Wedge resection is equal to segmental resection for pulmonary typical carcinoid patients at localized stage: a population-based analysis

Background Medical institutions worldwide have not reached a consensus on what surgery is the most advisable for pulmonary typical carcinoid (TC) patients at the localized stage. This research focuses on exploring whether wedge resection or segmental resection is the superior option. Methods The demographic and clinical information of 1,887 TC patients diagnosed at the localized stage from 2004 to 2015 was collected from the Surveillance, Epidemiology, and End Results (SEER) Program. Patient prognosis was evaluated by KM curves. The chi-square test was used to examine the variation between different groups that would be eliminated by propensity score matching (PSM). Univariate and multivariate Cox proportional hazard model analyses were used to evaluate prognostic values of relative factors. Results The prognosis of TC was the most favorable for patients suffering from pulmonary squamous cell carcinoma (SCC), adenocarcinoma (ADC), and pulmonary carcinoids (PCs). The choice to have surgery, not the type of surgery chosen, was the most significant independent prognostic factor correlated with overall survival (OS) and lung cancer-special survival (LCSS). The prognostic result of the comparison between wedge resection and segmental resection was not statistically significant before or after PSM. In subgroup analysis, the inference still held.


INTRODUCTION
Typical carcinoid (TC) is one of four major pathological classifications of neuroendocrine tumor (NET) of the lung categorized by the World Health Organization (WHO). The others are atypical carcinoid (AT), large-cell neuroendocrine carcinoma (LCNEC), and small-cell lung cancer (SCLC) (Caplin et al., 2015). Data shows that bronchopulmonary or lung NETs constitute approximately 25 percent of carcinoma of the lungs, while the prevalence of primary NETs is 20-25% (Ramirez et al., 2017;Yao et al., 2008). Compared with other types of NET, the incidence of TC and AT is lower and accounts for 1-2% of pulmonary malignant tumor (Chen, Travis & Krug, 2006;Travis, 2010). Another study indicates that there are 10 times more TC than AT patients (Gosain et al., 2018;Rekhtman, 2010). Until now, surgery has been the main treatment for TC, the low-grade lung NET, at the localized stage (Wolin, 2015;Wolin, 2017). The recommendation of adjuvant therapy, especially radiation and chemotherapy, remains controversial (Wolin, 2017).
For early stage patients with operable TC, there is no general agreement on the optimal operation mode (Fox et al., 2013;Mezzetti et al., 2003;Raz et al., 2015;Rea et al., 2007). Lobectomy and sublobectomy (wedge resection and segmental resection) are the major surgical options for localized diseases. Traditionally, anatomic resection (lobectomy and segmental resection) is considered the best option for patients with peripheral lung tumors, while segmentectomy and wedge resection are widely used for patients who have limited pulmonary function (Caplin et al., 2015). Recently, Taher Abu Hejleh et al. (Furqan et al., 2018), after collecting and researching the SEER data, suggested that the prognoses for lobectomy and sublobectomy were comparable. Some studies have shown segmental resection to be superior to wedge resection in liver cancer and non-small cell lung cancer (NSCLC) (El-Sherif et al., 2007;DeMatteo et al., 2000). Other studies found no differences between wedge resection and segmental resection in lung neoplasm (Altorki et al., 2016;Romano & Mark, 1992). However, currently no research compares the two operations, wedge resection and segmental resection, which constitute the sublobectomy, in TC patients. Although both surgeries are sublobectomies, wedge resection causes a smaller operative wound than segmental resection. Now that surgery is the main choice of treatment for TC patients, and given that other therapeutic methods are not powerful, the smaller operative wound is a major consideration after guaranteeing the survival rate. The objective of this research is to determine if there is a difference between wedge resection and segmental resection for early stage TC patients.

Data source
We have been given access to the SEER databases in October 2018 with the ''numbered'' SEER*Stat account(12991-Nov2017). The retrospective research collected the data from the Surveillance, Epidemiology and End Results (SEER) by using SEER*Stat 8.3.5 in December 2018. The inclusion criteria were as follows: (I) TC patients (code 8240) at the localized stage; (II) patients treated via no surgery at the primary site (code 00), wedge resection (code 21), segmentectomy (code 22), or lobectomy (code 33); (III) patients with only one primary tumor. The exclusion criteria included: (I) diagnosis before 2004, (II) survival of less than one month or unknown, (III) lung cancer-special survival (LCSS) missing, (IV) tumor size unknown, (V) scope of regional lymph node surgery unknown, (VI) radiation therapy unknown, and (VII) regional, distant, or unknown summary stage.
Presently squamous cell cancer and adenocarcinoma are common forms of lung cancer, accounting for 78% of this disease. Lung carcinoid is composed of typical carcinoid and atypical carcinoid. To acquire general prognosis information about TC patients, we compared the cumulative survival rate for lung carcinoid and its components with SCC and ADC. The clinical information was recorded in Table 1. Obviously, carcinoid has the more favorable prognosis compared with SCC and ADC in overall survival (OS) and lung cancer-specific survival (LCSS) (Fig. 1), as follows: hazard ratio (HR) = 0.149 (0.141-0.519), P < 0.001 of carcinoid versus ADC in OS; HR = 0.092 (0.085-0.100), P < 0.001 of carcinoid versus ADC in LCSS; , P < 0.001 of SCC versus ADC in OS; HR = 1.135 (1.123-1.147), P < 0.001 of SCC versus ADC in LCSS.
In conclusion, typical carcinoid has a more favorable prognosis than the other three types of lung cancers.

Baseline information of TC patients from SEER database
There were 1,887 TC patients in our cohort: 235 patients with no surgery, 465 undergoing wedge resection, 113 receiving segmental resection, and 1,074 going under the lobectomy. Some demographic and clinical information, such as age, gender, race, etc., was been recorded (Table 2). Notably, 71.8% patients were female and most patients were white. Compared with the no surgery group, the percentages of old patients (age ≥ 70 years)  are less than 30% of the other three surgery groups: 58.3% in no surgery, 27.1% in wedge resection, 15.9% in segmental resection, and 18.1% in lobectomy. Thus, it can be seen that age may be an important factor in choosing a treatment option. In addition, the number of female patients is much larger than that of male patients no matter what treatment they receive. As for the factor ''scope of the lymph node,'' wedge resection results in less regional lymph node resection compared to segmental resection or lobectomy: the percentage of no regional lymph node resection is 68.6% in wedge resection and 38.9% in segmental resection.

Prognosis for TC patients treated with no surgery, wedge resection, segmental resection, and lobectomy
The influence of surgery and surgery type on overall survival (OS) and lung cancer-specific survival (LCSS) for pulmonary TC patients were evaluated using KM curves (Figs. 2A, 2D).
There were significant differences between the group with no surgery and the other three groups, both in OS and LCSS: The 5-year survival rate of the no surgery, wedge resection, segmental resection, and lobectomy, respectively, were 73%, 90%, 94%, and 94% for OS (P < 0.001). There is an even greater gap between patients with no surgery and groups who underwent surgery when the 10-year survival rate is examined.
For further research about the impact of surgery type, the three groups were defined here as pairs to be assessed with KM curves (Fig. 3). Although the comparison between lobectomy and wedge resection indicated that the difference in patient prognoses was statistically significant for OS (P < 0.001), the propensity score match (PSM) eliminated

Univariate and multivariate COX regression analyses of TC patients in the localized stage
Relative risk factors for OS were analyzed using univariate and multivariate COX regression hazard models (Table 3). The factors whose P < 0.05, both in univariate and multivariate analysis, are: age (P < 0.001 in multivariate analysis), gender (P = 0.037 in multivariate   (Table 4). Data showed that age (P = 0.008 in multivariate analysis), tumor size (P = 0.012 in multivariate analysis), and surgery (P = 0.004 in multivariate analysis) are independent prognostic factors whose P < 0.05 both in univariate and multivariate analysis. As for LCSS, the HR of independent prognostic factors is as follows: 2.261 (95% CI [1.232-4.149], ≥70 y versus  between these two groups (Table 5). To reach a more objective conclusion, factors such as age, laterality, tumor size, and scope of regional lymph node must be balanced. Compared with wedge resection, segmental resection was more often offered to people of older age, with larger tumor sizes, who had a larger region of lymph node resected. To avoid being influenced by factors' disproportionate distribution and data bias, we used PSM (an analysis function of the software, Statistical Package for the Social Sciences) to acquire an adequate sample set. First, we found the unbalanced variates by t -test. Then we chose the PSM to solve those imbalances by matching analogical cases from two groups. After 1:1 PSM, 100 patients who underwent wedge resection and 100 patients with segmental resection were involved in the final research. Meanwhile, all factors, including age, laterality, tumor size, were balanced to some extent ( Table 5).
The prognoses for 100 pairs of patients after exactly matching were evaluated using KM curves (Figs. 2C, 2F). The difference between wedge resection and segmental resection did not achieve statistical significance: for OS, the 5-year survival rates of wedge resection and segmental resection were 90.07% and 94.17% respectively, and the 10-year survival rates of wedge resection and segmental resection were 83.48% and 84.78%, respectively (P = 0.320); for LCSS, the 5-year survival rates of wedge resection and segmental resection were 97.89% and 98.75%, respectively, and the 10-year survival rates of wedge resection and segmental resection were 93.67% and 98.75%, respectively (P = 0.342).

Univariate COX regression analyses of TC patients who underwent wedge resection and segmental resection before PSM
Relative risk factors for OS and LCSS were analyzed using a univariate COX regression hazard model (Table 6). The study showed that age was only a prognostic factor for OS and LCSS: HR

Subgroup analysis of TC patients who received wedge or segmental resection for OS and LCSS before PSM
The results showed no statistically significant difference between wedge and segmental resection in any subgroup of age, race, gender, tumor size, or scope of regional lymph node, which was rendered in KM plots . Cox regression analysis for OS also supported that view (Table 7). The results of Cox regression analysis for LCSS are not shown because death was rare after sectionalization.

DISCUSSION
The study showed that the prognosis of TC patients was best among those suffering from ADC and SCC. Further univariate and multivariate Cox analysis showed that the surgery option was independent of the prognosis for OS and LCSS in TC patients. As for the impact of surgery type (wedge and segmental resection) on survival, the result was not statistically significant either before or after PSM. In conclusion, the important discovery of this report is that wedge resection is likely to be equal to segmental resection for TC patients at the localized stage.
As the low-grade lung NET, TC has a favorable prognosis, where the 5-year and 10-year survival rate for OS and LCSS were highest compared with NSCLC and AC (Table 1). Some studies also showed that 5-year OS of TC was more than 87 percent (Filosso et al., 2013;Gosain et al., 2018). The excellent prognosis for TC may be attributed to a handful of Figure 5 Prognostic analysis for subgroup of race and Scope of Regional Lymph Node by KM plots. Overall survival and lung cancer specific survival in subgroups of race (white and others) and Scope of regional Lymph node (0-3 and ≥3) between wedge resection and segmental resection. There was no significant difference between two surgical options in OS: (A) white; (B) others; (C) 0-3; (D) ≥3. Besides, there was no significant difference between two surgical options in LCSS: (E) white; (F) 0-3; (G) ≥3. No death case was found in Others (subgroup of race) in LCSS so that the KM curves was missing. OS, overall survival; LCSS, Lung cancer-special survival; Scope of Reg. LN Sur., Scope of Regional Lymph Node.
Full-size DOI: 10.7717/peerj.7519/ fig-5 mitoses, no necrosis, and a low Ki-67 labeling index (Caplin et al., 2015;Rindi et al., 2014). However, other research has indicated that the survival rate for Pulmonary carcinoids (PCs) has dropped for the past 30 years (Gustafsson et al., 2008). In its milder forms, it may still be a major problem increasing mortality and delaying diagnosis. Compared with other lung carcinomas, only 10.4% of primary TC is located in the bronchi, so respiratory symptoms, such as chest infections, cough, hemoptysis, and chest pain are relatively rare (Gustafsson et al., 2008;Caplin et al., 2015). Similarly, lung neoplasms were larger than five cm resected from the TC patients at the localized stage because the symptoms were so mild. For TC patients who were at the localized stage, age, gender, and surgery option were independent prognostic factors for OS in the multivariate Cox analysis. However, the factor ''gender'' was replaced by the factor ''tumor size'' in assessment for LCSS in the multivariate analysis. Same with James W. Vaupel 's finding, Women live longer than men today (Zarulli et al., 2018). Long life span of female may contribute to the cause differences in analysis of OS. Besides, unlike OS, people were classified as having died from LCSS only when the cause was related to lung cancer. Therefore, gender may not influence the progress of death from TC. The significant difference between the group with no surgery and the group who underwent surgery indicated that surgery was essential for the treatment of TC even though it had low malignancy. In a Cox regression model with time dependence, age was statistically significant (data not shown). The HR increases as patients age. Like the results of Taher Abu Hejleh et al. (Furqan et al., 2018), research on surgery type showed that TC patients at the localized stage would benefit equally from lobectomy or sublobectomy. However, Taher Abu Hejleh et al. overlooked that sublobectomy contained both wedge and segmental resection. Comparisons between sublobectomy and lobectomy respectively were not done. Thus, superiority of lobectomy cannot be proved by being contrasted with sublobectomy, similar anatomy resection (segmentectomy), or wedge resection.
The emphasis of this research is on whether the benefits of wedge resection can equal those of segmentectomy. The raw data indicated that the choice between the two procedures did not influence the prognosis of patients who underwent sublobectomy. Furthermore, after 1:1 PSM, the gaps between the OS and LCSS survival rate of two surgery procedures closed, providing additional evidence for the uniformity of the benefits from treatment, whether treatment consisted of wedge resection or segmentectomy produced in TC therapy. Besides, the Cox analysis of TC patients who underwent those two surgeries before PSM showed that only age was a relevant prognostic factor for OS or LCSS, and surgery type was no exception. The therapeutic effect of sublobectomy eliminated the impact of factors other than age on prognosis and both two operating methods were effective for TC. Furthermore, this held true for many subgroups of TC patients, including groups that were younger or older, or male or female, verified by the subgroup analyses. A survey that involved 172 institutions worldwide showed that only 11 percent of participants regarded wedge resection as an appropriate surgical option for peripheral PCs (Caplin et al., 2015). However, the results of this research showed that patients did not benefit more from segmental resection than wedge resection. Compared to the anatomic resection, the operation wound from wedge resection was milder and caused less damage to pulmonary function. Therefore, the results proved that wedge resection, as well as anatomic resection, should be considered a conventional treatment for TC patients at the localized stage. Preserving better pulmonary function and leaving a smaller surgical wound will promote better quality of life for patients undergoing the surgery.
Inevitably, there are several limitations similar to those of most retrospective studies based on the SEER database used for this study. First of all, information on patients, such as complications, recurrence, follow-up treatment received, TMN stage, and so on was not complete. This reduced the level of accuracy and detail of the prognoses. Secondly, the number of cases was limited because the disease is rare. The results of the research will be more convincing when additional cases are studied. Third, this research may include more bias than a prospective study or a randomized trial.