Abstract
Introduction
The EML4–anaplastic lymphoma kinase (ALK) translocation is a recognized oncogenic driver in non-small cell lung cancer. We investigated immunohistochemistry (IHC) screening with fluorescence in-situ hybridization (FISH) confirmation for ALK detection and estimated the prevalence of ALK-positivity in our patient cohort of never smokers, together with differences in clinical outcomes and prognostic factors for patients with ALK-positive and ALK-negative tumors.
Methods
We designed a three-phase study (training, validation, and testing) in 300 never-smokers with lung adenocarcinoma from the observational Mayo Clinic Lung Cancer Cohort. Tumor samples were tested using IHC and FISH, and concordance between the methods was assessed. Clinical outcomes were assessed via 5-year progression or recurrence-free survival from diagnosis. Prognostic factors for ALK-positive tumors and metastases were also investigated.
Results
ALK-positive patients were significantly (p<0·05) younger and had higher-grade tumors than ALK-negative patients. ALK-positivity was 12.2% by IHC and confirmed at 8.2% of tumors by FISH, with complete concordance between IHC 3+/0 and FISH+/− assessments, respectively. Five-year risk of progression or recurrence was doubled for patients with ALK-positive compared with ALK-negative tumors; ALK-positive tumors also appeared to be associated with a higher risk of brain and liver metastases.
Conclusions
Our findings suggest that ALK-positivity is associated with a significantly poor outcome in non-smoking-related adenocarcinoma, and that ALK-positive tumors may be associated with an increased risk of brain and liver metastases compared with ALK-negative disease. Consequently, an unmet medical need exists in ALK-positive lung cancer patients, and effective ALK-specific therapies are needed now.
Keywords: EML4-anaplastic lymphoma kinase (ALK), non small cell lung cancer (NSCLC), immunohistochemistry (IHC), fluorescence in-situ hybridization (FISH), progression and recurrence free survival (PFS/RFS)
Introduction
The echinoderm microtubule-associated protein-like 4 (EML4)–anaplastic lymphoma kinase (ALK) translocation was discovered in non-small cell lung cancer (NSCLC) in 2007. (1) The EML4–ALK translocation can result in constitutive ALK kinase activity, and represents an oncogenic addiction pathway in lung cancer. (2) Other ALK fusion variants have also been reported, with uncertain oncokinase functions in NSCLC. (3-5) Several studies have shown ALK gene rearrangement (ALK-positivity) to correlate with never or light/former smoking status, younger age, adenocarcinoma histology, and to rarely coincide with EGFR or KRAS mutation. (6-11)Retrospective studies have reported prevalence estimates for ALK-positivity ranging from 1·6%(7)to 13%;(8) the variability resulting from factors including the methodology used to detect ALK gene rearrangement (e.g. polymerase chain reaction [PCR] with limited probe sets) and patient or tissue selection criteria. Therefore, several questions remain to be answered, including the true prevalence of ALK-positivity in unselected patient populations, and whether prevalence varies by disease stage, geography, or ethnicity. Moreover, the natural history and clinical outcome of ALK-positive versus ALK-negative patients needs to be fully defined.
In an effort to find a cost-effective method for selecting ALK-positive NSCLC patients who may benefit from ALK inhibitor therapies, we studied the role of scored immunohistochemistry (IHC) as a screening test. Given the high concordance between IHC score and ALK status by fluorescence in-situ hybridization (FISH), we proposed an algorithm for FISH confirmation. (12) In this study, we have further validated and tested the algorithm and performed a controlled comparison of clinical outcomes for ALK-positive versus ALK-negative adenocarcinoma patients. Our four specific objectives were to: (1) assess whether IHC can be a practical tool for ALK screening in adenocarcinoma with a confirmatory FISH test; (2) estimate the prevalence of ALK-positivity in our enriched study cohort by either or both IHC and FISH tests; (3) describe clinical characteristics of cases by IHC score, FISH status, and combined IHC and FISH results; and (4) most importantly, evaluate clinical outcomes of ALK-positive versus ALK-negative cases, while controlling for clinically important prognostic factors, in a naturally followed observational patient cohort.
Materials and Methods
Study design, sampling and data collection
In order to evaluate the ALK test concordance and estimate ALK positivity prevalence, we designed a three-phase study in an ALK-positive-enriched patient population: approximately 100 patients were included in each phase as a training (n=100), validation (n=99), and testing set (n=101), respectively (eFigure 1).
All patients were never smokers, defined as having smoked zero to 99 cigarettes during their lifetime, and were selected from the Mayo Clinic Lung Cancer Cohort, an observational follow-up study. (13, 14) The sample enrichment strategy included only never-smokers with adenocarcinoma diagnosed between 1997 and 2008 (none treated with crizotinib), with surgically resected samples, preserved as formalin-fixed paraffin-embedded tissue. These samples were from surgical treatments (i.e., lobectomy, pneumonectomy, and sleeve resection) or biopsies (i.e., wedge, pleural, and/or thoracotomy-based). In total, 300 cases were eligible for inclusion in the analyses. All eligible cases had banked tissue samples from surgical resections consisting of wedge resection or more extensive surgeries; available slides from each case were reviewed by two pathologists (ESY and MCA) to verify the diagnosis and to select the representative block containing the most architecturally intact tumor cells.
Follow-up is accomplished through detailed medical record review and patient questionnaires at within 6 months post diagnosis and then annually. For deceased patients, the follow-up questionnaire is sent to the next-of-kin. Additional details regarding patient identification and follow-up have been described previously.(15, 16) Information abstracted from the medical record for each patient included age, gender, race, smoking status, history of tobacco exposure, other medical conditions, symptoms at presentation, and Eastern Cooperative Oncology Group Performance Status (ECOG-PS). For the initial diagnosis of primary NSCLC, date of diagnosis, clinical stage, histopathological type and grade of tumor, type of surgical resection, and the use of adjuvant therapy were recorded. The evidence of progression and/or recurrent disease was abstracted from Mayo Clinic medical records or other reliable sources. Additional sources of data collection included lung cancer follow-up questionnaires, tumor registry questionnaires routinely sent to the patient from Mayo Clinic, and correspondence letters or copied medical records received from outside clinicians. When the presence or absence of progression and/or recurrent disease was documented at another institution, data were included only when the information was specific and considered reliable.(15) In order to standardize the quality of follow-up information across the all patients, follow-up information completed as of December 31, 2010 was included in the analysis.
Immunohistochemistry and fluorescence in-situ hybridization for ALK rearrangement
Immunohistochemistry
IHC using ALK1 monoclonal antibody (Dako, Carpinteria, California, USA) was performed as described previously by Yi et al. (12) An IHC score was assigned to each case according to the following criteria with at least 10% of tumor cells showed the designated staining pattern: 3+, intense, granular cytoplasmic staining; 2+, moderate, smooth cytoplasmic staining; 1+, faint cytoplasmic staining; and 0, no staining. IHC scoring was performed without the knowledge of FISH results.
Fluorescence in-situ hybridization for ALK rearrangement
Interphase molecular cytogenetic studies using a commercially available ALK probe (Vysis, Des Plaines, Illinois, USA) were performed as described previously by Yi et al. (12) FISH for ALK locus rearrangement was considered positive if 15% or more of at least 100 cells counted showed splitting of the florescent probes flanking the ALK locus. All FISH interpretation was performed without the knowledge of immunohistochemistry results for ALK.
Study endpoints
Survival was assessed as 5-year progression-free or recurrence-free survival (PFS/RFS) from diagnosis (events occurring after 5 years were censored and cases died before occurrence of recurrence/progression counted as events). To control for confounding effects of known predictors for lung cancer progression or recurrence, the following variables were matched or adjusted for: age at diagnosis, sex, lung cancer stage (I, II, III, and IV), and mode of treatment (surgery only, surgery and chemo/radiation, other/none/unknown).
Statistical analyses
Descriptive statistics were summarized and compared for a number of factors, including age at diagnosis, sex, ethnicity (U.S. white vs. other), grade of tumor differentiation (well differentiated, moderately differentiated, poorly or undifferentiated), stage, and treatment modality. Pre-progression treatment included all treatment received before primary progression, recurrence or development of second primary tumors. Both univariate and multivariate survival analyses were conducted by ALK status using IHC, FISH, and both combined IHC and FISH test scores comparing patients with ALK-positive tumors with patients with ALK-negative tumors. Five levels of analyses were performed: (1) comparison of patients' characteristics by ALK status using Kruskal–Wallis tests (continuous variables) and chi-square tests or Fisher's exact tests (categorical variables); (2) IHC and FISH tests concordance of ALK status analyzed by chi-square tests for homogeneity; (3) the prevalence of ALK-positive cases estimated by IHC test only, by FISH test only, and by both tests; (4) PFS/RFS by ALK status using adjusted (unmatched) and matched (a nested case–case) Cox proportional hazard models in which the hazard ratio or relative risk of the endpoint was estimated; and (5) comparison of detailed events of progression and recurrence between ALK-positive and ALK-negative cases. The ‘full model’ used all covariates (age, sex, stage, and pre-progression treatment) and the ‘select model’ employed a stepwise model selection procedure to pick the significant covariates. A Cox proportional hazards model was used to estimate the partial effects of stage; specifically for each individual, two survival curves were estimated for each value of stage, with individual covariates remaining fixed. The adjusted survival curves are then calculated as the within-stage mean estimated survival of all study subjects.
Results
Descriptive analysis
Patient characteristics by FISH status, IHC score, and both IHC score and FISH status combined are shown in Table 1. In total, 300 patients had IHC scores, 216 had FISH results, and 300 had either of the two tests. All cases with IHC1+, 2+, or 3+ have been tested for FISH and the 84 cases without FISH were IHC score 0. Across the three groups, a statistically significant difference between ALK-positive and ALK-negative cases was observed for age at diagnosis, tumor cell differentiation (grade), and treatment modality. In addition, stage was significantly different between FISH-positive and FISH-negative groups. Patients with ALK-positive tumors were younger and had more aggressive histologic grade and more advanced disease stage than those with ALK-negative tumors.
Table 1. Patient characteristics by IHC score, FISH status, and IHC score/FISH status combined. Shaded variables are statistically different (p<0.05) among the comparison groups.
FISH = fluorescence in-situ hybridization. IHC = immunohistochemistry. PFS/RFS = progression- or recurrence-free survival.
Not all patients had surgically-resected lung cancer.
| Characteristics of 300 patients with IHC scores | Characteristics of 216 patients by FISH status | Characteristics of 300 patients by IHC score/FISH status (FISH+/IHC3+/2+ vs FISH–/IHC0/1+) | ||||||
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| IHC3+ (n=18) | IHC2+ (n=14) | IHC1+ (n=65) | IHC0 (n=203) | Positive (n=22) | Negative (n=194) | Positive (n=34) | Negative (n=266) | |
| Age at diagnosis, years | ||||||||
| Median | 59·0 | 56·5 | 70·0 | 70·0 | 62·5 | 70·0 | 58·0 | 70·0 |
| Range | 36·0–77·0 | 40·0–84·0 | 41·0–91·0 | 17·0–89·0 | 36·0–77·0 | 32·0–91·0 | 36·0–84·0 | 17·0–91·0 |
| Gender, n (%) | ||||||||
| Female | 11 (61·1) | 13 (92·9) | 46 (70·8) | 155 (76·4) | 14 (63·6) | 147 (75·8) | 26 (76·5) | 199 (74·8) |
| Male | 7 (38·9) | 1 (7·1) | 19 (29·2) | 48 (23·6) | 8 (36·4) | 47 (24·2) | 8 (23·5) | 67 (25·2) |
| Race, n (%) | ||||||||
| Caucasian | 18 (100·0) | 11 (78·6) | 58 (89·2) | 176 (86·7) | 21 (95·5) | 172 (88·7) | 31 (91·2) | 232 (87·2) |
| Other | 0 (0·0) | 3 (21·4) | 7 (10·8) | 27 (13·3) | 1 (4·5) | 22 (11·3) | 3 (8·8) | 34 (12·8) |
| Grade of differentiation, n (%) | ||||||||
| Well differentiated | 4 (22·2) | 5 (35·7) | 27 (41·5) | 125 (61·6) | 5 (22·7) | 105 (54·1) | 10 (29·4) | 151 (56·8) |
| Moderately differentiated | 10 (55·6) | 8 (57·1) | 31 (47·7) | 62 (30·5) | 12 (54·5) | 75 (38·7) | 19 (55·9) | 92 (34·6) |
| Poorly/undifferentiated | 4 (22·2) | 1 (7·1) | 7 (10·8) | 16 (7·9) | 5 (22·7) | 14 (7·2) | 5 (14·7) | 23 (8·6) |
| Stage, n (%) | ||||||||
| I | 6 (33·3) | 7 (50·0) | 32 (49·2) | 130 (64·0) | 7 (31·8) | 122 (62·9) | 14 (41·2) | 161 (60·5) |
| II | 2 (11·1) | 2 (14·3) | 2 (3·1) | 10 (4·9) | 2 (9·1) | 7 (3·6) | 4 (11·8) | 12 (4·5) |
| III | 5 (27·8) | 4 (28·6) | 16 (24·6) | 37 (18·2) | 7 (31·8) | 37 (19·1) | 10 (29·4) | 52 (19·5) |
| IV | 5 (27·8) | 1 (7·1) | 15 (23·1) | 26 (12·8) | 6 (27·3) | 28 (14·4) | 6 (17·6) | 41 (15·4) |
| Treatment modality, n (%) | ||||||||
| Only surgery | 5 (27·8) | 9 (64·3) | 37 (56·9) | 136 (67·0) | 7 (31·8) | 122 (62·9) | 15 (44·1) | 172 (64·7) |
| Surgery and chemo/radiation | 9 (50·0) | 5 (35·7) | 15 (23·1) | 44 (21·7) | 11 (50·0) | 49 (25·3) | 15 (44·1) | 58 (21·8) |
| Other/none/unknown | 4 (22·2) | 0 (0·0) | 13 (20·0) | 23 (11·3) | 4 (18·2) | 23 (11·9) | 4 (11·8) | 36 (13·5) |
| Pre-progression treatment, n (%) | ||||||||
| Only surgery | 8 (44·4) | 11 (78·6) | 40 (61·5) | 147 (72·4) | 10 (45·4) | 137 (70·6) | 20 (58·8) | 186 (69·9) |
| Surgery and chemo/radiation | 5 (27·8) | 3 (21·4) | 12 (18·5) | 32 (15·8) | 7 (31·8) | 34 (17·5) | 9 (26·5) | 43 (16·2) |
| Other/none/unknown | 5 (27·8) | 0 (0·0) | 13 (20·0) | 24 (11·8) | 5 (22·7) | 23 (11·9) | 5 (14·7) | 37 (13·9) |
| Median PFS/RFS | 2·05 | 3·33 | 3·89 | 4·40 | 2·18 | 4·48 | 2·66 | 4·27 |
Not all resected tissue came from curative surgery.
Some tissue samples were from wedge, pleural, or thoracotomy-based biopsies.
“Unknown” treatments were for patients who received treatment outside our institution and the data are still being verified.
“Other” treatments included e.g., TAC of pleura and removal of solitary brain metastasis.
Concordance analysis
There were 225 samples tested by both assays from 221 distinct patients (four patients each had two samples tested). Concordance rates between IHC scores and FISH status are presented in eTable 1 (available online). Complete concordance (100%) was observed between IHC3+ scores and FISH-positive status and between IHC0 scores and FISH-negative status. Relatively high concordance was seen between IHC1+ scores and FISH-negative status (96·9% of IHC1+ was FISH-negative); while reasonable concordance was noted between IHC2+ scores and FISH-negative status (85·7% of IHC2+ was FISH-negative).
ALK-positivity prevalence estimation
The prevalence of ALK tumors was estimated using five methods (eTable 2, available online). (1) Using IHC as a screening test, the prevalence of ALK-positivity was 10·7% for IHC3+/2+, 6·0% for IHC3+, and 4·7% for IHC2+; (2) by FISH, the prevalence was 8·2%; (3) using IHC as a screening test and FISH as a confirmatory ‘gold standard’ test, the prevalence was 8·2% for FISH-positive or IHC3+; (4) the prevalence was 12·2% for FISH-positive or IHC3+ or IHC2+ in the training and validation sets; and (5) counting all FISH-positive and all IHC3+/2+ cases in the entire prevalence cohort, the prevalence was 11·3%.
Survival analysis
PFS/RFS was evaluated using two comparative analyses. In the unmatched analysis, three models (unadjusted, full, and select) considered samples as ALK-positive using three alternative definitions: IHC3+, FISH+, or both FISH and IHC3+. The select model adjusted for tumor stage only. As shown in Table 2, results were consistent across all models: the risk of experiencing lung cancer progression or recurrence within 5 years after diagnosis was over two-fold greater in ALK-positive cases compared with ALK-negative cases. Stage-adjusted survival curves for the three ALK-positive definitions are shown in Figure 1.
Table 2. Analysis of 5-year PFS/RFS by Cox proportional hazards models.
| HR (95% CI) | P value | |
|---|---|---|
| PFS/RFS, IHC3+ (n=18) vs IHC0/1+ (n=264) | ||
| Unadjusted | 3.33 (1.79, 6.17) | <0.0001 |
| Full model1 | 2.14 (1.12, 4.07) | 0.0210 |
| Age at diagnosis (per 1 year) | 0.99 (0.97, 1.02) | 0.5658 |
| Sex (males) | 1.47 (0.89, 2.43) | 0.1345 |
| Stage (vs. I) | ||
| II | 1.40 (0.41, 4.78) | 0.5931 |
| III | 3.74 (1.97, 7.09) | <.0001 |
| IV | 3.19 (1.48, 6.82) | 0.0029 |
| Treatment (vs. only surgery) | ||
| Surgery and chemotherapy or radiation | 1.68 (0.89, 3.18) | 0.1101 |
| Other/none/unknown | 2.32 (1.09, 4.91) | 0.0287 |
| Select model2 | 2.37 (1.27, 4.46) | 0.0071 |
| Stage (vs. I) | ||
| II | 1.40 (0.42, 4.46) | 0.5862 |
| III | 5.26 (3.07, 9.03) | <.0001 |
| IV | 5.04 (2.71, 9.36) | <.0001 |
| PFS/RFS, FISH+ (n=22) vs FISH– (n=191) | ||
| Unadjusted | 3.10 (1.75, 5.52) | 0.0001 |
| Full model1 | 1.95 (1.07, 3.58) | 0.0303 |
| Age at diagnosis (per 1 year) | 1.00 (0.98, 1.02) | 0.7575 |
| Sex (males) | 1.36 (0.80, 2.34) | 0.2588 |
| Stage (vs. I) | ||
| II | 0.99 (0.23, 4.27) | 0.9872 |
| III | 3.49 (1.78, 6.83) | 0.0003 |
| IV | 3.46 (1.43, 8.36) | 0.0058 |
| Treatment (vs. only surgery) | ||
| Surgery and chemotherapy or radiation | 1.40 (0.72, 2.71) | 0.3183 |
| Other/none/unknown | 2.24 (0.91, 5.53) | 0.0800 |
| Select model2 | 2.11 (1.17, 3.82) | 0.0134 |
| Stage (vs. I) | ||
| II | 1.02 (0.24, 4.36) | 0.9741 |
| III | 4.48 (2.55, 7.87) | <.0001 |
| IV | 5.59 (2.86, 10.93) | <.0001 |
| PFS/RFS, FISH+/IHC3+ (n=22) vs FISH–/IHC0/1+ (n=274) | ||
| Unadjusted | 3.37 (1.92, 5.90) | <0.0001 |
| Full model1 | 2.00 (1.11, 3.61) | 0.0208 |
| Age at diagnosis (per 1 year) | 0.99 (0.97, 1.01) | 0.3387 |
| Sex (males) | 1.34 (0.82, 2.19) | 0.2491 |
| Stage (vs. I) | ||
| II | 1.23 (0.37, 4.14) | 0.7347 |
| III | 3.78 (2.04, 7.00) | <.0001 |
| IV | 3.45 (1.65, 7.19) | 0.0010 |
| Treatment (vs. only surgery) | ||
| Surgery and chemotherapy or radiation | 1.51 (0.82, 2.79) | 0.1881 |
| Other/none/unknown | 2.15 (1.03, 4.46) | 0.0408 |
| Select model2 | 2.31 (1.31, 4.10) | 0.0040 |
| Stage (vs. I) | ||
| II | 1.22 (0.37, 4.04) | 0.7421 |
| III | 5.03 (2.99, 8.46) | <.0001 |
| IV | 5.24 (2.88, 9.53) | <.0001 |
Adjusted for age at diagnosis, sex, stage, and treatment pre-progression.
Adjusted for stage.
FISH, fluorescence in-situ hybridization; IHC, immunohistochemistry; PFS/RFS, progression or recurrence-free survival.
Figure 1.
(A) Stage-adjusted progression- or recurrence-free survival (PFS/RFS) survival curves for immunohistochemistry (IHC)3+ (positive) versus IHC0/1+ (negative) cases.a (B) Stage-adjusted PFS/RFS survival curves for fluorescence in-situ hybridization (FISH)-positive and FISH-negative cases.a (C) Stage-adjusted PFS/RFS survival curves for FISH+/IHC3+ (positive) and FISH-negative/IHC0/1+ (negative) casesa
aTable 2 Select model data
The matched analysis used three similar scenarios based on IHC/FISH tests, and on the number of matches of ALK-negative cases to each ALK-positive case. Results were consistent, showing a greater than two-fold increased risk of progression and/or recurrence in patients with ALK-positive tumors (eTable 3, available online).
Detailed comparison of progression and recurrence events between ALK-positive and ALK-negative cases
Location of metastases at diagnosis by ALK status was available for 47 advanced stage (local, regional, and remote metastasis) lung cancer cases and is presented in eTable 4 (available online) as descriptive information. The six patients with ALK-positive tumors appeared to have a greater frequency of chest cavity, brain, bone, and liver metastases than ALK-negative patients. However, no formal statistical test was performed due to small sample size.
Table 3 provides descriptions of the location and number of events in the survival cohort (n=299) by four ALK status subgroups. With the exception of adrenal glands, the ALK-positive group reported higher percentages of events for all individual locations than the combined ALK-negative groups, whereas the 12 IHC2+ cases (considered as ALK-negative unless also FISH-positive) had an event occurrence pattern which fell between patterns seen for the ALK-positive and the two other ALK-negative subgroups (IHC0, IHC1+). More detailed event data by specific ALK test results are provided in eTables 5 and 6 (available online). Specified adjusted Cox model results indicated that a significantly higher risk of extrathoracic events (brain and liver) was observed in patients harboring ALK-positive tumors versus ALK-negative tumors (Table 4).
Table 3. Number of lung cancer recurrences and/or progressions by location and ALK statusa.
| ALK+ (n=22)b | ALK– (n=277) | |||
|---|---|---|---|---|
|
| ||||
| IHC2+ (n=12) | IHC1+ (n=63) | IHC0 (n=202)b | ||
| No recurrence or progression, n (%) | 6 (30·0) | 7 (58·3) | 46 (73·0) | 153 (76·1) |
| Number in intrathoracic regionc, n (%) | ||||
| 1 | 4 (20·0) | 1 (8·3) | 8 (12·7) | 17 (8·5) |
| 2+ | 4 (20·0) | 3 (25·0) | 6 (9·5) | 21 (10·4) |
| Number in same lung, n (%) | ||||
| 1 | 4 (20·0) | 3 (25·0) | 9 (14·3) | 26 (12·9) |
| 2+ | 2 (10·0) | 0 (0·0) | 4 (6·4) | 8 (4·0) |
| Number in other lung, n (%) | ||||
| 1 | 1 (5·0) | 2 (16·7) | 3 (4·8) | 14 (7·0) |
| 2+ | 2 (10·0) | 1 (8·3) | 3 (4·8) | 7 (3·5) |
| Number in chest cavityd, n (%) | ||||
| 1 | 3 (15·0) | 2 (16·7) | 2 (3·2) | 6 (3·0) |
| 2+ | 0 (0·0) | 0 (0·0) | 0 (0·0) | 3 (1·5) |
| Number in extrathoracic regione, n (%) | ||||
| 1 | 6 (30·0) | 2 (16·7) | 6 (9·5) | 16 (8·0) |
| 2+ | 3 (15·0) | 1 (8·3) | 3 (4·8) | 9 (4·5) |
| Number in brain, n (%) | ||||
| 1 | 4 (20·0) | 3 (25·0) | 3 (4·7) | 9 (4·5) |
| 2+ | 1 (5·0) | 0 (0·0) | 0 (0·0) | 1 (0·5) |
| Number in bonesf, n (%) | ||||
| 1 | 2 (10·0) | 1 (8·3) | 3 (4·8) | 8 (4·0) |
| 2+ | 0 (0·0) | 0 (0·0) | 1 (1·6) | 4 (2·0) |
| Number in liver, n (%) | ||||
| 1 | 2 (10·0) | 0 (0·0) | 0 (0·0) | 3 (1·5) |
| 2+ | 0 (0·0) | 0 (0·0) | 1 (1·6) | 0 (0·0) |
| Number in kidneys, n (%) | ||||
| 1 | 1 (5·0) | 0 (0·0) | 0 (0·0) | 0 (0·0) |
| Number in adrenal glands, n (%) | ||||
| 1 | 0 (0·0) | 0 (0·0) | 0 (0·0) | 2 (1·0) |
| Number in other sitesg, n (%) | ||||
| 1 | 3 (15·0) | 0 (0·0) | 1 (1·6) | 6 (3·0) |
| 2+ | 0 (0·0) | 0 (0·0) | 1 (1·6) | 2 (1·0) |
n=299 cases in the survival cohort.
Anaplastic lymphoma kinase (ALK)+ is fluorescence in-situ hybridization (FISH)+ and/or immunohistochemistry (IHC)3+, and ALK– is FISH– and/or IHC0/1+ (included in ALK+ if FISH+ and IHC1+).
Two ALK+ cases and one IHC0 case have unknown locations for their recurrences and/or progressions; reported percentages are out of n=20 ALK+ cases and n=201 IHC0 cases.
Intrathoracic – includes same lung, other lung, and chest cavity.
Chest cavity – includes mediastinum, pleural effusion, pleural nodes, and pericardium.
Extrathoracic – includes brain, bones (including ribs), liver, kidneys, adrenal glands, and other sites.
Bones – includes all bones of the body, including ribs.
Other sites – includes all areas of the body not otherwise specified.
Table 4. Five-year time-to-recurrence and/or progression Cox proportional hazards models for location of events.
| Location of recurrence/progression | HR (95% CI) for ALK+ | p value |
|---|---|---|
| Intrathoracic | ||
| Unadjusted | 1·98 (0·90 to 4·35) | 0·09 |
| Full model | 0·71 (0·30 to 1·69) | 0·44 |
| Select modela | 0·76 (0·33 to 1·76) | 0·53 |
| Same lung | ||
| Unadjusted | 1·57 (0·62 to 3·96) | 0·34 |
| Full model | 0·74 (0·24 to 2·32) | 0·61 |
| Select modelb | 1·16 (0·45 to 2·99) | 0·76 |
| Other lung | ||
| Unadjusted | 1·57 (0·48 to 5·17) | 0·46 |
| Full model | 1·20 (0·20 to 7·16) | 0·85 |
| Select modelc | 2·46 (0·72 to 8·32) | 0·15 |
| Chest cavity | ||
| Unadjusted | 3·56 (1·00 to 12·65) | 0·05 |
| Full model | 9·56 (1·54 to 59·40) | 0·02 |
| Select modeld | 4·06 (0·98 to 16·90) | 0·05 |
| Extrathoracic | ||
| Unadjusted | 3.80 (1.82 to 7.92) | <0·001 |
| Full model | 2.53 (1.11 to 5.76) | 0·03 |
| Select modele | 2.44 (1.12 to 5.36) | 0·03 |
| Brain | ||
| Unadjusted | 4.75 (1.74 to 12.98) | 0·002 |
| Full model | 4.46 (1.37 to 14.51) | 0·01 |
| Select modelf | 4.55 (1.50 to 13.78) | 0·007 |
| Liver | ||
| Unadjusted | 6·90 (1·26 to 37·69) | 0·03 |
| Full model | Cannot be calculated | – |
| Select modelg | 8·22 (1·06 to 63·51) | 0·04 |
| Bones, kidneys, or adrenal glands | Samples too small | – |
| Other sites | ||
| Unadjusted | 4·92 (1·33 to 18·24) | 0·02 |
| Full model | 3·25 (0·30 to 35·11) | 0·33 |
| Select modelh | 1·14 (0·16 to 8·12) | 0·89 |
Adjusted for stage and extrathoracic.
Adjusted for treatment, other lung, chest cavity, and other sites.
Adjusted for same lung.
Adjusted for same lung and liver.
Adjusted for sex, stage, treatment, and intrathoracic.
Adjusted for stage, same lung, and bones.
Adjusted for treatment, other lung, and bones.
Adjusted for sex, stage, treatment, same lung, and kidneys.
Definitions: Intrathoracic – includes same lung, other lung, and chest cavity.
Chest cavity – includes mediastinum, pleural effusion, pleural nodes, and pericardium.
Extrathoracic – includes brain, bones, liver, kidneys, adrenal glands, and other sites.
Other sites – includes all areas of the body not otherwise specified.
Treatment includes all treatments before the first event of second primary, recurrence, or progression.
Full model = anaplastic lymphoma kinase (ALK) status, age at diagnosis, sex, stage, treatment, and remaining regions (e.g., same lung event is adjusted for other lung, chest cavity, brain, bones, liver, kidneys, adrenal glands, other sites, etc.).
Select model = ALK status and a stepwise model selection for the above covariates.
Event with unknown location – all location variables were set to ‘no’ to preserve sample size.
Model information: Time = time to first recurrence or progression; Event = event location (e.g., intrathoracic, same lung, brain, etc.) at any time. For example, the first recurrence could be intrathoracic, but if they had an extrathoracic recurrence 2 months later, extrathoracic would be an event and the time for the extrathoracic event would be the same as the intrathoracic event. Extrathoracic event would be a covariate in the intrathoracic model and vice versa.
Discussion
This study confirms earlier results by our group that IHC scores 3 and 0 were 100% concordant with FISH-positive and FISH-negative status, respectively, whereas IHC scores 1 and 2 may require further confirmatory testing. (12) In addition, we report the first ALK-positive prevalence estimate in an enriched sample of never smokers with adenocarcinoma. Prevalence rates ranged from 6·0% (by IHC3+) to 12·2% (FISH-positive/IHC3+/2+). This study also suggests, through controlled and matched analyses, a less favorable clinical outcome for ALK-positive compared with ALK-negative cases. Of note, we found the risk of lung cancer progression or recurrence in the 5 years post- diagnosis doubled in ALK-positive cases compared with ALK-negative cases (as defined by IHC, FISH, or combined IHC and FISH). Furthermore, despite the small sample size of ALK-positive patients and the small number of events, patients with ALK-positive tumors appeared to have a higher risk of developing tumor progression and/or recurrence in the brain and liver than patients with ALK-negative tumors.
We initially hypothesized that a high proportion of IHC2+ patients would also be FISH-positive. However, IHC–FISH concordance in this group was low, with only 14·3% of the 14 IHC2+ cases being FISH-positive. Therefore, we recommend confirmatory testing with FISH for any case showing IHC2+. As the patient and clinical characteristics of the IHC2+ group appeared intermediate between those for the IHC3+ and IHC1+ groups (see Tables 1 and 3, and eTable 5 [available online]), we grouped IHC3+ and IHC2+ patients together as an exploratory scenario for IHC–ALK-positivity. Patients with IHC1+ had a very low rate of FISH-positivity, with only 3.1% of IHC1+ cases showing ALK rearrangement. For practical purposes, IHC1+ cases can be considered ALK-negative. However, in an attempt to capture all patients with ALK rearrangement, targeted FISH testing could be considered for some adenocarcinoma patients with IHC1+, especially if they have other characteristics associated with ALK-positivity (never smoker status, young age, signet ring morphology, negativity for EGFR and KRAS mutations). The biological basis for the observed discordance in cases showing IHC2+ and FISH-negativity is under investigation by our group, and may be due to non-specific IHC staining, a unique ALK fusion variant or mutation not identified by one or both of the FISH probes used, or could be the result of normal ALK protein aberrantly expressed by some other mechanisms. These cases could represent a ‘transitional’ phase of an oncogenic process, and it may be important to determine whether patients with an IHC score of 2+ may also benefit from ALK-targeted therapy.
Although several studies have shown ALK-positive cases were more likely to have never or light/former smoking status and adenocarcinoma histology, (6, 8, 10) ALK-positive cases have been reported in current smokers, (1, 2, 17, 18) and in non-adenocarcinoma histology. (9, 17, 19) Furthermore, ALK-positivity has been reported along with the EGFR exon 19(10) and exon 20 mutations(20), despite ALK-positivity and EGFR mutations being mutually exclusive in all other studies to date. Therefore, other evaluation procedures should be explored to maximize the opportunity for these ‘exceptional’ cases to also benefit from ALK-targeted therapy.
Recent studies, attempting to elucidate the natural history of ALK-positive NSCLC in terms of response to standard therapies, have demonstrated no statistically significant difference in platinum-based chemotherapy response rates in ALK-positive compared with ALK-negative patients. (2, 8) Although not statistically significantly different, response rates were numerically smaller in ALK-positive patients. Shaw et al(8) reported a chemotherapy response rate for ALK-positive patients (n=12) of 25% versus 50% for EGFR-mutated (n=8) and 35% for ALK/EGFR wild-type patients (n=34), suggestive of a trend towards a poorer response to chemotherapy in the ALK-positive patients; however, this analysis did not adjust for between-group differences in age and smoking history. Likewise, Koh et al(2) reported an unadjusted first-line response rate to platinum-based chemotherapy of 18·8% in ALK-positive versus 34·4% in ALK-negative patients (p=0·088). Importantly, these two studies found that patients with ALK-positive NSCLC did not respond well to EGFR tyrosine kinase inhibitors, an association that was statistically significant. (2, 8) More recently, Camidge et al(21) and Lee et al(22) have presented data suggesting that ALK is predictive for favorable outcome with pemetrexed-based therapy. Shaw et al demonstrated significantly prolonged overall survival in ALK-positive NSCLC patients treated within a clinical trial of the experimental ALK inhibitor, crizotinib, when indirectly compared to non-trial patients with ALK-positive NSCLC who were treated with standard therapy.(23) Taken together, these studies show that patients with ALK-positive tumors do not have a more favorable clinical outcome with existing standard therapies, perhaps with the exception of pemetrexed-based therapy. ALK-specific inhibitors may prove to prolong overall survival, and ALK-positivity is a negative predictive marker for EGFR tyrosine kinase inhibitor treatment outcomes.
Our current study adds further evidence that ALK-positive NSCLC patients do not have a more favorable prognostic course of disease, as is the case in ALK-rearranged anaplastic large cell lymphoma. (24, 25) Rather, our data clearly demonstrate a less favorable prognosis, as measured by 5-year PFS/RFS. These findings may be due to complete clinical data ascertainment and the thorough consideration of potential confounding variables in all analyses. On the other hand, as with other retrospective studies of this rare gene rearrangement, the small sample size, the observational assessment of progression or recurrence, unknown EGFR status in some cases, and the use of a broad category of treatment modality may have affected our estimates. Our ongoing analyses will assess treatment and stage-specific outcomes with respect to known EGFR status as well as the correlation of signet ring cell morphology, ALK status, and clinical outcomes.
Because we have included an all inclusive patient cohort of never smokers with testable tissue sample, proportionally we have had a cohort with more early-stage or surgically-treated patients, as opposed to most other ALK-related studies focusing on late-stage patients. As a consequence, much longer follow-up time is required to obtain mature data for overall survival analysis. In the near future, we will achieve a complete view of treatment modalities and their effects on treatment outcomes including overall survival stratified by ALK status.
In conclusion, our results suggest that ALK-specific therapies are needed for patients with ALK-positive tumors due to significant worse 5-year PFS/RFS survival. In addition, if our data regarding a greater risk of brain or liver metastases are corroborated with more robust data, it will be even clearer that there is, indeed, an unmet and urgent medical need in the ALK-positive NSCLC patient population.
Supplementary Material
Acknowledgments
This study was partially supported by U.S. National Institutes of Health grants (R01 CA 84354, R01 CA 80127, and R01 CA115857), Mayo Foundation funds, and Pfizer. All data collection and analyses have been exclusively carried out by Mayo Clinic staff. Non-scientific editorial support and manuscript preparation were provided by Susan M. Ernst at Mayo Clinic and Stephen Jones and Martin Quinn at ACUMED (Tytherington, UK, funded by Pfizer Inc).
Footnotes
Contributors: PY co-led the entire project from study design, data collection, data analysis, results interpretation to manuscript preparation. KK participated in study design, results interpretation, and manuscript preparation. JMB was responsible for recruitment of cases and controls, and participated in manuscript preparation. MEJ conducted laboratory testing and participated in manuscript preparation. AO directed all laboratory testing and participated in manuscript preparation. JW was responsible for data retrieval, cleaning, and analysis. AJ was responsible for recruitment of cases and controls, and participated in manuscript preparation. CD was responsible for recruitment of cases and controls, and participated in manuscript preparation. RM was responsible for recruitment of cases and controls, and participated in manuscript preparation. CF conducted medical record data abstraction. SS conducted medical record data abstraction. FN was responsible for recruitment of cases and controls, and participated in manuscript preparation. JM was responsible for recruitment of cases and controls, and participated in manuscript preparation. MCA participated in study design and pathological review. HT provided advice in statistical analysis and participated in results discussion and interpretation. ESY co-led the whole project from study design, pathological review to manuscript preparation.
Conflicts of interest: KK is an employee of Pfizer, Inc.
Partial results of this study have been presented at the following lung cancer conferences: ESMO-EMCTO 2011 and IASLC-WCLC 2011.
Ethics approval: The protocol for this study was reviewed and approved by the Mayo Foundation Institutional Review Board and Biospecimens Committee.
References
- 1.Soda M, Choi YL, Enomoto M, et al. Identification of the transforming EML4-ALK fusion gene in on-small-cell lung cancer. Nature. 2007;448:561–566. doi: 10.1038/nature05945. [DOI] [PubMed] [Google Scholar]
- 2.Koh Y, Kim DW, Kim TM, et al. Clinicopathologic characteristics and outcomes of patients with anaplastic lymphoma kinase-positive advanced pulmonary adenocarcinoma: suggestion for an effective screening strategy for these tumors. J Thorac Oncol. 2011;6:905–912. doi: 10.1097/JTO.0b013e3182111461. [DOI] [PubMed] [Google Scholar]
- 3.Rikova K, Guo A, Zeng Q, et al. Global survey of phosphotyrosine signaling identifies oncogenic kinases in lung cancer. Cell. 2007;131:1190–1203. doi: 10.1016/j.cell.2007.11.025. [DOI] [PubMed] [Google Scholar]
- 4.Takeuchi K, Choi YL, Togashi Y, et al. KIF5B-ALK, a novel fusion oncokinase identified by an immunohistochemistry-based diagnostic system for ALK-positive lung cancer. Clin Cancer Res. 2009;15:3143–3149. doi: 10.1158/1078-0432.CCR-08-3248. [DOI] [PubMed] [Google Scholar]
- 5.Wong DW, Leung EL, Wong SK, et al. A novel KIF5B-ALK variant in nonsmall cell lung cancer. Cancer. 2011;117:2709–2718. doi: 10.1002/cncr.25843. [DOI] [PubMed] [Google Scholar]
- 6.Rodig SJ, Mino-Kenudson M, Dacic S, et al. Unique clinicopathologic features characterize ALK-rearranged lung adenocarcinoma in the western population. Clin Cancer Res. 2009;15:5216–5223. doi: 10.1158/1078-0432.CCR-09-0802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Takahashi T, Sonobe M, Kobayashi M, et al. Clinicopathologic features of non-small-cell lung cancer with EML4-ALK fusion gene. Ann Surg Oncol. 2010;17:889–897. doi: 10.1245/s10434-009-0808-7. [DOI] [PubMed] [Google Scholar]
- 8.Shaw AT, Yeap BY, Mino-Kenudson M, et al. Clinical features and outcome of patients with non-small-cell lung cancer who harbor EML4-ALK. J Clin Oncol. 2009;27:4247–4253. doi: 10.1200/JCO.2009.22.6993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wong DWS, Leung ELH, So KKT, et al. The EML4-ALK Fusion Gene is Involved in Various Histologic Types of Lung Cancer From Nonsmokers With Wild-type EGFR and KRAS. Cancer. 2009:1723–1733. doi: 10.1002/cncr.24181. http://www.interscience.wiley.com. [DOI] [PubMed]
- 10.Zhang X, Zhang S, Yang X, et al. Fusion of EML4 and ALK is associated with development of lung adenocarcinomas lacking EGFR and KRAS mutations and is correlated with ALK expression. Mol Cancer. 2010;9:188. doi: 10.1186/1476-4598-9-188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sakairi Y, Nakajima T, Yasufuku K, et al. EML4-ALK fusion gene assessment using metastatic lymph node samples obtained by endobronchial ultrasound-guided transbronchial needle aspiration. Clin Cancer Res. 2010;16:4938–4945. doi: 10.1158/1078-0432.CCR-10-0099. [DOI] [PubMed] [Google Scholar]
- 12.Yi ES, Boland JM, Maleszewski JJ, et al. Correlation of IHC and FISH for ALK Gene Rearrangement in NSCLC: IHC Score Algorithm for FISH. J Thorac Oncol. 2011;6:459–465. doi: 10.1097/JTO.0b013e318209edb9. [DOI] [PubMed] [Google Scholar]
- 13.Yang P, Allen MS, Aubry MC, et al. Clinical Features of 5,628 Primary Lung Cancer Patients: Experience at Mayo Clinic from 1997-2003. Chest. 2005;128:452–462. doi: 10.1378/chest.128.1.452. [DOI] [PubMed] [Google Scholar]
- 14.Visbal AL, Williams BA, Nichols FC, et al. Gender Differences in Non-Small Cell Lung Cancer Survival: An Analysis of 4,618 Patients Diagnosed Between 1997-2002. Ann Thorac Surg. 2004;78(1):209–215. doi: 10.1016/j.athoracsur.2003.11.021. [DOI] [PubMed] [Google Scholar]
- 15.Sugimura H, Nichols FC, Yang P, et al. Survival Following Recurrent Non-Small Cell Lung Cancer after Complete Pulmonary Resection. Annals of Thoracic Surgery. 2007;83:409–418. doi: 10.1016/j.athoracsur.2006.08.046. [DOI] [PubMed] [Google Scholar]
- 16.Williams BA, Sugimura H, Endo C, et al. Predicting Postrecurrence Survival Among Completely Resected Nonsmall-Cell Lung Cancer Patients. Annals of Thoracic Surgery. 2006;81:1021–1027. doi: 10.1016/j.athoracsur.2005.09.020. [DOI] [PubMed] [Google Scholar]
- 17.Boland JM, Erdogan S, Vasmatzis G, et al. Anaplastic lymphoma kinase immunoreactivity correlates with ALK gene rearrangement and transcriptional up-regulation in non-small cell lung carcinomas. Hum Pathol. 2009;40:1152–1158. doi: 10.1016/j.humpath.2009.01.012. [DOI] [PubMed] [Google Scholar]
- 18.Martelli MP, Sozzi G, Hernandez L, et al. EML4-ALK rearrangement in non-small cell lung cancer and non-tumor lung tissues. Am J Pathol. 2009;174:661–670. doi: 10.2353/ajpath.2009.080755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Inamura K, Takeuchi K, Togashi Y, et al. EML4-ALK fusion is linked to histological characteristics in a subset of lung cancers. J Thorac Oncol. 2008;3:13–17. doi: 10.1097/JTO.0b013e31815e8b60. [DOI] [PubMed] [Google Scholar]
- 20.Camidge DR, Kono SA, Flacco A, et al. Optimizing the detection of lung cancer patients harboring anaplastic lymphoma kinase (ALK) gene rearrangements potentially suitable for ALK inhibitor treatment. Clin Cancer Res. 2010;16:5581–5590. doi: 10.1158/1078-0432.CCR-10-0851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Camidge DR, Kono SA, Lu X, et al. Anaplastic lymphoma kinase gene rearrangements in non-small cell lung cancer are associated with prolonged progression-free survival on pemetrexed. J Thorac Oncol. 2011;6:774–780. doi: 10.1097/JTO.0b013e31820cf053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lee JO, Kim TM, Lee SH, et al. Anaplastic Lymphoma Kinase Translocation: A Predictive Biomarker of Pemetrexed in Patients with Non-small Cell Lung Cancer. J Thorac Oncol. 2011;6:1474–1480. doi: 10.1097/JTO.0b013e3182208fc2. [DOI] [PubMed] [Google Scholar]
- 23.Shaw AT, Yeap BY, Solomon BJ, et al. Impact of crizotinib on survival in patients with advanced, ALK-positive NSCLC compared with historical controls. J Clin Oncol. 2011;29 Suppl; Abstract # 7507. [Google Scholar]
- 24.Savage KJ, Harris NL, Vose JM, et al. ALK- anaplastic large-cell lymphoma is clinically and immunophenotypically different from both ALK+ ALCL and peripheral T-cell lymphoma, not otherwise specified: report from the International Peripheral T-Cell Lymphoma Project. Blood. 2008;111:5496–5504. doi: 10.1182/blood-2008-01-134270. [DOI] [PubMed] [Google Scholar]
- 25.ten Berge RL, Oudejans JJ, Ossenkoppele GJ, et al. ALK expression in extranodal anaplastic large cell lymphoma favours systemic disease with (primary) nodal involvement and a good prognosis and occurs before dissemination. J Clin Pathol. 2000;53:445–450. doi: 10.1136/jcp.53.6.445. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.