If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Crizotinib Associated with Ground-Glass Opacity Predominant Pattern Interstitial Lung Disease: A Retrospective Observational Cohort Study with a Systematic Literature Review
Crizotinib, an oral tyrosine kinase inhibitor that targets anaplastic lymphoma kinase, has proven to offer sustained progression-free survival in anaplastic lymphoma kinase–rearranged non–small-cell lung cancers. Occurrence of severe interstitial lung disease (ILD) was one of the crucial adverse events reported in randomized clinical trials and case reports.
Methods:
In September 2011, we observed a crizotinib-associated ILD case. Following this index case, we reviewed the clinical and computed tomographic scan features of all patients treated with crizotinib in our department, between October 2010 and July 2013, comparing patients with and without ILD. A systematic literature review was performed.
Results:
During this period, 29 patients were treated with crizotinib, five of whom developed ILD, in addition to the index case. Two types of adverse lung reactions may be observed in patients undergoing crizotinib therapy. The first is a severe, usually fatal, ILD that occurs during the first month of treatment (n = 1). The second is a less severe ILD, occurring later in time (n = 5). It occurs gradually with only few clinical symptoms, but predominant ground-glass opacities on computed tomography, along with an intensive lymphocytic alveolitis in bronchoalveolar lavage fluid. These cases had a longer response with a median progression-free survival duration at 19.9 months (17.9–23.5) compared with 6.2 months (1.2–13.6) for controls (p = 0.04).
Conclusion:
Forty-nine cases of crizotinib-associated ILD have been identified by the systematic review of the literature, including our six cases. Two types of adverse lung reactions may be observed with different presentation, prognosis, and treatment. Their potential mechanisms should be clarified. Nine patients with the less severe form of ILD were safely retreated.
An aberrant gene fusion echinoderm microtubule–associated protein-like 4-anaplastic lymphoma kinase occurs in 2% to 7% of worldwide non–small-cell lung cancers (NSCLC), encoding a cytoplasmic chimeric protein with constitutive kinase activity. In France, ALK rearrangement detected by fluorescence in situ hybridization (FISH) is observed in 3.7% of patients with advanced nonsquamous NSCLC.
Biomarkers (BM) France: results of routine EGFR, HER2, KRAS, BRAF, PI3KCA mutations detection and EML4-ALK gene fusion assessment on the first 10,000 non-small cell lung cancer (NSCLC) patients (pts).
The prevalence of ALK rearrangement is increased in young nonsmoker/light-smoker and adenocarcinoma patients. Crizotinib, an oral small-molecule tyrosine kinase inhibitor targeting ALK, MET, and ROS1 tyrosine kinases, has proven to exhibit a marked antitumor activity in ALK-rearranged NSCLC. Its efficacy has been evaluated in three multinational trials, namely two uncontrolled phase I/II and II studies (PROFILE 1001
) and a randomized controlled phase III study comparing crizotinib with single-agent pemetrexed or docetaxel in previously treated patients (PROFILE 1007
). The drug’s efficacy has been recently confirmed in a randomized controlled phase III study comparing crizotinib with cisplatin/pemetrexed doublet chemotherapy in first-line setting (PROFILE 1014).
These results have led to crizotinib approval for first-line and second- or more-line treatment of ALK-positive advanced NSCLC patients in the United States and Europe, respectively.
Crizotinib was generally well tolerated in the PROFILE studies, most adverse events being of grade 1 or 2 severity.
Occurrence of severe interstitial lung disease (ILD) was one of the crucial adverse events described. Across clinical trials (n = 1397), 34 patients (2.4%) had any grade ILD, 13 patients (0.9%) had grade 3 or 4, and seven patients (0.5%) had fatal cases. These cases generally occurred within 2 months after the initiation of treatment.
Food and Drug Administration therefore recommended to monitor patients for pulmonary symptoms indicative of ILD/pneumonitis, to exclude other potential causes of ILD/pneumonitis, and to permanently discontinue crizotinib in patients diagnosed with drug-related ILD/pneumonitis.
In September 2011, we observed a case of crizotinib-associated ILD that led to temporary crizotinib interruption, followed by reintroduction. After this index case analysis, we decided to review clinical characteristics and thoracic computed tomographic (CT) scans of all patients treated with crizotinib at our institution between October 2010 and July 2013. This article sought to compare patients with and without ILD and to perform a systematic review of the literature.
PATIENTS AND METHODS
In September 2011, a case of crizotinib-associated ILD was diagnosed in our Chest Department. All patients with diagnosed ALK-positive advanced-stage NSCLC detected using FISH (Vysis LSI ALK Dual Color, Break Apart Rearrangement Probe, Abbott Molecular, Des Plaines, IL) (M.A.) and receiving crizotinib between October 2010 and July 2013 in our institution were retrieved through the pharmaceutical database (J.F.F.). Crizotinib was delivered by our hospital pharmacy in the setting of the PROFILE 1005, 1007, and 1014 trials (n = 11), with individual and cohort temporary authorizations for use (n = 11) and European Medicines Agency authorization (Xalkori, Pfizer, New York, NY) (n = 7). The following characteristics were collected for each patient before crizotinib exposure: age, gender, smoking habits, performance status (World Health Organization), comorbidity (especially previously known ILD before crizotinib exposure), prior chemotherapy regimens, and other concomitant therapies.
All patients underwent baseline tumor assessment with body CT scan, including the brain, in the month before crizotinib initiation, followed by further assessments every 6 to 12 weeks until progression or death, in accordance with per-protocol or good clinical practice recommendations. A fluorodeoxyglucose-positron emission tomography (FDG-PET) CT was also carried out before treatment and in the event of suspected disease progression. The patients were monitored until January 31, 2014.
Definition of Cases
The diagnostic criteria of crizotinib-associated ILD were as follows: (1) there should be a history of drug exposure; (2) the clinical, imaging, and pathological pattern of lung involvement should conform to earlier observations with the drug; (3) other etiologies of ILD should be ruled out; (4) improvement should follow discontinuation of the suspected drug; and (5) symptoms should recur upon re-challenge.
All lung CT scans of each patient were reviewed by consensus by two specialists (P.C. and J.C.). ILD was defined by the occurrence of new and diffuse lung parenchymal abnormalities, including ground-glass opacities (GGO), consolidations, and reticulations (fine intralobar and coarse), on sequential CT scan examinations, according to the recommendations of the Fleischner Society.
Extensive assessments aimed at identifying the cause of ILD were conducted, whether cardiogenic, infectious, inflammatory, or related to tumor progression. Echocardiography as well as extensive microbiologic and autoimmune evaluations (including antinuclear antibody, anti-DNA antibody, antineutrophil cytoplasmic antibody, C3 and C4 complement components, rheumatoid factor, anti-cyclic citrullinated peptide antibody, and anti-aminoacyl-tRNA-synthetase antibody including precipitin line (PL) 1, 7, and 12) were performed in all patients. The serologies for domestic exposure hypersensitivity pneumonitis were also performed. Some patients received antibiotics as a therapeutic test before fiberoptic bronchoscopy was carried out. Fiberoptic bronchoscopy with bronchial and transbronchial biopsies and bronchoalveolar lavage (BAL) was conducted when possible to further characterize ILD and assess whether NSCLC lung progression had occurred. All other drugs given concomitantly to crizotinib were reviewed for each of the patients and their potential attributed lung toxicities after querying the Pneumotox database (http://www.pneumotox.com).
The study was conducted in accordance with the French law, which does not require approval of an Institutional Review Board or individual patients’ consent for such observational, noninterventional, and retrospective analyses of medical records.
Statistical Analyses
Analyses were performed using SAS Version 9.3 (SAS Institute, Cary, NC). Median and interquartile range values were provided. The threshold value for statistical significance was set at 0.05. Univariate analyses comparing the case and control groups were done using Wilcoxon–Mann–Whitney test for continuous variables and Fisher’s exact test for categorical data. Cumulative long-term survival curves were analyzed by means of the Kaplan–Meier method. Multivariate logistic regression models were constructed to identify independent risk factors associated with ILD.
Systematic Literature Review
A systematic search of Medline accessed through PubMed, Embase, Cochrane databases, and Google Scholar was performed from inception through January 2015. We included all case reports of crizotinib-associated ILD published as full-text articles or as meeting abstracts. There was no restriction on the language of publication when searching the electronic databases. Articles were identified using a combination of the following search terms: “crizotinib” AND “Lung Diseases, Interstitial.”
RESULTS
From October 2010 to July 2013, 29 patients were treated with crizotinib at our institution and included in the study cohort. Five additional cases linked to the index case of crizotinib-associated ILD were identified (6 of 29; 20.7%).
Description of the Index Case
In February 2011, a 36-year-old nonsmoker male was diagnosed in our Chest Department with ALK FISH-positive Stage IV NSCLC. Total body CT scan revealed bilateral lung, bone, intra-abdominal lymph nodes, and multiple brain metastases. In February 2012, crizotinib (250 mg twice a day) was initiated due to disease progression after a six-cycle carboplatin–paclitaxel–bevacizumab regimen followed by a 10-cycle bevacizumab maintenance therapy. This patient did not underwent any thoracic or brain radiotherapy. Crizotinib was well tolerated and resulted in partial tumor response. In September 2012, crepitant rales were found on chest auscultation, and the patient developed a nonproductive cough. Compared with the last examination, CT scan showed GGO occurrence in the lingula and left lower lobe, while tumor targets remained unchanged (Fig. 1). A BAL was performed with the following results: 540,000 cells/ml including macrophages 20%, lymphocytes 78% (CD3 cells 81%, CD4 58%, and CD8 38%), polymorphonuclear neutrophils 1%, and eosinophils 1%. Transbronchial biopsies did not find any tumoral lesions, and FDG-PET confirmed the persistence of a partial tumor response. The serologies for domestic exposure hypersensitivity pneumonitis were negative, as were the results of autoimmunity tests (Patients and Methods section). To confirm the diagnosis of drug-associated ILD, crizotinib was discontinued for 3 weeks. At this time, GGO slightly improved on CT, whereas the intense lymphocytic alveolitis persisted. Crizotinib was then reintroduced at the same posology and without corticosteroids. During the 16-month follow-up, the patient remained clinically asymptomatic while taking crizotinib 250 mg twice a day. Sequential CT scan confirmed the tumor partial response; GGO decreased in the left lower lung, disappeared in the lingula, and then occurred in right upper lobe, before finally disappearing from this lobe by the last evaluation.
FIGURE 1CT scans of index case. A and B, First CT scan at the diagnosis of ground-glass opacity and the presence of ground-glass opacity in the lingula. C, Complete disappearance of GGO in the lingula. D, Occurrence of GGO in the right upper lobe but not in the lingula. CT, computed tomography; GGO, ground-glass opacity.
Patients were stage IV and displayed two or more metastatic localizations. They received one to four chemotherapy lines before crizotinib treatment was started at 250 mg twice a day. No patient underwent any thoracic radiotherapy. At the cutoff date, after a median follow-up of 34 (24–42) months, 14 of 29 patients had died, with a median overall survival (OS) of 21 (10.7–36) months and a progression-free survival (PFS) of 8.5 (1.4–17.5) months on crizotinib therapy. Baseline characteristics did not differ between the six cases and 23 patients without ILD (Table 1), yet the median duration of crizotinib exposure was longer in cases than others (19.9 [17.9–23.5] versus 6.2 [1.2–13.6] months for patients with and without ILD, respectively, p = 0.03). In January 2014, three cases (50%) and seven patients without ILD (30.4%) were still treated with crizotinib (p = 0.63). There was no difference in the frequency and severity of adverse effects between patients with and without ILD. At this time, one case (16.7%) and 13 patients without ILD (56.5%) had died (p = 0.17). The median OS duration was not reached for cases and was at 19.1 months (10.3–32) for others. Median PFS duration on crizotinib was 19.9 months (17.9–23.5) and 6.2 months (1.2–13.6) for cases and others, respectively (p = 0.04) (Fig. 2). A multivariate logistic regression model did not identify independent risk factors associated with ILD.
TABLE 1Baseline Characteristics of the Study Population
Characteristics
Global Cohort (n = 29)
Patients with ILD (n = 6)
Patients without ILD (n = 23)
p (1 vs. 2)
1
2
Age at diagnosis in yr, median [IQR]
54 [45–65]
50 [37–65]
57 [45–65]
0.5
Gender, n (%)
Male
13 (44.8)
3 (50)
10 (43.5)
1
Female
16 (55.2)
3 (50)
13 (56.5)
Smoking, n (%)
Current and former smoker
9 (31)
1 (16.7)
8 (34.8)
0.63
Non smoker
20 (69)
5 (83.3)
15 (65.2)
ECOG performance status, n (%)
0–1
26 (89.7)
6 (100)
20 (87)
1
2–3
3 (10.3)
0 (0)
3 (13)
Underlying pulmonary disease, n (%)
COPD
1 (3.4)
0
1 (4.5)
0.31
ILD
1 (3.5)
1 (16.7)
0
Pulmonary metastasis
17 (58.6)
4 (66.7)
13 (56.6)
Pleural metastasis
10 (34.5)
4 (66.7)
6 (26)
No. of treatment lines before crizotinib, median [IQR]
All crizotinib-associated ILD cases displayed a favorable outcome without requiring permanent crizotinib discontinuation, except for one patient who died precociously. The death occurred in a 74-year-old nonsmoker female who presented with reversible ILD after receiving erlotinib therapy without any tumor response, before ALK rearrangement was identified (case 6). On day 28 after crizotinib initiation, she suddenly experienced acute respiratory distress syndrome (ARDS); an autopsy could not be conducted.
The five other cases exhibited similar features, summarized in Table 2. None of them presented ILD before crizotinib initiation. ILD occurred several months after crizotinib initiation (median, 8.5 [6.5–11.5] months). Diagnosis was based on follow-up CT scan in three patients who were clinically asymptomatic. One patient suffered from dry cough without dyspnea and fever, whereas another presented with a rapid onset, mimicking acute pneumonia. In this latter case, cutaneous vascularitis, arthralgia, and bilateral pleural effusion were also present and led to the diagnosis of drug-induced systemic lupus erythematosus (case 2, see Table 2 for autoimmunity). In the four other patients, autoimmune evaluation remained negative (Table 2). Lung CT scan revealed bilateral GGO in all patients (Fig. 1) and consolidations in one, without superimposed reticulation. Only one patient exhibited bilateral pleural effusion (case 2). Over time, the opacities tended to improve more or less completely and migrate from one lobe to another. One patient refused to undergo extensive pulmonary evaluation. BAL, carried out for the four other patients, showed a T-lymphocytic alveolitis with a predominant CD4 cell subset. Microbiological evaluation proved negative for virus, bacteria, fungi, and parasites. Tumor evaluation, including a total body CT scan and FDG-PET scan, did not reveal tumor progression to occur concomitantly to ILD diagnosis. Table S1 (Supplemental Digital Content 1, http://links.lww.com/JTO/A841) listed all other drugs given concomitantly to crizotinib for each of the patients and their potential attributed lung toxicities after querying the Pneumotox database (http://www.pneumotox.com). We have also excluded environmental exposures (data not shown).
TABLE 2Description of Patients with Ground-Glass Opacity Pattern Interstitial Lung Disease
Crizotinib was temporarily stopped in three patients and then reintroduced in all three, along with corticosteroids in one of them (case 2, 5 mg prednisone for lupus). Crizotinib was continued in two patients, in association with corticosteroids in one of them. Antibiotics when given had no effect (Table 2). No patient received diuretics for hydrosaline depletion. Long-term follow-up revealed three GGO relapses.
Results of Systematic Review
After a systematic review of the literature, we identified 43 other cases of crizotinib-associated ILD, among which 34 were notified in the PROFILE studies and nine case reports described in more detail
Acute interstitial lung disease in a patient with anaplastic lymphoma kinase-positive non-small-cell lung cancer after crizotinib therapy Disponible sur.
Regression of ILD; reintroduction of crizotinib at 250 mg twice a day then increased at 400 mg twice a day, relapse of ILD and temporary stop of crizotinib
Acute interstitial lung disease in a patient with anaplastic lymphoma kinase-positive non-small-cell lung cancer after crizotinib therapy Disponible sur.
In the present series, we reported six cases of crizotinib-associated ILD (20.7%) among 29 patients with advanced ALK-rearranged NSCLC treated by crizotinib. This frequency was much greater than in crizotinib registration studies (2.4%). Our opinion is that asymptomatic patients with occurrence of GGO on CT scan were not reported as crizotinib-associated ILD in these studies. In our cohort, there were finally only two cases of ≥3 grade ILD that is 7%. Case 6 was a fatal ILD and case 2 was a grade 3 ILD, in the context of a drug-induced systemic lupus erythematosus. By contrast, the four other cases were diagnosed by a secondary re-reading of the thoracic CT scans. The patients were asymptomatic or had very few symptoms, as it has been recently described by Asai et al.
Two types of crizotinib-associated ILD seem to exist. The first is severe, irreversible, requires the immediate stop of crizotinib, and is fatal in most of the cases, in line with one of our cases (case 6). Previous reports have already documented severe ILD/ARDS resulting in death early (<2 months) after crizotinib initiation, which suggests a potential direct drug-toxic effect
(Table 3). CT scan showed a rapid and bilateral widespread of GGO, with a final pathology diagnosis of diffuse alveolar damage in most of the cases (Table 3).
The second type is less severe, reversible, and may not need a definitive crizotinib withdrawal, in line to our index case and four of our other cases. It occurred later after crizotinib introduction (median, 8.5 [6.5–11.5] months). Patients were either mildly symptomatic or clinically asymptomatic. CT scan showed a predominant/exclusive GGO pattern of ILD (GGO-ILD), which was localized and faint. Other reported cases had few symptoms like ours (Table 3).
Acute interstitial lung disease in a patient with anaplastic lymphoma kinase-positive non-small-cell lung cancer after crizotinib therapy Disponible sur.
Acute interstitial lung disease in a patient with anaplastic lymphoma kinase-positive non-small-cell lung cancer after crizotinib therapy Disponible sur.
This latter being the only which presented a relapse. In our study, GGO-ILD favorably evolved without requiring long-term crizotinib cessation or high-dose corticosteroid use. In total, nine patients (four described in the literature
Acute interstitial lung disease in a patient with anaplastic lymphoma kinase-positive non-small-cell lung cancer after crizotinib therapy Disponible sur.
and five of our cases) were safely retreated. Finally, GGO-ILD seems to be associated with long-term control of disease progression, which made us reflect on the mechanism(s) of GGO-ILD occurrence.
A GGO pattern within the lung is not a common feature of ALK-rearranged NSCLC.
Anaplastic lymphoma kinase gene rearrangements in patients with advanced-stage non-small-cell lung cancer: CT characteristics and response to chemotherapy.
In a recent CT scan series of 36 ALK-positive patients, masses and nodules were observed in more than 40% of cases, and lymphadenopathy in 86.1%, while only one patient displayed GGO lesions. Similarly, other groups
found that most ALK-positive tumors showed a solid growth and lymphangitic pattern without GGO on CT. In our patients, the GGO lesions were not near the tumoral lesions, but at distance in normal lung, and tended to migrate over time. We also carefully investigated our patients by PET CT scan and fiberoptic bronchoscopy with BAL to eliminate disease progression. In addition, we did not identify any concomitant pulmonary infection, environmental hypersensitivity pneumonitis, or interstitial pneumonia with autoimmune features
(Table 2). We made a close study of other concomitant drugs to exclude other causes of drug-related ILD (http://www.pneumotox.com) (Table S1, Supplemental Digital Content 1, http://links.lww.com/JTO/A841). Finally, GGO decreased at least partially after stopping crizotinib with or without corticosteroids administration, which strongly suggests crizotinib imputability. Altogether, these findings indicate that GGO-ILD was associated with crizotinib exposure (Table S2, Supplemental Digital Content 1, http://links.lww.com/JTO/A841).
Other arguments suggest that crizotinib-associated ILD could be accounted for by a crizotinib-related hypersensitivity pneumonitis. The relatively long delay of occurrence (necessary for drug sensitization), resolution after crizotinib cessation, and re-occurrence after crizotinib reintroduction emphatically point to a hypersensitivity mechanism. In case 2, the observation of a systemic erythematous lupus with bilateral pleural effusion concomitant to GGO-ILD recognition was also very indicative of a drug-related immune mechanism.
Moreover, the GGO pattern on thoracic CT scan and the intense lymphocytic alveolitis observed in BAL fluid, without any pathogens, are both hallmarks for hypersensitivity pneumonitis, whether environmental or drug-related.
Risk of interstitial lung disease with gefitinib and erlotinib in advanced non-small cell lung cancer: a systematic review and meta-analysis of clinical trials.
Alternatively, several arguments could suggest that GGO-ILD reflected an effective immune antitumor response that was favored or elicited by crizotinib treatment. In all cases, this phenomenon occurred at a late time point after crizotinib introduction. The tumors were responding to crizotinib at the time of ILD occurrence. The duration of response seemed to be longer than that in patients without ILD, with a higher PFS (19.9 versus 6.2 months, p = 0.04). The mortality also seemed to be lower, with a median OS duration not achieved for the five cases at the cutoff time, after a 1.5- to 7-year postdiagnosis follow-up. However, these findings must be interpreted with caution in regard to the retrospective character of the analysis and the low number of patients included in the cohort. Moreover, an effective immune antitumor response was suggested by circulation of specific anti-ALK, B cells, and cytotoxic T cells observed in patients with ALK-positive anaplastic large cell lymphoma but not in those with ALK-negative lymphoma.
Furthermore, such a specific immune response was associated with a more limited disease at diagnosis, a lower number of circulating tumor cells, and a lower probability of relapse after remission.
Correlation of the autoantibody response to the ALK oncoantigen in pediatric anaplastic lymphoma kinase-positive anaplastic large cell lymphoma with tumor dissemination and relapse risk.
Finally, we hypothesize that GGO-ILD is a late event occurring in patients with ALK-rearranged NSCLC responding to crizotinib and could reflect an effective immune response to ALK-rearranged tumor cells.
CONCLUSION
In conclusion, our observations and the review of the literature suggest that at least two types of adverse lung reactions may occur in patients treated by crizotinib. On the one hand, GGO has an early onset after exposure to the drug. GGO is diffuse and spreads rapidly in both lungs finally resulting for an ARDS, usually fatal despite discontinuation of the drug. Corticosteroids are ineffective. On the other hand, GGO has a late onset, is rather localized faint in normal lung far from tumoral lesions, and migrate with time. The course is usually favorable after discontinuation of the drug. The interest of corticosteroids to accelerate the healing is uncertain. These two forms are described as “diffuse alveolar damage pattern” for the first and as “hypersensitivity pneumonia pattern” for the second. This GGO-ILD may be rather common if searched for by the clinician. Crizotinib in this setting should therefore be continued with careful lung monitoring, if no tumor progression has been identified after a full assessment.
ACKNOWLEDGMENTS
We thank Eliane Bertrand and Juliette Polselli (clinical research associate) for their help in monitoring patients included in clinical trials.
Biomarkers (BM) France: results of routine EGFR, HER2, KRAS, BRAF, PI3KCA mutations detection and EML4-ALK gene fusion assessment on the first 10,000 non-small cell lung cancer (NSCLC) patients (pts).
Acute interstitial lung disease in a patient with anaplastic lymphoma kinase-positive non-small-cell lung cancer after crizotinib therapy Disponible sur.
Anaplastic lymphoma kinase gene rearrangements in patients with advanced-stage non-small-cell lung cancer: CT characteristics and response to chemotherapy.
Risk of interstitial lung disease with gefitinib and erlotinib in advanced non-small cell lung cancer: a systematic review and meta-analysis of clinical trials.
Correlation of the autoantibody response to the ALK oncoantigen in pediatric anaplastic lymphoma kinase-positive anaplastic large cell lymphoma with tumor dissemination and relapse risk.
Disclosure: Dr. Cadranel declares the following relevant financial activities outside the submitted work: Board membership for Lilly and Boerhinger Ingelheim, and consultancy for Pzifer, Roche, and Novartis. Dr Gounant declares the following relevant financial activities outside the submitted work: Board membership for Roche and Lilly, and payment for development of educational presentations from Pfizer.
#Service d’Anatomie Pathologique, AP-HP, Hôpital Tenon, Paris, France
33541-3&id=gr1.jpg){kind=link}
33541-3&id=gr2.jpg){kind=link}