group were predominantly metachronous (58.7%) and right-sided (57%). On univariate analysis, younger age at diagnosis of second cancer (p ¼ 0.011), shorter latency period (p < 0.001), tumor size 2 cm or smaller (p < 0.001), and earlier disease stage (stages I and II, p < 0.001) were significantly associated with the receipt of surgical resection (Table 2). A comparison of crude odds ratios before and after the exclusion of patients with missing information is provided in Supplemental Table S3. In a multivariate model, patients with stage I/II disease (p ¼ 0.001) or tumor size 2 cm or smaller (p ¼ 0.002) were more likely to undergo surgical resection, whereas increasing latency interval was associated with a decreased likelihood of operation (p ¼ 0.001) (Table 3). The median OS was 39, 20, and 6 months for patients receiving resection alone, RT alone, and no treatment, respectively. Patients who underwent resection had higher 1-year and 3-year survival (79% and 52%, respectively) than did patients receiving RT (70% and 33%, respectively) and no treatment (32% and 9%). For patients who underwent surgical procedures (n ¼ 63), the overall rate of 1-month mortality and 3-month mortality was 11.1% and 12.7%. The 1-month and 3-month mortality rate for patients receiving lobectomy was 14.3% compared with 10.7% and 12.5% for sublobar resection (Table 4). The median OS after resection for metachronous second tumors was 40 months compared with 28 months for metastatic cancers (p ¼ 0.754). Patients who underwent sublobar resection had a higher OS than did those who underwent lobectomy (median, 42 months vs 18 months), but this was not statistically significant (p ¼ 0.244) (Table 4).

Comment Currently, limited data are available on the surgical treatment of patients in whom a second lung cancer develops after pneumonectomy. To date, approximately 100 surgically treated cases have been described in the literature (Table 5). In summary, an operative mortality ranging from 6.2% to 28.6% and 3-year survival rates from 20% to 61% have been reported previously [3–8, 12]. Metastatic tumors, N2 disease, and a second operation within 2 years of pneumonectomy have been found to be associated with poor prognosis [3, 8]. With the use of the SEER database, this study reports the largest cohort of patients with contralateral NSCLC

4



Table 2. Univariate Analysis of Factors Associated With Receipt of Surgical Resection (n ¼ 170) Variable

Received Surgical Procedure (n ¼ 63)

Mean age at contralateral cancer diagnosis (SD) Male sex White race Latency, years (SD) Size, cm 2 >2 Stage I/II III/IV Type Metachronous Metastatic Radiotherapy after pneumonectomy Yes No

65.1 35 59 3.5

(8.7) (55.6%) (93.7%) (3.4)

No Surgical Procedure (n ¼ 107) 68.9 71 92 7.0

p-value

(9.9) (66.4%) (86.0%) (5.0)

0.011 0.160 0.125