Examining labelling guidelines for AI-based software as a medical device: A review and analysis of dermatology mobile applications in Australia.

In recent years, there has been a surge in the development of AI-based Software as a Medical Device (SaMD), particularly in visual specialties such as dermatology. In Australia, the Therapeutic Goods Administration (TGA) regulates AI-based SaMD to ensure its safe use. Proper labelling of these devices is crucial to ensure that healthcare professionals and the general public understand how to use them and interpret results accurately. However, guidelines for labelling AI-based SaMD in dermatology are lacking, which may result in products failing to provide essential information about algorithm development and performance metrics. This review examines existing labelling guidelines for AI-based SaMD across visual medical specialties, with a specific focus on dermatology. Common recommendations for labelling are identified and applied to currently available dermatology AI-based SaMD mobile applications to determine usage of these labels. Of the 21 AI-based SaMD mobile applications identified, none fully comply with common labelling recommendations. Results highlight the need for standardized labelling guidelines. Ensuring transparency and accessibility of information is essential for the safe integration of AI into health care and preventing potential risks associated with inaccurate clinical decisions.

Minimum labelling requirements for dermatology artificial intelligence-based Software as Medical Device (SaMD): A consensus statement.

BACKGROUND/OBJECTIVES: Artificial intelligence (AI) holds remarkable potential to improve care delivery in dermatology. End users (health professionals and general public) of AI-based Software as Medical Devices (SaMD) require relevant labelling information to ensure that these devices can be used appropriately. Currently, there are no clear minimum labelling requirements for dermatology AI-based SaMDs. METHODS: Common labelling recommendations for AI-based SaMD identified in a recent literature review were evaluated by an Australian expert panel in digital health and dermatology via a modified Delphi consensus process. A nine-point Likert scale was used to indicate importance of 10 items, and voting was conducted to determine the specific characteristics to include for some items. Consensus was achieved when more than 75% of the experts agreed that inclusion of information was necessary. RESULTS: There was robust consensus supporting inclusion of all proposed items as minimum labelling requirements; indication for use, intended user, training and test data sets, algorithm design, image processing techniques, clinical validation, performance metrics, limitations, updates and adverse events. Nearly all suggested characteristics of the labelling items received endorsement, except for some characteristics related to performance metrics. Moreover, there was consensus that uniform labelling criteria should apply across all AI categories and risk classes set out by the Therapeutic Goods Administration. CONCLUSIONS: This study provides critical evidence for setting labelling standards by the Therapeutic Goods Administration to safeguard patients, health professionals, consumers, industry, and regulatory bodies from AI-based dermatology SaMDs that do not currently provide adequate information about how they were developed and tested.

Assessing the genetic risk of nodular melanoma using a candidate gene approach.

BACKGROUND: Nodular melanoma (NM) is a challenge to diagnose early due to its rapid growth and more atypical clinical presentation, making it the largest contributor to melanoma mortality. OBJECTIVES: Our study aim was to perform a rare-variant allele (RVA) analysis of whole-exome sequencing of patients with NM and non-NM (minor allele frequency ≤ 1% non-Finnish European) for a set of 500 candidate genes potentially implicated in melanoma. METHODS: This study recruited 131 participants with NM and 194 with non-NM from South-east Queensland and patients with NM from Victoria to perform a comparative analysis of possible genetic differences or similarities between the two melanoma cohorts. RESULTS: Phenotypic analysis revealed that a majority of patients diagnosed with NM were older males with a higher frequency of fair skin and red hair than is seen in the general population. The distribution of common melanoma polygenic risk scores was similar in patients with NM and non-NM, with over 28% in the highest quantile of scores. There was also a similar frequency of carriage of familial/high-penetrant melanoma gene and loss-of-function variants. We identified 39 genes by filtering 500 candidate genes based on the greatest frequency in NM compared with non-NM cases. The genes with RVAs of greatest frequency in NM included PTCH1, ARID2 and GHR. Rare variants in the SMO gene, which interacts with PTCH1 as ligand and receptor, were also identified, providing evidence that the Hedgehog pathway may contribute to NM risk. There was a cumulative effect in carrying multiple rare variants in the NM-associated genes. A 14.8-fold increased ratio for NM compared with non-NM was seen when two RVAs of the 39 genes were carried by a patient. CONCLUSIONS: This study highlights the importance of considering frequency of RVA to identify those at risk of NM in addition to known high penetrance genes.

Do age at diagnosis, tumour thickness and tumour site explain sex differences in melanoma survival? A causal mediation analysis using cancer registry data.

Women diagnosed with melanoma have better survival than men, but little is known about potential intervention targets to reduce this survival gap by sex. We conducted a population-based study using Victorian Cancer Registry data including 5833 women and 6780 men aged 15 to 70 years when diagnosed with first primary melanoma between 2007 and 2015. Deaths to the end of 2020 were identified through linkage to the Victorian and national death registries. We estimated the effect of age at diagnosis, tumour thickness and tumour site on reducing the melanoma-specific survival gap by sex (ie, interventional indirect effects [IIEs]) on risk difference (RD) scale. Compared to women, there were 211 (95% CI: 145-278) additional deaths per 10 000 in men within 5 years following diagnosis. We estimated that 44% of this gap would be reduced by a hypothetical intervention shifting the distribution of melanoma thickness in men to be the same as that observed for women (IIEthickness RD 93 [95% CI: 75-118] per 10 000) and 20% by an intervention on tumour site (head and neck/trunk vs upper limb/lower limb; IIEsite RD 42 [95% CI: 15-72] per 10 000), while an intervention on age at diagnosis would have a negligible effect. Tumour thickness, tumour site and age at diagnosis mediated 65% of the effect of sex on 5-year melanoma survival in Victoria. Of these factors, tumour thickness had the most considerable mediating effect, suggesting that effective promotion of earlier detection of melanoma in men could potentially nearly halve the gap in melanoma-specific survival by sex.

ACEMID cohort study: protocol of a prospective cohort study using 3D total body photography for melanoma imaging and diagnosis.

INTRODUCTION: Three-dimensional (3D) total body photography may improve early detection of melanoma and facilitate surveillance, leading to better prognosis and lower healthcare costs. The Australian Centre of Excellence in Melanoma Imaging and Diagnosis (ACEMID) cohort study will assess long-term outcomes from delivery of a precision strategy of monitoring skin lesions using skin surface imaging technology embedded into health services across Australia. METHODS AND ANALYSIS: A prospective cohort study will enrol 15 000 participants aged 18 years and above, across 15 Australian sites. Participants will attend study visits according to their melanoma risk category: very high risk, high risk or low/average risk, every 6, 12 and 24 months, respectively, over 3 years. Participants will undergo 3D total body photography and dermoscopy imaging at study visits. A baseline questionnaire will be administered to collect sociodemographic, phenotypic, quality of life and sun behaviour data. A follow-up questionnaire will be administered every 12 months to obtain changes in sun behaviour and quality of life. A saliva sample will be collected at the baseline visit from a subsample. ETHICS AND DISSEMINATION: The ACEMID cohort study was approved by the Metro South Health Human Research Ethics Committee (approval number: HREC/2019/QMS/57206) and the University of Queensland Human Research Ethics Committee (approval number: 2019003077). The findings will be reported through peer-reviewed and lay publications and presentations at conferences. TRIAL REGISTRATION NUMBER: ACTRN12619001706167.

Study protocol for a randomised controlled trial to evaluate the use of melanoma surveillance photography to the Improve early detection of MelanomA in ultra-hiGh and high-risk patiEnts (the IMAGE trial).

INTRODUCTION: Melanoma surveillance photography (MSP) is a comprehensive surveillance method that comprises two- or three-dimensional total body photography with tagged digital dermoscopy, performed at prescribed intervals. It has the potential to reduce unnecessary biopsies and enhance early detection of melanoma, but it is not yet standard care for all high-risk patients in Australia. This protocol describes a randomised controlled trial (RCT) designed to evaluate the clinical impact and cost-effectiveness of using MSP for the surveillance of individuals at ultra-high or high risk of melanoma from a health system perspective. METHODS AND DESIGN: This is a registry-based, unblinded, multi-site, parallel-arm RCT that will be conducted over 3 years. We aim to recruit 580 participants from three Australian states: Victoria, New South Wales and Queensland, via state cancer registries or direct referral from clinicians. Eligible participants within 24 months of a primary cutaneous melanoma diagnosis will be randomised 1:1 to receive either MSP in addition to their routine clinical surveillance (intervention group) or routine clinical surveillance without MSP (control group). Most participants will continue surveillance with their usual care provider, and the frequency of follow-up visits in both groups will depend on the stage of their primary melanoma and risk factors. The primary outcome measure of the study is the number of unnecessary biopsies (i.e. false positives, being cases where a lesion is biopsied due to suspected melanoma on clinical examination, either with or without MSP, but the resulting histopathology finding is negative for melanoma). Secondary outcomes include the evaluation of health economic outcomes, quality of life and patient acceptability. Two sub-studies will explore the benefit of MSP in high-risk patients prior to a melanoma diagnosis and the diagnostic performance of MSP in the teledermatology setting compared to the en face clinical setting. DISCUSSION: This trial will determine the clinical efficacy, cost-effectiveness and affordability of MSP to facilitate policy decision-making at the national and local levels, across primary and specialist care. TRIAL REGISTRATION: ClinicalTrials.gov NCT04385732 . Registered on May 13, 2020.

Trends and patterns of care of sentinel node biopsy in cutaneous melanoma: a population-based study in Queensland.

BACKGROUND: Sentinel node biopsy (SNB) has evolved from offering staging and prognostication to a procedure that guides therapeutic management. The aim was to evaluate the rate of SNB for patients with high-risk melanoma and assess factors that may have impacted on the procedure being performed. METHODS: Data of patients with primary invasive cutaneous melanoma from 01 January 2009 to 31 December 2019 were obtained from the Queensland Oncology Repository. High-risk melanoma was defined as ≥0.8 mm thick or < 0.8 mm with ulceration present (AJCC eighth edition pT1b -pT4 ). RESULTS: 14 006 (33.8%) of 41 412 patients diagnosed with cutaneous invasive melanoma were in the high-risk group. 2923(20.9%) patients had SNB, with the rate increasing from 14.2% (2009) to 36.8% (2019) (P = 0.002), and an increasing proportion being performed in public hospitals over the 11 year period (P = 0.02). Older age (OR0.96 (0.959-0.964) (P < 0.001)), female (OR0.91 (0.830-0.998) (P = 0.03)), head and neck primary (OR0.38 (0.33-0.45) (P < 0.001)), and pT1b (OR0.22 (0.19-0.25) (P < 0.001)) were factors associated with SNB not being performed. Travel out of the Hospital and Health Services of residence for SNB occurred in 26.2%. Although the travel rate decreased from 24.7% (2009) to 23.0% (2019) (P = 0.04), the absolute number increased due to the increase in SNB rate. Those most likely to travel were younger, from remote areas, or from affluent backgrounds. CONCLUSION: In this first Australian population-based study, there was an increased adherence to SNB guideline, although overall SLNB rates remain low, with nearly 2/3 of eligible cases not having the procedure in 2019. Although travel rates decreased marginally, the overall number increased. This study highlights the crucial need to further improve access to SNB for melanoma surgery for the Queensland population.

Cancer Treatment Patterns and Factors Affecting Receipt of Treatment in Older Adults: Results from the ASPREE Cancer Treatment Substudy (ACTS).

INTRODUCTION: Cancer treatment planning in older adults is complex and requires careful balancing of survival, quality of life benefits, and risk of treatment-related morbidity and toxicity. As a result, treatment selection in this cohort tends to differ from that for younger patients. However, there are very few studies describing cancer treatment patterns in older cohorts. METHODS: We used data from the ASPirin in Reducing Events in the Elderly (ASPREE) trial and the ASPREE Cancer Treatment Substudy (ACTS) to describe cancer treatment patterns in older adults. We used a multivariate logistic regression model to identify factors affecting receipt of treatment. RESULTS: Of 1893 eligible Australian and United States (US) participants with incident cancer, 1569 (81%) received some form of cancer treatment. Non-metastatic breast cancers most frequently received treatment (98%), while haematological malignancy received the lowest rates of treatment (60%). Factors associated with not receiving treatment were older age (OR 0.94, 95% CI 0.91-0.96), residence in the US (OR 0.34, 95% CI 0.22-0.54), smoking (OR 0.57, 95% CI 0.40-0.81), and diabetes (OR 0.56, 95% CI 0.39-0.80). After adjustment for treatment patterns in sex-specific cancers, sex did not impact receipt of treatment. CONCLUSIONS: This study is one of the first describing cancer treatment patterns and factors affecting receipt of treatment across common cancer types in older adults. We found that most older adults with cancer received some form of cancer treatment, typically surgery or systemic therapy, although this varied by factors such as cancer type, age, sex, and country of residence.

Evaluation of dynamic dermoscopic features of melanoma and benign naevi by sequential digital dermoscopic imaging and total body photography in a high-risk Australian cohort.

BACKGROUND/OBJECTIVES: Sequential digital dermoscopic imaging (SDDI) and total body photography (TBP) are recommended as a two-step surveillance method for individuals at high-risk of developing cutaneous melanoma. Dermoscopic features specific to melanoma have been well described, however, dynamic changes on serial imaging are less understood. This study aims to identify and compare dermoscopic features in developing melanomas and benign naevi that underwent SDDI and TBP to understand which dermoscopic features may be associated with a malignant change. METHOD: Histopathology reports from a private specialist dermatology clinic from January 2007 to December 2019 were reviewed. Histopathologically confirmed melanoma and benign naevi that underwent SDDI and TBP with a minimum follow-up interval of 3 months were included. RESULTS: Eighty-nine melanomas (38.2% invasive, median Breslow thickness 0.35 mm, range: 0.2-1.45 mm) and 48 benign naevi were evaluated by three experienced dermatologists for dermoscopic changes. Features most strongly associated with melanoma included the development of neovascularisation, asymmetry and growth in pigment network, additional colours, shiny white structures, regression, structureless areas and change to a multi-component pattern. The presence of atypical vessels (p = 0.02) and shiny white structures (p = 0.02) were significantly associated with invasive melanoma. CONCLUSION: Evaluation for certain evolving dermoscopic features in melanocytic lesions monitored by SDDI and TBP is efficient in assisting clinical decision making. SDDI with TBP is an effective tool for early detection of melanoma.

The role of surveillance imaging for resected high-risk melanoma.

BACKGROUND: Recommendations for surveillance imaging for resected melanoma vary considerably. This study examined the utility of imaging in patients with a high-risk primary melanoma undergoing a protocolized imaging schedule. METHODS: This retrospective study involved data collection regarding imaging, recurrence, and outcome characteristics for patients referred to the Victorian Melanoma Service from January 2016-April 2020 and managed for resected stage IIC or III melanoma. Patients with a T4b tumor who did not undergo a sentinel lymph node biopsy were included (T4bNX). Recurrences were "clinically detected" if they were primarily detected by patient symptoms or physical examination, or 'imaging-detected' if the patient was asymptomatic. Cox regression models including time-varying co-variates were used to assess the impact of imaging-detected versus clinically-detected recurrence on overall survival. RESULTS: Over a median follow-up time of 2.7 years, 199 patients underwent surveillance imaging (T4bNX:22, IIC:33, IIIA:22, IIIB:60, IIIC:61, IIID:1), and 44% (n = 88) experienced disease recurrence. Imaging detected over half (53%) of all recurrences. In adjusted analyses, mortality risk was reduced after an imaging-detected compared to clinically-detected recurrence at any given time from the start of surveillance (hazard ratio 0.25, 95% confidence interval 0.10-0.66, p = .005). CONCLUSION: Our study indicates that routine imaging in the early follow-up period of resected T4bNX, stage IIC and III melanoma plays an important role in the detection of asymptomatic recurrences. Imaging-detected recurrence may be associated with a survival benefit and studies with more prolonged follow-up are required to confirm these findings.

Informing a position statement on the use of artificial intelligence in dermatology in Australia.

Artificial Intelligence (AI) is the ability for computers to simulate human intelligence. In dermatology, there is substantial interest in using AI to identify skin lesions from images. Due to increasing research and interest in the use of AI, the Australasian College of Dermatologists has developed a position statement to inform its members of appropriate use of AI. This article presents the ACD Position Statement on the use of AI in dermatology, and provides explanatory information that was used to inform the development of this statement.

Immune checkpoint inhibitors and the risk of major atherosclerotic cardiovascular events in patients with high-risk or advanced melanoma: a retrospective cohort study.

BACKGROUND: Immune checkpoint inhibitors (ICI) are associated with immune-mediated adverse effects, potentially involving any organ. ICI has also been associated with an increased risk of cardiovascular disease in cancer populations. OBJECTIVE: To characterize the incidence and risk of major atherosclerotic cardiovascular events associated with ICI use in a high-risk and advanced melanoma population. METHODS: We conducted a retrospective cohort study of patients with high-risk or advanced melanoma (AJCC stage II, III or IV) presenting to an academic tertiary hospital between 2015-2020. The main outcome was major atherosclerotic cardiovascular events (MACE) including acute myocardial infarction, ischemic stroke, acute limb ischemia and coronary revascularization. RESULTS: The study cohort consisted of 646 patients, including 289 who had been treated with ICI. The incidence of MACE was higher in the ICI treated group (3.6 vs. 0.9 events per 100-person years). After adjusting for age, sex, smoking history and prior BRAF and/or MEK inhibitor use, ICI treatment was associated with an increased risk of MACE (HRadj 2.8, 95% CI 1.1-6.9, p = 0.03). Elevated risk was especially pronounced in patients with a past history of MACE (HR 14.4, 95% CI 1.9-112.3, p = 0.01). CONCLUSION: Patients with high-risk or advanced melanoma are at an increased risk of atherosclerotic cardiovascular events following ICI treatment, particularly those with a history of cardiovascular disease.

Association Between Low-Dose Methotrexate Exposure and Melanoma: A Systematic Review and Meta-analysis.

Importance: Methotrexate is widely used for the treatment of inflammatory disorders, including rheumatoid arthritis. Studies suggest that methotrexate may be associated with an increased risk of melanoma. Objective: To determine whether methotrexate exposure is associated with an increased risk of cutaneous melanoma. Data Sources: MEDLINE, Embase, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov were searched from inception to May 12, 2022, for eligible studies. Study Selection: Case-control studies, cohort studies, or randomized clinical trials (RCTs) were included if they examined the odds or risk of cutaneous melanoma in individuals exposed to low-dose methotrexate in comparison with individuals unexposed. No language limitations were applied. Data Extraction and Synthesis: Two independent reviewers extracted data on study characteristics and outcome data. The Meta-analysis of Observational Studies in Epidemiology guidelines were followed. To assess study quality, the Cochrane risk of bias tool was used for RCTs, and the Joanna Briggs Institute Checklist was used for cohort and case-control studies. Odds ratio from case-control studies and relative risk or hazard ratio from cohort studies or RCTs were pooled, and a random-effects model meta-analysis was conducted. Main Outcomes and Measures: Prespecified outcome was the odds ratio, hazard ratio, or risk ratio of cutaneous melanoma comparing low-dose methotrexate exposure with nonexposure. Results: Seventeen studies (8 RCTs, 5 cohort studies, 4 case-control studies) were eligible for inclusion, and of these, 12 studies with 16 642 cases of melanoma were pooled in the primary analysis. Indications for methotrexate included rheumatoid arthritis, psoriasis, psoriatic arthritis, and inflammatory bowel disease and were unknown in 5 studies. Compared with unexposed individuals, study participants with methotrexate exposure had a small increased risk of melanoma (pooled relative risk, 1.15; 95% CI, 1.08-1.22), but this did not persist in a sensitivity analysis excluding the largest study (pooled relative risk, 1.11; 95% CI, 1.00-1.24). Subgroup analyses according to comparator group (comparing methotrexate exposure with either immunomodulator alone vs immunomodulator and methotrexate) or the indication for methotrexate being rheumatoid arthritis provided similar risk estimates. Using geographical population melanoma incidence rates, a number needed to harm of 18 630 was calculated in Australia, and 41 425 in North America. Conclusions and Relevance: In this systematic review and meta-analysis, low-dose methotrexate exposure was associated with an increased melanoma risk, but the absolute risk increase could be considered negligible.

Association between hypertension and cutaneous melanoma, and the effect of aspirin: extended follow-up of a large randomised controlled trial.

BACKGROUND: The association between hypertension and melanoma is unclear, and previous analyses of data from the ASPirin in Reducing Events in the Elderly (ASPREE) study demonstrated a reduced number of invasive melanoma events amongst aspirin-exposed hypertensive individuals. METHODS: Data from the ASPREE study which included (1) the intervention period with a median follow-up of 4.7 years, and (2) the observational period with an additional 2 years follow-up, were combined for this analysis. Logistic regression analyses examined the association between baseline hypertension and treatment status and past melanoma history. Survival analyses examined the association between hypertension and melanoma risk, and the effect of aspirin across hypertension groups. Cox proportional hazards models were used to compare incidence across groups. RESULTS: 19,114 participants (median age of 74 years) were randomised to daily 100 mg aspirin or placebo. At baseline, hypertension and past melanoma history were recorded in 14,195 and 685 individuals, respectively. After adjustment for confounders, hypertension was significantly associated with past melanoma history (OR=1.34, 95%CI: 1.11-1.62). In a prospective analysis, baseline hypertension was not associated with melanoma risk. However, aspirin was associated with a reduced risk of incident melanoma amongst individuals with uncontrolled hypertension (blood pressure ≥140/90 mmHg; HR=0.63, 95%CI 0.44-0.89), but not in those with controlled hypertension (HR=1.04, 95%CI 0.74-1.46). CONCLUSION: Our results support a reduced melanoma incidence amongst individuals with uncontrolled hypertension exposed to aspirin. Additional studies are required to confirm these findings.

Sensitivity of two Australian melanoma risk tools to identify high-risk individuals among people presenting with their first primary melanoma.

AIMS: Regular skin examinations for early detection of melanoma are recommended for high-risk individuals, but there is minimal consensus regarding what constitutes 'high-risk'. Melanoma risk prediction models may guide this. We compared two online melanoma risk prediction tools: Victorian Melanoma Service (VMS) and Melanoma Institute Australia (MIA) risk tools; to assess classification differences of patients at high-risk of a first primary melanoma. METHODS: Risk factor data for 357 patients presenting with their first primary melanoma were entered into both risk tools. Predicted risks were recorded: 5-year absolute risk (VMS tool and MIA tool); 10-year, lifetime, and relative risk estimates (MIA tool). Sensitivities for each tool were calculated using the same high-risk thresholds. The MIA risk tool showed greater sensitivity on comparison of 5-year absolute risks (90% MIA vs 78% VMS). Patients had significantly higher odds of being classified as high or very-high risk using the MIA risk tool overall, and for each patient subgroup. Using either tool, patients of male gender or with synchronous multiple first primary melanomas were more likely to be correctly classified as high- or very-high risk using 5-year absolute risk thresholds; but tumour invasiveness was unrelated to risk. Classification differed when using the MIA risk categories based on relative risk. CONCLUSIONS: Both melanoma risk prediction tools had high sensitivity for identifying individuals at high-risk and could be used for optimising prevention campaigns. The choice of which risk tool, measure, and threshold for risk stratification depends on the intended purpose of risk prediction, and ideally requires information on specificity.