Clinical and technical evidence
A literature search was carried out for this briefing in accordance with the interim process and methods statement. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting email@example.com.
There are 4 studies summarised in this briefing, involving data from around 15,000 patients.
The studies show the diagnostic accuracy of the EarlyCDT‑Lung test and its ability to pick up early stage lung cancer when compared with standard care. The clinical evidence and its strengths and limitations is summarised in the overall assessment of the evidence.
The evidence for the EarlyCDT‑Lung test is of good quality but limited in its scale and scope. There is no long-term follow up of patients past 2 years and the frequency of testing has not been established.
The available evidence shows a variation in values for specificity and sensitivity across the studies and across lung cancer stages. This variation is probably because of different thresholds being applied in each study.
The evidence shows that EarlyCDT‑Lung can improve diagnostic yield and prediction of the risk of lung cancer.
EarlyCDT‑Lung test compared with lung cancer risk models or nodule size (4 mm to 20 mm).
Presence of lung cancer was confirmed either based on clinician's diagnosis only (inclusive cohort) or based on the availability of CT and pathology reports for people from the inclusive cohort (exclusive cohort). After applying exclusion criteria, 296 patients (221 without cancer and 75 with cancer) were left in the inclusive cohort and 269 (217 without cancer and 52 with cancer) in the exclusive cohort.
Results of model-free analysis showed that for the exclusive cohort, 22 of 68 patients with a positive test result and 30 of 201 of patients with a negative test result had lung cancer, giving a relative risk (RR) of 2.2 (95% confidence interval [CI] 1.3 to 3.5, positive prediction value [PPV] 32%). In the EarlyCDT‑Lung test, the RR increased to 2.3 (95% CI 1.3 to 3.9, PPV 40%). For the 4 mm to 20 mm group, addition of a positive EarlyCDT test gave an RR of 2.7 (95% CI 1.3 to 5.7), equal to an increase in absolute risk from 13% to 24%. Results of the nodule-based risk model showed that, for the inclusive cohort, the Gould and Brock models respectively overestimated and underestimated the RR to the actual rates observed (p<0.001). The Mayo estimates were closest to those observed (p=0.54). For the 4‑mm to 20‑mm nodules, the receiver operating characteristic curves for the models showed that, as specificity passed 90%, the sensitivity decreased to 30% or less. On adding EarlyCDT‑Lung, specificity increased and sensitivity decreased.
The study shows how the test is likely to be used in clinical practice. It was funded by the company. There was no blinding of test assessors, and test results appear to have been interpreted with knowledge of the reference standard. The study authors noted that no formal protocols were followed.
Based on 6‑month follow up of 99% of patients who were tested positive and 93% of those who tested negative, 61 (4%) were identified as having lung cancer, 25 of whom tested positive using EarlyCDT‑Lung. This shows a sensitivity of 41%. Of these, 57% when reported were stage 1 or stage 2. A positive test result was associated with a 5.4‑fold increase in lung cancer compared with a negative result. The specificity of the test at 6 months was 87%.
A relatively large sample size, which shows the diagnostic power of the test at 6 months. Further follow up would have been useful to show longer-term diagnostic performance. The study would have benefited from a comparator arm that had standard care alone. Some of the authors had conflicts of interest ranging from direct financial or working relationships with the company to consultancy work and research grants.
Case control study involving 4 separate groups of patients with newly diagnosed lung cancer. There were 122 from a single UK centre with small-cell lung cancer (group 1), 249 from multiple European centres (group 2), 122 from a single centre in Vancouver, Canada (group 3), 82 with no location specified (group 4).
In 3 of the 4 groups (groups 2 to 4) patients with lung cancer were, as far as possible, individually matched by gender, age, and smoking history to control individuals with no previous history of malignant disease.
These populations were combined with an early validation data set to give 1,077 patients with lung cancer and 1,296 matched controls.
Sensitivity 57%, specificity not reported (group 1); sensitivity 34%, specificity 87% (group 2); sensitivity 31%, specificity 84% (group 3); sensitivity 43%, specificity 89% (group 4); sensitivity 38%, specificity 88% (overall); sensitivity 34%, specificity 88% for non-small-cell lung cancer; and sensitivity 50%, specificity 88% for small-cell lung cancer. There was a higher sensitivity for small-cell lung cancer compared with non-small-cell lung cancer (p≤0.001) but no difference in sensitivity between non-small-cell lung cancer subtype (p=0.35).
A large sample and matched control involving several centres across Europe and the US, including a group from the UK. Results include a previous validation sample, with limited reporting. Matching was done on a number of important variables, but there was no reporting on the closeness of matching and consideration of other key variables when matching may also have been appropriate. The study was part funded by the company, 2 of the authors were consultants to the company and 1 of these was also a shareholder.
EarlyCDT‑Lung (positive tests were followed by chest X‑ray and CT imaging) compared with standard care.
During the study, 127 lung cancers were diagnosed (56 in the test group and 71 in the control arm). Of the test group, 9.8% had positive EarlyCDT‑Lung test results and 3% (n=18) of these people were diagnosed with lung cancer. The rate of late-stage (3 and 4) lung cancer diagnosis was lower in the test group than the intervention group (58.9% compared with 73.2%). More early-stage cancers were diagnosed in the test group (23 compared with 19). The EarlyCDT‑Lung test was positive for 12 of the 23 early cancers (sensitivity 52.2%, 95% CI 30.6% to 73.2%) and for 6 of the 33 late-stage cancers (sensitivity 18.2%, 95% CI 7.0% to 35.5%). Although the study was not powered to detect difference in mortality, lung cancer-specific mortality was lower in the intervention arm (17 compared with 24). Further information from the study authors note that EarlyCDT‑Lung test showed high overall specificity (90.3%), and moderate sensitivity (52.2%) for detecting early stage lung cancer at 2 years. Sensitivity at 6 months and at 1 year was 77.8% and 69.2% respectively for stage 1 or stage 2 disease, and 38.5% and 30.0% respectively for stage 3 or stage 4 disease.
The company notes that the main body and inserts of the kit are cardboard, which can be recycled. Reagent bottles and other kit contents which may encounter human serum are treated as contaminated waste and cannot be recycled. The company states that this is beyond its control.
No specialist equipment is needed to run the test. For patients with positive findings, there may be a reduction in follow-up CT scans.
There is no published evidence to support these claims.
Lung cancer screening study with low-dose CT scan and blood biomarker. ClinicalTrials.gov identifier: NCT01700257. Status: was recruiting, no results published. Indication: lung cancer. Devices: EarlyCDT‑Lung test. Location: US.