Dr Katie Sutton

DERM is an AI algorithm for detecting skin cancer. Developed by Skin Analytics, the algorithm analyses images of skin lesions and determines if the lesion is likely to be cancerous. DERM is already being used by some NHS providers, and is integrated with existing skin cancer pathways, triaging patients before being assessed by dermatologists. The Department of Health and Social Care (DHSC) funded deployment and a real-world evaluation of DERM as part of the AI in Health and Care Award. The evaluation aimed to inform future implementation by assessing key domains such as safety, accuracy, effectiveness, value and sustainability. Working as a collaborative Technology Specific Evaluation Team (TSET), Unity Insights and University of Surrey conducted an evaluation of DERM in four NHS sites, across 9,649 patients between February 2022 and April 2023. The resulting evaluation report is presented in full.

12009 Background: Older and frail patients with cancer (Ca) often receive less aggressive treatment and as a result have worse survival. Current methods of assessing fitness (performance status) for intensive treatment such as chemotherapy are inadequate. The complexity of geriatric assessments, lack of training and time pressures in busy clinics, mean that better solutions are needed. A UK initiative - the electronic frailty index (eFI) - is derived from a cumulative deficit frailty model and provides a measure of frailty alongside pre-existing conditions such as issues with mobility, fractures, falls, memory, sight, hearing , anaemia, tremor, diabetes, heart, thyroid, skin, respiratory, cerebrovascular, circulation, social vulnerability, and polypharmacy (36 fields) (Clegg et al). Patients are classified into the following groups: no frailty, mild, moderate or severe frailty. We used the same methodology to investigate whether eFI predicts adverse outcomes of chemotherapy in frail patients with Ca. Methods: The study conducted retrospective data analysis of Ca patients treated with chemotherapy from Public Health England (PHE) Systemic Chemotherapy Dataset (SACT) years 2015-2018. Eligible patients had stage II - III breast Ca, stage III colon Ca or stage IIIB–IV non-small-cell lung Ca (NSCLC). The data from SACT was linked with hospital episodes' statistics (obtained from NHS-Digital) to calculate 30-day SACT mortality, overall survival and hospital admissions. EFI was calculated from the above 35 fields; polypharmacy was not available from NHS-Digital data. Results: The eFI was calculated for 78799 patients: colorectal 17951, lung 22052, and breast 38796. 20388 patients were ≥ 70y. o. and 58411 were < 70y.o. Most patients were fit with an eFI score of 0-69% (54563), 19% (15,295) had mild frailty, 7.7% (6104)- moderate, and 3.6% (2837) had severe frailty. 4.2% (3356) of patients died within 30 days of SACT. For colorectal Ca patients the risk of dying within 30 days of SACT in patients ≥70y.o was twice that of the < 70y.o (OR 2.04 -CI 1.58 - 2.64); patients with mild eFI did not differ- OR: 1.07 (CI 0.78-1.45), moderate frailty: 1.6 (CI 1.1-2.33) and severe frailty: 2.13 (CI 1.34-3.39). In breast Ca patients, 30-day mortality for ≥70y.o. was 6.38 times higher than for < 70y.o (95% CI 4.29-9.49); eFI for mild frailty- OR of 1.45 (95% CI 0.78-2.71), moderate frailty-OR of 3.5 (95%CI 1.82-6.75) and severe frailty 5.73 (95% CI 2.66-12.3). The 30-day mortality in lung cancer in ≥70y.o and < 70y.o did not differ with OR 0.95 (95% CI 0.88-1.03) for ≥70y.o. Patients with mild frailty had OR for 30-day mortality of 1.17 (95% CI 1.07-1.28), moderate frailty-OR of 1.28 (95%CI 1.15-1.44) and severe frailty 1.48 (95% CI 1.28-1.77). Conclusions: The eFI closely predicts poor outcomes from SACT, particularly in early breast and colon cancer, and requires further evaluation in a prospective study.

The intention of this study was to offer an alternative perspective to the quantitative findings of larger randomised controlled trials by using a phenomenological approach to explore the impact on Quality of Life (QoL) for people with brain tumours of entering a research trial involving new anti-cancer agents. Given the subjective nature of the proposed topic, a phenomenological approach was adopted. Sample size was limited to five participants. A semi-structured interview technique was used. Interviews were digitally audio recorded with permission from those involved. In order to guide data analysis for this study, Colaizzi's framework was utilised. As a result of data analysis, two major themes were identified. These were ‘Hope and optimism’ and ‘The therapeutic relationship’. Three minor themes were also found. These were ‘A complex symptom profile’, ‘The importance of non-medical coping strategies’ and ‘Impressions of the QoL tools used’. This phenomenological study has highlighted key themes relating to QoL which are not addressed in some of the widely used assessment tools such as the EORTC QLQ C30 and BN20. They generally focus on health status, and do not capture issues identified in this study as being of significant importance to the QoL of participants such as hope and optimism, and the importance of the therapeutic relationship. They also omit reference to coping and management strategies.