Association of Shorter Time to Recurrence and Recurrence-Free Survival with Transthoracic Lung Biopsy in Stage I Lung Cancer
Article information
Abstract
Purpose
We aim to determine whether preoperative percutaneous needle aspiration or biopsy (PCNA/Bx) increases recurrence risk and reduces survival in stage I lung cancer patients, using a nationwide lung cancer registry.
Materials and Methods
We retrospectively included 3,452 patients diagnosed with stage I lung cancer who underwent curative surgery between 2014 and 2019, as recorded in the Korean Association of Lung Cancer Registry. To balance the characteristics of patients with and without PCNA/Bx, we applied inverse probability of treatment weighting. We used cumulative incidence plots and a weighted subdistribution hazard model to analyze time to recurrence. Recurrence-free survival and overall survival were analyzed using Kaplan-Meier curves and weighted Cox proportional hazard ratio models.
Results
In patients with adenocarcinoma, the use of PCNA/Bx was associated with a 1.9-fold increase (95% confidence interval [CI], 1.5 to 2.4) in the risk of recurrence and a 1.7-fold decrease (95% CI, 1.3 to 2.2) in recurrence-free survival. Subgroup analysis based on pathologic pleural invasion revealed that the risk of recurrence increased when PCNA/Bx was performed, with 2.1-fold (95% CI, 1.5 to 2.8) in patients without pleural invasion and 1.6-fold (95% CI, 1.0 to 2.4) in those with pleural invasion. No association was found between the use of PCNA/Bx and overall survival.
Conclusion
Preoperative PCNA/Bx was associated with increased recurrence risks in stage I adenocarcinoma, regardless of pathologic pleural invasion status. In early lung cancer cases where adenocarcinoma is strongly suspected and curative surgery is feasible, the use of transthoracic biopsy should be approached with caution.
Introduction
Lung cancer is the leading cause of cancer-related mortality worldwide, with an estimated 2.2 million diagnoses and 1.8 million deaths recorded annually [1]. The detection of early-stage lung cancer using computed tomography (CT) screening is effective in reducing lung cancer mortality [2,3], and approximately two-thirds of screen-detected lung cancer cases are located in the lung periphery [4]. Percutaneous needle aspiration or biopsy (PCNA/Bx) is a well-established, less-invasive technique for the histologic diagnosis of peripheral lung cancer. PCNA/Bx offers excellent diagnostic accuracy for lung cancer and is generally safe [5,6]. The most common immediate complications associated with PCNA/Bx are pneumothorax and hemoptysis, which occur in 25% and 5% of cases, respectively, and life-threatening events are rare [7,8].
One of the persistent concerns regarding preoperative PCNA/Bx for lung cancer is the risk of pleural recurrence, a significant long-term complication that necessitates caution when used for early-stage lung cancer patients undergoing curative resection. The relationship between preoperative PCNA/Bx and ipsilateral pleural recurrence has shown inconsistent results across various studies [9-12]. Notably, a recent patient-level meta-analysis that included 2,329 subjects from six original studies indicated a correlation between PCNA/Bx and an increased rate of ipsilateral pleural recurrence in stage I lung cancer [13]. However, these findings were based on a relatively small sample from a limited number of institutions, highlighting the need for further research to confirm their generalizability. This study seeks to evaluate the impact of preoperative PCNA/Bx on recurrence and survival using data from the Korean Association of Lung Cancer Registry (KALC-R), which encompasses a national cohort of lung cancer patients.
Materials and Methods
1. Study population
The KALC-R is a nationwide registry encompassing approximately 10% of all lung cancer cases in Korea [14]. Detailed information about KALC-R is provided in the Supplemental Material. From the KALC-R, we selected patients who were diagnosed with pathologic stage IA or IB lung cancer and underwent curative surgery from 2014 through 2019. The exclusion criteria were as follows: (1) missing data on body mass index (BMI), smoking history, tumor diameter, and pleural invasion; (2) uncertain recurrence status; (3) surgery performed more than 6 months after tissue diagnosis; (4) recurrence within 2 months of diagnosis; and (5) loss to follow-up.
The collected data included patient sex, age, smoking status, BMI, pathologic stage (following the eighth edition of the TNM International Staging System), pathologic tumor size, presence of pleural invasion, histological tumor type, and whether PCNA/Bx was conducted. Information on the timing of diagnostic procedures, operation, recurrence, and death was also obtained, with a follow-up extending until December 31, 2021.
2. Definition of outcomes
The primary outcome of the study was the time to recurrence, while the secondary outcomes included recurrence-free survival and overall survival. The KALC-R recorded the year and month of events such as diagnosis, recurrence, and last follow-up, without specifying exact dates. However, exact dates were provided for deaths.
Time to recurrence was defined as the period between the diagnostic procedure for lung cancer and the recurrence. All dates were set as the 15th of the month, except when recurrence occurred in the month of death, in which case the recurrence date was set as the 1st of the month. Additionally, if no recurrence was observed during follow-up, the time to recurrence was defined as the number of days between the last follow-up and the date of the diagnostic procedure. In cases where death occurred without recurrence, the time to recurrence was calculated as the number of days between the date of the diagnostic procedure and the date of death, with death considered a competing event of recurrence.
Recurrence-free survival was defined as the time from the date of the diagnostic procedure for lung cancer until either recurrence or death, while overall survival was defined as the time from the diagnostic procedure for lung cancer until death. If a patient died during the month of the tissue examination, or if follow-up concluded in the month of the tissue examination without a record of death, the survival duration was recorded as 1 day. If there was no recorded death by the end of follow-up, overall survival was calculated as the number of days from the procedure date to the last follow-up date.
3. Statistical analysis
To address potential imbalances in patient characteristics between those who underwent preoperative PCNA/Bx and those who did not, we utilized propensity score-based inverse probability of treatment weighting (IPTW) [15]. The likelihood of undergoing preoperative PCNA/Bx was calculated using a logistic regression model that included the following baseline characteristics: age, sex, BMI, smoking status, histologic tumor type, presence of pleural invasion, pathologic tumor size, and pathologic tumor stage. We then applied stabilized weights in our analysis [16]. We used descriptive statistics to summarize the baseline characteristics of patients who underwent PCNA/Bx and those who did not. The balance between the two groups was assessed using standardized mean differences (SMD) both before and after the application of IPTW. An absolute SMD greater than 10% was considered indicative of a meaningful imbalance [17].
For the analysis of time to recurrence, cumulative incidence plots were examined. Time to progression was estimated using a weighted form of the Fine-Gray subdistribution hazard ratio (sHR) model, with death before the onset of recurrence considered as a competing event [18]. For the analysis of recurrence-free survival and overall survival, Kaplan-Meier survival curves and a weighted Cox proportional hazard ratio (HR) model were employed.
Factors significant at the 0.1 level in the univariable analysis were considered candidate covariates for the multivariable analysis. The final model was determined through backward elimination. Preoperative diagnostic procedures (whether undergoing PCNA/Bx or not) and age were retained in the final model regardless of their statistical significance, based on the results of previous meta-analyses. Additionally, interaction terms between the diagnostic procedure and covariates were examined in the multivariable analysis. Both statistically and clinically significant interactions were included in the final model.
For the sensitivity analysis, we performed the same analyses without applying IPTW. We assessed the assumption of proportionality using Schoenfeld residuals. An expert statistician conducted all analyses using SAS ver. 9.4 (SAS Institute) and R ver. 4.1.2 (R Foundation for Statistical Computing).
Results
1. Patient demographics
After excluding certain individuals from the initially curated group of 3,633 patients, the study population consisted of 3,452 patients diagnosed with pathologic stage IA or IB lung cancer who underwent curative surgery (Fig. 1). Of these, 1,892 patients (54.8%) were male, with an average age of 64.2±10.0 years. Preoperative PCNA/Bx was performed on 1,183 patients (34.3%). Among the 2,269 patients (65.7%) who did not undergo PCNA/Bx, 379 (16.7%) were diagnosed through bronchoscopic biopsy, eight (0.4%) through sputum cytology, and the remaining 1,882 patients (82.9%) underwent upfront surgical lung biopsy. Those who underwent PCNA/Bx were older on average (SMD, 0.14), had larger tumors (SMD, 0.59), a higher rate of pleural invasion (SMD, 0.31), and a higher pathologic stage (SMD, 0.38). After applying IPTW, the absolute SMDs for all baseline characteristics were within 0.1, indicating a comparable distribution of baseline characteristics between the patients who underwent preoperative PCNA/Bx and those who did not (Table 1).
Patient selection diagram. BMI, body mass index; KALC-R, Korean Association of Lung Cancer Registry; PCNA/Bx, preoperative percutaneous needle aspiration or biopsy.
Baseline characteristics of the included patients
2. Time to recurrence
Recurrence was observed in 15.2% (180 out of 1,183) of patients who underwent PCNA/Bx, compared to 8.1% (183 out of 2,269) of those who did not undergo PCNA/Bx. In the univariate analysis, the PCNA/Bx group experienced a higher frequency of recurrence during follow-up, both before and after applying IPTW (Fig. 2). Additionally, recurrences were more common among male patients, those with a history of smoking, cases diagnosed with squamous cell carcinoma (SqCC), individuals with pleural invasion, and in stage IB (S1 Fig.).
Cumulative incidence of cancer recurrence by diagnostic procedure. (A) Without applying inverse probability of treatment weighting (IPTW), lung cancer recurrence at 12 and 36 months was 3.9% and 15.2% for preoperative percutaneous needle aspiration or biopsy (PCNA/Bx), significantly higher than the other diagnostic procedures (2.2% and 8.7%). (B) Applying IPTW, PCNA/Bx showed 12- and 36-month recurrence rates of 3.4% and 13.5%, still higher other diagnostic procedure rates of 2.4% and 9.6%.
The factors selected for multivariate analysis, both before and after IPTW application, included smoking history, stage, histological diagnosis, presence of pleural invasion, and the interaction between the implementation of PCNA/Bx and histological diagnosis. Consequently, the effect of PCNA/Bx varied depending on the histological diagnosis. After adjusting for age, smoking history, stage, and pleural invasion, the sHR for recurrence in cases of adenocarcinoma was 2.13 (95% confidence interval [CI], 1.65 to 2.76) before IPTW and 1.88 (95% CI, 1.46 to 2.41) after IPTW when PCNA/Bx was performed, compared to cases where it was not. However, for SqCC or other diagnoses, the effect of PCNA/Bx was not statistically significant. Additionally, in scenarios where PCNA/Bx was not performed, the recurrence risk for SqCC was estimated to be higher than that for adenocarcinoma, with sHRs of 2.82 (95% CI, 1.97 to 4.04) before IPTW and 2.44 (95% CI, 1.73 to 3.43) after IPTW. In contrast, when PCNA/Bx was conducted, there was no significant difference in recurrence rates based on histological diagnosis (Table 2, S2 Table).
Univariate and multivariate analysis for time to recurrence adjusted using IPTW
3. Subgroup analysis of time to recurrence based on the presence of pleural invasion
Regardless of pleural invasion, the group underwent PCNA/Bx experienced recurrence more frequently than the non-PCNA/Bx group (Fig. 3). In an analysis limited to patients without pleural invasion (84.1% of stage I cases), univariate analysis indicated more frequent recurrences among male patients, smokers, those diagnosed with SqCC, and those with stage IB disease. Multivariate analysis, similar to that for the overall stage I population, identified factors such as smoking, stage, tissue diagnosis, PCNA/Bx, and an interaction term between tissue diagnosis and PCNA/Bx as significant. After adjusting for age, smoking, and stage, adenocarcinoma cases demonstrated a significantly higher risk of recurrence when PCNA/Bx was performed (sHR, 2.37 [95% CI, 1.74 to 3.24] before applying IPTW; sHR, 2.08 [95% CI, 1.53 to 2.83] after applying IPTW). However, PCNA/Bx did not significantly impact recurrence rates in cases of SqCC or other diagnoses. In cases without PCNA/Bx, the time to recurrence for SqCC diagnoses was considerably shorter than for adenocarcinoma (sHR, 3.13 [95% CI, 2.08 to 4.73] before applying IPTW; sHR, 2.99 [95% CI, 2.01 to 4.44] after applying IPTW). Conversely, when PCNA/Bx was performed, there was no significant difference in recurrence rates based on tissue diagnosis (Table 3, S3 Table).
Cumulative incidence of recurrence by the presence of pleural invasion and diagnostic procedure in stage I patients. (A) When inverse probability of treatment weighting (IPTW) was not applied, patients with pleural invasion who underwent preoperative percutaneous needle aspiration or biopsy (PCNA/Bx) exhibited the highest lung cancer recurrence at 12 and 36 months (7.0% and 25.3%, respectively). Subsequently, the groups with pleural invasion but without PCNA/Bx (4.9% and 16.6%), those without pleural invasion who underwent PCNA/Bx (3.0% and 12.0%), and finally, those without pleural invasion who did not undergo PCNA/Bx (1.8% and 7.6%), demonstrated progressively lower rates of cancer recurrence. (B) After the application of IPTW, consistent trends remained. Among patients with pleural invasion, those who underwent PCNA/Bx displayed the most significantly elevated lung cancer recurrence at 12 and 36 months (5.9% and 23.4%, respectively). Subsequently, the groups with pleural invasion but without PCNA/Bx (4.5% and 16.5%, respectively), those without pleural invasion who underwent PCNA/Bx (2.8% and 11.3%, respectively), and finally, those without pleural invasion who did not undergo PCNA/Bx (2.0% and 8.2%, respectively), demonstrated progressively lower rates of cancer recurrence.
Time to recurrence in stage I patients with/without pleural invasion adjusted using IPTW
In patients with pleural invasion (15.9% of stage I cases), higher recurrence rates were associated with SqCC or other tissue diagnoses and stage IB in univariate analysis. However, multivariate analysis, adjusted for age, revealed that adenocarcinoma cases had a significantly higher risk of recurrence when PCNA/Bx was performed than when it was not performed (sHR, 1.65 [95% CI, 1.06 to 2.56] before applying IPTW; sHR, 1.56 [95% CI, 1.01 to 2.42] after applying IPTW) (Table 3, S3 Table).
4. Recurrence-free survival
The findings for recurrence-free survival according to whether PCNA/Bx was performed are shown in Fig. 4. Recurrence-free survival was significantly shorter in both PCNA/Bx groups, regardless of whether IPTW was applied, than in patients who did not undergo PCNA/Bx. Additionally, lower recurrence-free survival rates were observed in male patients, older individuals, those with a history of smoking, a BMI below 18.5 kg/m2, cases diagnosed with SqCC or other types, those with pleural invasion, and patients at pathologic stage IB (S4 Fig.).
Recurrence-free survival by diagnostic procedure. (A) Before applying inverse probability of treatment weighting (IPTW), the recurrence-free survival rates at 12 and 36 months for preoperative percutaneous needle aspiration or biopsy (PCNA/Bx) were 4.4% and 17.0%, respectively, which were significantly higher than those observed with other diagnostic procedures (3.0% and 10.8%, respectively). (B) After applying IPTW, the 12- and 36-month recurrence-free survival rates for PCNA/Bx were 3.8% and 15.0%, respectively, which remained higher than other diagnostic procedure (3.3% and 11.8%, respectively).
The factors selected for multivariate analysis were sex, BMI, stage, histological diagnosis, presence of pleural invasion, and the interaction between the implementation of PCNA/Bx and histological diagnosis. After adjusting for sex, age, BMI, pleural invasion, and stage, the recurrence-free survival for cases diagnosed with adenocarcinoma was estimated to be shorter when PCNA/Bx was performed than when it was not performed (HR, 1.93 [95% CI, 1.52 to 2.45] before applying IPTW; HR, 1.70 [95% CI, 1.33 to 2.18] after applying IPTW). In cases where PCNA/Bx was not performed, recurrence-free survival was estimated to be shorter for SqCC compared to adenocarcinoma (HR, 3.00 [95% CI, 2.24 to 4.02] before applying IPTW; HR, 2.71 [95% CI, 1.99 to 3.70] after applying IPTW). However, when PCNA/Bx was performed, there was no significant difference in recurrence-free survival based on the histopathological diagnosis (Table 4, S5 Table).
Univariate and multivariate analysis for recurrence-free survival adjusted using IPTW
5. Overall survival
During the follow-up period, 387 patients died (387/3,452, 11.2%), and the overall survival rates are depicted in Fig. 5 and S6 Fig. based on the PCNA/Bx procedure and baseline characteristics. The univariate analysis showed no significant differences in survival rates according to whether PCNA/Bx was performed, both before and after applying IPTW. However, lower survival rates were observed in male patients, older patients, and those with a history of smoking, BMI less than 18.5 kg/m2, a tissue diagnosis of SqCC or other types, pleural invasion, and pathologic stage IB (Table 5).
Overall survival by diagnostic procedure. (A) Without applying inverse probability of treatment weighting (IPTW), the 3- and 5-year survival probabilities for preoperative percutaneous needle aspiration or biopsy (PCNA/Bx) were 92.4% and 56.6%, respectively, which did not differ significantly from the observed with other diagnostic procedures (93.5% and 68.1%, respectively). (B) With IPTW applied, PCNA/Bx demonstrated 3- and 5-year survival rates of 93.2% and 57.9%, respectively, which remained comparable to the rates of 92.8% and 67.4%, respectively, associated with other diagnostic procedures.
Univariate and multivariate analysis for overall survival adjusted using IPTW
The factors included in the multivariate analysis remained consistent both before and after IPTW, including sex, BMI, tissue diagnosis, and pleural invasion. After adjusting for age, sex, BMI, tissue diagnosis, and pleural invasion, no significant difference was observed in the risk of death associated with the performance of PCNA/Bx, both before (p=0.902) and after IPTW (p=0.860) (Table 5, S7 Table).
Discussion
This study investigated the association of PCNA/Bx with recurrence and survival among patients from the nationwide multicenter lung registry in South Korea. The analysis adjusted for potential imbalances in patient characteristics using IPTW. The results showed that after adjusting for age, smoking history, stage, and pleural invasion, patients with adenocarcinoma who underwent PCNA/Bx experienced a 1.9-fold increase (HR, 1.88) in recurrence risk and a 1.7-fold increase in recurrence-free survival. Subgroup analysis revealed that these patients had nearly twice the likelihood of recurrence when PCNA/Bx was performed: a 2.08-fold increase in the absence of pleural invasion and a 1.56-fold increase in its presence. Further stratification showed a gradient in recurrence rates, with the highest observed in the group with pleural invasion undergoing PCNA/Bx, followed by those with pleural invasion but no PCNA/Bx, then those without pleural invasion but undergoing PCNA/Bx, and the lowest in the group without pleural invasion and no PCNA/Bx (Fig. 3). Although the registry did not specify the site of recurrence, preventing a focused examination of pleural recurrence, our findings highlight the potential long-term risks associated with PCNA/Bx in the diagnosis of early lung cancer.
Direct seeding through the needle tract during PCNA/Bx has been suspected as a cause of pleural seeding [19,20]. Furthermore, pleural injury during the biopsy procedure may create a favorable environment for tumor cell implantation [9]. An experimental study by Sawabata et al. [21] observed a significant increase in positive cytology for malignancy (60% of post-PCNA/Bx pleural washings vs. 10% of prePCNA/Bx washings). Despite these findings, debate continues regarding whether PCNA/Bx truly increases recurrence in early lung cancer, as several studies have reported that it does not raise the risk [11,12,22-24]. This debate is likely influenced by the high mortality rate associated with lung cancer and the relatively low incidence of pleural recurrence, which require long-term follow-up and a substantial patient cohort to draw definitive conclusions. This study, leveraging data from over 3,000 patients from a nationwide multicenter lung cancer registry, is the most extensive to date. It confirms an increase not only in recurrence but also in recurrence-free survival associated with PCNA/Bx. These results support the hypothesis that PCNA/Bx may cause pleural injury, which in turn heightens the risk of recurrence, even in cases where pleural invasion is already present. Based on these results, to minimize the potential detrimental effect of PCNA/Bx, it is advisable to avoid the indiscriminate use of PCNA/Bx for the pathologic diagnosis of pulmonary lesions. Alternative methods, such as a bronchoscopic biopsy-especially with advanced techniques like radial bronchoscopy-should be considered. Furthermore, for nodules that are strongly suspected to be malignancy, upfront surgical resection may also be a viable option as recommended by the British Thoracic Society [25].
In this study, PCNA/Bx was not associated with the risk of recurrence in SqCC, but it significantly raised the risk of recurrence in adenocarcinoma. While this discrepancy remains underexplored, a potential explanation could involve the glandular structure of adenocarcinoma [26], which could facilitate the shedding of tumor cells into the pleural space after pleural damage, thereby increasing the likelihood of pleural seeding after PCNA/Bx. Additionally, adenocarcinomas originate more frequently in the peripheral regions of the lungs than SqCC [4]. The proximity to the pleura may make them more susceptible to pleural spread if a pleural injury takes place. Given these observations, it might be prudent to reconsider the routine use of preoperative PCNA/Bx in stage I cancer cases when adenocarcinoma is strongly suspected and the patient is a candidate for curative resection. It is also important to note that the number of SqCC cases included in our analysis was relatively small (n=540, 15.6%) compared to adenocarcinoma (n=2,751, 79.7%). This disparity may have affected the statistical significance of our findings. Consequently, further studies incorporating a larger number of SqCC cases are necessary to validate the differential impact of PCNA/Bx on recurrence based on cancer subtype.
Several studies have examined the impact of preoperative PCNA/Bx on survival in lung cancer patients. Wisnivesky et al. [27], analyzing SEER data, found that preoperative PCNA/Bx did not affect survival in stage I lung cancer. Similarly, Zhang et al. [28] reported no impact in stage I adenocarcinoma patients. Conversely, a recent meta-analysis by Hao et al. [29] suggested reduced overall survival among patients who underwent PCNA/Bx. Additionally, Hong et al. [13] observed decreased overall survival and lung cancer-specific survival in the group aged 55 and below. In our study, after adjustments for age, sex, BMI, tissue diagnosis, and pleural invasion, we did not observe decreased survival associated with PCNA/Bx. Nevertheless, an intriguing finding presented in Fig. 4 shows a more frequent occurrence of mortality 48 months after PCNA/Bx, highlighting the necessity for long-term follow-up extended beyond 5 years. These mixed findings underscore the need for larger-scale studies with extended follow-up periods to more definitively assess the long-term impact of PCNA/Bx on survival.
A strength of this study is its analysis of extensive data available from a nationwide lung cancer registry, but there are several limitations. The primary limitation is the inability to determine the site of recurrence, especially for pleural recurrence. However, based on previous research findings, it is suspected that these recurrences associated with PCNA/Bx predominantly involve the pleura. Additional limitations include the lack of detailed information about the distance between the pleura and the tumor, the pattern or differentiation of cancer, as well as the procedural factors of biopsies. Notably, no specific PCNA/Bx techniques were associated with a reduced rate of pleural recurrence [30]. These gaps in data prevent more detailed analyses that could offer deeper insights into the factors influencing recurrence and survival. Another significant limitation of this study is the relatively short follow-up period, with a median of 36.0 months, extending only until 2021. This limited duration may not adequately capture long-term outcomes. Future studies should aim to include longer follow-up periods and incorporate detailed CT imaging and biopsy data to enhance the robustness of the findings. Lastly, a methodological limitation arises from the structure of the registry data, which is designed to protect patient privacy; specific day information for event occurrences is not available. This necessitates the assignment of an arbitrary date, typically the 15th of the month, for each recorded event. This approximation might potentially affect the accuracy of time-to-event analyses.
In conclusion, preoperative transthoracic lung biopsy was associated with an increased risk of recurrence in stage I adenocarcinoma, regardless of pleural invasion. In instances of early lung cancer where adenocarcinoma is strongly suspected and curative surgery is feasible, the use of transthoracic biopsy should be approached with caution. Clinicians should carefully weigh the benefits of PCNA/Bx against the potential risks to optimize the chances of curative resection without increasing the risk of recurrence.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
This retrospective study using anonymized data was exempted by the institutional review board (IRB No. 2212-063-1384), and informed consent was waived as we used anonymized data.
Author Contributions
Conceived and designed the analysis: Chae KJ, Park H, Yoon SH.
Collected the data: Chae KJ, Hong H, Park H, Yoon SH.
Contributed data or analysis tools: Chae KJ, Hong H, Park H, Yoon SH.
Performed the analysis: Hong H.
Wrote the paper: Chae KJ, Hong H, Yoon SH.
Conflict of Interest
Soon Ho Yoon holds stocks and stock options of MEDICAL IP, outside this work.