Prognostic significance of vascular invasion in intermediate-grade subtype of lung adenocarcinoma

Abstract

Objectives

Although the recent reclassification of histological subtypes of lung adenocarcinoma reflects disease prognosis better, the prognosis of papillary and acinar-predominant adenocarcinoma, which are highly prevalent, is heterogeneity. The present study aimed to identify the prognostic indicators for papillary and acinar-predominant adenocarcinoma.

Methods

This retrospective study included 315 consecutive patients with completely resected pathological N0 lung adenocarcinoma tumors ≤3 cm from two institutions. Tumors were classified according to histologically predominant subtypes as low-grade (adenocarcinoma in situ, minimally invasive adenocarcinoma or lepidic predominant), intermediate-grade (papillary or acinar predominant) or high-grade (solid or micropapillary predominant). Prognostic factors in intermediate-grade group were assessed among clinicopathological factors of age, gender, surgical procedure, tumor size, pleural, lymphatic and vascular invasion using Cox proportion hazards analyses.

Results

There were 174 patients in the low-grade group, 109 in the intermediate-grade group and 32 in the high-grade group. The 3-year recurrence-free survival rates were 98.1%, 86.3% and 74.8% for these groups, respectively (P < 0.001). In the intermediate-grade group, the presence of vascular invasion was an independent prognostic factor on multivariate Cox regression analysis of recurrence-free survival (hazard ratio, 3.48; 95% confidence interval, 1.26–9.57, P = 0.01). Classification of intermediate-grade group based on vascular invasion revealed a clear division into favorable and unfavorable prognostic subgroups.

Conclusions

Consideration of the vascular invasion status in addition to the predominant subtype could provide a more accurate assessment of malignant aggressiveness and prognosis of patients with early-stage lung adenocarcinoma.

Introduction

Recent developments in imaging technology and the widespread use of computed tomography (CT) for screening have increased the probability of detecting small-sized non-small cell lung cancers (1). However, the survivals rates are not always high after the complete resection of such cancer (2). Therefore, reliable predictors of recurrence need to be identified to select patients who should receive adjuvant therapy.

Lung adenocarcinomas are clinically, molecularly, radiologically and pathologically heterogeneous (3). The 2015 World Health Organization (WHO) classification of lung tumors recently proposed a new histological classification for lung adenocarcinoma (4,5). Many studies have found that the prognosis of lung adenocarcinoma significantly correlates with tumor subtype defined according to this classification (6–14). In early-stage lung adenocarcinoma, the prognosis of patients with adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA) or lepidic-predominant types is good (3,9), whereas that of micropapillary or solid-predominant adenocarcinoma is poor (3,7,10) and that of patients with show papillary or acinar-predominant tumors is heterogeneity. Therefore, additional investigation of these two tumor types is important to identify patients at high risk of recurrence. The present two cancer center study aimed to identify additional prognostic factors in patients with papillary or acinar-predominant early-stage lung adenocarcinoma.

Patients and methods

Patients

We retrospectively reviewed data from 315 consecutive patients with completely resected pathological N0 lung adenocarcinoma with pathological tumor sizes ≤3 cm at Hiroshima University Hospital (Hiroshima, Japan) or Kanagawa Cancer Center (Yokohama, Japan) between April 2006 and December 2010. The inclusion criteria included curative surgery without neoadjuvant chemotherapy or radiotherapy, because neoadjuvant therapy could cause alterations in the predominant pattern. All patients were preoperatively assessed by using high-resolution CT and fluorodeoxyglucose-positron emission tomography/CT. Tumors were reviewed according to the tumor, node, and metastasis (TNM) Classification of Malignant Tumors, seventh edition (2). The clinicopathological data were retrospectively collected from medical records. Sub-lobar resection was considered in patients with peripheral tumors, which could be completely resected with appropriate surgical margins. Patients with ground glass opacity tumors determined by high-resolution CT that were regarded as node negative and non-invasive in a prospective study (15) underwent wedge resection without lymph node assessment. All patients who were treated by lobectomy and segmentectomy underwent systematic nodal dissection. All patients were followed up from the day of surgery. Post-operative follow-up procedures included a physical examination and chest radiography every 3 months and a CT examination every 6 months for the first 2 years. Thereafter, the patients were assessed by performing physical examinations and chest radiography every 6 months and an annual chest CT examination. Recurrence was defined when suspicious lesions are diagnosed as recurrent tumors by biopsy. If suspicious lesions were not diagnosed in this manner, recurrence was clinically and comprehensively defined from radiographic CT and fluorodeoxyglucose-positron emission tomography/CT findings.

The Institutional Review Boards of both participating institutions approved the present two cancer center analysis and waived the requirement for informed consent from individual patients.

Pathological investigation

Surgically resected specimens were routinely fixed in 10% buffered formalin, serially cut into 5–7-mm-thick slices, and macroscopically assessed. Specimens were histologically diagnosed using hematoxylin-eosin staining and if necessary, immunohistochemical staining. Each tumor was reviewed using comprehensive histological subtyping, and the ratio of each histological component was recorded in 5% increments according to the 2015 WHO classification (4) . The predominant subtype in each tumor was determined as the subtype with the highest occupancy and classified as AIS, MIA, lepidic, acinar, papillary, micropapillary or solid-predominant invasive adenocarcinoma. Tumor grades were also defined as low (AIS, MIA or lepidic predominant), intermediate (papillary or acinar predominant) and high (solid or micropapillary predominant) grades. Visceral pleural invasion (PL) is described as being either present (at least one tumor-cell cluster visualized in respective vessels) or absent according to Elastica van Gieson staining. Lymphatic invasion (LY) was assessed by staining with D2–40, and vascular invasion (V) was assessed using Elastica van Gieson or Victoria blue-hematoxylin-eosin staining (Fig. 1). These considered present if at least one tumor-cell cluster was visualized in the respective vessels.

Statistical analysis

Summarized data are presented as numbers or means ± standard deviation. Categorical variables and continuous variables were compared using the χ² test and an unpaired t-test, respectively. Recurrence-free survival (RFS) was defined as elapsed time from surgery to recurrence, death or the last follow-up. The 3-year RFS rate was calculated because the median follow-up did not reach 5 years. Survival data were calculated using Kaplan–Meier curves and compared using the log-rank test. P values and hazard ratios (HRs) in univariate and multivariate analyses were calculated using a Cox regression model and RFS was assessed by using both univariate and multivariate analyses. All data were statistically analyzed using EZR (Saitama Medical Centre, Jichi Medical University), which is a graphical user interface for R (The R Foundation for Statistical Computing, version 2.13.0).

Results

Clinicopathological characteristics and prognosis for all patients

Table 1 shows the characteristics of the 315 eligible patients. The median follow-up period was 40.3 (range, 4.9–70.5) months. A total of 153 patients underwent sub-lobar resection; 98 underwent wedge resection and 55 underwent segmentectomy. The status of PL, LY and V was positive in 38 (12%), 28 (9%) and 42 (13%) patients, respectively. The histological subtypes included 56 (18%) AIS, 15 (5%) MIA, 103 (33%) lepidic, 56 (18%) papillary, 53 (17%) acinar, 28 (9%) solid and 4 (1%) micropapillary-predominant invasive adenocarcinomas.

Figure 2 shows survival curves according to predominant subtypes. The 3-year RFS rates were 100%, 93.3%, 97.8%, 90.5%, 81.8% 74.7% and 75.0% for AIS, MIA, lepidic-, acinar-, papillary-, solid and micropapillary-predominant tumors, respectively (P < 0.001, overall; Fig. 2A). Although none of the patients with lepidic-predominant tumors, MIA or AIS developed recurrence, 1 of 15 patients with MIA and 2 of 105 patients with lepidic-predominant tumors died of other diseases. The 3-year RFS rates were 98.1%, 86.3% and 74.8% for low-, intermediate- and high-grade groups, respectively (P < 0.001, overall; Fig. 2B).

Multivariate analysis for RFS of all patients selected intermediate- or high-grade predominant subtypes [HR, 7.94; 95% confidence interval (CI), 2.15–29.32; P = 0.001; HR, 12.34; 95% CI, 2.88–52.85, P < 0.001, respectively) as independent prognostic factors (Table 2).

Prognostic factors for patients with intermediate-grade predominant subtypes

Table 3 shows the patients characteristics according to low- (n = 174; 55%) intermediate- (n = 109; 35%) and high- (n = 32; 10%) grade subtypes. The positivity rates for PL, LY and V were the highest among patients with the high-grade subtype, followed by intermediate- and low-grade subtypes in that order (P < 0.001 for all).

The RFS of patients with intermediate-grade subtypes was assessed using univariate and multivariate Cox regression analyses (Table 4). Univariate analysis for RFS identified the presence of PL, LY or V as poor prognostic factors. Multivariate analysis for RFS showed that only the presence of V was an independent predictor of poor prognosis (HR, 3.33; 95% CI, 1.13–9.82; P = 0.02). On the other hand, multivariate analysis of low- and high-grade subtypes did not identify any independent pathological prognostic factors including V (data not shown). Patients grouped according to the three subtypes were then categorized as having favorable or unfavorable outcomes groups according to the presence of V. The 3-year RFS of patients without and with V in low-, intermediate- or high-grade subtypes were 98.1% vs. 100% (P = 0.81), 92.4% vs. 64.8% (P < 0.001) and 76.5% vs. 73.3% (P = 0.83), respectively (Fig. 3). The combined 3-year RFS of patients who had intermediate-grade tumors with V and high-grade tumors was poorer than that of patients with intermediate-grade tumors without V and low-grade tumors (70.7% vs. 96.2%, P < 0.001; Fig. 4).

Discussion

The present study showed that V is significant prognostic factor for patients with intermediate-grade subtypes. Considering the V status, the prognosis of early-stage lung adenocarcinoma can be further sub-classified as favorable and unfavorable group. When assessing the prognosis of early-stage lung adenocarcinoma, the predominant subtype should firstly be emphasized, and then other factors should be considered according to predominant subtype. As with previous study (3,7,9,10), the present study also found that the prognoses were excellent for patients with low-grade small lung adenocarcinoma and unfavorable for those with high-grade tumors. However, the 3-year RFS was 86.3% for the group with intermediate-grade tumors, and these patients could be further categorized into groups with a favorable or an unfavorable prognosis. Multivariate analysis of patients with intermediate-grade tumors showed that the presence of V was an independent prognostic factor. This finding enabled a clear classification of patients with early-stage lung adenocarcinoma into favorable and unfavorable prognostic subgroups.

More accurate identification of patients at high risk of recurrence is important to select patients who should receive adjuvant therapy. However, adjuvant therapy for pathological stage I lung adenocarcinoma is controversial. In Japan, tegafur-uracil has proven useful for treating pathological stage IA lung cancers >2 cm, as well as stage IB (16,17). On the other hand, adjuvant chemotherapy is not commonly administered for stage IA, and stage IB with low risk according to the National Comprehensive Cancer Network guideline (18). However, the present study found that the prognosis for patients even with tumors <3 cm, intermediate-grade tumors with vascular invasion or high-grade tumors was unfavorable (3-year RFS, 70.7%). Patients with such tumors might be candidates for adjuvant therapy. Since previous studies (19–23) of adjuvant therapy did not consider predominant subtypes or malignant grade, the prognosis of patients with stage I lung cancer was heterogeneity. Therefore, adjuvant therapy might not confer a significant benefit upon these patients according only to the TNM classification. Consideration that not only the TNM classification but also the predominant subtype and/or other pathological factors are significantly associated with prognosis is important.

Interestingly, the present study also found that the impact of V on prognosis differed according to the predominant subtype. The presence of V is significant for patients with intermediate-grade tumors but not for those with low- or high-grade tumors. Vascular invasion notably affected the survival of patients with intermediate-grade tumors, whereas the prognosis did not significantly differ between patients with and without V in low- and high-grade tumors. Although previous studies (24–27) have identified the presence of V as an unfavorable prognostic factor, the association between V and predominant subtype has remained elusive. Only 3 of 174 patients with low-grade tumors had V, and none of them recurred. The behavior of low-grade tumors was clinically homogenous and the prognosis was excellent, therefore, no further classification was required. In addition, vascular invasion might have minimally affected the survival of patients with high-grade tumors. One possible explanation is that the histological subtype of solid- or micropapillary-predominant invasive adenocarcinomas that correspond to high-grade tumors are highly malignant and aggressive regardless of pathological malignant grade including lymphatic, blood vessel and PL. No pathological prognostic factors were identified for high-grade tumors in this study. However, since some high-grade tumors did not recur, further analysis is required to identify factors that could classify such tumors.

Neither PL nor LY were independent prognostic factors, which conflicted with previous findings (28,29). These factors are fundamental features of lymph node metastasis. Since this study targeted patients with small pathological N0 tumors, even if they were positive for PL or LY, invasion might have been minimal compared with more advanced tumors, such as those with lymph node involvement. We also considered that small tumors with absent or slight lymphatic or PL might be controlled by lymph node dissection. However, V is considered a fundamental feature of distant metastasis, which was not controllable only by surgery.

In summary, if vascular invasion status is considered in addition to predominant subtypes, prognosis or malignancy could be more accurately assessed in patients with early-stage lung adenocarcinoma. Although further prospective large cohort studies are required because this study was limited by a small cohort and a retrospective design, this finding is important for the stratification of patients at high risk of recurrence to identify candidates for post-operative adjuvant therapy.

Funding

This study was not funded by any source.

Conflict of interest statement

None declared.

References

Appendix

In addition to the authors listed on the first page, the following author contributed equally to this study.

Yasuhiro Tsutani and Takeshi Mimura: Department of Surgical Oncology, Hiroshima University, Hiroshima.