Abstract
Background: The chances of pulmonary resection for small-sized lung cancer have increased because of the development of thin-slice computed tomography (CT). Though sublobar resection could be indicated for ground glass opacity (GGO)-dominant adenocarcinoma with low-grade behaviour, the malignant potential of solid-type, small-sized lung cancer has not been sufficiently assessed. We aimed to address the clinical outcomes of resected solid-type c-stage IA non-small cell lung cancer (NSCLC) smaller than 2 cm. Methods: A retrospective observational study involving 118 patients who had undergone a complete resection for lung cancer smaller than 2 cm with solid component more than 50% on CT was conducted, and their postoperative survival and recurrence pattern were analysed. Results: Thirty-five patients with solid component-dominant lesion (SCDL) and 83 patients with pure solid lesion (PSL) without GGO were enrolled. Lymph node involvement was found in 15 patients with PSL (18%). The 5-year disease-free survival (DFS) was 100% in SDCL patients and 83% in PSL patients. Multivariate analysis of PSL patients showed that lymph node metastasis and pleural invasion were independent negative prognostic predictors. The 5-year DFS was 88%, 80% and 46% in p-N0, p-N1 and p-N2 patients, respectively. The 5-year DFS was 33% for patients with pleural invasion, which was significantly worse than that for patients without pleural involvement. Postoperative recurrence was mainly observed as intrathoracic lesions within 3 years. Conclusions: A proportion of solid-type NSCLC has malignant potential, even for tumours smaller than 2 cm. Periodic intrathoracic evaluation is required following complete resection.
1 Introduction
Lung cancer is a leading cause of death from cancer. Several clinical trials with multimodal therapy, including adjuvant chemotherapy, have been conducted to improve the outcome of advanced disease, while the efficacy of induction therapy is still controversial [1–4]. Diagnostic imaging such as computed tomography (CT), magnetic resonance imaging (MRI) and positron-emission tomography CT (PET–CT) is now widely available clinically. The recent development of thin-slice and high-resolution CT or low-dose CT screening has particularly increased the chance of treating small-sized lung cancer.
Bronchioloalveolar adenocarcinoma, which is usually detected as ground glass opacity (GGO) on CT imaging, has been identified as a different type of non-small-cell lung cancer (NSCLC) with low-grade malignancy [5]. It has been reported that sublobar resection, such as segmentectomy or wide-wedge resection, is feasible for GGO-dominant lesions [6], while a lobectomy is still a standard procedure for NSCLC. A clinical trial of randomised lobectomy versus sublobar resection for peripheral small-sized lung cancer is ongoing in the USA (CALGB 140503), and a similar trial has been proposed by the Japan Clinical Oncology Group.
However, patients with small-sized NSCLC occasionally have locally advanced disease with lymph node metastases [7–9]. Most such cases have a lesion showing a solid component or solid-dominant GGO on preoperative CT imaging. These solid-type lesions could be a different entity among small-sized lung cancer, most of which is early disease.
The present study focussed on c-stage IA patients with solid component-dominant lesion (SCDL) or pure solid lesion (PSL) small-sized NSCLC, especially patients with PSL to clarify postoperative outcomes and consider the optimal surgical procedure. Disease-free survival (DFS) and the relapse pattern during postoperative follow-up were retrospectively analysed.
2 Materials and methods
2.1 Patients
The records of all 188 consecutive patients with c-stage IA peripheral lung cancer smaller than 2 cm, who underwent a complete resection between 1992 and 2007 at the Osaka University Hospital, Osaka, Japan, were reviewed. Institutional Review Board at the Osaka University approved this study, and patient consent was waived. Of these, 118 NSCLC patients with a solid or solid component-dominant lesion, in which the GGO area was less than 50% on thin-slice CT (Table 1 ), were selected. Follow-up of more than 5 years was completed in 100 patients, and the median follow-up time for all patients was 60.8 months. Preoperative diagnosis was performed by chest radiography and CT imaging, as well as biopsy using fibre optic bronchoscopy or percutaneous core needle biopsy. Core needle biopsy was performed under CT guidance in our institute. Lymph nodes larger than 1 cm in the short axis on CT were clinically defined as metastasis-positive. An 18-fluorodeoxyglucose positron-emission tomography (FDG-PET) and PET–CT were also available for nodal staging from 1997 and 2005, respectively. Mediastinoscopy was performed in two patients with mediastinal lymph node swelling observed in CT. Brain CT or MRI, abdominal CT and bone scintigraphy were used to detect distant metastases for clinical staging. The preoperative serum carcinoembryonic antigen (CEA) level was measured in 113 patients and included in the survival analysis as a variable. We performed a lobectomy with nodal dissection in principle, while sublobar resection was indicated in high-risk cases. Postoperative staging was performed according to the tumour, node, metastasis (TNM) classification. Pleural invasion was histologically evaluated using haematoxylin–eosin (H–E) staining. There was no parietal pleural invasion and only visceral pleural invasion was observed in this study. Adjuvant chemotherapy was indicated in p-stage II and III patients with good status since 2005, and as a result, given in nine patients. One patient had postoperative radiation. Chest and abdominal CT scan, brain CT or MRI and bone scintigraphy were used for postoperative follow-up. Initial recurrence sites and the pattern following complete resection were also evaluated. Variables used for evaluation were age, sex, tumour diameter, operative procedure, histology, serum CEA level, lymph node metastasis and pleural invasion.
Characteristics of 118 patients with c-stage IA solid-type non-small cell lung cancer smaller than 2 cm in diameter.
Characteristics of 118 patients with c-stage IA solid-type non-small cell lung cancer smaller than 2 cm in diameter.
2.2 Statistical analyses
DFS and overall survival (OS) were calculated using the Kaplan–Meier method, and the prognostic effects of variables on DFS and OS were analysed using the log rank test and a Cox regression model [10,11].
3 Results
The results of postoperative pathological examination are shown in Table 1. Adenocarcinoma is dominant in this category of solid-type, small-sized NSCLC in our institute. Of the 118 patients, 101 (86%) had accurate clinical staging, though the remaining 14% were underestimated. Hilar and mediastinal lymph node metastases were found in 15 patients (13%) with PSL or SCDL NSCLC smaller than 2 cm. The primary tumour showed pure solid attenuation on CT in these patients with nodal involvement, and 15 (18%) of 83 patients with PSL smaller than 2 cm had lymph node metastases.
The 5-year DFS was 100% for c-stage IA patients with SCDL and 83% for c-stage IA patients with PSL (Fig. 1 ). Postoperative relapses were observed within 3 years after surgery except in one case, who had recurrence at the resected margin after segmentectomy 10 years after surgery. The 5-year OS was 100% for patients with SCDL and 87% for patients with PSL. Since no recurrence was found and the outcome was excellent in patients with SCDL in the present study, we focussed on the patients with PSL for further analyses.
Disease-free survival (DFS) for c-stage IA patients with solid-type, small-sized, NSCLC who underwent a complete resection. The 5-year DFS is 100% for c-stage IA patients with SCDL and 83% for c-stage IA patients with PSL. SCDL, solid component-dominant lesion; PSL, pure solid lesion.
Disease-free survival (DFS) for c-stage IA patients with solid-type, small-sized, NSCLC who underwent a complete resection. The 5-year DFS is 100% for c-stage IA patients with SCDL and 83% for c-stage IA patients with PSL. SCDL, solid component-dominant lesion; PSL, pure solid lesion.
Patients with PSL underwent a lobectomy in 63, segmentectomy in 17 and wedge resection in three cases. By analysing patients with PSL by pathological lymph node status, the 5-year DFS was 88% for patients with p-N0, 80% for patients with p-N1 and 46% for patients with p-N2 (Fig. 2 ). The postoperative outcome of patients with node metastases was worse than that for p-N0 patients (p = 0.0002). Similar results were obtained in the overall analysis, and the 5-year OS was 95% for p-N0 patients, 50% for p-N1 patients and 56% for p-N2 patients (p = 0.0009). Lymph node involvement was a significant prognostic factor for postoperative recurrence and survival in patients with pure solid, small-sized, lung cancer. The distribution of metastatic lymph node by the tumour location was upper + hilar/interlober/peripheral zones for eight tumours in the right upper lobe, upper + AP + hilar/peripheral zones for three tumours in the left upper lobe and upper + subcarinal + lower + hilar/interlober/peripheral zones for four tumours in the right lower lobe.
Disease-free survival (DFS) for c-stage IA patients with pure solid, small-sized tumour by histological lymph node status. The 5-year DFS for patients with p-N0, p-N1, and p-N2 was 88%, 80%, and 46%, respectively.
Disease-free survival (DFS) for c-stage IA patients with pure solid, small-sized tumour by histological lymph node status. The 5-year DFS for patients with p-N0, p-N1, and p-N2 was 88%, 80%, and 46%, respectively.
Survival was also assessed according to histological pleural invasion. The 5-year DFS for PSL patients was 87% with negative pleural invasion and 33% for positive pleural invasion (Fig. 3 ); the difference in DFS was significant (p = 0.0002), and pleural invasion had a clinically negative impact on postoperative recurrence. The overall survival analysis showed that 5-year OS for PSL patients was 91% with negative pleural invasion and 63% with positive pleural invasion (p = 0.06). Thus, pleural invasion was a significant prognostic factor for postoperative recurrence.
Disease-free survival (DFS) for c-stage IA patients with pure solid, small-sized tumour by pleural invasion. The 5-year DFS was 87% for patients with negative pleural invasion and 33% for patients with positive pleural invasion.
Disease-free survival (DFS) for c-stage IA patients with pure solid, small-sized tumour by pleural invasion. The 5-year DFS was 87% for patients with negative pleural invasion and 33% for patients with positive pleural invasion.
Elevated preoperative CEA levels (>5 ng ml−1) were observed in three of 33 SCDL patients (9.1%) and in 35 of 80 PSL patients (43.8%) (p = 0.0001, chi-square test). Among PSL patients, the 5-year DFS for patients with a high CEA level was 75%, which was slightly worse than that for patients with a normal CEA (88%), though the difference was not significant (p = 0.13).
Univariate analysis using age, sex, tumour size, operative procedure, histology, preoperative CEA level, lymph node metastasis and pleural invasion as the variables revealed that lymph node metastasis was a prognostic factor for both OS and DFS (Table 2 ) following complete resection of pure solid, small-sized, NSCLC using a proportional hazard model, though pleural invasion was a significant factor only for DFS (Table 2). Multivariate analysis was performed using tumour size and serum CEA level, which was identified to be prognostic factors in small-sized lung cancer in our previous study [12], in addition to lymph node metastasis and pleural invasion showing prognostic tendency or significance with univariate analysis in the present study. There was no significant correlation between lymph node metastasis and pleural invasion in the present study (p = 0.35 by chi-square test). While only node involvement significantly influenced overall survival (Table 3 ), both lymph node metastasis and pleural invasion were detected to be independent risk factors for recurrence (Table 3).
Univariate analyses of potential prognostic factors in patients with c-stage IA, pure solid, non-small cell lung cancer smaller than 2 cm.
Univariate analyses of potential prognostic factors in patients with c-stage IA, pure solid, non-small cell lung cancer smaller than 2 cm.
Multivariate analysis of potential prognostic factors in patients with c-stage IA, pure solid, non-small cell lung cancer smaller than 2 cm.
Multivariate analysis of potential prognostic factors in patients with c-stage IA, pure solid, non-small cell lung cancer smaller than 2 cm.
The initial recurrence site was also assessed during postoperative follow-up (Table 4 ). Of 15 patients, 13 (87%) had intrathoracic relapse, including hilar and mediastinal lymph nodes, pleural dissemination and pulmonary resected margin. The patients suffering relapse at the resected margin included a radical segmentectomy for tumour 1 cm in diameter in the anterior segment of the right upper lobe and a wedge resection for two patients. Distant metastasis was detected in the brain and bone in two patients as an initial relapse site. Two patients had simultaneous recurrence (lymph node + dissemination + lymphangitis, lymph node + pulmonary metastasis). Among 28 patients, who underwent a sublobar resection (23 segmentectomy and five wedge resection), resected margin recurrences were found in three (11%), while no relapse with lymph node metastasis was observed. All three patients suffering resected margin relapse underwent a completion lobectomy and survived for 31, 41 and 138 months without disease.
Initial recurrence site on postoperative follow-up.
Initial recurrence site on postoperative follow-up.
4 Discussion
Recently, radical sublobar resection, such as a segmentectomy or wide-wedge resection, has been recognised as an option for small-sized lung cancer [13–15], though a lobectomy with nodal dissection is still the gold standard for the treatment of resectable lung cancer [16]. However, patients with locally advanced disease with small-sized primary lesions, which usually show a solid component on CT, are occasionally encountered. Bronchioloalveolar adenocarcinoma showing a pure GGO lesion on CT imaging, which is histopathologically classified as Noguchi A or B type, is well understood to be a low-grade lung cancer [17,18]. A limited resection, such as segmentectomy or wide-wedge resection, could be curative for such slow-growing lung cancers. However, it has been reported that 10–15% of patients with small-sized lesions had nodal involvement [7–9], while we reported the node metastasis ratio as 15% among patients with NSCLC smaller than 2 cm [12]. Thus, we intended to address the characteristics of c-stage IA solid-type NSCLC, which might be also a candidate for sublobar resection, and identified the risk of lymph node metastasis and postoperative local recurrence in patients with a solid tumour. When conducting sublobar resection for such PSL patients, careful preoperative and intra-operative assessments for nodal metastasis and pleural invasion are mandatory.
Multivariate analysis revealed that both node involvement and pleural invasion were independent prognostic factors for recurrence in pure solid, small-sized NSCLC, though the previous study including small-sized lung cancer with GGO lesions showed a significant correlation between lymph node metastasis and pleural invasion [12]. The clinical influence of pleural invasion has been previously investigated and found to be a significant predictor of worse survival in T1 lung cancer [19]. In the soon-to-arrive TNM classification, lung cancer smaller than 2 cm is classified as T1a, though tumour invading visceral pleura is categorised as T2 [20]. The high risk of recurrence for patients with tumour invading visceral pleura (Fig. 3) provides support for the new TNM staging system, although the pleural factor was not identified as an independent predictor for OS in the present study. Pleural involvement might influence relapse rather than OS in small-sized pure solid lung cancer. Pleural invasion should be assessed using preoperative CT and given careful inspection during surgery. Macroscopically, severe pleural indentation or alteration could be considered pleural involvement and intra-operative rapid diagnosis with frozen section could be an optional to decide an extent of resection. Thus, a standard lobectomy can be recommended for pure solid, small-sized, lung cancer with pleural invasion, except in the clinical trial setting.
Radical sublobar resection requires intrapulmonary or hilar lymph node sampling during surgery to avoid underestimating nodal metastases. We usually dissect the lymph node adjacent to the segmental bronchus, which is considered to be the sentinel node of lymph drainage. The analyses of metastatic nodes according to tumour location showed the usual distribution in small-sized lung cancer; upper zone from upper lobe and subcarinal or lower zone from lower lobe in addition to hilar and intrapulmonary nodes. Fortunately, none of the patients in the present study treated with sublobar resection had lymph node recurrence. However, the concept of sentinel lymph node is still controversial because of the presence of variant lymphatic vessels that sometimes cause skip mediastinal metastases, especially in cases with pleural invasion [21,22]. Intra-operative evaluation of hilar and mediastinal nodes should be considered for sublobar resection as a minimum requirement in cases with pleural involvement.
With regard to the preoperative serum CEA level, no significant impact on DFS was found in PSL patients, while the rate of serum CEA elevation was evidently higher in PSL patients than in SCDL patients. Though our previous study analysing small-sized NSCLC including GGO lesions indicated the prognostic impact of prethoracotomy CEA level on postoperative survival and a relationship with lymph node metastasis [12], the clinical significance of the CEA level is unclear in patients with pure solid small-sized lesions.
Recurrence was found mainly in the thorax, including lymph nodes, pulmonary metastases or lymphangitis, pleural dissemination and resected margin within 3 years after operation. These results indicate that solid small-sized lung cancer might be cured when sufficient intervention to achieve local control is performed. According to ‘NCCN Clinical Practice Guidelines in Oncology, Non-small Cell Lung Cancer’, a contrast-enhanced chest CT every 6 months for 2 years, followed by non-contrast-enhanced CT annually, is recommended in completely resected stage I–IIIA NSCLC. This guideline may also be appropriate in solid small-sized NSCLC, in which the main focus of recurrence is the loco-regional thorax. Adjuvant therapy following surgery could be discussed in NSCLC patients with a small-sized solid tumour, though a further prospective clinical study is needed.
In conclusion, the characteristics of c-stage IA lung cancer with a pure solid or solid component-dominant lesion smaller than 2 cm were addressed using survival analysis of completely resected cases. A proportion of solid-type NSCLC has malignant potential with nodal or pleural involvement. Because sublobar resection for such solid small-sized lesion might have the risk of underestimation of nodal status, the feasibility should be evaluated in a prospective clinical trial. In addition, periodic surveys for local recurrence are recommended for 3 years following surgery in patients with solid-type lung cancer.
