Risk factors of recurrence for resected T1aN0M0 invasive lung adenocarcinoma: a clinicopathologic study of 177 patients

Lung carcinoma is one of leading causes of cancer death worldwide, and its incidence rate continues to increase [1]. Adenocarcinoma is the most common histologic type of lung cancer, especially among Chinese women [2]. The widespread use of computed tomography (CT) screening encouraged by positive results of the National Lung Screening Trial is detecting more and more early-staged adenocarcinomas [3, 4]. T1aN0M0 constitutes the earliest stage of invasive lung cancer. As the standard of care, follow-up and surveillance without adjuvant therapy are recommended after complete resection [5]. However, some patients experience recurrence and die of lung cancer. According to the seventh lung cancer staging project by the International Association for the Study of Lung Cancer (IASLC), T1a non-small cell lung cancer (NSCLC) has a 5-year survival of about 80% [6]. Therefore, there has been a need to identify these small but aggressive tumors for more careful follow-up and/or adjuvant therapy.

Many efforts have investigated the use of clinical, radiologic, and histologic factors to refine the prognosis of adenocarcinomas after surgical resection. Studies have reported prognostic significance of gender [7, 8], smoking history [9], tumor size [10], degree of differentiation [11], visceral pleural invasion [12], lymphovascular invasion (LVI) [13, 14], ground-glass opacity/solid ratio [15], size of invasive tumor [16], maximum uptake on positron emission tomography (PET) scan [10, 15]. Given the fact that prognostic factors of lung cancer may vary with tumor size [17], whether these factors have prognostic value in T1aN0M0 adenocarcinomas is still unknown.

Remarkable advances in understanding lung adenocarcinoma have led to a new classification system, sponsored by the IASLC, the American Thoracic Society (ATS), and the European Respiratory Society (ERS) [18]. Several independent studies have validated the prognostic significance of new adenocarcinoma histologic subtypes [16, 19, 20], but little attention has been paid to T1a tumors after radical resection. However, previous studies of adenocarcinoma ≤ 2 cm failed to include new classification into the analysis [13, 21, 22]. The prognostic factors for completely resected T1aN0M0 adenocarcinomas are still an open question. It has been suggested that prognostic factors for lung cancer may vary with tumor size [17].

The main purpose of our study was to determine the risk factors for recurrence in a cohort of Chinese patients with radically resected invasive T1aN0M0 adenocarcinoma.

To the best of our knowledge, the current study is the first clinicopathologic prognostic analysis of T1aN0M0, incorporating clinicopathologic and new IASLC/ATS/ERS classification variables [13, 22]. The only published study investigating the prognostic utility of the classification for patients with ≤ 2 cm adenocarcinomas focused on the association of the percentage of micropapillary component with recurrence, and between lobectomy versus limited resection. We carefully designed this retrospective study, trying to identify risk factors that can reflect accurately the behaviors of early-staged lung adenocarcinomas.

Because adenocarcinoma in situ (AIS), and minimally invasive adenocarcinoma (MIA) have 100% or near 100% disease specific survival after complete resection [18], they were excluded in our study of risk factors for recurrence. Visceral pleural invasion (VPI) was also excluded since it is not only determined by aggressiveness of tumor, but also by its anatomic location. Although one study which focused on adenocarcinomas < 2 cm found no impact of VPI on recurrence or survival [32], a nation-wide registry study confirmed the association of worse survival with VPI in NSCLC of all sizes [33], and the adverse impact of VPI was hypothesized as dissemination through parietal sub-pleural lymphatic drainage [34]. Moreover, we chose patients after radical lobectomy and lymph node dissection, so that surgical margin-related recurrence would not influence the analysis [35].

Our cohort of T1aN0M0 lung adenocarcinoma patients had a 5-year RFS of 83.7%, similar to previous reports [22, 36]. The distribution of IASLC/ATS/ERS subtypes varies considerably in the literature, which may have resulted from differences between geographical regions and ethnic populations (east versus west) [16, 31], patient groups (all stages versus stage I) [29], and the expertise of the pathologists [37]. The frequency of lepidic predominant ranged from 26.7% reported by a Japanese study to 5.6% in a US cohort, while that for papillary predominant ranged from 40.9% to 4.7%. The percentage of micropapillary ranged from 15.2% to 0 out of 191 adenocarcinomas in a Japanese study. In the current study, the most frequent histologic subtype was acinar predominant, comprising 58 out of 177 cases (32.8%), followed by lepidic predominant, 25.4%. Papillary-, micropapillary-, and solid predominant adenocarcinomas accounted for 37 (20.9%), 19 (10.7%), and 18 (10.2%) specimens respectively.

In common with previous reports [7, 13, 16, 19, 28‐31, 38, 39], our data confirmed the prognostic value of the IASLC/ATS/ERS histologic classification, along with LVI, by log-rank test. The 5-year RFS of lepidic predominant was 94.9%; papillary predominant, 91.9%; acinar predominant, 83.1%; solid predominant, 67.4%, and micropapillary predominant, 57.6%. To gain more statistical power, previous studies combined subtype of similar prognosis to construct a two-tiered or three-tiered grading scheme. However, difference remains regarding the designation of papillary predominant adenocarcinomas. Interestingly, most studies on Asian patients grouped this subtype into lower- or intermediate-grade [16, 19, 28, 30, 31], while most studies from western countries, grouped it into the high-grade group [27, 29] except that by Yoshizawa et al. [28]. Comparison of these two grouping schemes on our cohort favored the former way of grouping (Figure 2, P = 0.005 versus P = 0.181), suggesting that the behaviors of papillary predominant subtype may have ethnical differences. This hypothesis can be supported by the findings in Asian patients of the association between EGFR mutation and papillary predominant adenocarcinomas [39, 40], but not in an Australian cohort [27].

The observation that new histologic subtype grouping remained the only risk factor for recurrence in patients of T1aN0M0 disease on multivariate Cox regression, and reported correlation of clinicopathologic variables with histologic subtypes, promoted us to investigate the relationship of these variables with subtype groupings. We chose low-grade or high-grade as binary dependent variables instead of each histologic subtype, for the reason that the present study focused on recurrence risks, which are more significantly related to subtype groupings. The correlation analysis confirmed the association of male gender and presence of LVI (borderline significant) with high-grade group (Table 3). Other researchers have reported similar associations. Hung et al. found that the lepidic predominant subtype is associated with less smoking exposure, smaller tumor size, absence of LVI, and well/moderately differentiated histologic grade, while the high-grade group member, solid predominant, is associated with male gender, smoking exposure, larger tumor size, and poorly differentiated histologic grade [19]. Yanagawa and colleagues reported higher frequency of smokers in solid predominant than in other subtypes [16]. Our results, along with these observations, suggested that these clinicopathologic variables, such as gender, smoking history and LVI, might not be independent prognostic factors, which may challenge the independency of these prognostic factors, such as female gender [7].

The local infiltration ability and metastatic potential of solid- and micropapillary predominant subtypes in the high-grade group, may explain the increased risk of recurrence. A recently published comparison of small (≤2 cm) adenocarcinoma patients who underwent sublobar resection versus lobectomy found that tumors with ≥ 5% micropapillary component had more locoregional recurrences after limited resection, especially those with small surgical margins, but not in cases who underwent lobectomies [35]. The authors suggested a greater capacity for local infiltration of micropapillary subtype. Likewise, a study of adenocarcinomas with N2 metastasis suggested greater metastatic potential of micropapillary and solid predominant subtypes, even though they may not be the predominant subtype [27]. In a study of adenocarcinoma of post-operative adenocarcinomas at all stages, Russell reported the highest incidence of N2 metastases and invasion of lymphovascular spaces and visceral pleura for micropapillary predominant adenocarcinoma [20]. Indirect evidence also comes from a subtyping analysis of stages I to IV adenocarcinomas, which showed the highest rate of nodal metastases was for micropapillary predominant subtype (76%), followed by solid predominant (51%), whereas for lepidic, the rate was only 7% [29]. However, with respect to prognosis, the predominant subtype was the main determinant [27, 29].

There is growing interest in considering sublobar resection for early-stage lung cancer. To date, limited evidence only supports the use of sublobar resection for subsolid lesions, a radiologic feature of preinvaisve or less invasive tumors [5, 41]. Histologic subtypes of adenocarcinoma according to IASLC/ATS/ERS classification also stratifies invasiveness and may have potential implication for selecting limited resection histologically, as suggested by Nitadori [35]. However, the accuracy of reporting adenocarcinoma subtypes on frozen section is still an open question.

We acknowledge several limitations of our retrospective study, which has a relatively small cohort of patients with sub-optimal follow-up periods. The fact that even indolent ground-glass opacity lesions can develop local recurrence more than five years after resection [42] implied the necessity of very long follow-up for these very early-staged cases. Moreover, not all recurrences were biopsy-confirmed, which is more accurate than radiologic evidence.

Our results showed that the solid- and micropapillary predominant subtype of IASLC/ATS/ERS classification remains the only risk factor for post-operative recurrence of T1aN0M0 adenocarcinomas, suggesting that they can be indicators of aggressive tumor behaviors. We agree with the proposal by Warth et al. that a tumor grading system based on the predominant pattern might assist clinical decision-making for pulmonary adenocarcinoma management [29].

FY, KC and XL are staff members of Department of Thoracic Surgery, Peking University People’s Hospital. YL is a PhD student of Peking University Health Science Center. KS and DB are staff members of Department of Pathology, Peking University People’s Hospital.