This study is a single-centre prospective observational cohort study designed to assess and compare the sensitivity and specificity of a Lunit INSIGHT MMG assisted human reading to the standard care double human reading process within mammography review at a "one-stop" breast clinic (non-inferiority study).

The current imaging reporting process is a sequential double read of mammography and ultrasound (if available) images, by consultant radiologists or radiographers. The first reader produces a report which is then sent to the second reader who reviews it. If the second reader agrees with the first, this is reflected in the second reader's report which translates into a decision for further action; in the event of disagreement, a third reader arbitrates and produces the final report.

In the past, breast clinics have had to resort to single reader reporting due to staff shortages and high demand. This results in delays to any further assessments that may be required. It is worth noting however that despite difficulties in meeting the target, the current clinical pathway has proven to be cost effective.

The Lunit INSIGHT MMG tool could generate benefits and potential efficiencies if it were introduced to the clinical service as an assistant reader within the mammography reporting process, by replacing one of the two human readers in the current standard of care. Before this can be assessed however, its non-inferiority in combination with a human reader in comparison to standard of care (double human reading) must first be established. This study will aim to address this issue in the first instance, maintaining standard of care for all patients seen within the 2 week wait pathway, by introducing the use of Lunit INSIGHT MMG into one of two arms within this prospective, observational parallel cohort study.

The study has three main objectives, which will be met through separate sub-studies and analyses.

First, it will investigate and compare the performance of the Lunit INSIGHT MMG tool in combination with a human reader, against the current standard of care (two human readers). Quantitative data will be collected from a prospective, observational cohort study which will take place during routine clinical care at the participating site in parallel to standard care. The main aims will be:

1. To assess and compare the sensitivity and specificity of Lunit INSIGHT MMG combined with a single human reader for non-inferiority compared to double human reading
2. To assess and compare the rate of escalation to arbitration (following non-agreement) for non-inferiority of Lunit INSIGHT MMG combined with a single human reader against double human reading
3. To assess clinician reported agreement with the Lunit INSIGHT MMG report
4. To address any differences in performance characteristics as a function of patient age 6 To assess how Lunit score relates to tumour grade on biopsy

Second, the acceptability and perceptions surrounding the AI will be investigated using qualitative data obtained from patient and workforce interviews and questionnaires from the latter. The main aims will be to assess the acceptability of the Lunit INSIGHT MMG tool to patients and the workforce and explore patient and workforce perceptions surrounding it and its implementation in clinical pathways.

Third, a health economic evaluation will be carried out in a cost effectiveness analysis to assess the value of introducing the Lunit INSIGHT MMG tool at a "one-stop" breast clinic. The main aims will be two. The implementation costs of the Lunit Insight MMG tool and costs of a scan read by it, in combination with a human reader, will be compared against the cost of double human readings based on standard practices. In addition, a valuation of outcomes will also be undertaken following the lifetime of the cohort with the use of the Lunit INSIGHT MMG tool compared with standard of care.

Inclusion Criteria:

Non-Inferiority Study

• Women 18 years of age or older.
• Referred to a "one-stop" breast clinic following GP appointment for breast symptoms.

Patient and workforce qualitative studies Patients: as above. Workforce: Clinical staff and other key stakeholders involved in the implementation of Lunit INSIGHT MMG both male and female

• Able to undergo mammography.

Exclusion Criteria:

Non-Inferiority Study

• Men will not be included in this study.
• Women who are not able to tolerate a mammogram or are unable to have one due to physical difficulties in engagement with the technology caused by immobility or being wheelchair bound.

Patient and workforce qualitative studies Patients: as above. Workforce: Has no connection to Lunit INSIGHT MMG implementation irrespective of being male or female

The symptomatic breast population will be recruited from the breast service 2 week pathway referred to a 'one-stop' breast clinic

No patient identifiable data will be shared outside of the direct care team for the main study. The qualitative interview data will only be available to the King's College Research team which also be anonymized after transcription.