- Email: [email protected]
Characteristics and Outcome of ROS1-Positive Non–Small Cell Lung Cancer Patients in Routine Clinical Practice
ORIGINAL ARTICLE
Characteristics and Outcome of ROS1-Positive Non–Small Cell Lung Cancer Patients in Routine Clinical Practice Sehhoon Park, MD,a Beung-Chul Ahn, MD,b Sung Won Lim, MD,a Jong-Mu Sun, MD, PhD,a Hye Ryun Kim, MD, PhD,b Min Hee Hong, MD,b Se-Hoon Lee, MD, PhD,a Jin Seok Ahn, MD, PhD,a Keunchil Park, MD, PhD,a Yoon La Choi, MD, PhD,c Byoung Chul Cho, MD, PhD,b Myung-Ju Ahn, MD, PhDa,* a
Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea b Yonsei Cancer Center, Division of Medical Oncology, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea c Department of Pathology and Translational Genomics, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea Received 22 April 2018; revised 11 May 2018; accepted 12 May 2018 Available online - 5 June 2018
ABSTRACT Introduction: ROS1-rearranged NSCLC is classified as a distinct molecular subset of NSCLC with a therapeutic target. ROS1 rearrangement is most often identified in never-smokers with adenocarcinoma and EGFR and ALK receptor tyrosine kinase gene (ALK) wild type. Treatment with tyrosine kinase inhibitors (TKIs), which target the ROS1 kinase domain, is considered the standard of care. TKIs have been shown to have a robust and durable response. However, information regarding the clinical outcomes of TKI treatment, including brain metastasis, remains limited. Methods: We identified 103 consecutive cases of ROS1positive NSCLC by using break-apart fluorescence in situ hybridization (n ¼ 84), next-generation sequencing (n ¼ 23), or both (n ¼ 3). Information regarding fusion breakpoints was available for eight patients. Clinical data, including patient characteristics, incidence of brain metastasis, response to chemotherapy, or to TKIs, were retrospectively analyzed. Results: The median patient age was 56 years, and 58.9% of the patients were female. Most of the patients (75.7%) were never-smokers. Adenocarcinoma was predominant (98.1%), and two cases with pleomorphic carcinoma were identified. Sixty percent of patients had an extrathoracic metastatic lesion, and 22% had an intracranial lesion at the initial presentation or at the time of recurrence. The median time to development of brain metastases was 12.0 months (range 2.1–84.1). The most common fusion partner was CD74 molecule gene (CD74), followed by syndecan 4 gene (SDC4),
ezrin gene (EZR), tropomyosin 3 gene (TPM3), TRK-fused gene (TFG), zinc finger CCHC-type containing 8 gene (ZCCHC8), sacrolemma associated protein gene (SLMAP), and myosin VC gene (MYO5C). All of these fusion partners preserved the tyrosine kinase domain of ROS1. The median overall survival time was 52.1 months (95% confidence interval [CI]: 23.6–not reached). In the 90 patients who were treated with pemetrexed-based chemotherapy, the overall response rate and progression-free survival time were 53.3% and 8.0 months (95% CI: 6.4–11.7), respectively. The overall response rate and progression-free survival time were 70.7% and 12.7 months (95% CI: 8.1–21.8), respectively, for the 50 patients treated with TKIs. Brain metastasis was more often observed during TKI treatment (15.5%) than during pemetrexed-based chemotherapy (6.7%).
*Corresponding author. Drs. Park, Ahn, and Lim equally contributed to this work. Disclosure: Dr. Lee reports personal fees from AstraZeneca/ MedImmune, Roche, Bristol-Myers Squibb, Merck, and Novartis outside the submitted work. Dr. Ahn reports personal fees from Amgen, Pfizer, AstraZeneca, Menarini, Roche, Eisai, Boehringer Ingelheim, BMS, and Janssen outside the submitted work. The remaining authors declare no conflict of interest. Address for correspondence: Myung-Ju Ahn, MD, PhD, Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul 06351, Republic of Korea. E-mail: silkahn@skku. edu or [email protected]. ª 2018 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 https://doi.org/10.1016/j.jtho.2018.05.026
Journal of Thoracic Oncology
Vol. 13 No. 9: 1373-1382
1374 Park et al
Journal of Thoracic Oncology
Conclusions: ROS1-positive NSCLC has distinct clinical characteristics, with an effective and durable response to both TKIs and pemetrexed-based chemotherapies. Regardless, given its novel characteristics and distinct clinical responses to conventional chemotherapies and TKIs, the treatment strategy for ROS1-positive NSCLC remains to be further developed. 2018 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: ROS1; Crizotinib; Non–small cell lung cancer
Pemetrexed;
Carcinoma;
Vol. 13 No. 9
therapy allowed up to 9 months of PFS in ROS1-positive NSCLC.14,15 ROS1-positive patients have lower expression of thymidylate synthetase (TS) than ROS1-negative patients did; this expression pattern is thought to explain the high efficacy of pemetrexed in ROS1-positive patients.14 In contrast, patients with ALK-positive NSCLC with similar levels of TS expression have a shorter PFS (range 4.2–7.0 months).16,17 Despite the many clinical investigations of TKIs in patients with ROS1-rearranged NSCLC, information regarding its clinical outcomes remain limited.18,19 In this study, we have therefore reported the comprehensive analysis of clinical data from ROS1-positive patients in Korea.
Introduction ROS1 is a receptor tyrosine kinase of the insulin receptor family. Its rearrangement is recognized as a targetable oncogenic driver that accounts for up to 2% of NSCLCs.1,2 The ROS1 fusion protein maintains the kinase domain in the 30 region of ROS1. It has diverse counterparts, including CD74 molecule, ezrin, syndecan 4, solute carrier family 34 member 2, tropomyosin 3, interleukin 4 induced 1 (also known as FIG1), cell division cycle 6, KDEL endoplasmic reticulum protein retention receptor 2 at 50 terminus.3,4 Interestingly, the ROS1 and ALK receptor tyrosine kinase (ALK) domains have highly conserved adenosine triphosphate–binding domains, with nearly 77% homology of their amino acid sequences.5,6 This structural similarity enables the ALK inhibitor to bind to the ROS1 protein even more strongly than it does to the ALK protein.6 This phenomenon explains the promising outcomes demonstrated by use of an ALK inhibitor in the treatment of ROS1-rearranged NSCLC. In the PROFILE 1001 study, crizotinib, which is a potent TKI for ALK, ROS1, and mesenchymal-epithelial transition, demonstrated an overall response rate (ORR) of 72%, with a median of 19.2 months of progression-free survival (PFS).7 Ceritinib, which is a more potent ALK TKI, has demonstrated an ORR of 62% with a PFS of 19.3 months in crizotinib-naive patients.8 A number of TKIs with ROS1 inhibition are currently in clinical development.9–11 On the basis of a few clinical experiences, there have been reports of development of an acquired resistance mechanism after crizotinib therapy in the PROFILE 1001 study. One such mechanism of resistance is ROS1 G2032R in the CD74 molecule–ROS1 fusion protein located at the tyrosine kinase domain.12 Another characteristic of ROS1-positive NSCLC is a durable response to pemetrexed. Pemetrexed and platinum doublets followed by pemetrexed maintenance therapy has allowed approximately 4 months of PFS in patients with advanced nonsquamous NSCLC with no targetable genomic alterations.13 In contrast, the same
Methods Patients and Samples We identified 103 patients with NSCLC with ROS1 gene rearrangements who were treated at Samsung Medical Center and Severance Hospital between January 2001 and February 2018. Clinical data, including patient characteristics, the incidence of brain metastasis, and responses to chemotherapy or TKIs, were retrospectively analyzed. Because of local regulation by the health authority, according to which off-label use of chemotherapeutic agents (including targeted agents) in not allowed, TKIs are available only for clinical trials. Hence, the sequence of treatment was decide by the availability of clinical trials. Patients were evaluated for treatment response either by simple chest radiography or chest computed tomography scan. Chest posterior-anterior radiography was performed after every cycle and chest computed tomography was performed after every two or three cycles of treatment. In addition to regular followup, additional imaging was conducted at the physician’s discretion. The response rate was calculated by adding the percentage of patients with a partial response and the percentage with a complete response to the treatment. This study was approved by an institutional review board (institutional review board approval No. 2018-04-012-002 and No. 4-2018-0195). Informed consent was waived.
Statistical Analysis PFS was calculated from the start of treatment to radiographic progression or death. Otherwise, patients were censored from analysis. Overall survival (OS) was defined as the interval from the time of treatment initiation to the date of death. The OS and PFS times were calculated from the initiation of TKI- or pemetrexedbased treatment. To characterize OS on the basis of treatment exposure, we also measured OS from the date of diagnosis to the date of death. Kaplan-Meier curves
Characteristics of ROS1 NSCLC
September 2018
and risk tables were used to tabulate the treatment failure. The Cox proportional hazard progression model was used to calculate the hazard ratio. p Values of 0.05 or lower were considered statistically significant. The Response Evaluation Criteria in Solid Tumors version 1.1 were used to assess treatment response.9 STATA software (version 12.1, StataCorp LP, College Station, TX) was used for statistical analysis.
Identification of ROS1 Rearrangement ROS1 rearrangement was identified by using fluorescence in situ hybridization (FISH) (POSEIDON ROS1 Dual Color Break Apart Probe, Kreatech, Inc., Amsterdam, the Netherlands) or next-generation sequencing (NGS). As previously described, patients were considered ROS1 positive if there was a split or isolated signal in a kinase domain in at least 15% of tumor cells (after 50 tumor cells were counted in each sample).1 The internal normal control was identified in the corresponding chromosome targets. A total of 22 samples were sequenced by using the CancerScan panel (LabGenomics, Seongnam, Republic of Korea) to identify the breakpoint of ROS1.20
Results Clinicopathologic Characteristics of the Study Population A total of 103 patients with ROS1-positive NSCLC were identified. Most of these patients were female (68.9%) and never-smokers (75.7%). We identified two cases of pleomorphic carcinoma. The remainder of the patients (98.1%) had adenocarcinoma. Interestingly, four patients (4.8%) had EGFR mutations coexisting with their ROS1 rearrangements, as follows: EGFR exon 20 polymorphism c.2361G>A; exon 20 c2319_2320 insertion AACCCCCAC; exon 19 deletion; and exon 21 L833V and L858R. Of the patients, 80 (77.7%) initially presented at an advanced stage. Most patients (93.2%) received palliative chemotherapy, whereas 7.8% patients received definitive concurrent chemoradiotherapy. At their initial presentation, 39.8% of patients had lesions confined to the thorax. As of the time of data lock, 43 patients (41.8%) were alive, 34 (33.0%) had died, and 26 (25.2%) were lost to followup. The median duration of follow-up of the study population was 22.1 months. The pemetrexed-based regimen included pemetrexed and cisplatin followed by pemetrexed maintenance. This regimen was administered in 87.4% of patients, and TKIs were given to 56.3% of patients. The ROS1 rearrangement was identified by FISH (n ¼ 84), NGS (n ¼ 23), or both (n ¼ 3). A ROS1 breakpoint was identified in 9 of the 23 patients tested by NGS (Table 1).
1375
Table 1. Baseline Demographics (N ¼ 103) Characteristic
Value
Median age at diagnosis, y (range) Male sex, n (%) Female sex, n (%) Smoking, n (%) Nonsmoker Current smoker Past smoker ECOG PS, n (%) 0 1 2 Histologic type, n (%) Adenocarcinoma Pleomorphic carcinoma Genomic alteration, n (%) EGFR MT present EGFR MT absent ALK alteration present ALK alteration absent ALK alteration not tested Treatment method, n (%) Neoadjuvant CTx Adjuvant Adjuvant CCRT Definitive CCRT Palliative Initial disease status, n (%) Initially diagnosed in nonoperable setting, n (%) Stage I Stage II Stage III Stage IV CTx regimen, n (%) Pemetrexed-based CTx EGFR TKI Docetaxel Gemcitabine/carboplatin Immunotherapy TKI Metastasis, n (%) Intrathoracic only Extrathoracic ROS1 test results, n (%) Pts with NGS results Pts with FISH results Pts with both results
56 (28–85) 32 (31.1%) 71 (68.9%) 78 (75.7%) 7 (6.8%) 18 (17.5%) 39 (37.9%) 60 (58.2%) 4 (3.9%) 101 (98.1%) 2 (1.9%) 4 (3.9%) 99 (96.1%) 0 (0.0%) 100 (97.1%) 3 (2.9%) 3 (2.9%) 12 (11.7%) 3 (2.9%) 8 (7.8%) 96 (93.2%) 80 (77.7%) 6 (5.8%) 4 (3.9%) 24 (23.3%) 69 (67.0%) 90 21 20 18 12 58
(87.4%) (20.4%) (19.4%) (17.5%) (11.7%) (56.3%)
41 (39.8%) 62 (60.2%) 23 (22.3%) 83 (80.6%) 3 (2/9%)
ECOG PS, Eastern Cooperative Oncology Group performance status; MT, mutation; ALK, ALK receptor tyrosine kinase; CTx, chemotherapy; CCRT, concomittant chemotherapy and radiotherapy; TKI, tyrosine kinase inhibitor; Pts, patients; NGS, next-generation sequencing; FISH, fluorescence in situ hybridization.
Treatment Outcomes The median OS (calculated from the date of diagnosis to the date of death) for the entire study population was 52.1 months (95% confidence interval [CI]: 23.6–not reached [NR]) (Fig. 1A). The median OS for patients
1376 Park et al
Journal of Thoracic Oncology
Vol. 13 No. 9
Figure 1. Kaplan-Meier plot of the median overall survival (mOS) analysis. (A) Overall survival for the entire study population calculated from the date of relapse or metastasis to the last follow-up date; overall survival of patients treated with a tyrosine kinase inhibitor (TKI) (B), pemetrexed-based cytotoxic chemotherapy (C), and pemetrexed without a TKI (D). CI, confidence interval; NR, not reached; cTx, chemotherapy; mon, month.
treated with TKIs (n ¼ 58) was 64.9 months (95% CI: 26.3–NR), compared with 20.7 months (95% CI: 8.4– 54.3) for those who did not receive TKI treatment (n ¼ 45) (hazard ratio [HR] ¼ 0.44, 95% CI: 0.25–0.79) (Fig. 1B). Most of the study population was treated with pemetrexed-based chemotherapy (n ¼ 90). However, it was difficult to interpret the actual effect of pemetrexed treatment, especially on OS, because some of the patients (n ¼ 39) also received TKI either before or after the pemetrexed-based chemotherapy. The OS for the patients treated with pemetrexed-based chemotherapy was 54.1 months (95% CI: 24.0–NR) with an HR of 0.55 (95% CI: 0.23–1.31) (Fig. 1C). The OS in patients treated with pemetrexed but without crossover to a TKI (n ¼ 51) was 60.1 months (95% CI; 25.6–NR) (Fig. 1D). The median PFS for those treated with a tyrosine kinase inhibitor was 12.7 months (95% CI: 8.1–21.8). The median OS for those treated with a tyrosine kinase inhibitor was 23.2 months (95% CI: 12.2–NR) (Fig. 2A and B). Of the four different TKIs targeting the ROS1 adenosine triphosphate–binding site, ceritinib had the longest PFS (17.3 months). However, PFS was not reached in the case of TPX-0005, given the short-term follow-up. The PFS for pemetrexed-based therapy was 8.0 months (95% CI: 6.4–11.7) (Fig. 2C), and the OS for
pemetrexed-based therapy was 42.5 months (95% CI: 21.9–NR) (Fig. 2D). The ORR for treatment with a TKI was 70.7%, which was higher than the ORR with pemetrexed-based treatment (53.3%). The treatment durations and best responses based on treatment with a TKI are shown in Figure 3 and Table 2. Among the patients, there was unique subgroup treated with second TKIs and immune checkpoint inhibitors. Six patients were re-treated with a TKI after failure of treatment with an initial TKI. Of these six patients, one patient showed a partial response to the treatment. However, these patients showed a comparably short PFS (range 1.2–3.4 months) (Supplementary Table 1). Twelve patients received immune checkpoint inhibitors, with three of them showing a partial response to the treatment. The PFS of the patients treated with immune checkpoint inhibitors ranged from 1.1 to 10.7 months (Supplementary Table 2).
Patterns of Brain Metastases in ROS1-Positive NSCLC In all, 23 patients (22.3%) had intracranial metastases at the time of their initial diagnosis or at the time of recurrence. Among the 80 patients without brain
Characteristics of ROS1 NSCLC
September 2018
1377
Figure 2. Kaplan-Meier plot for the median progression-free survival (mPFS) and median overall survival (mOS) from the beginning of treatment. (A) mPFS of patients treated with a tyrosine kinase inhibitor. (B) mOS without a tyrosine kinase inhibitor. (C) mPFS with pemetrexed-based cytotoxic chemotherapy. (D) mOS with pemetrexed-based cytotoxic chemotherapy. CI, confidence interval; NR, not reached; mon, month.
metastasis, 24 experienced development of brain metastasis during follow-up. The median time to brain metastasis was 12.0 months (95% CI: 8.5–19.1) (Fig. 4). The most common site of metastasis with pemetrexedbased treatment was intrathoracic; with TKI treatment, the most common form was intracranial progression (15.5%).
ROS1 Breakpoint and TKI Response NGS was performed for 23 patients. We identified the following fusion partners of ROS1: TRK-fused gene (TFG), tropomyosin 3 gene (TPM3), syndecan 4 gene (SDC4), CD74 molecule gene (CD74), ezrin gene (EZR), zinc finger CCHC-type containing 8 gene (ZCCHC8), sacrolemma associated protein gene (SLMAP), and myosin VC gene (MYO5C) (Fig. 5). Among eight samples, five had a breakpoint between exons 31 and 33; this breakpoint has little effect in the transmembrane and tyrosine kinase domains of the ROS1 protein. Three samples had a breakpoint at exon 35, which disrupts the structures of the transmembrane domain. However, we did not identify any clinical correlation between the different fusion partners and the TKI treatment outcome (Supplementary Fig. 2. and Supplementary Table 3). Resistance developed in one patient who was treated
with crizotinib for 32.0 months. Repeat biopsy in the patient demonstrated a G2032R mutation in the CD79 molecule (CD79)-ROS1 fusion gene. This patient subsequently received palliative lorlatinib with partial response.
Discussion Target therapy and the predictive genomic markers have been shown to be efficacious in patients with EGFR-mutated and ALK-positive NSCLC. The ROS1 gene rearrangement is the third genomic alteration in patients with NSCLC for which TKI therapy has been approved by the U.S. Food and Drug Administration. In this study, we found that ROS1-positive patients have characteristics similar to those of ALK-positive patients.21,22 The patients were mostly female, nonsmokers, and approximately 50 years old. The ROS1 rearrangement is identified predominantly in patients with adenocarcinoma. However, it has a different clinical presentation with regard to tropism. In most patients (60.2%), at least one extrathoracic lesion was present at initial diagnosis. Brain tropism is less prominent in ROS1-positive patients than in patients without this mutation. Only 22.3% of patients presented with brain
1378 Park et al
Figure 3. Treatment response with a tyrosine kinase inhibitor. Response indicates the best response to the treatment. The number in the blue bar indicates progression-free survival in months. PR, partial response; CR, complete response; PD, progressive disease.
metastases; however, approximately 30% of patients who did not have intracranial lesions experienced development of brain metastasis during follow-up. Interestingly, intracranial progression was more common in patients treated with a TKI than in those who received chemotherapy. This finding suggests a need to identify novel agents with high penetration of the bloodbrain barrier.
Journal of Thoracic Oncology
Vol. 13 No. 9
The long duration of PFS with pemetrexed-based treatment is a well-known phenomenon in patients with ROS1 rearrangement.15,18,23–25 We found that pemetrexed-treated patients (n ¼ 90) had 8.0 months of PFS and 42.5 months of OS from the date of initiation of pemetrexed therapy. These findings were comparable to previous data. The prolonged OS in patients treated with pemetrexed in our data set can be explained by the fact that 81.1% of the population received pemetrexed as the first-line treatment, and 43.3% eventually crossed over to TKI therapy. The OS in patients treated with pemetrexed only is the OS from the date of initiation of pemetrexed therapy for patients treated with pemetrexed without crossover to TKI therapy. This value was still surprising (50.0 months [95% CI: 23.3–NR]); it may be explained by the comparably low level of TS expression in ROS1-positive patients. A similar result was shown in patients with NSCLC who tested negative for TS expression.14,26,27 Overall, this observation suggests that pemetrexed is a treatment option (of cytotoxic chemotherapy) for patients who cannot access TKIs. We studied four different TKIs in our patient population, namely, crizotinib, ceritinib, TPX-0005, and entrectinib. It is noteworthy that the ORR was at least 65% and the DCR was at least 80% for all four TKIs. However, the median PFS for TKIs in this study (12.7 months) was shorter than the median PFS times of the representative phase 1 crizotinib trial (19.2 months), phase 2 ceritinib trial (19.3 months), and entrectinib studies.7,28,29 Given the small number of patients for each regimen and various lines of therapy, further investigation is needed to evaluate the efficacy of the different agents. As previously described, we identified one case in which a patient acquired resistance to crizotinib as a result of a G2032R mutation. Further studies are needed to determine the mechanisms of resistance to other novel agents and the next optimal treatment options on the basis of the respective mechanisms of resistance. The most common fusion partner of the ROS1 gene was CD74. This finding was consistent with those of prior studies. However, we also identified the following unique fusion partners of the ROS1 gene that have not been reported previously: ZCCHC8, SLMAP, MYO5C, and TFG. How the ROS1 fusion protein gains its constitutive activation signal is unclear; however, the correlation between clinical efficacy and type of fusion partner has not yet been observed.7 We also did not find any difference in the clinical outcomes in response to TKIs among the various fusion partners (see Supplementary Table 3). Furthermore, little is known regarding the oncogenicity of the fusion proteins. What is known, however, is that most of the fusion gene does not contain dimerization domains. In addition, the ROS1 kinase
September 2018
Table 2. Overall Response Based on Treatment Treatment Response Treatment Line
Best Response ORR DCR
n
Tyrosine kinase inhibitor
58
Crizotinib
15
NE: 1 Pt (1.7%) CR: 3 Pts (5.2%) PR: 38 Pts (65.5%) Stable disease: 9 Pts (15.5%) PD: 7 Pts (12.1%) ORR: 70.7%/DCR: 86.2% ORR: 73.3%/DCR: 80.1%
Ceritinib
23
ORR: 65.2%/DCR: 86.9%
TPX-0005
7
ORR: 85.7%/DCR: 100.0%
Entrectinib
13
ORR: 69.2%/DCR: 84.6%
TKI after TKI
6
ORR: 16.7%/DCR: 50.0%
Immune checkpoint inhibitor Pemetrexedbased treatment
12
ORR: 25.0%/DCR: 50.0%
90
1 line: 5 Pts (8.6%) 2 lines: 32 Pts (55.2%) >3 lines: 21 Pts (36.2%)
1 line: 73 Pts (81.1%) 2 lines: 7 Pts (7.8%) >3 lines: 10 Pts (11.1%)
NE: 2 Pts (2.2%) CR: 1 Pts (1.1%) PR: 47 Pts (52.2%) Stable disease: 31 Pts (34.5%) PD: 9 Pts (10.0%) ORR: 53.3%/DCR: 87.8%
Median DOT (range)
Median PFS (95% CI)
Median OS (95% CI)
7.3 mo (0.4–48.9)
12.7 mo (8.1–21.8)
5.5 mo (1.0–48.9) 10.3 mo (1.8–45.9) 5.3 mo (1.7–7.4) 6.3 mo (0.6–25.7) 1.4 mo (0.1–3.1) 2.0 mo (0.0–11.0) 6.0 mo (0.0–85.6)
PD Event
Lung
Bone
Extrathoracic
Brain
23.2 mo (12.2–NR)
32 (55.2%)
22 (37.9%)
4 (6.9%)
7 (12.1%)
9 (15.5%)
13.1 mo (4.4–NR) 17.3 mo (8.1–23.2) NR
15.1 mo (5.4–NR) NR
8 (53.3%)
7 (46.7%)
1 (6.7%)
1 (6.7%)
1 (6.7%)
16 (69.6%)
11 (47.8%)
3 (13.0%)
4 (17.4%)
4 (17.4%)
NR
0
0
0
0
0
10.0 mo (4.6–22.4) 1.3 mo (0.16–NR) 1.6 mo (1.0–6.5)
17.7 mo (5.1–NR) 1.3 mo (6.9–NR) 23.4 mo (1.5–NR)
8 (61.5%)
4 (30.8%)
0
2 (15.4%)
4 (30.8%)
3 (50.0%)
3 (50.0%)
1 (16.7%)
2 (33.3%)
1 (16.7%)
12 (100.0%) 11 (91.7%)
1 (8.3%)
2 (16.7%)
1 (8.3%)
8.0 mo (6.4–11.7)
42.5 mo (21.9–NR)
71 (78.9%)
7 (7.8%)
12 (13.3%)
6 (6.7%)
50 (55.6%)
ORR, overall response rate; DCR, disease control rate; DOT, duration of treatment; PFS, progression-free survival; CI, confidence interval; PD, progressive disease; Pts, patients; NE, not evaluable; NR, not reached; CR, complete response; PR, partial response.
Characteristics of ROS1 NSCLC
Treatment
PD Site
1379
1380 Park et al
Journal of Thoracic Oncology
Vol. 13 No. 9
Figure 5. Overall landscape of ROS1 fusion partners detected by next-generation sequencing (n ¼ 23). TFG, TRK-fused gene; TPM3, tropomyosin 3 gene; SDC4, syndecan 4 gene; CD74, CD 74 molecule gene; EZR, ezrin gene; ZCCHC8, zinc finger CCHC-type containing 8 gene; SLMAP, sacrolemma associated protein gene; MYO5C, myosin VC gene.
Figure 4. (A) Number of brain metastases when the cancer was first diagnosed. (B) Number of brain metastases that developed during the palliative treatment. (C) Kaplan-Meier plot for the time to brain metastases for patients without initial brain lesions (n ¼ 24). The median time to progression was 12.0 months (95% confidence interval: 8.5–19.1).
domain is always retained in the fusion protein, which was also shown in our eight samples (Supplementary Fig. 2).2,5 Our result are very similar to the those reported on the basis of real clinical data rather than on the basis of outcomes of clinical trials.19 Recently, Gainor et al.,19 reported a pooled analysis of ROS1-positive patients
(n ¼ 39) treated in a clinical setting. The ROS1-positive patients had 11.0 months of PFS, which was similar to the 12.7 months of PFS with treatment with a TKI from our data set. We also observed similarity in tropism in the ROS1-positive population to that of those without this mutation. At the time of diagnosis, 19.4% of patients with ROS1 rearrangements had brain metastasis, whereas 59.0% had extrathoracic metastases. These data are very similar to ours, in which 22.3% of ROS1-positive patients had brain metastases and 60.2% had extrathoracic metastases at the time of the initial diagnosis. Once again, our study supports the unique clinical characteristics of ROS1-positive NSCLC represented by tropism in distant metastases. There is no standard method for detecting ROS1 rearrangement. Although FISH is the most common method, it cannot provide information regarding fusion partners. Reverse-transcriptase polymerase chain reaction is another widely used method; however, it can miss unknown fusion genes. ROS1 immunohistochemistry is simple and considered to have high concordance with FISH. It is also difficult to interpret, given the variable location of ROS1 fusions and large amount of background staining. In contrast, NGS can provide detailed information regarding the fusion partners; however, it is limited by its high cost, long turnaround time, and need for a large amout of tissue.28,30 There is one reported case of crizotinib being used to treat a patient who was
Characteristics of ROS1 NSCLC
September 2018
CD74-ROS1–negative according to FISH but CD74-ROS1– positive according to NGS.31 In contast, in a case of ROS1 positivity according to FISH but ROS1 negativity on the basis of NGS results, the patient had a partial response to first-line pemetrexed but no response to second-line crizotinib. Both of these rare cases reflect the need for careful consideration of ROS1 positivity that is detected with a single method. This study involved the largest series of clinical data on ROS1-rearranged NSCLC to date; regardless, it does have several limitations. Given the retrospective analysis, this study was subject to potential biases. Because of the low rate of discovery of ROS1 positivity, it was difficult to perform a direct comparison between drugs. Last but not least, the comparison between patients treated with a TKI and then chemotherapy and those treated with chemotherapy and then a TKI has limitations owing to the small number of patients treated with a TKI as first-line treatment (Supplementary Table 4). In conclusion, our data reveal the unique clinical characteristics and outcomes of patients with ROS1positive NSCLC in Korea. Although ROS1-rearranged NSCLC is rare, it has a robust response to TKIs. Therefore, further efforts are needed to standardize the diagnostic method, develop new therapeutic strategies, and improve patient access.
9.
10.
11.
12.
13.
14.
Supplementary Data
15.
Note: To access the supplementary material accompanying this article, visit the online version of the Journal of Thoracic Oncology at www.jto.org and at https://doi. org/10.1016/j.jtho.2018.05.026.
16.
References 1. Bergethon K, Shaw AT, Ou SH, et al. ROS1 rearrangements define a unique molecular class of lung cancers. J Clin Oncol. 2012;30:863–870. 2. Takeuchi K, Soda M, Togashi Y, et al. RET, ROS1 and ALK fusions in lung cancer. Nat Med. 2012;18:378–381. 3. Davies KD, Doebele RC. Molecular pathways: ROS1 fusion proteins in cancer. Clin Cancer Res. 2013;19:4040–4045. 4. Murphy DA, Ely HA, Shoemaker R, et al. Detecting gene rearrangements in patient populations through a 2-step diagnostic test comprised of rapid IHC enrichment followed by sensitive next-generation sequencing. Appl Immunohistochem Mol Morphol. 2017;25:513–523. 5. Gainor JF, Shaw AT. Novel targets in non-small cell lung cancer: ROS1 and RET fusions. Oncologist. 2013;18:865–875. 6. Huber KV, Salah E, Radic B, et al. Stereospecific targeting of MTH1 by (S)-crizotinib as an anticancer strategy. Nature. 2014;508:222–227. 7. Shaw AT, Ou SH, Bang YJ, et al. Crizotinib in ROS1rearranged non-small-cell lung cancer. N Engl J Med. 2014;371:1963–1971. 8. Lim SM, Kim HR, Lee JS, et al. Open-label, multicenter, phase II study of ceritinib in patients with non-small-cell
17.
18.
19.
20.
21.
22.
1381
lung cancer harboring ROS1 rearrangement. J Clin Oncol. 2017;35:2613–2618. Solomon BJ, Bauer TM, Felip E, et al. Safety and efficacy of lorlatinib (PF-06463922) from the dose-escalation component of a study in patients with advanced ALKþ or ROS1þ non-small cell lung cancer (NSCLC) [abstract]. J Clin Oncol. 2016;34:9009. Ardini E, Menichincheri M, Banfi P, et al. Entrectinib, a Pan-TRK, ROS1, and ALK inhibitor with activity in multiple molecularly defined cancer indications. Mol Cancer Ther. 2016;15:628–639. Drilon A, Siena S, Ou SI, et al. Safety and antitumor activity of the multitargeted pan-trk, ros1, and alk inhibitor entrectinib: combined results from two phase I trials (ALKA-372-001 and STARTRK-1). Cancer Discov. 2017;7:400–409. Awad MM, Katayama R, McTigue M, et al. Acquired resistance to crizotinib from a mutation in CD74-ROS1. N Engl J Med. 2013;368:2395–2401. Paz-Ares L, de Marinis F, Dediu M, et al. Maintenance therapy with pemetrexed plus best supportive care versus placebo plus best supportive care after induction therapy with pemetrexed plus cisplatin for advanced non-squamous non-small-cell lung cancer (PARAMOUNT): a double-blind, phase 3, randomised controlled trial. Lancet Oncol. 2012;13: 247–255. Song Z, Su H, Zhang Y. Patients with ROS1 rearrangement-positive non-small-cell lung cancer benefit from pemetrexed-based chemotherapy. Cancer Med. 2016;5:2688–2693. Chen YF, Hsieh MS, Wu SG, et al. Efficacy of pemetrexedbased chemotherapy in patients with ROS1 fusion-positive lung adenocarcinoma compared with in patients harboring other driver mutations in East Asian populations. J Thorac Oncol. 2016;11:1140–1152. Shaw AT, Kim DW, Nakagawa K, et al. Crizotinib versus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med. 2013;368:2385–2394. Solomon BJ, Mok T, Kim DW, et al. First-line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J Med. 2014;371:2167–2177. Mazieres J, Zalcman G, Crino L, et al. Crizotinib therapy for advanced lung adenocarcinoma and a ROS1 rearrangement: results from the EUROS1 cohort. J Clin Oncol. 2015;33:992–999. Gainor JF, Tseng D, Yoda S, et al. Patterns of metastatic spread and mechanisms of resistance to crizotinib in ROS1-positive non-small-cell lung cancer. JCO Precis Oncol. 2017;2017. Shin HT, Choi YL, Yun JW, et al. Prevalence and detection of low-allele-fraction variants in clinical cancer samples. Nat Commun. 2017;8:1377. Soria JC, Tan DSW, Chiari R, et al. First-line ceritinib versus platinum-based chemotherapy in advanced ALK-rearranged non-small-cell lung cancer (ASCEND-4): a randomised, open-label, phase 3 study. Lancet. 2017;389:917–929. Peters S, Camidge DR, Shaw AT, et al. Alectinib versus crizotinib in untreated ALK-positive non-small-cell lung cancer. N Engl J Med. 2017;377:829–838.
1382 Park et al 23. Franchina T, Russo A, Ricciardi GR, et al. Long time response with chemotherapy in ROS1 NSCLC patient with unusual metastatic site. Cancer Biol Ther. 2016:1–5. 24. Liang Y, Wakelee HA, Neal JW. Relationship of driver oncogenes to long-term pemetrexed response in non– small-cell lung cancer. Clin Lung Cancer. 2015;16: 366–373. 25. Riess JW, Padda SK, Bangs CD, et al. A case series of lengthy progression-free survival with pemetrexedcontaining therapy in metastatic non–small-cell lung cancer patients harboring ROS1 gene rearrangements. Clin Lung Cancer. 2013;14:592–595. 26. Nicolson MC, Fennell DA, Ferry D, et al. Thymidylate synthase expression and outcome of patients receiving pemetrexed for advanced nonsquamous non-small-cell lung cancer in a prospective blinded assessment phase II clinical trial. J Thorac Oncol. 2013;8:930–939. 27. Sun JM, Ahn JS, Jung SH, et al. Pemetrexed plus cisplatin versus gemcitabine plus cisplatin according to
Journal of Thoracic Oncology
28.
29.
30.
31.
Vol. 13 No. 9
thymidylate synthase expression in nonsquamous nonsmall-cell lung cancer: a biomarker-stratified randomized phase II trial. J Clin Oncol. 2015;33:2450–2456. Rossi G, Jocolle G, Conti A, et al. Detection of ROS1 rearrangement in non-small cell lung cancer: current and future perspectives. Lung Cancer (Auckl). 2017;8: 45–55. Ahn M, Cho BC, Siena S, et al. Entrectinib in patients with locally advanced or metastatic ros1 fusion-positive non-small cell lung cancer (NSCLC). J Thorac Oncol. 2017;12:S1783. Bubendorf L, Buttner R, Al-Dayel F, et al. Testing for ROS1 in non-small cell lung cancer: a review with recommendations. Virchows Arch. 2016;469:489–503. Drilon A, Wang L, Arcila ME, et al. Broad, hybrid capturebased next-generation sequencing identifies actionable genomic alterations in lung adenocarcinomas otherwise negative for such alterations by other genomic testing approaches. Clin Cancer Res. 2015;21:3631–3639.
Characteristics and Outcome of ROS1-Positive Non–Small Cell Lung Cancer Patients in Routine Clinical Practice Sehhoon Park, MD,a Beung-Chul Ahn, MD,b Sung Won Lim, MD,a Jong-Mu Sun, MD, PhD,a Hye Ryun Kim, MD, PhD,b Min Hee Hong, MD,b Se-Hoon Lee, MD, PhD,a Jin Seok Ahn, MD, PhD,a Keunchil Park, MD, PhD,a Yoon La Choi, MD, PhD,c Byoung Chul Cho, MD, PhD,b Myung-Ju Ahn, MD, PhDa,* a
Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea b Yonsei Cancer Center, Division of Medical Oncology, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea c Department of Pathology and Translational Genomics, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea Received 22 April 2018; revised 11 May 2018; accepted 12 May 2018 Available online - 5 June 2018
ABSTRACT Introduction: ROS1-rearranged NSCLC is classified as a distinct molecular subset of NSCLC with a therapeutic target. ROS1 rearrangement is most often identified in never-smokers with adenocarcinoma and EGFR and ALK receptor tyrosine kinase gene (ALK) wild type. Treatment with tyrosine kinase inhibitors (TKIs), which target the ROS1 kinase domain, is considered the standard of care. TKIs have been shown to have a robust and durable response. However, information regarding the clinical outcomes of TKI treatment, including brain metastasis, remains limited. Methods: We identified 103 consecutive cases of ROS1positive NSCLC by using break-apart fluorescence in situ hybridization (n ¼ 84), next-generation sequencing (n ¼ 23), or both (n ¼ 3). Information regarding fusion breakpoints was available for eight patients. Clinical data, including patient characteristics, incidence of brain metastasis, response to chemotherapy, or to TKIs, were retrospectively analyzed. Results: The median patient age was 56 years, and 58.9% of the patients were female. Most of the patients (75.7%) were never-smokers. Adenocarcinoma was predominant (98.1%), and two cases with pleomorphic carcinoma were identified. Sixty percent of patients had an extrathoracic metastatic lesion, and 22% had an intracranial lesion at the initial presentation or at the time of recurrence. The median time to development of brain metastases was 12.0 months (range 2.1–84.1). The most common fusion partner was CD74 molecule gene (CD74), followed by syndecan 4 gene (SDC4),
ezrin gene (EZR), tropomyosin 3 gene (TPM3), TRK-fused gene (TFG), zinc finger CCHC-type containing 8 gene (ZCCHC8), sacrolemma associated protein gene (SLMAP), and myosin VC gene (MYO5C). All of these fusion partners preserved the tyrosine kinase domain of ROS1. The median overall survival time was 52.1 months (95% confidence interval [CI]: 23.6–not reached). In the 90 patients who were treated with pemetrexed-based chemotherapy, the overall response rate and progression-free survival time were 53.3% and 8.0 months (95% CI: 6.4–11.7), respectively. The overall response rate and progression-free survival time were 70.7% and 12.7 months (95% CI: 8.1–21.8), respectively, for the 50 patients treated with TKIs. Brain metastasis was more often observed during TKI treatment (15.5%) than during pemetrexed-based chemotherapy (6.7%).
*Corresponding author. Drs. Park, Ahn, and Lim equally contributed to this work. Disclosure: Dr. Lee reports personal fees from AstraZeneca/ MedImmune, Roche, Bristol-Myers Squibb, Merck, and Novartis outside the submitted work. Dr. Ahn reports personal fees from Amgen, Pfizer, AstraZeneca, Menarini, Roche, Eisai, Boehringer Ingelheim, BMS, and Janssen outside the submitted work. The remaining authors declare no conflict of interest. Address for correspondence: Myung-Ju Ahn, MD, PhD, Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul 06351, Republic of Korea. E-mail: silkahn@skku. edu or [email protected]. ª 2018 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 https://doi.org/10.1016/j.jtho.2018.05.026
Journal of Thoracic Oncology
Vol. 13 No. 9: 1373-1382
1374 Park et al
Journal of Thoracic Oncology
Conclusions: ROS1-positive NSCLC has distinct clinical characteristics, with an effective and durable response to both TKIs and pemetrexed-based chemotherapies. Regardless, given its novel characteristics and distinct clinical responses to conventional chemotherapies and TKIs, the treatment strategy for ROS1-positive NSCLC remains to be further developed. 2018 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: ROS1; Crizotinib; Non–small cell lung cancer
Pemetrexed;
Carcinoma;
Vol. 13 No. 9
therapy allowed up to 9 months of PFS in ROS1-positive NSCLC.14,15 ROS1-positive patients have lower expression of thymidylate synthetase (TS) than ROS1-negative patients did; this expression pattern is thought to explain the high efficacy of pemetrexed in ROS1-positive patients.14 In contrast, patients with ALK-positive NSCLC with similar levels of TS expression have a shorter PFS (range 4.2–7.0 months).16,17 Despite the many clinical investigations of TKIs in patients with ROS1-rearranged NSCLC, information regarding its clinical outcomes remain limited.18,19 In this study, we have therefore reported the comprehensive analysis of clinical data from ROS1-positive patients in Korea.
Introduction ROS1 is a receptor tyrosine kinase of the insulin receptor family. Its rearrangement is recognized as a targetable oncogenic driver that accounts for up to 2% of NSCLCs.1,2 The ROS1 fusion protein maintains the kinase domain in the 30 region of ROS1. It has diverse counterparts, including CD74 molecule, ezrin, syndecan 4, solute carrier family 34 member 2, tropomyosin 3, interleukin 4 induced 1 (also known as FIG1), cell division cycle 6, KDEL endoplasmic reticulum protein retention receptor 2 at 50 terminus.3,4 Interestingly, the ROS1 and ALK receptor tyrosine kinase (ALK) domains have highly conserved adenosine triphosphate–binding domains, with nearly 77% homology of their amino acid sequences.5,6 This structural similarity enables the ALK inhibitor to bind to the ROS1 protein even more strongly than it does to the ALK protein.6 This phenomenon explains the promising outcomes demonstrated by use of an ALK inhibitor in the treatment of ROS1-rearranged NSCLC. In the PROFILE 1001 study, crizotinib, which is a potent TKI for ALK, ROS1, and mesenchymal-epithelial transition, demonstrated an overall response rate (ORR) of 72%, with a median of 19.2 months of progression-free survival (PFS).7 Ceritinib, which is a more potent ALK TKI, has demonstrated an ORR of 62% with a PFS of 19.3 months in crizotinib-naive patients.8 A number of TKIs with ROS1 inhibition are currently in clinical development.9–11 On the basis of a few clinical experiences, there have been reports of development of an acquired resistance mechanism after crizotinib therapy in the PROFILE 1001 study. One such mechanism of resistance is ROS1 G2032R in the CD74 molecule–ROS1 fusion protein located at the tyrosine kinase domain.12 Another characteristic of ROS1-positive NSCLC is a durable response to pemetrexed. Pemetrexed and platinum doublets followed by pemetrexed maintenance therapy has allowed approximately 4 months of PFS in patients with advanced nonsquamous NSCLC with no targetable genomic alterations.13 In contrast, the same
Methods Patients and Samples We identified 103 patients with NSCLC with ROS1 gene rearrangements who were treated at Samsung Medical Center and Severance Hospital between January 2001 and February 2018. Clinical data, including patient characteristics, the incidence of brain metastasis, and responses to chemotherapy or TKIs, were retrospectively analyzed. Because of local regulation by the health authority, according to which off-label use of chemotherapeutic agents (including targeted agents) in not allowed, TKIs are available only for clinical trials. Hence, the sequence of treatment was decide by the availability of clinical trials. Patients were evaluated for treatment response either by simple chest radiography or chest computed tomography scan. Chest posterior-anterior radiography was performed after every cycle and chest computed tomography was performed after every two or three cycles of treatment. In addition to regular followup, additional imaging was conducted at the physician’s discretion. The response rate was calculated by adding the percentage of patients with a partial response and the percentage with a complete response to the treatment. This study was approved by an institutional review board (institutional review board approval No. 2018-04-012-002 and No. 4-2018-0195). Informed consent was waived.
Statistical Analysis PFS was calculated from the start of treatment to radiographic progression or death. Otherwise, patients were censored from analysis. Overall survival (OS) was defined as the interval from the time of treatment initiation to the date of death. The OS and PFS times were calculated from the initiation of TKI- or pemetrexedbased treatment. To characterize OS on the basis of treatment exposure, we also measured OS from the date of diagnosis to the date of death. Kaplan-Meier curves
Characteristics of ROS1 NSCLC
September 2018
and risk tables were used to tabulate the treatment failure. The Cox proportional hazard progression model was used to calculate the hazard ratio. p Values of 0.05 or lower were considered statistically significant. The Response Evaluation Criteria in Solid Tumors version 1.1 were used to assess treatment response.9 STATA software (version 12.1, StataCorp LP, College Station, TX) was used for statistical analysis.
Identification of ROS1 Rearrangement ROS1 rearrangement was identified by using fluorescence in situ hybridization (FISH) (POSEIDON ROS1 Dual Color Break Apart Probe, Kreatech, Inc., Amsterdam, the Netherlands) or next-generation sequencing (NGS). As previously described, patients were considered ROS1 positive if there was a split or isolated signal in a kinase domain in at least 15% of tumor cells (after 50 tumor cells were counted in each sample).1 The internal normal control was identified in the corresponding chromosome targets. A total of 22 samples were sequenced by using the CancerScan panel (LabGenomics, Seongnam, Republic of Korea) to identify the breakpoint of ROS1.20
Results Clinicopathologic Characteristics of the Study Population A total of 103 patients with ROS1-positive NSCLC were identified. Most of these patients were female (68.9%) and never-smokers (75.7%). We identified two cases of pleomorphic carcinoma. The remainder of the patients (98.1%) had adenocarcinoma. Interestingly, four patients (4.8%) had EGFR mutations coexisting with their ROS1 rearrangements, as follows: EGFR exon 20 polymorphism c.2361G>A; exon 20 c2319_2320 insertion AACCCCCAC; exon 19 deletion; and exon 21 L833V and L858R. Of the patients, 80 (77.7%) initially presented at an advanced stage. Most patients (93.2%) received palliative chemotherapy, whereas 7.8% patients received definitive concurrent chemoradiotherapy. At their initial presentation, 39.8% of patients had lesions confined to the thorax. As of the time of data lock, 43 patients (41.8%) were alive, 34 (33.0%) had died, and 26 (25.2%) were lost to followup. The median duration of follow-up of the study population was 22.1 months. The pemetrexed-based regimen included pemetrexed and cisplatin followed by pemetrexed maintenance. This regimen was administered in 87.4% of patients, and TKIs were given to 56.3% of patients. The ROS1 rearrangement was identified by FISH (n ¼ 84), NGS (n ¼ 23), or both (n ¼ 3). A ROS1 breakpoint was identified in 9 of the 23 patients tested by NGS (Table 1).
1375
Table 1. Baseline Demographics (N ¼ 103) Characteristic
Value
Median age at diagnosis, y (range) Male sex, n (%) Female sex, n (%) Smoking, n (%) Nonsmoker Current smoker Past smoker ECOG PS, n (%) 0 1 2 Histologic type, n (%) Adenocarcinoma Pleomorphic carcinoma Genomic alteration, n (%) EGFR MT present EGFR MT absent ALK alteration present ALK alteration absent ALK alteration not tested Treatment method, n (%) Neoadjuvant CTx Adjuvant Adjuvant CCRT Definitive CCRT Palliative Initial disease status, n (%) Initially diagnosed in nonoperable setting, n (%) Stage I Stage II Stage III Stage IV CTx regimen, n (%) Pemetrexed-based CTx EGFR TKI Docetaxel Gemcitabine/carboplatin Immunotherapy TKI Metastasis, n (%) Intrathoracic only Extrathoracic ROS1 test results, n (%) Pts with NGS results Pts with FISH results Pts with both results
56 (28–85) 32 (31.1%) 71 (68.9%) 78 (75.7%) 7 (6.8%) 18 (17.5%) 39 (37.9%) 60 (58.2%) 4 (3.9%) 101 (98.1%) 2 (1.9%) 4 (3.9%) 99 (96.1%) 0 (0.0%) 100 (97.1%) 3 (2.9%) 3 (2.9%) 12 (11.7%) 3 (2.9%) 8 (7.8%) 96 (93.2%) 80 (77.7%) 6 (5.8%) 4 (3.9%) 24 (23.3%) 69 (67.0%) 90 21 20 18 12 58
(87.4%) (20.4%) (19.4%) (17.5%) (11.7%) (56.3%)
41 (39.8%) 62 (60.2%) 23 (22.3%) 83 (80.6%) 3 (2/9%)
ECOG PS, Eastern Cooperative Oncology Group performance status; MT, mutation; ALK, ALK receptor tyrosine kinase; CTx, chemotherapy; CCRT, concomittant chemotherapy and radiotherapy; TKI, tyrosine kinase inhibitor; Pts, patients; NGS, next-generation sequencing; FISH, fluorescence in situ hybridization.
Treatment Outcomes The median OS (calculated from the date of diagnosis to the date of death) for the entire study population was 52.1 months (95% confidence interval [CI]: 23.6–not reached [NR]) (Fig. 1A). The median OS for patients
1376 Park et al
Journal of Thoracic Oncology
Vol. 13 No. 9
Figure 1. Kaplan-Meier plot of the median overall survival (mOS) analysis. (A) Overall survival for the entire study population calculated from the date of relapse or metastasis to the last follow-up date; overall survival of patients treated with a tyrosine kinase inhibitor (TKI) (B), pemetrexed-based cytotoxic chemotherapy (C), and pemetrexed without a TKI (D). CI, confidence interval; NR, not reached; cTx, chemotherapy; mon, month.
treated with TKIs (n ¼ 58) was 64.9 months (95% CI: 26.3–NR), compared with 20.7 months (95% CI: 8.4– 54.3) for those who did not receive TKI treatment (n ¼ 45) (hazard ratio [HR] ¼ 0.44, 95% CI: 0.25–0.79) (Fig. 1B). Most of the study population was treated with pemetrexed-based chemotherapy (n ¼ 90). However, it was difficult to interpret the actual effect of pemetrexed treatment, especially on OS, because some of the patients (n ¼ 39) also received TKI either before or after the pemetrexed-based chemotherapy. The OS for the patients treated with pemetrexed-based chemotherapy was 54.1 months (95% CI: 24.0–NR) with an HR of 0.55 (95% CI: 0.23–1.31) (Fig. 1C). The OS in patients treated with pemetrexed but without crossover to a TKI (n ¼ 51) was 60.1 months (95% CI; 25.6–NR) (Fig. 1D). The median PFS for those treated with a tyrosine kinase inhibitor was 12.7 months (95% CI: 8.1–21.8). The median OS for those treated with a tyrosine kinase inhibitor was 23.2 months (95% CI: 12.2–NR) (Fig. 2A and B). Of the four different TKIs targeting the ROS1 adenosine triphosphate–binding site, ceritinib had the longest PFS (17.3 months). However, PFS was not reached in the case of TPX-0005, given the short-term follow-up. The PFS for pemetrexed-based therapy was 8.0 months (95% CI: 6.4–11.7) (Fig. 2C), and the OS for
pemetrexed-based therapy was 42.5 months (95% CI: 21.9–NR) (Fig. 2D). The ORR for treatment with a TKI was 70.7%, which was higher than the ORR with pemetrexed-based treatment (53.3%). The treatment durations and best responses based on treatment with a TKI are shown in Figure 3 and Table 2. Among the patients, there was unique subgroup treated with second TKIs and immune checkpoint inhibitors. Six patients were re-treated with a TKI after failure of treatment with an initial TKI. Of these six patients, one patient showed a partial response to the treatment. However, these patients showed a comparably short PFS (range 1.2–3.4 months) (Supplementary Table 1). Twelve patients received immune checkpoint inhibitors, with three of them showing a partial response to the treatment. The PFS of the patients treated with immune checkpoint inhibitors ranged from 1.1 to 10.7 months (Supplementary Table 2).
Patterns of Brain Metastases in ROS1-Positive NSCLC In all, 23 patients (22.3%) had intracranial metastases at the time of their initial diagnosis or at the time of recurrence. Among the 80 patients without brain
Characteristics of ROS1 NSCLC
September 2018
1377
Figure 2. Kaplan-Meier plot for the median progression-free survival (mPFS) and median overall survival (mOS) from the beginning of treatment. (A) mPFS of patients treated with a tyrosine kinase inhibitor. (B) mOS without a tyrosine kinase inhibitor. (C) mPFS with pemetrexed-based cytotoxic chemotherapy. (D) mOS with pemetrexed-based cytotoxic chemotherapy. CI, confidence interval; NR, not reached; mon, month.
metastasis, 24 experienced development of brain metastasis during follow-up. The median time to brain metastasis was 12.0 months (95% CI: 8.5–19.1) (Fig. 4). The most common site of metastasis with pemetrexedbased treatment was intrathoracic; with TKI treatment, the most common form was intracranial progression (15.5%).
ROS1 Breakpoint and TKI Response NGS was performed for 23 patients. We identified the following fusion partners of ROS1: TRK-fused gene (TFG), tropomyosin 3 gene (TPM3), syndecan 4 gene (SDC4), CD74 molecule gene (CD74), ezrin gene (EZR), zinc finger CCHC-type containing 8 gene (ZCCHC8), sacrolemma associated protein gene (SLMAP), and myosin VC gene (MYO5C) (Fig. 5). Among eight samples, five had a breakpoint between exons 31 and 33; this breakpoint has little effect in the transmembrane and tyrosine kinase domains of the ROS1 protein. Three samples had a breakpoint at exon 35, which disrupts the structures of the transmembrane domain. However, we did not identify any clinical correlation between the different fusion partners and the TKI treatment outcome (Supplementary Fig. 2. and Supplementary Table 3). Resistance developed in one patient who was treated
with crizotinib for 32.0 months. Repeat biopsy in the patient demonstrated a G2032R mutation in the CD79 molecule (CD79)-ROS1 fusion gene. This patient subsequently received palliative lorlatinib with partial response.
Discussion Target therapy and the predictive genomic markers have been shown to be efficacious in patients with EGFR-mutated and ALK-positive NSCLC. The ROS1 gene rearrangement is the third genomic alteration in patients with NSCLC for which TKI therapy has been approved by the U.S. Food and Drug Administration. In this study, we found that ROS1-positive patients have characteristics similar to those of ALK-positive patients.21,22 The patients were mostly female, nonsmokers, and approximately 50 years old. The ROS1 rearrangement is identified predominantly in patients with adenocarcinoma. However, it has a different clinical presentation with regard to tropism. In most patients (60.2%), at least one extrathoracic lesion was present at initial diagnosis. Brain tropism is less prominent in ROS1-positive patients than in patients without this mutation. Only 22.3% of patients presented with brain
1378 Park et al
Figure 3. Treatment response with a tyrosine kinase inhibitor. Response indicates the best response to the treatment. The number in the blue bar indicates progression-free survival in months. PR, partial response; CR, complete response; PD, progressive disease.
metastases; however, approximately 30% of patients who did not have intracranial lesions experienced development of brain metastasis during follow-up. Interestingly, intracranial progression was more common in patients treated with a TKI than in those who received chemotherapy. This finding suggests a need to identify novel agents with high penetration of the bloodbrain barrier.
Journal of Thoracic Oncology
Vol. 13 No. 9
The long duration of PFS with pemetrexed-based treatment is a well-known phenomenon in patients with ROS1 rearrangement.15,18,23–25 We found that pemetrexed-treated patients (n ¼ 90) had 8.0 months of PFS and 42.5 months of OS from the date of initiation of pemetrexed therapy. These findings were comparable to previous data. The prolonged OS in patients treated with pemetrexed in our data set can be explained by the fact that 81.1% of the population received pemetrexed as the first-line treatment, and 43.3% eventually crossed over to TKI therapy. The OS in patients treated with pemetrexed only is the OS from the date of initiation of pemetrexed therapy for patients treated with pemetrexed without crossover to TKI therapy. This value was still surprising (50.0 months [95% CI: 23.3–NR]); it may be explained by the comparably low level of TS expression in ROS1-positive patients. A similar result was shown in patients with NSCLC who tested negative for TS expression.14,26,27 Overall, this observation suggests that pemetrexed is a treatment option (of cytotoxic chemotherapy) for patients who cannot access TKIs. We studied four different TKIs in our patient population, namely, crizotinib, ceritinib, TPX-0005, and entrectinib. It is noteworthy that the ORR was at least 65% and the DCR was at least 80% for all four TKIs. However, the median PFS for TKIs in this study (12.7 months) was shorter than the median PFS times of the representative phase 1 crizotinib trial (19.2 months), phase 2 ceritinib trial (19.3 months), and entrectinib studies.7,28,29 Given the small number of patients for each regimen and various lines of therapy, further investigation is needed to evaluate the efficacy of the different agents. As previously described, we identified one case in which a patient acquired resistance to crizotinib as a result of a G2032R mutation. Further studies are needed to determine the mechanisms of resistance to other novel agents and the next optimal treatment options on the basis of the respective mechanisms of resistance. The most common fusion partner of the ROS1 gene was CD74. This finding was consistent with those of prior studies. However, we also identified the following unique fusion partners of the ROS1 gene that have not been reported previously: ZCCHC8, SLMAP, MYO5C, and TFG. How the ROS1 fusion protein gains its constitutive activation signal is unclear; however, the correlation between clinical efficacy and type of fusion partner has not yet been observed.7 We also did not find any difference in the clinical outcomes in response to TKIs among the various fusion partners (see Supplementary Table 3). Furthermore, little is known regarding the oncogenicity of the fusion proteins. What is known, however, is that most of the fusion gene does not contain dimerization domains. In addition, the ROS1 kinase
September 2018
Table 2. Overall Response Based on Treatment Treatment Response Treatment Line
Best Response ORR DCR
n
Tyrosine kinase inhibitor
58
Crizotinib
15
NE: 1 Pt (1.7%) CR: 3 Pts (5.2%) PR: 38 Pts (65.5%) Stable disease: 9 Pts (15.5%) PD: 7 Pts (12.1%) ORR: 70.7%/DCR: 86.2% ORR: 73.3%/DCR: 80.1%
Ceritinib
23
ORR: 65.2%/DCR: 86.9%
TPX-0005
7
ORR: 85.7%/DCR: 100.0%
Entrectinib
13
ORR: 69.2%/DCR: 84.6%
TKI after TKI
6
ORR: 16.7%/DCR: 50.0%
Immune checkpoint inhibitor Pemetrexedbased treatment
12
ORR: 25.0%/DCR: 50.0%
90
1 line: 5 Pts (8.6%) 2 lines: 32 Pts (55.2%) >3 lines: 21 Pts (36.2%)
1 line: 73 Pts (81.1%) 2 lines: 7 Pts (7.8%) >3 lines: 10 Pts (11.1%)
NE: 2 Pts (2.2%) CR: 1 Pts (1.1%) PR: 47 Pts (52.2%) Stable disease: 31 Pts (34.5%) PD: 9 Pts (10.0%) ORR: 53.3%/DCR: 87.8%
Median DOT (range)
Median PFS (95% CI)
Median OS (95% CI)
7.3 mo (0.4–48.9)
12.7 mo (8.1–21.8)
5.5 mo (1.0–48.9) 10.3 mo (1.8–45.9) 5.3 mo (1.7–7.4) 6.3 mo (0.6–25.7) 1.4 mo (0.1–3.1) 2.0 mo (0.0–11.0) 6.0 mo (0.0–85.6)
PD Event
Lung
Bone
Extrathoracic
Brain
23.2 mo (12.2–NR)
32 (55.2%)
22 (37.9%)
4 (6.9%)
7 (12.1%)
9 (15.5%)
13.1 mo (4.4–NR) 17.3 mo (8.1–23.2) NR
15.1 mo (5.4–NR) NR
8 (53.3%)
7 (46.7%)
1 (6.7%)
1 (6.7%)
1 (6.7%)
16 (69.6%)
11 (47.8%)
3 (13.0%)
4 (17.4%)
4 (17.4%)
NR
0
0
0
0
0
10.0 mo (4.6–22.4) 1.3 mo (0.16–NR) 1.6 mo (1.0–6.5)
17.7 mo (5.1–NR) 1.3 mo (6.9–NR) 23.4 mo (1.5–NR)
8 (61.5%)
4 (30.8%)
0
2 (15.4%)
4 (30.8%)
3 (50.0%)
3 (50.0%)
1 (16.7%)
2 (33.3%)
1 (16.7%)
12 (100.0%) 11 (91.7%)
1 (8.3%)
2 (16.7%)
1 (8.3%)
8.0 mo (6.4–11.7)
42.5 mo (21.9–NR)
71 (78.9%)
7 (7.8%)
12 (13.3%)
6 (6.7%)
50 (55.6%)
ORR, overall response rate; DCR, disease control rate; DOT, duration of treatment; PFS, progression-free survival; CI, confidence interval; PD, progressive disease; Pts, patients; NE, not evaluable; NR, not reached; CR, complete response; PR, partial response.
Characteristics of ROS1 NSCLC
Treatment
PD Site
1379
1380 Park et al
Journal of Thoracic Oncology
Vol. 13 No. 9
Figure 5. Overall landscape of ROS1 fusion partners detected by next-generation sequencing (n ¼ 23). TFG, TRK-fused gene; TPM3, tropomyosin 3 gene; SDC4, syndecan 4 gene; CD74, CD 74 molecule gene; EZR, ezrin gene; ZCCHC8, zinc finger CCHC-type containing 8 gene; SLMAP, sacrolemma associated protein gene; MYO5C, myosin VC gene.
Figure 4. (A) Number of brain metastases when the cancer was first diagnosed. (B) Number of brain metastases that developed during the palliative treatment. (C) Kaplan-Meier plot for the time to brain metastases for patients without initial brain lesions (n ¼ 24). The median time to progression was 12.0 months (95% confidence interval: 8.5–19.1).
domain is always retained in the fusion protein, which was also shown in our eight samples (Supplementary Fig. 2).2,5 Our result are very similar to the those reported on the basis of real clinical data rather than on the basis of outcomes of clinical trials.19 Recently, Gainor et al.,19 reported a pooled analysis of ROS1-positive patients
(n ¼ 39) treated in a clinical setting. The ROS1-positive patients had 11.0 months of PFS, which was similar to the 12.7 months of PFS with treatment with a TKI from our data set. We also observed similarity in tropism in the ROS1-positive population to that of those without this mutation. At the time of diagnosis, 19.4% of patients with ROS1 rearrangements had brain metastasis, whereas 59.0% had extrathoracic metastases. These data are very similar to ours, in which 22.3% of ROS1-positive patients had brain metastases and 60.2% had extrathoracic metastases at the time of the initial diagnosis. Once again, our study supports the unique clinical characteristics of ROS1-positive NSCLC represented by tropism in distant metastases. There is no standard method for detecting ROS1 rearrangement. Although FISH is the most common method, it cannot provide information regarding fusion partners. Reverse-transcriptase polymerase chain reaction is another widely used method; however, it can miss unknown fusion genes. ROS1 immunohistochemistry is simple and considered to have high concordance with FISH. It is also difficult to interpret, given the variable location of ROS1 fusions and large amount of background staining. In contrast, NGS can provide detailed information regarding the fusion partners; however, it is limited by its high cost, long turnaround time, and need for a large amout of tissue.28,30 There is one reported case of crizotinib being used to treat a patient who was
Characteristics of ROS1 NSCLC
September 2018
CD74-ROS1–negative according to FISH but CD74-ROS1– positive according to NGS.31 In contast, in a case of ROS1 positivity according to FISH but ROS1 negativity on the basis of NGS results, the patient had a partial response to first-line pemetrexed but no response to second-line crizotinib. Both of these rare cases reflect the need for careful consideration of ROS1 positivity that is detected with a single method. This study involved the largest series of clinical data on ROS1-rearranged NSCLC to date; regardless, it does have several limitations. Given the retrospective analysis, this study was subject to potential biases. Because of the low rate of discovery of ROS1 positivity, it was difficult to perform a direct comparison between drugs. Last but not least, the comparison between patients treated with a TKI and then chemotherapy and those treated with chemotherapy and then a TKI has limitations owing to the small number of patients treated with a TKI as first-line treatment (Supplementary Table 4). In conclusion, our data reveal the unique clinical characteristics and outcomes of patients with ROS1positive NSCLC in Korea. Although ROS1-rearranged NSCLC is rare, it has a robust response to TKIs. Therefore, further efforts are needed to standardize the diagnostic method, develop new therapeutic strategies, and improve patient access.
9.
10.
11.
12.
13.
14.
Supplementary Data
15.
Note: To access the supplementary material accompanying this article, visit the online version of the Journal of Thoracic Oncology at www.jto.org and at https://doi. org/10.1016/j.jtho.2018.05.026.
16.
References 1. Bergethon K, Shaw AT, Ou SH, et al. ROS1 rearrangements define a unique molecular class of lung cancers. J Clin Oncol. 2012;30:863–870. 2. Takeuchi K, Soda M, Togashi Y, et al. RET, ROS1 and ALK fusions in lung cancer. Nat Med. 2012;18:378–381. 3. Davies KD, Doebele RC. Molecular pathways: ROS1 fusion proteins in cancer. Clin Cancer Res. 2013;19:4040–4045. 4. Murphy DA, Ely HA, Shoemaker R, et al. Detecting gene rearrangements in patient populations through a 2-step diagnostic test comprised of rapid IHC enrichment followed by sensitive next-generation sequencing. Appl Immunohistochem Mol Morphol. 2017;25:513–523. 5. Gainor JF, Shaw AT. Novel targets in non-small cell lung cancer: ROS1 and RET fusions. Oncologist. 2013;18:865–875. 6. Huber KV, Salah E, Radic B, et al. Stereospecific targeting of MTH1 by (S)-crizotinib as an anticancer strategy. Nature. 2014;508:222–227. 7. Shaw AT, Ou SH, Bang YJ, et al. Crizotinib in ROS1rearranged non-small-cell lung cancer. N Engl J Med. 2014;371:1963–1971. 8. Lim SM, Kim HR, Lee JS, et al. Open-label, multicenter, phase II study of ceritinib in patients with non-small-cell
17.
18.
19.
20.
21.
22.
1381
lung cancer harboring ROS1 rearrangement. J Clin Oncol. 2017;35:2613–2618. Solomon BJ, Bauer TM, Felip E, et al. Safety and efficacy of lorlatinib (PF-06463922) from the dose-escalation component of a study in patients with advanced ALKþ or ROS1þ non-small cell lung cancer (NSCLC) [abstract]. J Clin Oncol. 2016;34:9009. Ardini E, Menichincheri M, Banfi P, et al. Entrectinib, a Pan-TRK, ROS1, and ALK inhibitor with activity in multiple molecularly defined cancer indications. Mol Cancer Ther. 2016;15:628–639. Drilon A, Siena S, Ou SI, et al. Safety and antitumor activity of the multitargeted pan-trk, ros1, and alk inhibitor entrectinib: combined results from two phase I trials (ALKA-372-001 and STARTRK-1). Cancer Discov. 2017;7:400–409. Awad MM, Katayama R, McTigue M, et al. Acquired resistance to crizotinib from a mutation in CD74-ROS1. N Engl J Med. 2013;368:2395–2401. Paz-Ares L, de Marinis F, Dediu M, et al. Maintenance therapy with pemetrexed plus best supportive care versus placebo plus best supportive care after induction therapy with pemetrexed plus cisplatin for advanced non-squamous non-small-cell lung cancer (PARAMOUNT): a double-blind, phase 3, randomised controlled trial. Lancet Oncol. 2012;13: 247–255. Song Z, Su H, Zhang Y. Patients with ROS1 rearrangement-positive non-small-cell lung cancer benefit from pemetrexed-based chemotherapy. Cancer Med. 2016;5:2688–2693. Chen YF, Hsieh MS, Wu SG, et al. Efficacy of pemetrexedbased chemotherapy in patients with ROS1 fusion-positive lung adenocarcinoma compared with in patients harboring other driver mutations in East Asian populations. J Thorac Oncol. 2016;11:1140–1152. Shaw AT, Kim DW, Nakagawa K, et al. Crizotinib versus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med. 2013;368:2385–2394. Solomon BJ, Mok T, Kim DW, et al. First-line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J Med. 2014;371:2167–2177. Mazieres J, Zalcman G, Crino L, et al. Crizotinib therapy for advanced lung adenocarcinoma and a ROS1 rearrangement: results from the EUROS1 cohort. J Clin Oncol. 2015;33:992–999. Gainor JF, Tseng D, Yoda S, et al. Patterns of metastatic spread and mechanisms of resistance to crizotinib in ROS1-positive non-small-cell lung cancer. JCO Precis Oncol. 2017;2017. Shin HT, Choi YL, Yun JW, et al. Prevalence and detection of low-allele-fraction variants in clinical cancer samples. Nat Commun. 2017;8:1377. Soria JC, Tan DSW, Chiari R, et al. First-line ceritinib versus platinum-based chemotherapy in advanced ALK-rearranged non-small-cell lung cancer (ASCEND-4): a randomised, open-label, phase 3 study. Lancet. 2017;389:917–929. Peters S, Camidge DR, Shaw AT, et al. Alectinib versus crizotinib in untreated ALK-positive non-small-cell lung cancer. N Engl J Med. 2017;377:829–838.
1382 Park et al 23. Franchina T, Russo A, Ricciardi GR, et al. Long time response with chemotherapy in ROS1 NSCLC patient with unusual metastatic site. Cancer Biol Ther. 2016:1–5. 24. Liang Y, Wakelee HA, Neal JW. Relationship of driver oncogenes to long-term pemetrexed response in non– small-cell lung cancer. Clin Lung Cancer. 2015;16: 366–373. 25. Riess JW, Padda SK, Bangs CD, et al. A case series of lengthy progression-free survival with pemetrexedcontaining therapy in metastatic non–small-cell lung cancer patients harboring ROS1 gene rearrangements. Clin Lung Cancer. 2013;14:592–595. 26. Nicolson MC, Fennell DA, Ferry D, et al. Thymidylate synthase expression and outcome of patients receiving pemetrexed for advanced nonsquamous non-small-cell lung cancer in a prospective blinded assessment phase II clinical trial. J Thorac Oncol. 2013;8:930–939. 27. Sun JM, Ahn JS, Jung SH, et al. Pemetrexed plus cisplatin versus gemcitabine plus cisplatin according to
Journal of Thoracic Oncology
28.
29.
30.
31.
Vol. 13 No. 9
thymidylate synthase expression in nonsquamous nonsmall-cell lung cancer: a biomarker-stratified randomized phase II trial. J Clin Oncol. 2015;33:2450–2456. Rossi G, Jocolle G, Conti A, et al. Detection of ROS1 rearrangement in non-small cell lung cancer: current and future perspectives. Lung Cancer (Auckl). 2017;8: 45–55. Ahn M, Cho BC, Siena S, et al. Entrectinib in patients with locally advanced or metastatic ros1 fusion-positive non-small cell lung cancer (NSCLC). J Thorac Oncol. 2017;12:S1783. Bubendorf L, Buttner R, Al-Dayel F, et al. Testing for ROS1 in non-small cell lung cancer: a review with recommendations. Virchows Arch. 2016;469:489–503. Drilon A, Wang L, Arcila ME, et al. Broad, hybrid capturebased next-generation sequencing identifies actionable genomic alterations in lung adenocarcinomas otherwise negative for such alterations by other genomic testing approaches. Clin Cancer Res. 2015;21:3631–3639.