Recurrence Dynamics of Pathological N2 Non–Small Cell Lung Cancer Based on IASLC Residual Tumor Descriptor

1. Patients

We included patients with pathologically confirmed N2 NSCLC who underwent curative intent surgery at the Asan Medical Center, Seoul, South Korea, between January 2004 and December 2021. Clinical information of included patients was extracted from our prospectively collected database. The data of all patients who underwent surgery at our institution were stored in electronic medical records, which were regularly updated by experienced researchers. The exclusion criteria of this study were (1) patients who are either elderly (> 80 years old) or very young (< 18 years old), (2) those with concurrent malignancies, (3) received neoadjuvant treatment before surgery, (4) had wedge resection, (5) presence of distant metastasis, and (6) mortality within 30 days (Fig. 1).

Fig. 1.

Flowchart of this study. IASLC, International Association for the Study of Lung Cancer; NSCLC, non–small cell lung cancer; R0, complete resection; R(un), uncertain resection; R1/2, incomplete resection.

2. Staging workup and treatment

Diagnosis, clinical staging, and surgical strategy were described in a previous study [8]. In summary, the workup for diagnosis and clinical staging was based on imaging modalities such as chest computed tomography (CT), positron emission tomography computed tomography (PET-CT), and brain magnetic resonance imaging. If clinical N2 NSCLC was suspected, mediastinal LN biopsy was selectively performed via mediastinoscopy, endobronchial ultrasound (EBUS), or endoscopic ultrasound. If biopsy-proven N2 metastasis was confirmed, a multidisciplinary team comprising oncologists, pathologists, radiologists, radiation oncologists, and surgeons decided on the treatment approach. If a single N2 metastasis is suspected, upfront surgery is preferred when complete surgical resection can be achieved without the need for pneumonectomy and in the absence of clinical extracapsular extension (ECE). The staging of the patients included in this study was retrospectively determined based on the eighth edition of the American Joint Committee on Cancer Staging Manual [11].

Surgical treatment for NSCLC included the anatomical resection of pulmonary parenchyma and systematic mediastinal lymph node dissection (LND), which refers to the en bloc removal of the ipsilateral mediastinal and hilar LNs, including the surrounding adipose tissue. Specifically, LND includes the 2R, 4R, 7, 8, 9, 10R, 11R, and 12 nodes for right-sided NSCLC, and the 5, 6, 7, 8, 9, 10L, 11L, 12 and selectively 4L (if metastasis was suspected) nodes for left-sided NSCLC. All LNs were designated by the operator during surgery. LN stations were classified according to their anatomical location according to the Mountain-Dresler correction of the American Thoracic Society (until 2008) [12] and the IASLC LN map (since 2008) [13].

Adjuvant chemotherapy was recommended for all patients with pathologically confirmed N2 NSCLC. However, patients with poor performance status or those who expressed no intention of receiving treatment were excluded after consultation with the multidisciplinary team. Systematic adjuvant chemotherapy consisted of a platinum-based regimen administered for four cycles starting from 4 to 6 weeks after surgery. Targeted therapy was widely used as an additional treatment option in cases of recurrence or tumor progression after adjuvant treatment. For adjuvant radiation therapy, a dose of 1.8 Gy per day was administered. Patients who underwent complete resection received a total dose of 50.4 Gy, and patients with positive margins received 55-60 Gy. Patients were considered receiving adjuvant therapy only when the full course of treatment was completed.

3. Follow-up and recurrence

Outpatient-based follow-up was performed at intervals of 3 months for the first 2 years after surgery, every 6 months for the following 3 years, and annually thereafter with chest CT. When recurrence was suspected during surveillance, a systemic imaging workup, including PET-CT, was performed. The date of recurrence was defined as the date of initial detection on the imaging study, and the site of initial recurrence was recorded. After a pathological or clinical diagnosis of recurrence, a treatment plan was established by a multidisciplinary team.

4. Definition of variables

The UICC R classification, which consists of R0 resection with no residual tumor, R1 with microscopic residual tumor, and R2 with macroscopic residual tumor, has demonstrated prognostic value [1]. Subsequently, the IASLC R classification introduced an R(un) between R0 and R1/2, which was validated by several studies. According to the IASLC R classification, for R0 resection, there must be no residual tumor at the resection margin and all of the following: (1) sufficient LN evaluation must be performed (≥ 3 N1 and N2 stations, ≥ 3 N1 and N2 LNs, and LN 7 station evaluated), (2) the highest mediastinal LN must be negative for malignancy, (3) no carcinoma in situ on the bronchial stump, and (4) the pleural lavage test must be negative for malignancy. If any of these conditions is not satisfied, it is classified as R(un) [2]. R1 resection is defined as the presence of microscopic invasion at the resection margin, ECE of LNs at the resection margin of the lung specimen, or those removed separately, as well as malignant pleural or pericardial effusions. R2 resection is defined as the presence of macroscopic remaining tumor after resection, or lack of resection of involved LNs.

The highest mediastinal LN station was defined as the lowest numerically numbered mediastinal LN station evaluated. While systematic LN dissection was routinely performed in our institution, routine dissection of the 4L LN in the case of left-sided NSCLC was not performed. Therefore, when evaluating the 4L LN, 4L was defined as the highest mediastinal LN station, and when not evaluating the 4L LN, 5 was designated. There was ongoing debate regarding the hierarchical order of LN stations due to the variability in LN drainage patterns [14]. However, in most of studies, the lowest numerically numbered LN station was defined as the highest station [5,15].

According to the IASLC recommendation, N2 disease is subclassified into N2a1, N2a2, and N2b. N2a1 is defined as a single N2 metastasis without N1 metastasis (skip metastasis), N2a2 is a single N2 metastasis with N1 metastases, and N2b indicates multiple N2 metastases [16].

The histologic grade was classified according to tumor differentiation and consisted of grade I (well differentiated), grade II (moderately differentiated), and grade III (poorly differentiated), and cases where the grade could not be assessed.

Although the concept of recurrence may be somewhat inconsistent in patients who underwent incomplete resection, in this study, disease progression in patients who underwent incomplete resection was included in the recurrence category.

Locoregional recurrence was defined as recurrence at the surgical site, bronchial stump, ipsilateral lung parenchyma, or mediastinal and hilar LNs. Distant metastasis was referred to as recurrence that extends beyond the locoregional site including metastasis to the pleura, pericardium, other LNs, or distant organs. If locoregional and distant recurrence were detected simultaneously at the time of diagnosis, it was defined as mixed recurrence. Second primary lung cancer was defined according to the criteria of Martini and Melamed [17].

Disease-free survival (DFS) was defined as the time duration from the date of surgery to the date of recurrence, disease progression, death of any cause, or last follow-up. Overall survival (OS) was defined as the time duration from the date of surgery to the date of death of any cause or last follow-up.

5. Statistical analysis

Continuous variables are shown as means and standard deviations, and categorical variables are shown as frequencies and percentages. To compare between IASLC R groups (R0, R[un], and R1/2), analysis of variance for continuous variables and the chi-square tests for categorical variables were used. For comparisons between two groups, Student’s t test or the Mann-Whitney U test for continuous variables and the chi-square tests for categorical variables were applied. DFS and OS were calculated using the Kaplan–Meier method to describe the time to recurrence and death, respectively; these were compared using the log-rank test. Bonferroni correction was used to calculate the p-values in the log-rank test for multiple comparisons of the survival curves. The hazard ratio (HR) of recurrence was evaluated using the life table method, which was the conditional probability of manifesting recurrence within a certain time interval. To display the underlying patterns easier, some instability owing to random variation in the hazard rate estimates was dealt with using the kernel smoothing method. Univariable and multivariable Cox proportional hazards models were applied to identify the prognostic factors for recurrence. After excluding the correlated variables, those with p < 0.05 in the univariable analyses were included in the initial multivariable Cox model. The final multivariable model was selected using the forward stepwise selection method (p ≤ 0.10 for entering the model and p ≤ 0.05 for remaining in the model).

All statistical analyses were performed using R ver. 4.0.2 (The R Foundation for Statistical Computing, Vienna, Austria). p < 0.05 was considered statistically significant.

1. Patient characteristics

A total of 1,373 patients were included in this study. Their mean age was 61.7±10.0 years, and 862 were male (62.8%). According to the UICC R classification, 1,250 (91.0%), 112 (8.2%), and 11 (0.8%) patients were classified as R0, R1, and R2, respectively. According to the IASLC R classification, 576 (42.0%), 286 (20.8%), and 511 (37.2%) were reclassified as R0, R(un), and R1/2, respectively. Among the UICC R0 patients, 576 (46.1%) patients were still IASLC R0 status, while 279 (22.3%) and 395 (31.6%) patients were reclassified as R(un) and R1/2, respectively (S1 Table). Of the 286 patients classified as R(un), 279 were reclassified from UICC R0 and seven were reclassified from UICC R1. Among the patients classified as IASLC R(un) from UICC R0, 25 (9.0%) patients had insufficient LN evaluation, including 19 patients (6.8%) with fewer than three N1 or N2 LN stations and six patients (2.2%) without evaluation of the LN 7 station. The remaining 254 patients (91.0%) were positive for malignancy in the highest mediastinal LN station (S2 Table). Among the patients classified as UICC R1, seven (6.2%) were reassigned to R(un) because of carcinoma in situ on the bronchial resection margin. Furthermore, 395 patients from UICC R0 were reclassified as IASLC R1/2 status because of ECE (S2 Table).

The baseline characteristics of the included patients are shown in Table 1. There were no differences in age (p=0.398), sex (p=0.623), smoking history (p=0.451), comorbidities (p=0.594), forced expiratory volume in 1 second (p=0.650), and diffusing capacity to carbon monoxide (p=0.322) between groups. In addition, there were no differences in the frequency of performing EBUS (p=0.424) and mediastinoscopy (p=0.762), histologic subtype (p=0.659), tumor size (p=0.073), and adjuvant therapy (p=0.095). However, in the R(un), left-sided NSCLC was more common (p=0.001), and a significantly lower number of harvested LNs was observed (p=0.001).

Table 1.

Baseline characteristics of NSCLC patients with pathologically confirmed N2 metastasis according to the IASLC R classification

2. Progression outcomes

The mean follow-up duration was 52.88±44.13 months, during which there were 710 cases of recurrence (51.7%) and 667 deaths (48.6%) (S3 Table). According to the IASLC R classification, the 5-year DFS rates were 43.2% for R0, 34.2% for R(un), and 21.1% for R1/2 (p < 0.001). There were significant differences between R0 and R(un) (p < 0.001), as well as between R(un) and R1/2 (p=0.004) (Fig. 2A). The 5-year OS rates were 62.4% for R0, 54.6% for R(un), and 43.4% for R1/2 (p < 0.001). Significant differences were observed between R0 and R(un) (p=0.040), as well as between R(un) and R1/2 (p=0.004) (Fig. 2B).

Fig. 2.

(A) Disease-free survival of patients with pathologic N2 disease according to the IASLC R classification. (B) Overall survival of patients with pathologic N2 disease according to the IASLC R classification. (C) Hazard rate for recurrence of patients with pathologic N2 disease according to the IASLC R classification. IASLC, International Association for the Study of Lung Cancer; R, residual tumor; R0, complete resection; R1/2, incomplete resection; R(un), uncertain resection.

Univariable and multivariable Cox proportional hazards analyses were performed to identify independent prognostic factors for recurrence in patients with pN2 NSCLC. In the univariable analysis, the IASLC R descriptor was a significant prognostic factor, and the HRs of R(un) and R1/2 compared to R0 were 1.42 (95% confidence interval [CI], 1.16 to 1.73; p=0.001) and 1.88 (95% CI, 1.59 to 2.22; p < 0.001), respectively. In the final multivariable Cox model, the IASLC R descriptor was still a significant prognostic factor, but the prognostic difference between R0 and R(un) became insignificant after adjustment for histologic type, subdivided N2 descriptor, subdivided pathological stage, and the performance of adjuvant therapy (HR, 1.18; 95% CI, 0.96 to 1.46; p=0.123) (Table 2).

Table 2.

Univariable and multivariable Cox analysis for recurrence in patients with pN2 non–small cell lung cancer

3. Recurrence dynamics

Fig. 2C shows the hazard rate for recurrence based on the IASLC R descriptor. Patients with R(un) had a 1.4-fold higher hazard rate of recurrence than those with R0. However, the hazard rate curves displayed similar patterns between the three groups, with a peak at approximately 12 months after surgery.

Subgroup analyses for recurrence sites were performed in R0 and R(un) patients, excluding R1/2 patients who underwent incomplete resection. Between R0 and R(un), there were significant differences in overall recurrence (43.1% vs. 54.5%, p=0.002) and distant recurrence (23.8% vs. 39.2%, p < 0.001), while locoregional (8.0% vs. 6.3%, p=0.451) and mixed (11.3% vs. 9.1%, p=0.385) recurrence were similar. Further analysis was performed after stratification with the IASLC N2 descriptor (Table 3). There was no significant difference in recurrence rate between R0 and R(un) in locoregional and mixed recurrence for patients with N2a1, N2a2, and N2b (Table 3). However, a significant difference was shown in distant recurrence for patients with N2a1 (18.0% vs. 32.7%, p=0.037) and in N2a2 (24.7% vs. 37.9%, p=0.044) (Table 3). On the other hand, a significant difference was not shown in distant recurrence for patients with N2b (31.9% vs. 41.7%, p=0.154) (Table 3).

Table 3.

Recurrence patterns stratified by subdivided N2 descriptors in the R0 and R(un) groups

Multivariable Cox analyses for recurrence were performed according to the recurrence site in R0 and R(un) patients. Several factors were adjusted to identify the prognostic significance of positivity for the highest mediastinal LN node station: age, sex, histology, pathologic stage, the IASLC N2 subclassification, and adjuvant treatment. In the overall recurrence, only the N2 subclassification was identified as an independent prognostic factor. Compared to N2a1, the HR of N2a2 was 1.48 (95% CI, 1.14 to 1.92; p=0.004), and that of N2b was 1.79 (95% CI, 1.34 to 2.40; p < 0.001). In the distant recurrence, both positivity for the highest mediastinal LN station and N2 subclassification were confirmed as independent prognostic factors; the HR of positivity for the highest mediastinal LN station was 1.44 (95% CI, 1.08 to 1.92; p=0.013) (Table 4). However, positivity for the highest mediastinal LN station was not a prognostic factor for locoregional and mixed recurrence.

Table 4.

Multivariable Cox analysis for recurrence according to the recurrence site in patients with R0 and R(un)