2 The diagnostic tests
Clinical need and practice
2.1 Lung cancer is one of the most common types of cancer in the UK. It causes symptoms such as persistent cough, coughing up blood, and feeling short of breath. People in the early stages of the disease may not have symptoms and so lung cancer is often diagnosed late. In 2018, more than 65% of lung cancers were diagnosed at stage 3 or 4 (Cancer Research UK). The NHS Long Term Plan sets out the NHS's ambition to diagnose 75% of all cancers at stages 1 or 2 by 2028.
2.2 Detecting abnormal lung features such as lung nodules, pleural effusion (excess fluid around the lung), collapsed lung or lung segments, destruction or erosion of bony structures such as the ribs, as well as masses or lymph nodes within the mediastinum could help find lung cancer early. Lung abnormalities can be seen on a chest X‑ray. The chest X‑ray may be done because of signs and symptoms that suggest lung cancer. Lung abnormalities that suggest cancer can also be detected incidentally on chest X‑rays done for reasons unrelated to lung cancer, such as trauma, heart problems, or investigation and monitoring of other lung conditions.
2.3 Software with artificial intelligence (AI)‑derived algorithms can be used to automatically detect lung abnormalities on chest X‑ray images. This could help a radiologist or reporting radiographer with review and interpretation of chest X‑ray images and support clinical decisions about the need for a CT scan or further investigations. Using the technology alongside clinician review may help to detect suspected lung cancer by:
-
identifying normal or abnormal images and highlighting suspected abnormalities
-
prioritising review of chest X‑rays and speeding up subsequent referral to CT scan
-
increasing the accuracy of suspected lung cancer detection by radiologists or reporting radiographers
-
reducing the time to review and report chest X‑rays.
The interventions
2.4 The software included in this evaluation have AI‑derived algorithms to detect lung abnormalities for suspected cancer. All software technologies in clinical settings use fixed algorithms. They cannot adapt in real time using data from the clinical practice setting in which they are used. In the NHS, AI‑based technologies are only used alongside healthcare professionals. The healthcare professional who reviews the chest X-ray using the software makes the final reporting decision. These software could be radiological computer-assisted triage (CAST) software, computer-aided detection (CADe) software or computer-aided diagnosis (CADx) software. CAST software is intended to help prioritisation and triage of medical images that are time sensitive for urgent review. CADe and CADx software are technologies that can detect or diagnose an abnormality on the X‑ray.
2.5 The technologies in this evaluation are:
-
AI‑Rad Companion Chest X‑ray (class 2a medical device, CADx, Siemens Healthineers)
-
Annalise CXR (class 2b medical device, CADe/CAST, Annalise ai)
-
Auto Lung Nodule Detection (class 2a medical device, CADe, Samsung)
-
ChestLink (class 2b medical device, CADe/CAST, Oxipit)
-
ChestView (class 2a medical device, CADe, Gleamer)
-
Chest X‑ray (class 2a medical device, CADe, Rayscape)
-
ClearRead Xray (class 2a medical device, CADe, Riveraintech)
-
InferRead DR Chest (class 2a medical device, CADe, Infervision)
-
Lunit INSIGHT CXR (class 2a medical device, CADe, Lunit)
-
Milvue Suite (class 2a medical device, CADe/CAST, Milvue)
-
qXR (class 2a medical device, CADe, Qure.ai)
-
Red dot (class 2a medical device, CADe/CADx/CAST, Behold.ai)
-
SenseCare‑Chest DR PRO (class 2b medical device, CADe, SenseTime)
-
VUNO Med‑Chest X‑Ray (class 2a medical device, CADe, VUNO).