Rebiopsy for patients with non-small-cell lung cancer after epidermal growth factor receptor-tyrosine kinase inhibitor failure.

Although third-generation epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKI) can overcome T790M-mediated resistance in non-small-cell lung cancer (NSCLC), rebiopsy to confirm T790M status is occasionally difficult. We aimed to investigate the current tendency and the limitations of rebiopsy in clinical practice. This study included 139 consecutive NSCLC patients with EGFR mutations, who had experienced progressive disease (PD) after EGFR-TKI treatment. We retrospectively reviewed patient characteristics, tumor progression sites and rebiopsy procedures. Of 120 patients (out of the original 139) who were eligible for clinical trials, 75 (63%) underwent rebiopsy for 30 pleural effusions, 32 thoracic lesions, four bone, two liver, and seven at other sites. Rebiopsy procedures included 30 thoracocentesis, 24 transbronchial biopsies, 13 computed tomography (CT)-guided needle biopsies and 8 other procedures. Of the 75 rebiopsied patients, 71 (95%) were pathologically diagnosed with malignancy; and 34 (45%) had available tissue samples for EGFR analyses. Of the 75 biopsied patients, 61 (81%) were analyzed for EGFR mutation, using tissue or cytology samples; T790M mutations were identified in 20 (33%) of the 61 patients. Of the 120 patients, 45 (38%) did not undergo rebiopsy, because of inaccessible tumor sites (n = 19), patient refusal (n = 6) or decision of physician (n = 10). In conclusion, among patients with EGFR mutations who had PD after EGFR-TKI treatment, 63% underwent rebiopsy. Most rebiopsy samples were diagnosed with malignancy. However, tissue samples were less available and T790M mutations were identified less frequently than in previous studies. Skill and experience with rebiopsy and noninvasive alternative methods will be increasingly important.

Non‐small‐cell lung cancer (NSCLC) is the leading cause of cancer‐related death worldwide. Although having limited efficacy, cytotoxic chemotherapies have been the primary therapeutic option for unresectable NSCLC. However, epidermal growth factor receptor‐tyrosine kinase inhibitors (EGFR‐TKI) have shown dramatic responses with more acceptable toxicity in patients with EGFR‐mutant NSCLC.1 However, patients with advanced NSCLC that harbors EGFR‐sensitive mutations tend to develop progressive disease (PD) after a median response of 12 months, mainly due to acquired resistance to EGFR‐TKI through secondary mutations.2, 3 The most frequently reported secondary mutation is the EGFR Thr790Met (T790M) point mutation within exon 20, which accounts for approximately half of acquired resistance to EGFR‐TKI.4, 5, 6 Recently, third‐generation EGFR‐TKI have been reported to be effective against T790M+ NSCLC, and they are accessible through clinical trials.7, 8 We can register some of these clinical trials if rebiopsy tissue samples in PD lesions are available. However, performing rebiopsy to confirm T790M status is occasionally impossible, and obtaining tissue samples by rebiopsy remains challenging. In the present study we aim to evaluate the current status of rebiopsy at our institution and consider how to overcome the issues of rebiopsy in the clinical setting.

Patients and Methods

Patients

We initially screened 139 consecutive patients with NSCLC harboring EGFR‐sensitive mutations who had experienced PD after any EGFR‐TKI treatment at Shizuoka Cancer Center between January 2014 and March 2015. Both tissue sampling and cytologic specimen sampling were defined as rebiopsy. Patient characteristics, tumor progression sites and rebiopsy procedure were retrospectively reviewed from medical records. We also identified reasons for not undergoing rebiopsy in the non‐rebiopsied patients. The primary objective of our study was to evaluate the frequency of rebiopsy and incidence of T790M mutation compared with previous studies. The second objective was to identify factors that make rebiopsy difficult. These factors included “inaccessible lesions,” which were defined as intracranial lesions or difficult target lesions (<20 mm) in tissue sampling. “Decision of physician” included patients who had other progressive lesions that required emergent radiation therapy (e.g. metastases in spinal cord and symptomatic brain metastases).

Processing method for pleural effusion cytology samples

Pleural effusion cytology samples were provided by Shizuoka Cancer Center. From 30 pleural effusion cytology samples collected from patients, 27 were diagnosed with NSCLC (adenocarcinoma) between January 2014 and March 2015 and confirmed by a pathologist to contain tumor cells. Samples were frozen within 10 and 30 min of sampling and stored at −80°C. Frozen samples were thawed at 37°C and refrozen rapidly to disrupt the cells and ensure an even distribution, then divided into equal aliquots in the laboratory. DNA was extracted at the laboratory using our own standard procedures and Cobas EGFR Mutation Test Kits (Roche Diagnostics K.K., Tokyo, Japan).

Epidermal growth factor receptor mutational analysis

Rebiopsies were conducted with various lesions at our institution. We used the Scorpion Amplification Refractory Mutation System (Scorpion ARMS method) in EGFR mutational analyses.9 Some patients received rebiopsies on several occasions or at multiple lesions. In these cases, positive results of EGFR mutations or T790M mutation had priority to be adopted. No other acquired resistant molecular mechanisms (e.g. MET amplification) were examined in the present study.

Statistical analysis

Statistical analyses were performed using JMP 9 (SAS Institute, Cary, NC, USA), and the χ2 and Mann–Whitney U‐tests were used to evaluate differences in categorical and continuous variables between the two groups, respectively. P < 0.05 was considered significant. This retrospective study was approved by the institutional review board of Shizuoka Cancer Center.

Results

Rebiopsy rate after epidermal growth factor receptor‐tyrosine kinase inhibitor failure

Among 139 patients who had experienced PD after EGFR‐TKI treatment, 19 patients were ineligible for clinical trials because of poor performance status (PS; n = 10), comorbidity (n = 7), or because they were 85 and 87 years old (n = 2). Among 120 patients, tumor progression sites included 36 pleural effusion, 57 thoracic primary/metastatic lesions, 26 brain metastases, 21 bone metastases, 15 lymph node metastases, 7 hepatic metastases and 8 other lesions. Of the 120 remaining patients, 75 (63%) underwent rebiopsy. Individual characteristics of 120 patients included in this study are shown in Table 1. The rebiopsy and non‐rebiopsy groups did not significantly differ in age, sex, smoking status, PS, EGFR mutation type or response to initial EGFR‐TKI treatment. Anatomical sites of rebiopsy were as follows: 30 pleural effusion, 32 thoracic lesions, four bone lesions, two hepatic lesions and seven other lesions (pericardial effusion [n = 2], adrenal lesion [n = 1], skin lesion [n = 1], brain lesion [n = 1], leptomeningeal lesion [n = 1] and ascites [n = 1]). Rebiopsy procedures included 30 thoracocentesis, 24 transbronchial biopsies, 14 CT‐guided needle biopsies and 7 other procedures (surgery of the bone lesion [n = 1] and brain lesion [n = 1], pericardiocentesis [n = 2], skin biopsy [n = 1], abdominocentesis [n = 1] and lumber puncture [n = 1]), as shown in Table 2. Of the 75 patients in the rebiopsy group, 71 (95%) were pathologically diagnosed with malignancy. Tissue samples for EGFR analyses were available in 34 (45%) of 75 patients, and EGFR mutational analyses were performed in 61 (81%) of 75 patients by using tissue or cytology samples. T790M mutations were identified in 20 (33%) of these 61 patients (Fig. 1). Subsequent chemotherapies after PD are shown in Table 3. Third‐generation EGFR‐TKI were administered significantly more frequently in the rebiopsy group (24%) than in the non‐rebiopsy group (9%; P = 0.04). EGFR‐TKI approved in Japan (gefitinib, erlotinib, afatinib) were administered significantly more frequently in the non‐rebiopsy group (40%) than in the rebiopsy group (20%; P = 0.01).

Table 1

Individual baseline characteristics (n = 120 in total)

	Rebiopsy group (n = 75)	Non‐rebiopsy group (n = 45)	P
Mean age (range) in years	68 (34–87)	71 (34–86)	0.23
Sex
Male	26	13	0.51
Female	49	32
Smoking status
Smoker	32	19	0.96
Non‐smoker	43	26
Performance status
0–1	55	33	1.00
2–4	20	12
Metastatic site
Brain	14	17	0.02
Bone	25	16	0.80
EGFR mutation
Exon 19 deletion	45	25	0.77
Exon 21 L858R	27	17
Others	3	3
Response to initial EGFR‐TKI treatment
Complete response	0	0	0.76
Partial response	47	32
Stable disease	22	11
Progressive disease	2	1

EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitor.

Table 2

Anatomical progressive diseasesites and methods for rebiopsy (n = 75)

	n (%)
Rebiopsy sites
Pleural effusion	30 (40)
Thoracic lesion	32 (43)
Bone lesion	4 (5)
Hepatic lesion	2 (3)
Others	7 (9)
Pericardial effusion	2 (3)
Adrenal lesion	1 (1)
Skin lesion	1 (1)
Brain lesion	1 (1)
Leptomeningeal lesion	1 (1)
Ascites	1 (1)
Rebiopsy methods
Thoracocentesis	30 (40)
Transbronchial biopsy	24 (32)
CT‐guided needle biopsy	14 (19)
Others	7(9)
Surgery (bone, brain)	2 (3)
Pericardiocentesis	2 (3)
Skin biopsy	1 (1)
Abdominocentesis	1 (1)
Lumber puncture	1 (1)

CT, computed tomography.

Figure 1

Overview of rebiopsies among patients with non‐small‐cell lung cancer (NSCLC) treated with early‐version epidermal growth factor receptor‐tyrosine kinase inhibitors (EGFR‐TKI). Patients were tested with the aim of enrolling them in trials for third‐generation EGFR‐TKI, which is notably effective against NSCLC with T790M mutations in the gene. , EGFR mutation carrier; PD, progressive disease; PS, performance status.

Table 3

Subsequent treatments for non–small‐cell lung cancer patients with PD after rebiopsy (n = 120)

	Rebiopsy group (%)	Non‐rebiopsy group (%)	P
	n = 75	n = 45
Subsequent regimens
Third‐generation EGFR‐TKI	18 (24)	4 (9)	0.04a
Approved EGFR‐TKI (gefitinib, erlotinib, afatinib)	15 (20)	18 (40)	0.01a
Chemotherapy other than EGFR‐TKI	19 (25)	6 (13)	0.12
BSC or observation (including beyond PD)	23 (31)	17 (38)	0.42

P<0.05. BSC, best supportive care; EGFR‐TKI, epidermal growth factor receptor‐tyrosine kinase inhibitor; PD, progressive disease.

Rebiopsy after first epidermal growth factor receptor‐tyrosine kinase inhibitor failure

Of 67 patients who had experienced PD after initial administration of EGFR‐TKI, 5 were judged as ineligible for clinical trials; 27 (44%) patients underwent rebiopsy, of whom 25 (93%) of the 27 patients were pathologically diagnosed with malignancy, using tissue or cytology samples. EGFR mutational analyses were performed for 23 (85%) of the 27 patients. T790M mutations were identified in 7 (30%) of the 23 patients before subsequent systemic therapy. T790M detection rates were similar between patients who underwent rebiopsy after their first EGFR‐TKI therapy (30%) and those who were rebiopsied after their second or subsequent TKI therapies (35%, P = 0.71).

Rebiopsy for primary/metastatic lesions

Thirty patients underwent rebiopsy for pulmonary lesions by CT‐guided needle biopsies or transbronchial biopsies: 18 patients for primary lesions and 13 patients for metastatic lesions. The detection rate of T790M mutations did not significantly differ between those rebiopsied for primary pulmonary lesions (4 of 18 patients [22%]) and those rebiopsied for metastatic lesions (including extrapulmonary lesions) (16 of 43 patients [37%]; P = 0.26).

Reasons for patients not undergoing rebiopsy

Among 120 patients eligible for clinical trials who had experienced PD after EGFR‐TKI treatment between January 2014 and March 2015, 45 (38%) patients did not undergo rebiopsy because of inaccessible tumor sites (n = 19: 18 with brain metastasis and 1 with small pulmonary lesion), decision of physician (n = 10), patient refusal (n = 6) or unknown reasons (n = 10). Patients affected by “decision of physician” could not undergo rebiopsy even after completing emergent therapies (e.g. palliative radiotherapy for spinal cord compression or symptomatic brain metastases), mainly because of deterioration in PS.

Discussion

In this retrospective study, 63% of 120 consecutive patients with NSCLC who harbored EGFR mutations and had PD after EGFR‐TKI treatment underwent rebiopsy, and the frequencies of T790M mutation at rebiopsy were 33% in the clinical setting at our institution. We conducted this retrospective study because of the increasing importance of rebiopsy after EGFR‐TKI failure. Third‐generation EGFR‐TKI can solve the major problem of acquired resistance after EGFR‐TKI treatments in NSCLC patients with EGFR sensitive mutations, by inhibiting both EGFR activating and resistance mutations while sparing wild‐type EGFR.10, 11 In recent studies, third‐generation TKI demonstrated promising response rates against tumors with acquired EGFR T790M.7, 8 Therefore, rebiopsies to detect T790M mutations are often required to enroll patients into clinical trials of third‐generation EGFR‐TKI. However, rebiopsy is occasionally challenging for both patients and physicians.

In our study, frequencies of rebiopsy for NSCLC patients harboring EGFR mutations who had PD after EGFR‐TKI treatment seemed to be lower than in previous studies (60–86%).12, 13, 14 These earlier studies may have been more selective than our study in terms of patient characteristics (e.g. more strict exclusion criteria) and rebiopsy methods (e.g. fewer cytological methods such as lumber punctures and thoracocentesis, and more tissue acquisition methods such as video‐assisted thoracic surgery, lung core biopsy, CT‐guided biopsy and transbronchial biopsy); our study might show rebiopsy frequencies in real clinical settings more precisely. One reason for the lower rebiopsy rate in our study is that our study included NSCLC patients with only inaccessible lesions (brain metastases or very small pulmonary lesions). In such cases, rebiopsy is not performed until other accessible lesions emerge. However, the rebiopsy may be limited to indolent cases because lung carcinoma often shows rapid growth. Another reason is decisions by physicians whose patients experienced PD. Indeed, patients who show PD as bone metastases or multiple brain metastases frequently need emergent radiation therapies to prevent concomitant symptoms, and physicians may hesitate to perform rebiopsy. In such cases, the decision to perform subsequent rebiopsies may depend on the efficacy of emergent therapies. In the present study, the T790M mutation frequency at rebiopsy (33%) was also low compared with previous studies (30–83%).13, 15, 16, 17, 18 The incidence of T790M mutation in Japanese patients with EGFR‐mutant NSCLC following EGFR‐TKI therapies might be lower than that in Caucasians.15, 16 T790M mutations are possibly induced by EGFR‐TKI exposure, and EGFR‐TKI treatment beyond PD or duration of EGFR‐TKI before rebiopsy might be related to the incidence of T790M mutation. However, little data is available that supports these hypotheses.

Third‐generation EGFR‐TKI were administered after PD significantly more frequently in the rebiopsy group than in the non‐rebiopsy group, while EGFR‐TKI approved in Japan were administered significantly more frequently in the non‐rebiopsy group (see Table 3). The result showed that some patients in the non‐rebiopsy group could receive sequential EGFR‐TKI therapies after PD: they will be able to receive third generation EGFR‐TKI if better rebiopsy methods or alternatives to detect T790M mutation are developed. Skill and experience with rebiopsy, especially of transbronchial biopsies and CT‐guided needle biopsies, will be also important in acquiring sufficient tissue samples. Yoon et al.14 report that repeat biopsy by CT‐guided needle biopsies can be performed safely with mild potential complications (such as pneumothorax and pulmonary hemorrhage) in patients with NSCLC. However, skilled experts are needed for CT‐guided biopsies, especially for extrapulmonary lesions such as hepatic metastases and bone metastases, which can make rebiopsy especially challenging. Furthermore, cytology sampling, such as pleural effusion, pericardial effusion and cerebrospinal fluid, is often more easily available than tissue sampling under the presence of skilled experts; in the present study, tissue samples were available in only 45% of the patients. Liquid biopsies are an alternative for determining T790M mutation status among patients who are unable to undergo tissue rebiopsy, if they have sufficient diagnostic yields.19, 20 Some reports show that the concordance of T790M mutation status between liquid biopsy and conventional biopsy methods to be 50–65%.20, 21 Recently, osimertinib was approved in the EU according to blood‐based test for circulating tumor DNA (ctDNA), and it might be meaningful for patients and physicians to have alternatives to detect T790M mutation. Our study also revealed that the T790M detection rate in primary lesions was not significantly higher than that in metastatic lesions. Therefore, we need not hesitate to perform rebiopsy, regardless of the rebiopsy sites.

This study is limited by its retrospective nature. Our results may also have been affected by our single‐institution cohort and colleagues, and rebiopsy or EGFR mutation testing was performed based on the decision of each physician. Furthermore, acquired resistance is not strictly defined as Jackman et al.22 propose in their analysis. Some patients received rebiopsies several times and their T790M mutation status changed heterochronously. The disappearance of T790M mutation may be due to low sensitivity of the EGFR mutation analysis method. However, the Scorpion ARMS method, which was used in the present study, showed reasonable sensitivity and specificity in previous reports.9 Some reports also suggested that T790M was spatiotemporally heterogeneous due to selective pressure from EGFR‐TKI, which might be another reason.18, 23, 24 Finally, we did not evaluate other resistance mechanisms in these rebiopsies, such as MET‐amplification, epithelial‐to‐mesenchymal transition, BRAF mutation, HER2‐amplification and PIK3CA‐mutations, as previously mentioned. This limits the accuracy of the relative frequency of the various mechanisms we present, and potentially underestimates the prevalence of overlap among the different mechanisms of resistance.

In conclusion, 63% of the NSCLC patients harboring EGFR mutations who had PD after EGFR‐TKI treatment underwent rebiopsy, and most of these rebiopsy samples were diagnosed with malignancy in the clinical setting at our institution. However, tissue samples were available for fewer than half of the patients and T790M mutations were identified in no more than 33% of the patients who underwent rebiopsy. Skill and experience with rebiopsy to obtain adequate tissue samples and noninvasive procedures, including analysis using cytology specimens and liquid biopsies, will be increasingly important in assessing NSCLC patients who develop PD after treatment with first‐generation or second‐generation EGFR‐TKI.

Disclosure Statement

The authors have no conflicts of interest to declare.