Thyroid transcription factor-1 (TTF-1) is expressed in thyroid follicle, parathyroid gland, alveolar epithelium, and diencephalon, and participates in the differentiation, development, and functional maintenance of these organs (1-3). A number of studies have shown that TTF-1 expression is associated with improved survival in patients with advanced non-squamous non-small cell lung cancer (non-sq-NSCLC) (4-6). More interesting, expression of TTF-1 has been shown to be associated with the efficacy of several antitumor drugs and progression-free survival (PFS) (7-15). Of note, non-sq-NSCLC patients with TTF-1 expression have been shown to have a better prognosis than those without expression, responding to pemetrexed (PEM)-containing chemotherapy (7-15). Due to its high response rate and low incidence of serious side effects, PEM has become one of the essential antineoplastic agents for non-sq-NSCLC in routine clinical practice (7-15). Some of these previous studies examined the correlation between the expression of TTF-1 and PFS of patients treated with PEM-containing chemotherapy including non-sq NSCLC patients with driver mutations, such as epidermal growth factor receptor (EGFR) mutation (8,10-12,14,15). However, none of them analyzed the association between TTF-1 expression and PEM response considering the presence or absence of the driver gene. Driver gene-positive NSCLC, such as those carrying an EGFR mutation or an ALK rearrangement, can be treated with specific responsive tyrosine kinase inhibitors (TKIs). In such driver gene-positive patients, TKIs are prescribed as the first-line therapy. Therefore, the treatment sequence for these patients is clearly different from that for patients with driver gene-negative NSCLC. It is also unclear how the presence or absence of the driver gene is related to response to PEM-containing chemotherapy.

Therefore, we conducted this study to clarify whether the presence or absence of the driver gene affects the relationship between TTF-1 expression and response to PEM in non-sq-NSCLC patients.

This study examined the medical records of all patients diagnosed with NSCLC between April 2009 and June 2022 at three tertiary hospitals: Mito Medical Center, University of Tsukuba, Ryugasaki Saiseikai General Hospital, and University of Tsukuba Hospital. NSCLC was pathologically diagnosed using the World Health Organization classification with tumor node metastasis staging (TNM Classification, eighth Edition) (16). Staging was performed from computed tomography/magnetic resonance imaging of the head, ultrasonography/computed tomography of the abdomen, and bone scans. Patient demographics including age, sex, histopathology, disease stage and survival were extracted from their medical records. We confirmed the presence or absence of driver genes. The study included patients with non-sq-NSCLC receiving first-line or later lines of PEM-containing chemotherapy.

Expression of TTF-1 was assessed by immunohistochemical staining using a rabbit monoclonal antibody (1:250; Abcam, Cambridge, UK). We evaluated the degree of positivity as follows: negative (-, 0%); very focally positive (+, <25% in hot spots); focally positive (++, <50% in hot spots); diffusely positive (+++, >50% and a uniform pattern). In this study, only ‘diffusely positive’ were treated as TTF-1-positive.

Statistical analysis. Nominal variables were compared using a chi-squared test and values with an unknown population variance were compared using a nonparametric Mann-Whitney test. PFS was defined as the period from randomization to disease progression or death from any cause. PFS was calculated with a Kaplan-Meier analysis and compared using a log-rank test. Cox proportional hazards modeling with the forward-backward stepwise method was used to identify the independent variables to be included in the final model, with PFS as the dependent variable. Multivariable analyses included only variables with a p-value of less than 0.1 in univariate analysis. All statistical analyses were conducted using SPSS version 23 (IBM Corporation, Armonk, NY, USA). A p-value of less than 0.05 was considered significant.

Ethics. This study complied with the Ethical Guidelines for Clinical Studies issued by the Ministry of Health, Labor, and Welfare of Japan. Written informed consent to participate in a non-interventional retrospective study was obtained from each patient. The Mito Medical Center-University of Tsukuba Hospital Ethics Committee approved the examination of medical records for the purpose of this study.

Clinicopathological features of the patients. Figure 1 shows the study flow chart. During the study period, 1,203 non-sq-NSCLC patients were diagnosed. Among them, 339 patients were treated with PEM-containing chemotherapy. Of these, 250 patients underwent TTF-1 immunostaining before treatment. Of the 250 patients, 185 were negative for any driver gene, and 65 were positive for a driver gene (54 EGFR gene positive, 11 other driver gene positive). Table I shows the clinicopathological features of TTF-1-positive and -negative driver gene-negative patients. There were 185 driver gene-negative patients, which were treated with PEM-containing chemotherapy. Among the 185driver gene-negative patients with non-sq NSCLC, 98 (53.0%) were TTF-1 positive. Fifty-three (81.5%) of 65 driver gene-positive patients were TTF-1 positive: 42 (77.8%) of 54 EGFR-positive patients and all six patients carrying the ALK rearrangement were TTF-1 positive.

In this study, we designated ‘driver gene- negative-TTF-1 positive patient group’, ‘driver gene- negative-TTF-1 negative patient group’, ‘driver gene- positive-TTF-1 positive patient group’, and ‘driver gene- positive-TTF-1 negative patient group’ as Group A, Group B, Group C, and Group D, respectively.

Comparison of PFS between driver dene negative patients with or without TTF-1. We first compared the clinicopathological features of Group A and Group B patients. As shown in Table I there was no statistically significant difference between them, so we compared PFS of these groups. Figure 2A shows the Kaplan-Meier curves of these two groups of patients. PFS of Group A patients was significantly longer than that of Group B patients (p=0.001) (median PFS: 9.0 and 4.0 months). In the uni- and multivariate analyses, TTF-1 positivity was confirmed to be a significant favorable factor in patients with driver gene-negative non-sq NSCLC (Table II).

Comparison of PFS between driver gene- positive patients with or without TTF-1. Table II shows the clinicopathological features of Group C and Group D patients. We confirmed that there was no significant difference in the features of these patients, and then compared PFS between driver gene positive patients with or without TTF-1. As shown in Figure 2B, there was no statistically significant difference in PFS between Group C and Group D patients (p=0.2805) (median PFS: 7.0 and 5.0 months). In 54 EGFR-positive patients, there was no statistically significant difference in PFS between TTF-1-positive and -negative patients (p=0.6139) (median PFS: 5.0 and 4.5 months). In univariate analyses, none of the factors favored PFS in patients with driver gene-positive non-sq NSCLC (Table II).

Comparison of PFS between driver gene- positive and negative patients with or without TTF-1. Table III shows the clinicopathological features of driver gene-negative (Group A and B) and driver gene-positive patients (Group C and D). There was no significant difference in the features of these patients except for sex and TTF-1. As ‘sex was not a factor affecting PFS in neither the driver gene-negative group nor the driver gene-positive group (Table II), we compared PFS of driver gene-negative patients with that of driver gene-positive patients. As shown in Figure 3A, there was no significant difference between these two groups (p=0.9955). No significant difference in PFS was observed between Groups A, B and C (p=0.1185). However, a statistically significant difference in PFS was confirmed between ‘Group A+Group B+Group C’ and Group D (p=0.001) (Figure 3B).

In this study, the following four results were obtained. First, in 165 driver gene-negative patients, the PFS of TTF-1-expressing patients treated with PEM-containing chemotherapy was significantly longer than that of TTF-1-negative patients, as observed in previous reports (7-15). In uni- and multivariate analysis, the expression of TTF-1 was confirmed as a significant favorable factor for PFS of patients treated with PEM-containing chemotherapy. Second, however, in 65 driver gene-positive patients, PFS of those treated with PEM-containing chemotherapy was not significantly different from that of TTF-1-negative patients. In 54 EGFR-positive patients, the PFS of TTF-1-positive patients treated with PEM-containing chemotherapy was not significantly different from that of TTF-1-negative patients. Third, there was no significant difference in PFS following treatment with PEM-containing chemotherapy between driver gene-negative and driver gene-positive patients. Fourth, the comparison of PFS of patients treated with PEM-containing chemotherapy and grouped according to the presence of driver gene and expression of TTF-1indicated that PFS was shorter in ‘driver gene-negative and TTF-1-negative’ patients.

Table IV shows the driver gene and positivity for TTF-1 in previous studies (8,10,11,13,14,17-27). The positivity rates varied since the TTF-1 antibody used and the threshold for positive staining were different in each study, but the positive rate in driver gene-negative patients is generally lower than that in driver gene-positive patients (8,10,11,13,14,17-27). This result is consistent with our present study. Regarding driver gene-positive patients, the TTF-1 expression was evaluated only in patients with EGFR mutation and those carrying the ALK rearrangement as shown in Table IV. To the best of our knowledge, there is no study on TTF-1 expression in patients with other driver genes than EGFR mutation and ALK rearrangement. Some studies have investigated the relationship between TTF-1 expression and PEM-containing chemotherapy in driver gene-negative patients, but no study examined this relationship in driver gene-positive patients.

In our analysis of 65 driver gene-positive patients (54 of whom were EGFR mutation-positive), no significant difference was found in PFS between TTF-1-positive (median PFS: 7.0 months) and TTF-1-negative patients (median PFS: 5.0 months) treated with PEM-containing chemotherapy. In a review by Han et al., the median PFS of patients treated with PEM-containing chemotherapy after EGFR-TKI failure was 5.09 months (28). Kaneda et al. reported that in first-line cisplatin plus PEM treatment of 78 EGFR-positive patients, the median PFS in the group carrying the L858R mutation and the group with Ex 19 del was 9.4 and 5.5 months, respectively (29). Considering these research results, PFS of patients treated with PEM-containing chemotherapy is considered to reflect real clinical practice. In addition, median PFS of driver gene-negative and driver gene-positive patients treated with PEM-containing chemotherapy was 6.0 and 6.0 months, respectively, and there was no significant difference between them. PEM has often been administered as first-line therapy in driver gene-negative patients, but it is also prescribed as second-line or later therapy in driver gene-positive patients. Even after driver-gene-specific TKI therapy, the efficacy of PEM-containing chemotherapy in patients with driver gene-positive non-sq NSCLC might be as good as that given as first-line to patients with driver gene-negative non-sq NSCLC.

TTF-1 has been associated with PFS of driver gene-negative patients treated with PEM-containing chemotherapy, but not in driver gene-positive patients. The reason for this difference is not clear. In our study, there was no statistically significant difference in PFS among Groups A, C, and D. There was a significant difference in PFS between these three groups and the driver gene-negative-TTF-1-negative group (Group B). Considering these results together, it is inferred that only Group B shows poor PFS following PEM-containing chemotherapy. In relation to poor PFS in driver gene-negative TTF-1 negative non-sq NSCLC patients, there are some interesting reports (30,31). Tanaka et al. reported that sergycin, a chondroitin sulfate proteoglycan, played a pivotal role in tumor-stromal interaction and reprogramming into an aggressive and immunosuppressive tumor microenvironment in TTF-1-negative lung adenocarcinoma (30). In a report by Matsubara et al., immunohistological expression of TFF-1 correlated with tumor spread through air spaces in the lung cancer lesion, and a poor prognosis in advanced stages (31). However, the mechanism is still unknown.

This study had some limitations. First, although there was no intentional selection bias, it must be pointed out that these were the results of the analysis of the data of patients who were evaluated for TTF-1, and not the results of all patients during the study period. Although patients from three institutions were collected the number is small and the study retrospective. Due to the relationship between the time when this study was conducted and when the driver gene tests became available, it is possible that we were not able to identify driver gene-positive patients other than those carrying the EGFR mutation or the ALK rearrangement. In TTF-1 immunohistochemical staining, only “uniform pattern of 50% or more” was regarded as “positive” in this study. This might have influenced the results. It is hypothesized that overall survival might differ between driver gene-positive and -negative patients due to differences in TTF-1, but this study did not analyze overall survival. Based on these limitations, it is expected that further studies will be conducted to confirm and expand our results.

In driver gene-positive non-sq NSLC patients, the presence or absence of TTF-1 was found to have little effect on the response to PEM-containing chemotherapy. In driver gene-positive non-sq NSLC patients, even second-line or later treatment after specific-TKI treatment, PEM-containing chemotherapy could be expected to have the same level of efficacy as first-line PEM-containing chemotherapy in driver gene-negative-TTF-1-positive patients. Regarding PEM-containing chemotherapy in driver-gene positive non-sq NSCLC patients, a similar effect could be expected regardless of the expression of TTF-1.

The Authors have no conflicts of interest to declare in relation to this study.

SO and HS contributed to the planning, and OS, MK, ST and HS collected data. SO and KM conducted the design and acquisition of data and drafting the manuscript. HS and NH supervised the manuscript. All Authors read and approved the final manuscript.