Artificial intelligence (AI)-derived computer-aided detection (CAD) software for detecting and measuring lung nodules in CT scan images

  • HealthTech guidance
  • HTG687
  • Published: 
Download guidance (PDF)

1 Recommendations

NICE is aware that companies are reviewing their CE marking in response to changes and advances in regulations for digital health technologies.

1.1

For people having a chest CT scan because of signs or symptoms that suggest lung cancer, there is not enough evidence to recommend artificial intelligence (AI)‑derived computer-aided detection (CAD) software alongside clinician review of CT scan images. For these people, the following AI-derived CAD software should only be used in research to help detect, measure and assess the growth of lung nodules:

  • AI-Rad Companion Chest CT (Siemens Healthineers)

  • AVIEW LCS+ (Coreline Soft)

  • ClearRead CT (Riverain Technologies)

  • contextflow SEARCH Lung CT (contextflow)

  • InferRead CT Lung (Infervision)

  • Lung AI (Arterys)

  • qCT-Lung (Qure.ai)

  • SenseCare-Lung Pro (SenseTime)

  • Veolity (MeVis)

  • Veye Lung Nodules (Aidence)

  • VUNO Med‑LungCT AI (VUNO).

1.2

For people having a chest CT scan for reasons not related to suspicion of lung cancer, there is not enough evidence to recommend AI‑derived CAD software alongside clinician review of CT scan images. For these people, the following AI-derived CAD software should only be used in research to help detect, measure and assess the growth of lung nodules:

  • AI-Rad Companion Chest CT (Siemens Healthineers)

  • AVIEW LCS+ (Coreline Soft)

  • ClearRead CT (Riverain Technologies)

  • contextflow SEARCH Lung CT (contextflow)

  • InferRead CT Lung (Infervision)

  • Lung AI (Arterys)

  • qCT‑Lung (Qure.ai)

  • SenseCare‑Lung Pro (SenseTime)

  • Veolity (MeVis)

  • Veye Lung Nodules (Aidence)

  • VUNO Med-LungCT AI (VUNO).

1.3

For people having a chest CT scan as part of targeted lung cancer screening, AI-derived CAD software technologies have the potential to be cost effective. But there is not enough evidence to determine which of them are the most clinically and cost effective. Centres using AI‑derived CAD software alongside clinician review as part of targeted lung cancer screening should generate further evidence. This is to make sure the potential benefits are realised in practice for people having screening and for clinicians using the software, and to allow comparisons between the different software (see recommendation 1.4).

1.4

Further evidence generation is recommended (see the section on further research) to assess:

  • the prevalence of nodules in people who have a chest CT scan because of signs or symptoms that suggest lung cancer

  • how using AI-derived CAD software alongside clinician review of CT scan images affects the accuracy of detecting, measuring and assessing the growth of lung nodules, including in people with underlying lung conditions and people whose family background means they are more likely to have subsolid nodules

  • how using AI-derived CAD software alongside clinician review affects clinical decision making

  • how using AI-derived CAD software alongside clinician review affects scan review and reporting time.

Why the committee made these recommendations

Using AI‑derived CAD software alongside clinician review of CT scans could improve detecting, measuring and assessing the growth of lung nodules.

When used in people having a CT scan because of suspected lung cancer or for reasons not related to lung cancer, using the software could lead to more people being identified with lung nodules that are not likely to be cancer. This could lead to people having CT surveillance they do not need, which may cause unnecessary anxiety. This is uncertain because there is not much evidence, so more research is needed.

Although there is more evidence in targeted lung cancer screening, it is too limited to show which technologies are the most clinically and cost effective. But the model results suggest that using the software alongside clinician review has the potential to be cost effective. So, although there is not yet enough evidence to recommend the software, centres may use it as part of targeted lung cancer screening. But they should generate evidence to make sure the potential benefits of using the software are realised in practice and to allow comparisons of the different technologies.