In the current study, the authors investigated whether thyroid transcription factor-1 (TTF-1) expression is correlated with the International Association for the Study of Lung Cancer (IASLC)/American Thoracic Society (ATS)/European Respiratory Society (ERS) classification and whether it stratifies patients with stage I lung adenocarcinoma with respect to disease recurrence.
Patients with stage I lung adenocarcinoma were classified according to the IASLC/ATS/ERS classification. Tissue microarrays were constructed and immunostaining for TTF-1 was performed. A total of 452 cases were available for analysis. Tumors were dichotomized based on the intensity of nuclear TTF-1 expression as negative (score of 0) or positive (score of 1–3). The cumulative incidence of recurrence (CIR) was used to estimate disease recurrence probabilities.
TTF-1 expression was identified in 92% of patients, including 100% of patients with minimally invasive or lepidic-predominant adenocarcinoma, 94% of patients with acinar-predominant adenocarcinoma, 98% of patients with papillary-predominant adenocarcinoma, 93% of patients with micropapillary-predominant adenocarcinoma, 86% of patients with solid-predominant adenocarcinoma, 67% of patients with colloid-predominant adenocarcinoma, and 47% of patients with invasive mucinous carcinoma. The CIR for patients with negative TTF-1 expression (n = 34 patients; 5-year CIR, 40%) was significantly higher than that for patients with positive TTF-1 expression (n = 418 patients; 5-year CIR, 15%) (P < .001). Among the patients with intermediate-grade tumors, the CIR for patients with negative TTF-1 expression (n = 16 patients; 5-year CIR, 45%) was significantly higher than that for patients with positive TTF-1 expression (n = 313 patients; 5-year CIR, 14%) (P < .001). On multivariate analysis, negative TTF-1 expression was found to be significantly correlated with an increased risk of disease recurrence (hazards ratio, 2.55; P = .009).
Lung cancer is the leading cause of death from cancer.1 Currently, tumor-lymph node-metastasis (TNM) stage is the most important prognostic factor for lung cancer.2 For patients with stage I lung cancer, however, survival outcomes remain variable.3 There is a need to refine the prognostic factors for early-stage lung cancer. The International Association for the Study of Lung Cancer (IASLC)/American Thoracic Society (ATS)/European Respiratory Society (ERS) has proposed a new classification for lung adenocarcinoma.4 Histologic subtyping according to this classification has been reported to have significant prognostic value.5–7
Thyroid transcription factor-1 (TTF-1), a homeodomain-containing nuclear transcriptional protein of the Nkx2 gene family, is expressed in epithelial cells of the fetal through the adult lung.8 TTF-1 is also expressed in lung carcinoma, and previous studies have reported a prognostic association with TTF-1 expression in patients with non-small cell lung cancer (NSCLC).9–27 The majority of these have demonstrated that a lack of TTF-1 expression is correlated with a worse prognosis. However, in these studies most cohorts were heterogeneous with regard to histology (adenocarcinoma and squamous cell carcinoma) and/or with regard to TNM stage (early and advanced). To the best of our knowledge, no study to date has specifically investigated the prognostic utility of TTF-1 expression using a uniform cohort of early-stage lung adenocarcinomas, and the association between TTF-1 expression and the IASLC/ATS/ERS classification.
In the current study, we attempted to determine whether TTF-1 expression is correlated with the IASLC/ATS/ERS classification and whether TTF-1 expression stratifies patients with respect to disease recurrence in cases of stage I lung adenocarcinoma.
MATERIALS AND METHODS
The current retrospective study was approved by the Institutional Review Board (WA0269-08). We reviewed all patients diagnosed with pathologic stage I solitary lung adenocarcinoma who underwent surgical resection at the study institution between 1995 and 2005. A total of 514 cases had tumor slides available for histologic evaluation. Among them, 471 cases had tumor blocks available for the construction of tissue microarrays. Clinical data were collected from the prospectively maintained database. Disease stage was based on the seventh edition of the American Joint Committee on Cancer staging manual.2
All available hematoxylin and eosin (H&E)-stained tumor slides (mean, 5 slides/case; range, 1-12 slides/case) were reviewed by 2 separate pathologists (KK and WDT) who were blinded to the patients' clinical outcomes using an Olympus BX51 microscope (Olympus Co., Tokyo, Japan) with a standard eyepiece measuring 22 mm in diameter. Discrepancies between the 2 pathologists in the assignment of the predominant subtype were later resolved by consensus on a multiple-headed microscope. The percentage of each histologic pattern was recorded in 5% increments. Tumors were classified according to the IASLC/ATS/ERS classification as adenocarcinoma in situ; minimally invasive adenocarcinoma (MIA); and invasive adenocarcinoma, which was subdivided into lepidic-predominant, acinar-predominant, papillary-predominant, micropapillary-predominant, solid-predominant, colloid-predominant, and invasive mucinous adenocarcinoma.4 Invasive mucinous adenocarcinoma was divided into 2 groups: pure mucinous (when having a >90% invasive mucinous pattern) and mixed mucinous/nonmucinous (when having at least 10% of each component).4 Tumors were grouped by architectural grading as low (adenocarcinoma in situ, MIA, or lepidic-predominant), intermediate (papillary-predominant or acinar-predominant), and high (micropapillary-predominant, solid-predominant, colloid-predominant, or invasive mucinous) grade.5,28
Nuclear features were examined with a high-power field (HPF) of × 400 magnification (0.237 mm2). Nuclear atypia was graded as previously reported: mild, moderate, and severe.29 Tumors were classified by mitotic count per 10 HPF as low (0-1); intermediate (2-4); and high (≥ 5).29 The following factors were also investigated: visceral pleural invasion2; lymphatic and vascular invasion; and the presence of necrosis. Lymphatic invasion was defined by the presence of tumor cells within an endothelium-lined space with lymphocytes. Vascular invasion was defined by the presence of tumor cells within blood vessels.
Formalin-fixed, paraffin-embedded tumor specimens were used for the construction of tissue microarrays. In brief, 4 representative tumor areas (2 from the most predominant histologic pattern and 2 from the second most predominant pattern) were marked on H&E-stained slides, and cylindrical 0.6-mm tissue cores were arrayed from the corresponding paraffin blocks into a recipient block by an automated tissue arrayer (ATA-27; Beecher Instruments, Sun Prairie, Wis), resulting in 7 tissue microarray blocks. In all, 452 patients had adequate cores available for immunohistochemical analysis.
Immunohistochemical Analysis and Scoring of TTF-1
In brief, 4 μm-thick sections from the microarray blocks were deparaffinized. Antigen retrieval was conducted using citrate buffer (pH 6.0). The standard avidin-biotin-peroxidase complex was used for immunostaining of anti–TTF-1 antibody (SPT24 [Novocastra Laboratories, Newcastle-upon-Tyne, UK], diluted at 1:50). Sections were stained using a Ventana Discovery XT automated immunohistochemical stainer (Ventana Medical Systems, Tucson, Ariz), in accordance with the manufacturer's guidelines. Normal lung tissues were stained as positive controls in parallel with the study tissues.
Because nuclear TTF-1 expression demonstrated a diffuse pattern and was expressed in at least 50% of the tumor area of each core, in the majority of cases (>90%) TTF-1 expression was evaluated based on the intensity of immunostaining. This intensity was scored as 0 (no expression), 1 (mild), 2 (intermediate), or 3 (strong) in each tumor core, as shown in Figure 1. The average intensity score for the tumor cores was considered to be the TTF-1 expression for each patient. On average, 3.2 tumor cores per patient were available for analysis.
Associations between clinicopathologic factors and TTF-1 expression were analyzed using the Fisher exact test for categorical variables and the Wilcoxon test for continuous variables.
Time-to-recurrence analyses were performed using competing risks methodology, which is the appropriate technique when a large percentage of patients die before experiencing disease recurrence. The Kaplan-Meier method estimates the probability of disease recurrence with the assumption that no deaths occur, which is unrealistic in this population of patients with early-stage disease, in whom a large number of deaths without documented recurrence were observed. Instead, cumulative incidence of recurrence (CIR) estimates the risk of disease recurrence by accounting for death as a competing event.30,31 Patients were followed from the time of surgery and censored if they were alive without documented disease recurrence at the time of the last follow-up.
We investigated the effect of clinicopathological factors and TTF-1 expression on CIR. Differences in CIR were assessed using the Gray method (for univariate nonparametric analyses) and the Fine-Gray method (for multivariate analyses).30,31 We first examined the univariate association between clinicopathologic factors and CIR to determine candidate variables for inclusion in a multivariate model.
All significance tests were 2-sided and used a 5% level of significance. Statistical analyses were conducted using SAS (version 9.2; SAS Institute Inc, Cary, NC) and R (R Development Core Team, Vienna, Austria) statistical software, including the “survival” and “cmprsk” packages.
Association Between Patient Clinicopathologic Factors and Disease Recurrence
The median age of all 452 patients was 69 years (range, 33 years-89 years). The majority of patients were women (63%) and had stage IA disease (68%). Approximately 84% of patients underwent lobectomy (Table 1).
Table 1. Association Between Clinicopathologic Factors and Disease Recurrence
Abbreviation: CIR, cumulative incidence of disease recurrence.
Significant P values (<.05) are shown in bold type.
Pathologic TNM stage
According to the histologic subtyping, 9 tumors were MIA (8 nonmucinous and 1 mixed mucinous/nonmucinous), 26 were lepidic-predominant tumors, 203 were acinar-predominant tumors, 126 were papillary-predominant tumors, 14 were micropapillary-predominant tumors, 56 were solid-predominant tumors, 3 were colloid-predominant tumors, and 15 were invasive mucinous tumors (6 pure mucinous and 9 mixed mucinous/nonmucinous tumors).
Seventy-three patients developed disease recurrence and 102 died of any cause without a documented recurrence. Twenty-two of the tumors recurred in the lung, 13 in the lymph nodes, and 38 in distant organs. The median follow-up for patients who did not develop disease recurrence was 57.3 months (range, 0.3 months-160.1 months). On univariate analysis, male sex (P = .010), sublobar resection (P = .008), higher stage of disease (stage IB; P < .001), higher architectural grade (P = .001), lymphatic invasion (P = .005), vascular invasion (P = .003), presence of necrosis (P < .001), greater nuclear atypia (P = .010), and higher mitotic count (P < .001) were associated with an increased risk of disease recurrence (Table 1).
Association Between TTF-1 and Histologic Subtype or Clinicopathologic Factors
With regard to the TTF-1 expression score, 34 tumors had a score of 0, 32 had a score of 1, 195 had a score of 2, and 191 had a score of 3 (Fig. 1). When stratifying tumors by TTF-1 score, the CIR for patients with a score of 0 was significantly higher (5-year CIR, 40%) than that for patients with a score of 1 (19%), 2 (17%), or 3 (12%) (P < .001) (Fig. 2A). On the basis of the 5-year CIR for each score, we decided to dichotomize TTF-1 expression into negative (score of 0) versus positive (score of 1-3).
All MIA and lepidic-predominant tumors demonstrated TTF-1 expression. TTF-1 expression was present in 94% of acinar-predominant tumors (190 of 203 tumors), in 98% of papillary-predominant tumors (123 of 126 tumors), in 93% of micropapillary-predominant tumors (13 of 14 tumors), in 86% of solid-predominant tumors (48 of 56 tumors), in 67% of colloid-predominant tumors (2 of 3 tumors), and in 47% of invasive mucinous tumors (7 of 15 tumors) (Fig. 3). Of the MIA tumors, 100% of nonmucinous MIA tumors (7 of 7 tumors) demonstrated strong TTF-1 expression, and 1 mixed mucinous/nonmucinous MIA tumor showed intermediate expression. Strong TTF-1 expression was identified in 65% of lepidic-predominant tumors (17 of 26 tumors), in 47% of acinar-predominant tumors (96 of 203 tumors), in 42% of papillary-predominant tumors (53 of 126 tumors), in 14% of micropapillary-predominant tumors (2 of 14 tumors), and in 23% of solid-predominant tumors (13 of 56 tumors). Of the invasive mucinous tumors, positive TTF-1 expression was identified in 50% of pure mucinous tumors (3 of 6 tumors) and in 44% of mixed mucinous/nonmucinous tumors (4 of 9 tumors). Strong TTF-1 expression was not identified in pure mucinous tumors, but it was identified in 11% of mucinous/nonmucinous tumors (1 of 9 tumors) and in 33% of colloid-predominant tumors (1 of 3 tumors). TTF-1 expression was most frequently identified in low-grade tumors (100%), followed by intermediate-grade tumors (95%) and high-grade tumors (80%) (P < .001) (Table 2).
Table 2. Association Between TTF-1 Expression and Clinicopathologic Factors
Significant P values (<.05) are shown in bold type.
Tumor size, cm
Pathologic TNM stage
Negative TTF-1 expression was found to be significantly correlated with a higher mitotic count (median, 7; range, 0-39) compared with positive expression (median, 2; range, 0–43) (P = .002) (Fig. 4).
Negative TTF-1 expression was also associated with larger tumor size (P = .003), higher stage of disease (stage IB; P = .020), and the presence of necrosis (P < .001) (Table 2).
Association Between TTF-1 Expression and CIR
The CIR for patients with negative TTF-1 expression (n = 34 patients; 5-year CIR, 40%) was found to be significantly higher than that for patients with positive TTF-1 expression (n = 418 patients; 5-year CIR, 15%) (P < .001) (Fig. 2B). This result was confirmed in a subgroup analysis limited to the 379 patients who underwent lobectomy. The CIR for patients with negative TTF-1 expression (n = 27 patients; 5-year CIR, 39%) was found to be significantly higher than that for patients with positive TTF-1 expression (n = 352 patients; 5-year CIR, 13%) (P < .001).
Among patients with intermediate-grade tumors, the CIR for those with negative TTF-1 expression (n = 16 patients; 5-year CIR, 45%) was significantly higher than that for patients with positive TTF-1 expression (n = 313 patients; 5-year CIR, 14%) (P < .001) (Fig. 2C). Among patients with high-grade tumors, the CIR for those with negative TTF-1 expression (n = 18 patients; 5-year CIR, 35%) was higher than that for patients with positive TTF-1 expression (n = 70 patients; 5-year CIR, 23%), although the difference was not statistically significant (P = .44).
Among patients with acinar-predominant tumors, the CIR for those with negative TTF-1 expression (n = 13 patients; 5-year CIR, 38%) was higher than that for patients with positive TTF-1 expression (n = 190 patients; 5-year CIR, 14%) (P = .008). However, with regard to the other histologic subtypes, the small sample sizes prevented comparisons.
On multivariate analysis patients with negative TTF-1 expression remained at a significantly increased risk of disease recurrence (hazards ratio [HR], 2.55; P = .009) (Table 3). Among patients with intermediate-grade tumors, TTF-1 expression was found to be an independent predictor of recurrence (HR, 3.84; P = .002). However, among patients with high-grade tumors, TTF-1 expression did not appear to influence the risk of recurrence (HR, 1.60; P = .36).
Table 3. Results of Multivariate Cox Proportional Hazards Model
Significant P values(<.05) are shown in bold type.
TTF-1 expression Negative vs positive
High vs intermediate
High vs low
Surgical procedure Sublobar vs lobar
Pathologic TNM stage IB vs IA
Lymphatic invasion Present vs absent
The results of the current study demonstrate that a lack of TTF-1 expression is identified more frequently in patients with high-grade tumors and is an independent predictor of disease recurrence in patients with stage I lung adenocarcinoma, especially those with intermediate-grade tumors.
Several studies have reported no association between TTF-1 expression and lung adenocarcinoma differentiation.17,20 However, in the current study, an inverse association was found between TTF-1 expression and architectural grade based on the predominant tumor subtype. To our knowledge, the association between TTF-1 expression and the IASLC/ATS/ERS classification has not been previously investigated in patients with stage I lung adenocarcinoma, although studies using small cohorts of patients have suggested an association between higher TTF-1 expression and the lepidic pattern.14,16,21 In addition, TTF-1 expression has been reported to be correlated with a lower Ki-67 proliferation index in patients with NSCLC.10,12 In the current study, TTF-1–positive tumors were found to be significantly correlated with a lower mitotic count and smaller tumor size.
Tumors formerly classified as mucinous bronchioloalveolar carcinoma, which is considered to be an invasive mucinous adenocarcinoma according to the IASLC/ATS/ERS classification, demonstrate no or less-frequent TTF-1 expression.14,32-35 Pure mucinous tumors may have no or very low TTF-1 expression, and mixed mucinous/nonmucinous tumors express TTF-1 more frequently than pure mucinous tumors.34 Similarly, in the current study, strong TTF-1 expression was not observed in patients with pure mucinous tumors.
To our knowledge, 19 studies investigating the association between TTF-1 expression and survival have been published to date.9-27 TTF-1–positive tumors were associated with better survival in 12 studies11,13,16,17,19-22,24–27 and with worse survival in 1 study,9 and were found to have no association with survival in 6 studies.10,12,14,15,18,23 The majority of these studies used cohorts that were heterogeneous with regard to histology and/or TNM stage. One study investigated the prognostic significance of TTF-1 expression in a more uniform cohort that was composed of patients with early-stage adenocarcinomas, although the study cohort was collected very selectively, including 50 patients with bronchioloalveolar carcinoma and 50 with conventional invasive adenocarcinoma.16 In a meta-analysis of 10 eligible studies published up to and including 2005, the combined HR for TTF-1 expression in patients with adenocarcinoma was 0.53 (95% confidence interval, 0.29–0.95).36 The current study, which is composed of a uniform, large cohort of patients with stage I lung adenocarcinoma, confirmed the prognostic significance of TTF-1 expression.
In previous studies in which survival analysis was performed, positive TTF-1 expression has been defined by various methods. Several studies used the percentage of positive tumor cells, with a cutoff value of 1% to 75%.9–13,15,17–19,23,26,27 Other studies used staining intensity alone21 or a combination of the percentage of positive tumor cells and staining intensity.16,24,25 Because of the high percentage of tumors that are positive for TTF-1, we used staining intensity only to classify the degree of TTF-1 expression, and we demonstrated the CIR differences using 4 groups of TTF-1 expression intensity, even though we finally dichotomized tumors into positive and negative. One limitation of the current study, which used tissue microarray analysis, is that TTF-1–negative tumors might be considered to be focally positive if stained with whole-tissue blocks. However, as we have previously reported, when whole-tissue blocks were used, TTF-1 positivity was predominantly bimodal: either diffusely positive (84%) or completely negative (11%). Furthermore, the total positive rate in the whole-tissue block study (89%) was similar to that in the current study using tissue microarray analysis (92%).37 Therefore, we believe that our conclusion would not change dramatically even if TTF-1 negativity were confirmed using whole-tissue block.
Several studies in which multivariate analysis was performed have demonstrated that a lack of TTF-1 expression is an independent predictor of worse prognosis.17,20–22,25,27 However, to our knowledge, it remains unclear whether a lack of TTF-1 expression remains an independent prognostic predictor, even after adjusting for the IASLC/ATS/ERS classification. In the current study, negative TTF-1 expression was found to be an independent predictor of disease recurrence after adjusting for the IASLC/ATS/ERS classification.
One limitation to using the IASLC/ATS/ERS classification is that the majority of patients (73% in the current study) are classified as having intermediate-grade tumors. Therefore, it is necessary to recognize poor prognostic factors for this group of patients. In the current study, TTF-1 expression was found to stratify intermediate-grade tumors into 2 groups with respect to disease recurrence.
In patients with lung adenocarcinoma, the morphologic feature (histologic subtype) appears to be correlated with a specific molecular expression (TTF-1). For patients with stage I lung adenocarcinoma, TTF-1 expression is an independent predictor of disease recurrence. Since the morphologic assessment of H&E-stained slides and immunohistochemical analysis have become routine in clinical practice, prognostic stratification using the IASLC/ATS/ERS classification and TTF-1 immunohistochemistry can be readily implemented in the treatment of patients with lung adenocarcinoma.
We thank Joe Dycoco for his help with the lung adenocarcinoma database in the Division of Thoracic Service in the Department of Surgery at Memorial Sloan-Kettering Cancer Center, Avani Giri and Louie Lopez for their help making the tissue microarray, Irina Linkov for her technical assistance with the immunohistochemical analysis, and David Sewell for his editorial assistance.
Supported in part by the International Association for the Study of Lung Cancer Young Investigator Award; National Lung Cancer Partnership/LUNGevity Foundation Research Grant; American Association for Thoracic Surgery Third Edward D. Churchill Research Scholarship; Mesothelioma Applied Research Foundation grant in memory of Lance S. Ruble; William H. Goodwin and Alice Goodwin, the Commonwealth Foundation for Cancer Research and the Experimental Therapeutics Center; New York State Empire Clinical Research Investigator Program; the National Cancer Institute (grants R21CA164568, R21CA164585, U54CA137788, and U54CA132378); and the US Department of Defense (grant LC110202 and PR101053).