Published ahead of print on June 23, 2006, doi:10.1164/rccm.200602-160OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200602-160OC
Diagnostic Percutaneous Transthoracic Needle Biopsy Does Not Affect Survival in Stage I Lung CancerDivisions of General Internal Medicine and Pulmonary, Critical Care Medicine, and Sleep Medicine, Mount Sinai School of Medicine; Department of Radiology, New York-Presbyterian Hospital-Weill Cornell Medical Center, New York, New York Correspondence and requests for reprints should be addressed to Juan P. Wisnivesky, M.D., M.P.H., Department of Medicine, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1087, New York, NY 10029. E-mail: juan.wisnivesky{at}mssm.edu
Rationale: Lung cancer dissemination has been cited as a potential adverse consequence of diagnostic percutaneous transthoracic needle biopsy (PTNB) of lung nodules. Despite multiple reports in the literature of malignant spread along the needle track, the potential effect of lung cancer dissemination after PTNB on survival is unknown. Objective: To evaluate whether diagnostic PTNB is associated with increased risk of lung cancer death. Methods: This study included 8,607 cases of surgically resected stage I non–small cell lung cancer diagnosed between 1991 and 1999 from the Surveillance, Epidemiology, and End Results (SEER) registry linked to Medicare records. Overall and lung cancer–specific survival of patients who had and did not have PTNB was compared using Kaplan-Meier curves. Stratified survival analyses and Cox regression were used to compare survival with adjustment for potential confounders. Results: Approximately 36% of patients underwent diagnostic PTNB. Overall and lung cancer–specific survival did not differ in patients that underwent PTNB as part of their cancer diagnostic work-up and those who did not (p = 0.57 and 0.46, respectively). In stratified and multivariate analysis, PTNB was not associated with an increased risk of death after controlling for age, race, income, access to care, comorbidities, tumor histology and size, and type of treatment received. Conclusions: In this large national sample, preoperative PTNB was not associated with increased risk of death. These data suggest that PTNB can be safely used for the work-up of pulmonary nodules when there is a suspicion of lung cancer.
Key Words: diagnosis • lung cancer • needle biopsy • survival Percutaneous transthoracic needle biopsy (PTNB) of the lung is a well-established and commonly used method for evaluation of pulmonary nodules, particularly if malignancy is suspected. PTNB is generally regarded as a safe procedure with limited associated morbidity (1). Pneumothorax remains the most frequent complication and a tube thoracostomy is occasionally required for treatment (2, 3). Fatal complications due to systemic air embolism, hemorrhage, or pericardial tamponade have been reported, but these are rare. Chest wall implantation, tumor spread through the needle tract, and dissemination of cancer cells into the pleural space, although less common, are potentially serious complications of this procedure (4–22). Malignant implantation after PTNB has been reported in the literature by multiple investigators and dissemination has been demonstrated in experimental models (4–6, 8–17, 22, 23). As a consequence, it has been suggested that patients with a solitary pulmonary nodule who are good operative candidates and amenable to a potentially curative resection would benefit from direct surgical exploration without preoperative percutaneous biopsy (4, 18, 24). Despite this theoretical limitation, PTNB is frequently performed in patients with solitary pulmonary nodules when there is a suspicious of lung cancer. The clinical implications of lung cancer dissemination after PTNB on survival, however, has yet to be defined. The purpose of our study was to evaluate whether PTNB is associated with increased risk of death by comparing survival of patients with surgically resected stage I non–small cell lung cancer who that underwent PTNB as part of the diagnostic work-up of their malignancy with a concurrent cohort of patients who did not undergo this procedure.
Cases were selected from the Surveillance, Epidemiology, and End Results (SEER) registry linked to Medicare records (25). We identified all patients older than 65 yr with primary non–small cell lung cancer diagnosed between 1991 and 1999 who were not identified at autopsy or from death certificates (26). Of these, 9,471 were resected stage I cases (according to American Joint Cancer Committee criteria) (27). We excluded patients enrolled in Health Maintenance Organizations or who were not covered by Medicare Part B, leaving a cohort of 8,607 patients with stage I disease. We identified PTNB use from Medicare claims (inpatient, outpatient, and physician files). All claims within 3 mo of cancer diagnosis were examined for codes indicating PTNB use (Current Procedural Terminology [CPT] codes 32405, 76003, and 76360, and the International Classification of Diseases, 9th revision [ICD-9], procedure code 3326) (28). We used a similar approach to determine fiberoptic bronchoscopy use (CPT codes 31622–31625, 31628, 31629, 31632, 31633, and 32602, and ICD-9 procedure codes 3322–3324 and 3327). Baseline demographic data were obtained from SEER. All patients in the study were insured by Medicare. Socioeconomic status was estimated based on the median income for the zip code of the patient's residence. We linked SEER-Medicare data with the Area Resource File to obtain information on physician availability in the patient's area of residence (29). To evaluate the burden of comorbid diseases, we calculated the Charlson comorbidity index using data from Medicare inpatient, outpatient, and physician records (30, 31). Cases were classified as resected if SEER data indicated that a surgical procedure (segmentectomy, wedge resection, lobectomy, or pneumonectomy) had been performed. Postoperative radiation therapy was ascertained from SEER data. Patients were considered as treated with adjuvant chemotherapy if data from Medicare claims indicated that the individual received chemotherapy within 3 mo of diagnosis (32). Survival was determined as the interval from cancer diagnosis to the SEER date of death. Those surviving past December 31, 1999, were classified as censored (alive at the end of follow-up). The cause of death was coded according to SEER, which uses state death certificates as a primary source.
Differences in the baseline characteristics of patients who underwent or did not undergo PTNB were evaluated using the Based on the number of deaths observed among patients in the cohort, we estimated that the study had an 80% power to detect an increased hazard of death after PTNB of 1.09 at a 0.05 significance level. Analyses were performed using SAS software (SAS Institute, Cary, NC). The study was exempt from institutional review board evaluation.
Patient Characteristics A total of 8,607 patients were included in the analysis; 3,082 (36%) underwent PTNB as part of the diagnostic work-up of lung cancer and 5,525 (64%) did not. Preoperative fiberoptic bronchoscopy was used in 4,506 (52%) patients for cancer diagnosis; approximately half of these procedures included a biopsy. The median follow-up time for the entire cohort was 30 mo. The clinicopathologic characteristics of study patients are reported in Table 1. PTNB was more frequently performed in women, elderly patients, adenocarcinoma cases, patients with tumors greater than 3 cm, and patients with multiple comorbidities. There were no significant differences in the proportion of patients undergoing PTNB relative to median estimated income, number of physicians in geographic area, surgical treatment (limited vs. full resection), and radiation therapy use.
Unadjusted and Stratified Analysis Overall and lung cancer–specific survival was similar among patients who underwent PTNB as part of their cancer work-up and those who did not (p = 0.57 and 0.46, respectively; Figure 1). Median overall survival was 52 mo (95% confidence interval [CI], 48–54 mo) for both groups. The 5-yr overall survival rate for patients who had PTNB was 44% (95% CI, 41–46%) compared with 45% (95% CI, 43–46%) for those patients who did not undergo this test. Similarly, there was no significant difference among groups in median lung cancer–specific survival or 5-yr cancer survival rates.
Stratified analysis showed no significant differences in overall and lung cancer–specific survival among patients who had and did not have PTNB across several relevant subgroups (Table 2). When stratifying patients according to the type of diagnostic work-up, patients undergoing PTNB had similar survival among those who had a bronchoscopy (with or without biopsy) and those who went directly to surgery.
Adjusted Analysis We performed Cox regression analysis to evaluate for possible differences in survival among patients who had or did not have preoperative PTNB after adjusting for age, race and ethnicity, median income, physician availability, tumor histology and size, comorbidity score, type of surgery (limited vs. full resection), and whether the patient received postoperative radiation or chemotherapy. As seen in Table 3, PTNB was not associated with an increased risk of death after controlling for imbalances in baseline characteristics (hazard ratio [HR], 0.99; 95% CI, 0.93–1.06). Similarly, the adjusted risk of lung cancer mortality was not different among patients who had and did not have PTNB as part of their diagnostic work-up (HR, 0.96; 95% CI, 0.88–1.04). Patients who underwent and did not undergo PTNB also had similar overall and lung cancer–specific survival when the analysis was restricted to patients with adenocarcinoma or to the subgroup of patients that did not have bronchoscopy (data not shown).
PTNB is a well-established technique for cytologic diagnosis of lung cancer. The primary objective of this evaluation is to accurately identify malignant nodules for surgical excision and to spare patients with benign nodules unnecessary surgical interventions. Although generally regarded as a safe procedure, one of the major arguments against PTNB is the possibility of dissemination of potentially curable cancers (4, 18, 24). In this nationally representative sample of 8,607 patients with surgically resected stage I non–small cell lung cancer, PTNB was not associated with increased long-term mortality risk. These data suggest that PTNB can be safely performed even for patients with clinically localized cancers. Lung cancer dissemination after PTNB has been reported by multiple investigators (4–18). Larger series have estimated that the incidence of this complication is less than 1%; however, the total number of biopsies in these series included patients with benign disease and patients with multiple procedures whose needle biopsies had been nondiagnostic (6, 14, 19, 34). In addition, case reports have been limited to patients who have evidence of disease in the chest wall. Thus, current estimates do not include cases of pleural dissemination, subclinical disease, or patients who were lost to follow-up or died of unrelated diseases before local spread was identified. Tumor dissemination appears to be more frequent with large bore needles or when the procedure involves several passes through the cancer; however, tumor-seeding has been also described when fine needles are used (8, 11, 14, 15, 22, 35). The mechanism of implantation metastasis is still unclear. Although Sawabata and colleagues showed tumor cells remaining in the pleural space after PTNB in a tumor model, the clinical implication of these findings is unknown (23). Struve-Christensen showed that malignant cells could be found in a high percentage of needle biopsy tracks but this produced a clinically significant lesion in a minority of cases (17). Given the large number of patients in the study and the long-term survival data, the study was powered to find relatively small differences in survival among patients undergoing PTNB. In addition, Kaplan-Meier analysis showed almost complete superposition of the survival curves of patients who had and did not have PTNB even more than 5 yr after diagnosis. The observed difference in the survival curves during the last year of follow-up was not statistically significant and likely related to the small number of patients alive at this time in both groups. As the estimated time to local recurrence after PTNB is 6 to 26 mo (6), these data strongly suggest that lung cancer dissemination after PTNB is a rare event having no significant impact on lung cancer curability or clinical outcomes. In addition, the analyses stratified by type of diagnostic test (bronchoscopy, with or without biopsy, or those who went directly to surgery) suggest that neither bronchoscopic biopsy affects survival. Several strengths and limitations should be noted. The generalizability of our findings should be strong. The SEER-Medicare registry contains population-based data and therefore is less affected by referral patterns that might be associated with hospital-based case series. The availability of Medicare claims data allowed us not only to identify patients who had PTNB or bronchoscopy but to control for socioeconomic status, health insurance, and comorbid conditions, all potential confounders of the impact of PTNB on survival. In addition, the large number of patients with resected stage I non–small cell lung cancer in the registry allows for precise estimation of survival rates of patients who had and did not have PTNB. Because this was a retrospective study, the selection of patients undergoing PTNB was not random. We observed differences in the baseline characteristics of patients who had and did not have PTNB and although we controlled for known confounding variables, imbalances may have persisted for unobserved variables. The physician's decision to have the patient undergo PTNB or other diagnostic test in actual practice can be quite complex, and may not be subject to quantification in a nonexperimental design. Thus, unmeasured confounders may explain in part the observed lack of association between PTNB and mortality. Despite this limitation, it is unlikely that a large randomized control trial comparing different diagnostic procedures for patients with stage I lung cancer will be conducted in the near future. Thus, in the absence of information from a randomized controlled trial, data from a large, population-based cohort are probably the best source for data to evaluate this question. We focused on Medicare beneficiaries who were 65 yr or older; thus, we could not explore whether these findings apply to younger patients with lung cancer. No data regarding the type or size of needle used for biopsies or the number of passes performed during the procedure is provided in the SEER-Medicare database. Thus, we could not assess whether risk of dissemination varies according to these factors. In summary, this study shows that preoperative PTNB is not associated with increased risk of mortality among patients with localized non–small cell lung cancer. Thus, lung cancer dissemination does not appear to be a common phenomenon among these patients. This information should reassure physicians considering the use of this procedure for the work-up of solitary pulmonary nodules when there is a suspicion of lung cancer.
The authors thank the Applied Research Branch, Division of Cancer Prevention and Population Science, National Cancer Institute; the Office of Information Services, and the Office of Strategic Planning, Center for Medicare and Medicaid Services; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) program tumor registries in the creation of the SEER-Medicare Database. The interpretation and reporting of the data from these agencies are the sole responsibilities of the authors. The authors thank Angela Fahey of IMS, Inc., for expert assistant in dataset construction.
Supported by the Agency for Healthcare Research and Quality (K08 HS013312). Originally Published in Press as DOI: 10.1164/rccm.200602-160OC on June 23, 2006 Conflict of Interest Statement: J.P.W. does not have a financial interest in a commercial entity that has an interest in the subject of this manuscript. C.I.H. does not have a financial interest in a commercial entity that has an interest in the subject of this manuscript. D.F.Y. is a physician consultant to PneumRx and received $5,000; he also owns stock in this company. Received in original form February 3, 2006; accepted in final form June 21, 2006
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2 test. The Kaplan-Meier method was used to estimate survival rates. We compared overall (all-cause mortality) and lung cancer–specific survival for patients who had PTNB and those who did not using the log-rank statistic. To estimate cancer-specific survival, deaths attributed to causes other than lung cancer were censored at the date of death. We assessed the impact of PTNB on patients' outcomes by comparing unadjusted survival rates, by conducting stratified analyses within relevant subgroups, and by determining the effect of PTNB on survival while controlling for important confounders using Cox regression (
