Characteristics and Survival Outcomes of Patients With Metastatic <i>RET</i> Fusion-Positive Solid Tumors Receiving Non-RET Inhibitor Standards of Care in a Real-World Setting.
Free full text in Europe PMC
Abstract
Purpose
RET fusions are oncogenic drivers across different solid tumors. However, the genomic landscape and natural history of patients with RET fusion-positive solid tumors are not well known. We describe the clinical characteristics of RET tyrosine kinase inhibitor (TKI)-naïve patients with RET fusion-positive solid tumors (excluding non-small-cell lung cancer [NSCLC]), treated in a real-world setting and assess the prognostic effect of RET fusions.Methods
Data for RET TKI-naïve patients with metastatic solid tumors (excluding NSCLC) who had ≥one Foundation Medicine comprehensive genomic profiling test (January 1, 2011-March 31, 2022) were obtained from a deidentified nationwide (US-based) clinicogenomic database. The primary objective of this study was to compare the overall survival (OS) of patients with RET fusion-positive tumors versus matched patients with RET wild-type (RET-WT) tumors. Patients with RET-WT solid tumors were matched (4:1) to patients with RET fusion-positive tumors on the basis of preselected covariates.Results
The study population included 26 patients in the RET fusion-positive cohort, 7,220 patients in the RET-WT cohort (before matching), and 104 patients in the matched RET-WT cohort. Co-occurring genomic alterations were rare in the RET fusion-positive cohort. Median OS was consistently lower in patients with RET fusion-positive tumors versus those with RET-WT tumors, using three different analyses (hazard ratios, 2.0, 1.7, and 2.2).Conclusion
These data suggest that RET fusions represent a negative prognostic factor in patients with metastatic solid tumors and highlight the need for wider genomic testing and use of RET-specific TKIs that could improve patient outcomes. Our study also highlights the value of real-world data when studying rare cancers or cancers with rare genomic alterations.Free full text
Characteristics and Survival Outcomes of Patients With Metastatic RET Fusion–Positive Solid Tumors Receiving Non-RET Inhibitor Standards of Care in a Real-World Setting
Abstract
PURPOSE
RET fusions are oncogenic drivers across different solid tumors. However, the genomic landscape and natural history of patients with RET fusion–positive solid tumors are not well known. We describe the clinical characteristics of RET tyrosine kinase inhibitor (TKI)-naïve patients with RET fusion–positive solid tumors (excluding non–small-cell lung cancer [NSCLC]), treated in a real-world setting and assess the prognostic effect of RET fusions.
METHODS
Data for RET TKI-naïve patients with metastatic solid tumors (excluding NSCLC) who had ≥one Foundation Medicine comprehensive genomic profiling test (January 1, 2011-March 31, 2022) were obtained from a deidentified nationwide (US-based) clinicogenomic database. The primary objective of this study was to compare the overall survival (OS) of patients with RET fusion–positive tumors versus matched patients with RET wild-type (RET-WT) tumors. Patients with RET-WT solid tumors were matched (4:1) to patients with RET fusion–positive tumors on the basis of preselected covariates.
RESULTS
The study population included 26 patients in the RET fusion–positive cohort, 7,220 patients in the RET-WT cohort (before matching), and 104 patients in the matched RET-WT cohort. Co-occurring genomic alterations were rare in the RET fusion–positive cohort. Median OS was consistently lower in patients with RET fusion–positive tumors versus those with RET-WT tumors, using three different analyses (hazard ratios, 2.0, 1.7, and 2.2).
CONCLUSION
These data suggest that RET fusions represent a negative prognostic factor in patients with metastatic solid tumors and highlight the need for wider genomic testing and use of RET-specific TKIs that could improve patient outcomes. Our study also highlights the value of real-world data when studying rare cancers or cancers with rare genomic alterations.
INTRODUCTION
Fusions of the rearranged during transfection (RET) gene are oncogenic drivers across a number of solid tumors.1-3 Treatment options for patients with RET-altered solid tumors were previously limited to multikinase inhibitors, but these can be associated with significant toxicities and high rates of dose reduction/discontinuation, and there may also be limited efficacy.1,4,5 Therefore, there is a need for better precision therapies that selectively target RET alterations and anticipated resistance mechanisms.
Two RET-specific tyrosine kinase inhibitors (TKIs), selpercatinib and pralsetinib, have demonstrated durable antitumor activity and manageable toxicity profiles and are approved for the treatment of advanced/metastatic RET fusion–positive non–small-cell lung cancer (NSCLC; both drugs in the United States and Europe), advanced/metastatic RET-altered thyroid cancer (both drugs in the United States; selpercatinib in Europe), and advanced/metastatic RET fusion–positive solid tumors (selpercatinib in the United States; neither in Europe).6-9
RET fusions, although rare,5,10,11 have been detected in a wide range of solid tumors.5,12 However, the genomic landscape and natural history of patients with RET fusion–positive tumors are not well documented. It is important to understand whether RET fusion–positive tumors behave differently to RET wild-type (RET-WT) tumors of the same histology and whether RET fusions are prognostic.
Examining the association between rare alterations and patient outcomes is challenging in randomized treatment trials because of limited patient numbers. Therefore, large-scale real-world data provide a valuable alternative for evaluating these associations.
The study objectives were to (1) describe the clinical characteristics and overall survival (OS) of patients with RET fusion–positive (for which there are limited data) and RET-WT solid tumors and (2) provide an example of how real-world data collected from routine clinical practice can be used to determine the prognostic value of rare alterations. This can then be used to evaluate targeted therapies in tumor-agnostic clinical trials.
METHODS
Study Design and Data Sources
This retrospective, observational study of RET TKI-naïve patients with metastatic, RET fusion–positive or RET-WT solid tumors used real-world data collected during routine clinical practice from the nationwide (US-based) deidentified Flatiron Health-Foundation Medicine clinicogenomic database (CGDB).
Deidentified data were obtained from approximately 280 Flatiron Health cancer clinics in the United States (approximately 800 care sites); retrospective longitudinal clinical (patient-level structured and unstructured) data were derived from electronic health records and curated via technology-enabled abstraction. These data were linked to genomic data derived from the Foundation Medicine comprehensive genomic profiling (CGP) tests in the Flatiron Health-Foundation Medicine CGDB using deidentified, deterministic matching.13 Ethics committee approval was not required because this study used anonymized patient data and did not directly enroll patients.
Study Population
Eligible patients had ≥one documented clinical visit in a Flatiron Health network center between January 1, 2011, and March 31, 2022, and underwent CGP testing by Foundation Medicine before April 1, 2022. Genomic alterations were identified via CGP of >300 cancer-related genes using FoundationOne, FoundationOneCDx, or FoundationOneHeme assays.14-16
The study population included patients with a de novo metastatic diagnosis who had not previously been treated with a RET TKI in any therapy line. As Flatiron CGDB retains detailed documentation of all medications received by a patient, regardless of their on-label or off-label status, all patients who received a next-generation RET inhibitor (such as pralsetinib or selpercatinib) or a study drug as part of a clinical trial were excluded to eliminate any bias in the description of standard of care. Patients diagnosed with RET fusion–positive NSCLC were also excluded as this represents a well-defined population with published data,17 and there are licensed drugs for this particular indication6-9; this exclusion is consistent with ongoing clinical trials.18 Other patient exclusion criteria included multiple cancer diagnoses or no initial diagnosis date, >one type of CGP or a CGP report date before the initial diagnosis, an initial diagnosis within 3 months before the data cutoff; death before 2012 (when Foundation Medicine's CGP was established), and a visit gap of >90 days after the initial diagnosis (patients may have been treated temporarily in a center outside of the Flatiron Health network postdiagnosis).
Determination of RET Status
RET positivity was defined as the presence of a fusion or rearrangement with a predicted known/likely functional status as defined by Foundation Medicine. This included a 3’ RET fusion (breakpoint between exon 1 and intron 11) with a protein-coding 5’ gene fusion partner, which was predicted to be in-frame with an intact kinase domain.19 Fusions with non–protein-coding gene partners or intragenic fusions were excluded. RET-WT status was determined when CGP was unable to detect any qualifying RET fusions. To reduce the probability of misclassifying patients with false-negative results as RET-WT, we only analyzed solid tumor samples (using FoundationOne, FoundationOneCDx, or FoundationOneHeme assays).
Study Design
We used data from eligible patients with known RET status in the Flatiron Health-Foundation Medicine database. We also performed a nested case-control study within this database. Patients with RET-WT tumors were matched with patients with RET fusion–positive tumors (4:1 ratio) using Mahalanobis distance matching20 on the basis of several covariates: age, sex, race, tumor type, practice type (academic v community), Eastern Cooperative Oncology Group performance status (ECOG PS) measured from 30 days before to 7 days after index date, year of CGP, time from initial diagnosis of metastatic disease to CGP report date, and number of oncologist-defined, rule-based treatment lines before CGP report date. An absolute mean difference of <0.1 was used as a quality metric to indicate negligible distance between groups.21
Statistical Analyses
The primary objective was to evaluate the association between OS and RET fusion status. OS was defined as time from the index date until death from any cause or the censoring date (ie, last follow-up date when last known to be alive). The index date could be the date of the CGP report or the date of metastatic diagnosis. We also examined patient characteristics, treatment patterns, and genomic alterations: tumor mutational burden (TMB), microsatellite instability (MSI), and co-occurring oncogenic functional alterations in ALK, BRAF, ERBB2, EGFR, NTRK, ROS1, MET, and KRAS.
Kaplan-Meier curves and Cox regression were used for the analysis of OS. R version 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria) and Python version 3.9.9 (Python Software Foundation, Wilmington, DE) were used for all analyses. A crude (unmatched) analysis used the CGP report date as the index for calculating time-to-event in the OS analysis as confirmation of biomarker status was a predefined inclusion criterion for study entry. Additionally, two different models analyzed OS on the basis of the nested case-control patients: model 1 used the CGP report date as the index (as for the crude analysis), whereas model 2 used the date of metastatic diagnosis as the index and also adjusted for immortal time bias with left truncation by considering the CGP report date.22
Time from metastatic diagnosis to CGP report may differ between patients, and the adjustment for immortal time bias accounts for patients having to survive long enough to receive a CGP report, during which time the outcome of interest (death) cannot occur (ie, they are immortal during that period).23 This immortal time bias can result in overestimation of the outcome event rate in the control group, underestimation of the event rate in the exposed group, or both.24 To address this issue, we used the date of metastatic diagnosis as the index and estimated survival using a left truncation model as proposed in the study by Mackenzie et al.22
Ethics
This study used deidentified patient data from the Flatiron Health-Foundation Medicine CGDB, a US-wide longitudinal database curated through technology-enabled abstraction, and did not directly enroll patients.
RESULTS
Patient Characteristics
Between January 1, 2011, and March 31, 2022, a total of 222 patients with RET fusion–positive tumors and ≥one documented clinical visit in a Flatiron Health network center were selected. In total, 196 patients were excluded on the basis of criteria described above, including 92 patients with NSCLC (Fig (Fig1A);1A); 26 patients constituted the RET fusion–positive cohort for this analysis. Baseline characteristics are presented in Table Table1.1. The majority of patients with RET fusion–positive tumors were male (57.7%) and had ECOG PS 0/1 (53.8%); the mean age was 65.3 years. The median follow-up time from CGP report date for the RET fusion–positive cohort was 5.7 (IQR, 7.2) months while the median time from initial diagnosis of metastatic disease to CGP report date was 3.5 (IQR, 8.8) months.
Cohort attrition for (A) the RET fusion–positive population and (B) the RET-WT population. Structured activity refers to a recording of vital information, a medication administration, a noncanceled drug order, or a reported laboratory test/result. aAlso excluded patients with no initial metastatic disease diagnosis date or a diagnosis within 3 months before the data cutoff, patients who died before 2012, patients with multiple cancer diagnoses, and patients with a CGP report date before the initial diagnosis, with no impact on attrition. bAlso excluded patients with no initial metastatic disease diagnosis date or a diagnosis within 3 months before the data cutoff, patients who died before 2012, and patients with a CGP report date before the initial diagnosis, with no impact on attrition. 90d, 90 days; CGDB, clinicogenomic database; CGP, comprehensive genomic profiling; CSD, clinical study drug; FMI, Foundation Medicine, Inc; NSCLC, non–small-cell lung cancer; RET, rearranged during transfection; RETi, RET inhibitor; WT, wild type.
TABLE 1.
Baseline Patient Characteristics
| Characteristic | RET Fusion–Positive (N = 26) | RET-WT (N = 7,220) | |
|---|---|---|---|
| Matched (n = 104) | Nonmatched (n = 7,116) | ||
| Sex,a No. (%) | |||
| 11 (42.3) | 43 (41.3) | 4,035 (56.7) | |
| 15 (57.7) | 61 (58.7) | 3,081 (43.3) | |
| Age, years, mean (SD) | 65.3 (10.3) | 61.8 (12.0) | 64.8 (9.9) |
| Race, No. (%) | |||
| 0 | 0 | 173 (2.4) | |
| 1 (3.8) | 4 (3.8) | 615 (8.6) | |
| 0 | 0 | 15 (0.2) | |
| 21 (80.8) | 89 (85.6) | 4,599 (64.6) | |
| 3 (11.5) | 9 (8.7) | 1,120 (15.7) | |
| 1 (3.8) | 2 (1.9) | 594 (8.3) | |
| Primary tumor type, No. (%) | |||
| 9 (34.6) | 37 (35.6) | 2,899 (40.7) | |
| 4 (15.4) | 15 (14.4) | 1,089 (15.3) | |
| 4 (15.4) | 16 (15.4) | 83 (1.2) | |
| 3 (11.5) | 12 (11.5) | 338 (4.7) | |
| 2 (7.7) | 8 (7.7) | 1,167 (16.4) | |
| 1 (3.8) | 4 (3.8) | 335 (4.7) | |
| 1 (3.8) | 4 (3.8) | 418 (5.9) | |
| 1 (3.8) | 4 (3.8) | 307 (4.3) | |
| 1 (3.8) | 4 (3.8) | 480 (6.7) | |
| ECOG PS,c No. (%) | |||
| 7 (26.9) | 30 (28.8) | 1,652 (23.2) | |
| 7 (26.9) | 31 (29.8) | 2,342 (32.9) | |
| 1 (3.8) | 4 (3.8) | 869 (12.2) | |
| 11 (42.3) | 39 (37.5) | 2,253 (31.7) | |
| Practice type, No. (%) | |||
| 4 (15.4) | 12 (11.5) | 1,002 (14.1) | |
| 22 (84.6) | 92 (88.5) | 6,114 (85.9) | |
| Serum albumin, g/dLd | |||
| 3.5 (0.5) | 3.8 (0.6) | 3.7 (0.6) | |
| 6 (23.1) | 18 (17.3) | 1,081 (15.2) | |
| Absolute neutrophil count, 109/L)d | |||
| 5.1 (2.5) | 5.4 (5.0) | 6.4 (59.7) | |
| 10 (38.5) | 28 (26.9) | 1,957 (27.5) | |
| Platelet count, 109/Ld | |||
| 243.7 (114.9) | 240.5 (120.1) | 247.9 (122.5) | |
| 7 (26.9) | 19 (18.3) | 1,234 (17.3) | |
| No. of prior lines of treatment, (%) | |||
| 2 (7.7) | 8 (7.7) | 807 (11.3) | |
| 6 (23.1) | 24 (23.1) | 2,580 (36.3) | |
| 7 (26.9) | 27 (26.0) | 1,933 (27.2) | |
| 11 (42.3) | 45 (43.3) | 1,796 (25.2) | |
| PD-L1 status at CGP report, No. (%) | |||
| 0 | 3 (2.9) | 70 (1.0) | |
| 1 (3.8) | 10 (9.6) | 502 (7.1) | |
| 5 (19.2) | 11 (10.6) | 1,315 (18.5) | |
| 20 (76.9) | 80 (76.9) | 5,229 (73.5) | |
| Documentation of any PD-L1 therapy, No. (%) | |||
| 0 | 3 (2.9) | 301 (4.2) | |
| 26 (100) | 101 (97.1) | 6,815 (95.8) | |
| PD-L1 therapy on or before CGP report date, No. (%) | |||
| 0 | 2 (1.9) | 111 (1.6) | |
| 26 (100) | 102 (98.1) | 7,005 (98.4) | |
| Year of CGP report, No. (%) | |||
| 0 | 0 | 4 (0.1) | |
| 0 | 0 | 64 (0.9) | |
| 2 (7.7) | 5 (4.8) | 312 (4.4) | |
| 3 (11.5) | 12 (11.5) | 492 (6.9) | |
| 2 (7.7) | 8 (7.7) | 529 (7.4) | |
| 4 (15.4) | 16 (15.4) | 799 (11.2) | |
| 3 (11.5) | 12 (11.5) | 1,133 (15.9) | |
| 7 (26.9) | 29 (27.9) | 1,401 (19.7) | |
| 2 (7.7) | 10 (9.6) | 1,367 (19.2) | |
| 3 (11.5) | 12 (11.5) | 1,015 (14.3) | |
| Follow-up time from CGP report, median (IQR) | 5.7 (7.2) | 6.2 (9.8) | 7.5 (13.6) |
| Time from initial diagnosise to CGP report date, months, median (IQR) | 3.5 (8.8) | 4.0 (9.4) | 4.5 (16.9) |
Abbreviations: CGP, comprehensive genomic profiling; ECOG PS, Eastern Cooperative Oncology Group performance status; RET, rearranged during transfection; SCLC, small-cell lung cancer; SD, standard deviation; WT, wild type.
The RET fusion–positive cohort consisted of nine distinct tumor/histology types, most commonly colorectal cancer (CRC; n = 9; 34.6%; Fig Fig2).2). Nine different RET fusion partners were detected, and the most common were NCOA4 (n = 12; 46.2%), CCDC6 (n = 6; 23.1%), and ERC1 (n = 2; 7.7%). CEP135, FAM13C, FGFR1OP, KIAA1217, KIF5B, and MACROD2 were detected in one patient each.
Tumor types in the RET fusion–positive cohort (N = 26). Patients with RET fusion–positive non–small-cell lung cancer were excluded. aNeuroendocrine tumors included one GI tumor and two unspecified anatomical locations. CUP, cancer of unknown primary; RET, rearranged during transfection; SCLC, small-cell lung cancer.
Fifteen patients had a CGDB record for prior antineoplastic therapy, of whom seven (46.7%) had received ≥two prior lines of therapy, six (40.0%) had received one prior line of therapy, and two (13.3%) were recorded as treatment naïve.
Of 62,456 patients with solid tumors in the CGDB, 7,220 patients with RET-WT solid tumors met the eligibility criteria (Fig (Fig1B).1B). These patients were matched for tumor type to the RET fusion–positive cohort to prevent inconsistencies from biasing the OS outcomes. After covariate matching for demographic and clinical characteristics, the matched RET-WT cohort included 104 patients. The RET fusion–positive and matched RET-WT cohorts had comparable baseline serum albumin levels and absolute neutrophil and platelet counts, although missing data ranged from 17.3% to 38.5% across the cohorts (Table (Table1).1). Most patients had missing PD-L1 status in the CGP report (76.9% in both the RET fusion–positive and matched RET-WT cohorts). Three patients (2.9%) in the matched RET-WT cohort had received prior PD-L1 therapy versus none in the RET fusion–positive cohort.
Presence of Co-Occurring Genomic Alterations
In the RET fusion–positive cohort, 17 patients (65.4%) had low TMB (<5.7 mut/Mb), and two patients (7.7%) were TMB-high (≥20 mut/Mb); 13 patients (50.0%) had low MSI, and one patient (3.8%) was MSI-high (Table (Table2).2). One patient in this cohort had an ERBB2 amplification (3.8%); no other assessed oncogenic coalterations were identified.
TABLE 2.
Co-Occurring Biomarkers and Molecular Characteristics
| Co-occurring Biomarkers/Molecular Characteristics | RET Fusion–Positive (N = 26), No. (%) | RET-WT (N = 7,220) | |
|---|---|---|---|
| Matched (n = 104) | Nonmatched (n = 7,116) | ||
| TMB status, No. (%) | |||
| 2 (7.7) | 0 | 215 (3.0) | |
| 5 (19.2) | 16 (15.4) | 1,323 (18.6) | |
| 17 (65.4) | 88 (84.6) | 5,578 (78.4) | |
| 2 (7.7) | 0 | 0 | |
| MSI-high, No. (%) | |||
| 1 (3.8) | 0 | 119 (1.7) | |
| 13 (50.0) | 86 (82.7) | 6,021 (84.6) | |
| 12 (46.2) | 18 (17.3) | 976 (13.7) | |
| Oncogenic alterations, No. (%) | |||
| 0 | 0 | 13 (0.2) | |
| 0 | 6 (5.8) | 389 (5.5) | |
| 1 (3.8) | 4 (3.8) | 300 (4.2) | |
| 0 | 0 | 53 (0.7) | |
| 0 | 0 | 13 (0.2) | |
| 0 | 0 | 11 (0.2) | |
| 0 | 0 | 9 (0.1) | |
| 0 | 38 (36.5) | 2,623 (36.9) | |
NOTE. Only variants of known or likely functional status were included.
Abbreviations: ALK, anaplastic lymphoma kinase; BRAF, proto-oncogene B-Raf; ERBB2, Erb-B2 receptor tyrosine kinase 2; EGFR, epidermal growth factor receptor; KRAS, Kirsten rat sarcoma viral oncogene homolog; MET, mesenchymal epithelial transition factor receptor; MSI, microsatellite instability; NTRK, neurotrophic tyrosine receptor kinase; RET, rearranged during transfection; ROS1, ROS proto-oncogene 1; TMB, tumor mutational burden; WT, wild type.
Most patients in the matched RET-WT cohort had low TMB (84.6%) and low MSI (82.7%); none had high TMB or MSI (Table (Table2).2). The only genomic alterations observed in this cohort were KRAS alterations (36.5%), BRAF alterations (5.8%), and ERBB2 amplifications (3.8%).
OS
In the crude analysis, the median OS was 6.0 months (95% CI, 1.6 to 9.9) for the RET fusion–positive cohort (N = 26) and 10.4 months (95% CI, 10.0 to 10.9) for the nonmatched RET-WT population (N = 7,220; Table Table3;3; Fig Fig3A).3A). The hazard ratio (HR) was 2.0 (95% CI, 1.3 to 3.1), indicating that patients in the RET fusion–positive cohort had a two-fold increase in risk of death compared with the nonmatched RET-WT cohort.
TABLE 3.
Analysis of OS
| Analysis | Cohort | No. of Deaths, (%) | Median OS, Months (95% CI) | HR (95% CI) |
|---|---|---|---|---|
| Crude | RET fusion–positive (N = 26) | 20 (76.9) | 6.0 (1.6 to 9.9) | 2.0 (1.3 to 3.1) |
| Nonmatched RET-WT (N = 7,220) | 4,838 (67.0) | 10.4 (10.0 to 10.9) | ||
| Model 1a | RET fusion–positive (N = 26) | 20 (76.9) | 6.0 (1.6 to 9.9) | 1.7 (1.0 to 2.9) |
| Matched RET-WT (n = 104) | 72 (69.2) | 9.4 (5.5 to 11.7) | ||
| Model 2b | RET fusion–positive (N = 26) | 20 (76.9) | 6.9 (1.6 to 9.6) | 2.2 (1.3 to 3.7) |
| Matched RET-WT (n = 104) | 72 (69.2) | 11.3 (7.7 to 17.1) |
Abbreviations: CGP, comprehensive genomic profiling; HR, hazard ratio; OS, overall survival; RET, rearranged during transfection; WT, wild type.
Kaplan-Meier estimates of OS (A) comparing the RET fusion–positive population (N = 26) with the crude RET-WT population (before matching; N = 7,220), using the CGP report date as the index date; (B) comparing the RET fusion–positive population with the matched RET-WT population (n = 104), using the CGP report date as the index date (model 1); (C) comparing the RET fusion–positive population with the matched RET-WT population (n = 104), using the initial diagnosis date as the index date (corrected for immortal time bias; model 2a). aThe number of patients in model 1 (B) and model 2 (C) is the same (ie, N = 26 for the RET fusion–positive cohort and n = 104 for the RET-WT cohort); however, in model 2, where left truncation is used to estimate survival, only six patients with RET fusion–positive tumors and 37 patients with RET-WT tumors had both a metastatic diagnosis and a CGP date at time zero. The remaining patients satisfied cohort entry criteria (ie, CGP report date) at later times, and this immortal time was taken into account when calculating OS. CGP, comprehensive genomic profiling; OS, overall survival; RET, rearranged during transfection; WT, wild type.
The results for models 1 and 2 (with matched controls) were consistent with the crude analysis. Using model 1 (CGP report date as the index date), the median OS was 6.0 months (95% CI, 1.6 to 9.9) in the RET fusion–positive cohort (N = 26) and 9.4 months (95% CI, 5.5 to 11.7) in the matched RET-WT cohort (N = 104; Table Table3;3; Fig Fig3B),3B), with an adjusted HR of 1.7 (95% CI, 1.0 to 2.9). Correspondingly, using model 2 (date of metastatic diagnosis as the index date, after adjusting for immortal time bias), the median OS was 6.9 months (95% CI, 1.6 to 9.6) in the RET fusion–positive cohort (N = 26) and 11.3 months (95% CI, 7.7 to 17.1) in the matched RET-WT cohort (N = 104; Table Table3;3; Fig Fig3C).3C). The adjusted HR was 2.2 (95% CI, 1.3 to 3.7), confirming that patients with RET fusion–positive tumors had approximately twice the risk of death compared with matched RET-WT controls.
DISCUSSION
This large-scale study investigated the prognostic association between RET fusion status and OS, and as far as we are aware, this is the first such study to include multiple tumor types. Other published studies have focused on NSCLC,17 which we exclude, and there has been one study each in CRC25 and medullary thyroid cancer.26 NSCLC was excluded from this analysis for several reasons. First, RET fusion–positive NSCLC represents a well-defined population, for which data on the association between RET status and outcomes have already been published.17 Second, if the 92 cases of NSCLC had been included alongside the 26 cases of other cancer types in the RET fusion–positive cohort, they would have biased the associations toward patterns in lung cancer while the focus of this analysis was on other and rare cancer types. Furthermore, there are already licensed drugs for RET fusion–positive lung cancer (namely selpercatinib and pralsetinib).6-9 We included CRC and thyroid cancer because they are uncommon among RET fusion–positive patients, and there is limited evidence on the prognostic associations for RET in these cancers.
Co-occurrence of other assessed oncogenic alterations was not observed in the RET fusion–positive cohort, except for one patient with an ERBB2 amplification. These data support the hypothesis that RET fusions are the primary oncogenic driver in these tumors.1,5,11 Patients in the matched RET-WT cohort had a high rate of KRAS alterations (36.5%) and some BRAF alterations (5.8%). There were no other notable baseline patient or molecular characteristics in the RET fusion–positive cohort.
In three different analyses (crude, model 1, model 2), patients with RET fusion–positive tumors consistently had a shorter median OS than those with RET-WT tumors. The approximate two-fold increase in risk of death suggests that RET fusions have a negative prognostic effect. A similar conclusion was reached in a study of patients with CRC (24 with RET fusion–positive and 291 with RET-WT tumors), in which the adjusted OS HR was 2.97 (95% CI, 1.25 to 7.07; P = .014).25
A study of patients with metastatic RET fusion–positive NSCLC suggested that RET positivity may be associated with improved survival.17 However, some analyses were not statistically significant, and therefore, the data were inconclusive. A retrospective study of patients with medullary thyroid cancer found that at time points between 6 and 48 months, progression-free survival rates after first-line therapy were similar between all-comers and patients who had RET mutations; however, median progression-free survival was higher in patients with RET mutations.26 It is possible that RET alterations have a different prognostic effect in different tumor types, which may justify examining them separately.
Our data indicate that patients with RET fusion–positive tumors may not be optimally treated by current standards of care, highlighting the unmet need for precision therapies that specifically target this biomarker.3,27 Retrospective analyses from a multicenter study of 218 patients with RET fusion–positive NSCLC showed significantly improved survival using RET-specific TKIs (such as pralsetinib and selpercatinib).28 Single-arm clinical trials have also shown good efficacy and safety with RET-specific TKIs in patients with RET-altered solid tumors.29-34 In these trials, response rates in patients with previously treated RET fusion–positive NSCLC or thyroid cancer ranged from 63% to 84% with pralsetinib29,30 and 61% to 77% with selpercatinib.32,34 Consequently, both drugs achieved grade 3 in the ESMO Magnitude of Clinical Benefit Score version 1.1, the highest grade for single-arm evidence on the basis of response rates of >60%.35 Taken together with our findings, these data suggest that RET-specific TKIs may represent better treatment options for patients with RET fusion–positive tumors than non-RET inhibitor standards of care.
RET fusions are rare in solid tumors,5,10,11 making it difficult to conduct randomized head-to-head trials of new versus established therapies. Standards of care also differ across histologies and treatment lines; therefore, matching a control arm with the same disease entities and types/numbers of prior treatments is unfeasible. Using high-quality real-world data can robustly evaluate patient outcomes contingent on certain oncogenic drivers and can address key questions around the biological plausibility of using genetic biomarkers as novel therapeutic targets36 or describe unmet needs for patients. A recent study in patients with NTRK fusion–positive solid tumors also used data from the Flatiron Health-Foundation Medicine CGDB and concluded that NTRK fusions may represent another negative prognostic factor in patients with locally advanced/metastatic tumors.37 Again, this reflects the value of real-world data as an integral resource for clinical evidence generation beyond the confines of conventional clinical trials,38-40 particularly when considering outcomes for patients with rare molecular alterations.
Strengths of our study include the use of real-world data from a large CGDB, with a broad network of sites, which has been used in published analyses of other biomarkers.37,41 Consistent results across three different analyses, including one to account for immortal time bias, support the robustness of our findings.
Limitations of this study include its retrospective nature and the relatively small number of patients with RET fusion–positive tumors, reflecting the rarity of these fusions.2 Our study cohort was heterogeneous, and it could not be determined if/how tumor types with relatively long survival or large censoring could influence the findings. We adjusted for several important factors in the regression analyses, but there may also be unmeasured confounding factors. In addition, some of these factors had a certain degree of missing data, and therefore, we could not totally assess their impact on the measured associations. It is unknown how well our data represent the wider spectrum of RET fusion–positive solid tumors as CGP is not yet a routine practice for all tumor types. Potential biases may exist if CGP was preferentially performed on patients who did not respond to treatment in the real-world setting.
In tumor-agnostic clinical trials, multiple tumor types with the same alteration are given an experimental targeted therapy. There may be a single control arm (including patients who are all given standard-of-care treatment) or no control arm. Ideally, each tumor type in the experimental arm would have corresponding control patients, and the trial would be powered for this comparison. In reality, the number of patients with a specific tumor type in the trial is too small for this to be feasible, especially when the overall prevalence of genetic alterations, such as RET or NTRK fusions,42,43 is already low. If patients with an alteration-positive tumor have improved prognoses versus those who are alteration-negative, the apparent increased efficacy seen in tumor-agnostic trials could be partly or wholly due to the prognostic effect of the alteration (ie, confounding) and not the treatment. The prognostic association cannot be determined from these trials because, by design, they exclude patients without these alterations. However, we show that patients with RET fusion–positive tumors have worse OS; in tumor-agnostic trials, any improvements in efficacy associated with experimental therapies targeted for RET fusions are more likely to be due to the treatment. Such treatments essentially must not only overcome the negative prognostic association but also provide extra benefit versus standard of care.
In conclusion, our study shows the value of real-world data by highlighting that large data sets are needed to yield a sufficient number of patients when studying rare cancers or genomic alterations. We focused on RET status, but there are plans to use the same CGDB to examine the prognostic performance of other uncommon alterations. Our study showed that patients with RET fusion–positive tumors, excluding NSCLC, have worse survival outcomes than patients with RET-WT tumors, highlighting the importance of these fusions as actionable drug targets and the need for widespread integration of CGP in routine clinical practice. This evidence may help interpret future clinical trials of tumor-agnostic therapies developed for RET fusions (other than in NSCLC); our findings indicate that the negative prognostic association would be unlikely to explain treatment benefits. This information could be used by researchers, regulators, and other decision makers. Combined efforts across industry, academia, health care authorities, and payers are needed to evaluate RET-specific TKIs for patients with RET fusion–positive tumors and potentially allow more patients to benefit from advances in precision oncology in the future.
ACKNOWLEDGMENT
Medical writing support for the development of this manuscript, under the direction and written contributions of the authors, was provided by Lietta Nicolaides, PhD, of Ashfield MedComms, an Inizio company, and funded by F. Hoffmann-La Roche Ltd.
Notes
Allan Hackshaw
Stock and Other Ownership Interests: Thermo Fisher Scientific, Illumina
Honoraria: Abbvie, UCB, Takeda, Daiichi Sankyo, Clovis Oncology, Ipsen, AstraZeneca, Kyowa Kirin International, Pfizer, Servier, Almirall, Boehringer Ingelheim
Consulting or Advisory Role: Roche, Abbvie, Grail
Research Funding: Roche (Inst), Boehringer Ingelheim (Inst), MSD (Inst), Autolus (Inst), Grail (Inst), AstraZeneca (Inst), Bristol Myers Squibb/Sanofi (Inst), Pfizer (Inst)
Otto Fajardo
Employment: Roche
Stock and Other Ownership Interests: Roche
Urania Dafni
Honoraria: Roche
Expert Testimony: AstraZeneca
Hans Gelderblom
Research Funding: Novartis (Inst), Ipsen (Inst), Deciphera (Inst), AmMax Bio (Inst)
Pilar Garrido
Employment: Teva
Consulting or Advisory Role: Roche Pharma AG, AstraZeneca, MSD Oncology, Bristol Myers Squibb, Takeda, Lilly, Pfizer, Novartis/Pfizer, Boehringer Ingelheim, Abbvie, Amgen, Bayer, Gebro Pharma, Nordic Group, Janssen Biotech, Janssen Oncology, GlaxoSmithKline, IO Biotech, Merck KGaA, Daiichi Sankyo Europe GmbH, Sanofi/Regeneron
Speakers' Bureau: Roche Pharma AG, Takeda, AstraZeneca, MSD Oncology, BMS, Pfizer, Novartis, Boehringer Ingelheim, Nordic Group, Janssen, Boehringer Ingelheim, Janssen Oncology, Medscape, touchIME
Travel, Accommodations, Expenses: AstraZeneca Spain, Roche Pharma AG
Other Relationship: Janssen Oncology, Novartis, IO Biotech
Salvatore Siena
Stock and Other Ownership Interests: Guardant Health, Myriad Genetics
Consulting or Advisory Role: Bayer, Bristol Myers Squibb, Daiichi Sankyo, Merck, Novartis, CheckmAb, Agenus, AstraZeneca, GlaxoSmithKline, MSD Oncology, Pierre Fabre, Seagen, T-One Therapeutics
Research Funding: MSD Oncology (Inst)
Patents, Royalties, Other Intellectual Property: Amgen
Travel, Accommodations, Expenses: Amgen, Bayer, Roche
Matthew H. Taylor
Honoraria: Bristol Myers Squibb Foundation, Eisai, Bayer, Merck, Pfizer, Regeneron, Roche, Blueprint Medicines
Consulting or Advisory Role: Bristol Myers Squibb, Eisai, Loxo, Bayer, Blueprint Medicines, Novartis, Sanofi, Cascade Prodrug, Merck, Pfizer, Exelixis, Immune-Onc Therapeutics, Regeneron
Speakers' Bureau: Bristol Myers Squibb, Eisai, Merck, Blueprint Medicines
Research Funding: Bristol Myers Squibb (Inst), Eisai (Inst), Pfizer (Inst), Merck (Inst), Moderna Therapeutics (Inst), Loxo/Bayer (Inst), Blueprint Medicines (Inst), Seagen (Inst)
Walter Bordogna
Employment: Roche
Stock and Other Ownership Interests: Roche
Christos Nikolaidis
Employment: Roche
Travel, Accommodations, Expenses: Roche
No other potential conflicts of interest were reported.
*A.H. and O.F. contributed equally to this work as co-lead authors.
DATA SHARING STATEMENT
For up-to-date details on Roche's Global Policy on the Sharing of Clinical Information and how to request access to related clinical study documents, see here: https://go.roche.com/data_sharing. The data that support the findings of this study have been originated by Flatiron Health, Inc and Foundation Medicine, Inc. These deidentified data may be made available upon request and are subject to a license agreement with Flatiron Health and Foundation Medicine; interested researchers should contact moc.noritalf@imf-bdgc and moc.noritalf@sseccaatad to determine licensing terms.
AUTHOR CONTRIBUTIONS
Conception and design: Otto Fajardo, Christos Nikolaidis
Provision of study materials or patients: Otto Fajardo, Christos Nikolaidis
Collection and assembly of data: Otto Fajardo, Christos Nikolaidis
Data analysis and interpretation: All authors
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.
Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).
Allan Hackshaw
Stock and Other Ownership Interests: Thermo Fisher Scientific, Illumina
Honoraria: Abbvie, UCB, Takeda, Daiichi Sankyo, Clovis Oncology, Ipsen, AstraZeneca, Kyowa Kirin International, Pfizer, Servier, Almirall, Boehringer Ingelheim
Consulting or Advisory Role: Roche, Abbvie, Grail
Research Funding: Roche (Inst), Boehringer Ingelheim (Inst), MSD (Inst), Autolus (Inst), Grail (Inst), AstraZeneca (Inst), Bristol Myers Squibb/Sanofi (Inst), Pfizer (Inst)
Otto Fajardo
Employment: Roche
Stock and Other Ownership Interests: Roche
Urania Dafni
Honoraria: Roche
Expert Testimony: AstraZeneca
Hans Gelderblom
Research Funding: Novartis (Inst), Ipsen (Inst), Deciphera (Inst), AmMax Bio (Inst)
Pilar Garrido
Employment: Teva
Consulting or Advisory Role: Roche Pharma AG, AstraZeneca, MSD Oncology, Bristol Myers Squibb, Takeda, Lilly, Pfizer, Novartis/Pfizer, Boehringer Ingelheim, Abbvie, Amgen, Bayer, Gebro Pharma, Nordic Group, Janssen Biotech, Janssen Oncology, GlaxoSmithKline, IO Biotech, Merck KGaA, Daiichi Sankyo Europe GmbH, Sanofi/Regeneron
Speakers' Bureau: Roche Pharma AG, Takeda, AstraZeneca, MSD Oncology, BMS, Pfizer, Novartis, Boehringer Ingelheim, Nordic Group, Janssen, Boehringer Ingelheim, Janssen Oncology, Medscape, touchIME
Travel, Accommodations, Expenses: AstraZeneca Spain, Roche Pharma AG
Other Relationship: Janssen Oncology, Novartis, IO Biotech
Salvatore Siena
Stock and Other Ownership Interests: Guardant Health, Myriad Genetics
Consulting or Advisory Role: Bayer, Bristol Myers Squibb, Daiichi Sankyo, Merck, Novartis, CheckmAb, Agenus, AstraZeneca, GlaxoSmithKline, MSD Oncology, Pierre Fabre, Seagen, T-One Therapeutics
Research Funding: MSD Oncology (Inst)
Patents, Royalties, Other Intellectual Property: Amgen
Travel, Accommodations, Expenses: Amgen, Bayer, Roche
Matthew H. Taylor
Honoraria: Bristol Myers Squibb Foundation, Eisai, Bayer, Merck, Pfizer, Regeneron, Roche, Blueprint Medicines
Consulting or Advisory Role: Bristol Myers Squibb, Eisai, Loxo, Bayer, Blueprint Medicines, Novartis, Sanofi, Cascade Prodrug, Merck, Pfizer, Exelixis, Immune-Onc Therapeutics, Regeneron
Speakers' Bureau: Bristol Myers Squibb, Eisai, Merck, Blueprint Medicines
Research Funding: Bristol Myers Squibb (Inst), Eisai (Inst), Pfizer (Inst), Merck (Inst), Moderna Therapeutics (Inst), Loxo/Bayer (Inst), Blueprint Medicines (Inst), Seagen (Inst)
Walter Bordogna
Employment: Roche
Stock and Other Ownership Interests: Roche
Christos Nikolaidis
Employment: Roche
Travel, Accommodations, Expenses: Roche
No other potential conflicts of interest were reported.
REFERENCES
Articles from JCO Precision Oncology are provided here courtesy of American Society of Clinical Oncology
Citations & impact
This article has not been cited yet.
Impact metrics
Alternative metrics
Discover the attention surrounding your research
https://www.altmetric.com/details/159473205
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Characteristics and outcomes of patients with RET-fusion positive non-small lung cancer in real-world practice in the United States.
BMC Cancer, 21(1):28, 05 Jan 2021
Cited by: 30 articles | PMID: 33402119 | PMCID: PMC7786962
Efficacy of immune checkpoint inhibitor therapy in patients with RET fusion-positive non-small-cell lung cancer.
Immunotherapy, 13(11):893-904, 18 Jun 2021
Cited by: 17 articles | PMID: 34139897
RET fusions as primary oncogenic drivers and secondary acquired resistance to EGFR tyrosine kinase inhibitors in patients with non-small-cell lung cancer.
J Transl Med, 20(1):390, 04 Sep 2022
Cited by: 7 articles | PMID: 36059009 | PMCID: PMC9441062
RET rearrangements in non-small cell lung cancer: Evolving treatment landscape and future challenges.
Biochim Biophys Acta Rev Cancer, 1877(6):188810, 04 Oct 2022
Cited by: 9 articles | PMID: 36202311
Review
Funding
Funders who supported this work.
Cancer Research UK (1)
Cancer Clinical Trials in the Cancer Research UK and UCL Cancer Trials Centre
Professor Allan Hackshaw, University College London
Grant ID: 25349