Artificial intelligence and machine learning algorithms for early detection of skin cancer in community and primary care settings: a systematic review.

Skin cancers occur commonly worldwide. The prognosis and disease burden are highly dependent on the cancer type and disease stage at diagnosis. We systematically reviewed studies on artificial intelligence and machine learning (AI/ML) algorithms that aim to facilitate the early diagnosis of skin cancers, focusing on their application in primary and community care settings. We searched MEDLINE, Embase, Scopus, and Web of Science (from Jan 1, 2000, to Aug 9, 2021) for all studies providing evidence on applying AI/ML algorithms to the early diagnosis of skin cancer, including all study designs and languages. The primary outcome was diagnostic accuracy of the algorithms for skin cancers. The secondary outcomes included an overview of AI/ML methods, evaluation approaches, cost-effectiveness, and acceptability to patients and clinicians. We identified 14 224 studies. Only two studies used data from clinical settings with a low prevalence of skin cancers. We reported data from all 272 studies that could be relevant in primary care. The primary outcomes showed reasonable mean diagnostic accuracy for melanoma (89·5% [range 59·7-100%]), squamous cell carcinoma (85·3% [71·0-97·8%]), and basal cell carcinoma (87·6% [70·0-99·7%]). The secondary outcomes showed a heterogeneity of AI/ML methods and study designs, with high amounts of incomplete reporting (eg, patient demographics and methods of data collection). Few studies used data on populations with a low prevalence of skin cancers to train and test their algorithms; therefore, the widespread adoption into community and primary care practice cannot currently be recommended until efficacy in these populations is shown. We did not identify any health economic, patient, or clinician acceptability data for any of the included studies. We propose a methodological checklist for use in the development of new AI/ML algorithms to detect skin cancer, to facilitate their design, evaluation, and implementation.

Interventions for basal cell carcinoma: abridged Cochrane systematic review and GRADE assessments.

BACKGROUND: Basal cell carcinoma (BCC) is the most common cancer affecting white-skinned individuals, and the worldwide incidence is increasing. Although rarely fatal, BCC is associated with significant morbidity and costs. OBJECTIVES: To assess the effects of interventions for primary BCC in immunocompetent adults. METHODS: We updated our searches of the following databases to November 2019: Cochrane Skin Group Specialised Register, CENTRAL, MEDLINE, Embase, CINAHL and LILACS. Certainty of evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation method. We used standard methodological procedures expected by Cochrane. RESULTS: We included 52 randomized controlled trials with 6990 participants (median age 65 years; range 20-95). Mean study duration was 13 months (range 6 weeks-10 years). Ninety-two per cent (n = 48/52) of studies exclusively included histologically low-risk BCC (nodular and superficial subtypes). The certainty of evidence was predominantly low or moderate for the outcomes of interest. Overall, surgical interventions have the lowest recurrence rates, and there may be slightly fewer recurrences with Mohs micrographic surgery over surgical excision for primary, facial BCC (high-risk histological subtype or located in the 'H-zone' or both) (low-certainty evidence). Nonsurgical treatments, when used for low-risk BCC, are less effective than surgical treatments, but recurrence rates are acceptable and cosmetic outcomes are probably superior. CONCLUSIONS: Surgical interventions have lower recurrence rates and remain the gold standard for high-risk BCC. Of the nonsurgical treatments, topical imiquimod has the best evidence to support its efficacy for low-risk BCC. Priorities for future research include agreement on core outcome measures and studies with longer follow-up.

'Efficacy and safety of imiquimod 5% cream for basal cell carcinoma: a meta-analysis of randomized controlled trial': a critical appraisal.

AIM: Jia and He aimed 'to compare the efficacy and safety of imiquimod with other treatments in patients with basal cell carcinoma' (BCC). DESIGN AND INCLUSION CRITERIA: Meta-analysis of studies that included patients with histologically confirmed BCC treated with imiquimod 5% cream compared with all other treatments, including vehicle, excisional surgery, cryosurgery, fluorouracil and methyl aminolaevulinate photodynamic therapy. OUTCOMES: The main outcome measures included histological and composite clearance rates, success rates, complete response rates, tumour-free survival and adverse events. RESULTS: Thirteen studies with a total of 4265 patients were included in the review. Pooled analyses comparing imiquimod with all or any of the listed comparators, including vehicle, demonstrated higher histological clearance rates [risk ratio (RR) 9·28, 95% confidence interval (CI) 5·56-15·5; P < 0·001], higher composite clearance rates (RR 34·2, 95% CI 21·3, 55·1; P = 0·001), no significant difference in success rates (RR 0·98, 95% CI 0·89-1·08; P = 0·73), higher complete response rates (RR 3·15, 95% CI 1·55-6·38; P = 0·001), no significant difference in tumour-free survival (RR 1·15, 95% CI 0·98-1·35; P = 0·088) and increased incidence of adverse events (RR 2·00, 95% CI 1·39-2·88; P < 0·001). CONCLUSIONS: The authors state that 'imiquimod significantly exhibited benefit effect in improving the histological/composite clearance rates' compared with other treatments, and they suggest it could be used as the first-choice treatment for patients with BCC. COMMENT: The main concerns related to the article by Jia and He are that the research question is replicative, it makes little sense to combine all BCC types in a meta-analysis, and it also makes no sense to combine an active treatment against a combination of vehicle and other active treatments. There are also concerns about bias related to the use of the same study data more than once in a meta-analysis. Furthermore, we have identified an example of covert duplicate publication, which further compounds the profusion of misleading systematic reviews.

Clinical characteristics, symptoms and burden of psoriasis and atopic dermatitis in adults.

BACKGROUND: There is debate as to whether psoriasis and atopic dermatitis (AD) belong to the same disease spectrum. OBJECTIVES: To describe and compare disease characteristics, lifestyle factors and disease burden in adult patients with psoriasis and AD. METHODS: We linked registry data with clinical and patient-reported outcomes from the Danish Skin Cohort, containing 3348 and 3834 adults with dermatologist-verified psoriasis or AD respectively, and 2946 adults from the general population. RESULTS: The participants were predominantly women and middle-aged. Patients with psoriasis mostly reported disease onset throughout adulthood, but with a distinct early incidence peak in those with a positive family history or severe disease. AD predominantly began in childhood, with only a very discrete incidence peak in adulthood. Scalp, extremity, chest and abdomen involvement was common to both diseases. Scalp/hairline, elbows, nails, intergluteal cleft, umbilicus, knees and legs were most frequently affected in patients with psoriasis. In AD, periocular, neck, antecubital fossae, back of the hands, interdigital areas and popliteal fossae were commonly affected. Patients with psoriasis (but not AD) were generally more overweight, obese and physically inactive, and had a positive smoking history, compared with the general population. Patients with both diseases experienced more frequent flares with increasing disease severity. Patients generally displayed uncontrolled disease despite being on systemic therapies. Itch and skin pain were much more severe in patients with AD, whereas joint pain was more common in patients with psoriasis. CONCLUSIONS: We identified important similarities and differences in the clinical characteristics of adults with psoriasis and AD; these should help clinicians to prioritize and improve patient management. What's already known about this topic? Psoriasis and atopic dermatitis in adults are increasingly being compared, and there is discussion as to whether they are part of the same disease spectrum. What does this study add? In this comparative study, patient-reported disease burden was markedly higher in atopic dermatitis than in psoriasis, whereas lifestyle-associated cardiometabolic risk factors were more frequent in psoriasis. In both disease groups, the condition in the majority of patients was uncontrolled even while they were on systemic therapy. The contrasting presentations highlight that these diseases are two distinct and different entities rather than belonging to the same spectrum.

What's new in psoriasis treatment? An analysis of systematic reviews published in 2015.

This review provides a summary of key findings from 27 systematic reviews of 51 articles first published or indexed during 2015, focusing on the treatment of psoriasis and on precision medicine in psoriasis. The evidence supports weight-loss interventions by dieting and exercise for improvement in disease severity in overweight and obese patients with psoriasis. No significant increased risk of serious infections was reported for the biologic therapies adalimumab, etanercept and ustekinumab compared with appropriate comparators. Evidence could not provide reliable estimates of rare adverse events, emphasizing the need for large prospective registries. Polymorphisms in the tumour necrosis factor (TNF)-α gene may confer improved responses to TNF inhibitor (TNFI) therapy, but the studies to date lack power to detect a true association. From the limited available evidence, multidisciplinary management is both more effective and more satisfactory for patients with psoriasis and psoriatic arthritis than conventional consultations. This summary of reviews provides a succinct guide for clinicians and patients wishing to remain up to date with high-quality evidence for the treatment of psoriasis.

What is the evidence base for atopic eczema treatments? A summary of published randomized controlled trials.

Atopic eczema (AE) is a common chronic inflammatory skin condition. While many AE treatment options are available, the evidence to support their efficacy varies in depth and quality. In 2000, a National Institute for Health Research (NIHR) Health Technology Assessment systematic review identified and evaluated existing randomized controlled trials (RCTs) of AE treatments. To ensure continuing utility, the NIHR commissioned an update to the review. Here, we present an overview of the updated report and its key findings. Systematic reviews and RCTs of AE treatments that included participants with AE (criteria based or diagnosed) were identified using Medline, Embase, CENTRAL, Latin American and Caribbean Health Sciences, Allied and Complementary Medicine Database, Cumulative Index to Nursing and Allied Health Literature and Cochrane Skin Group Specialised Register [searched to 31 August 2013 (RCTs) and 31 December 2015 (systematic reviews)]. Outcome measures included symptoms, AE severity, quality of life and adverse effects. Study quality was assessed using the Cochrane Collaboration risk of bias tool. Of the 287 new RCTs identified, only 22 (8%) were judged to have a low risk of bias. When combined with RCTs from the previous review (n = 254), we found 'reasonable evidence of benefit' for corticosteroids, calcineurin inhibitors, Atopiclair® , ciclosporin, azathioprine, ultraviolet radiation and education programmes. Interventions with reasonable evidence of 'no benefit' included some dietary interventions, ion exchange water softeners, multiple daily applications of topical corticosteroids and antibiotic-containing corticosteroids for noninfected AE. Many common treatments lack evidence of efficacy and warrant further evaluation. The evidence base for AE is still hampered by poor trial design and reporting. The trials included in this review were used to establish the Global Resource of EczemA Trials (GREAT) database.

Eczema and indoor environment: lessons from the International Study of Asthma and Allergies in Childhood (ISAAC) Phase 2.

BACKGROUND: Numerous studies have reported a positive association between damp housing conditions and asthma, but little is known about indoor environmental exposures in relation to childhood eczema. We aimed to specifically investigate the effect of indoor mould and dampness on eczema risk in the International Study of Asthma and Allergies in Childhood (ISAAC). METHODS: ISAAC Phase 2 is a cross-sectional study of 46 051 children aged 8-12 years from 20 countries. Information on demographics, eczema symptoms, and dampness was gathered with parental questionnaires. Children were examined for eczema and underwent skin prick testing. In a stratified subgroup, dust samples were collected to measure house dust mite exposure. Sex, maternal education, parental allergy, pet ownership, maternal smoking, having an older sibling, bedroom sharing, and cooking with fuels were explored as potential confounders or effect modifiers in logistic regression analysis. FINDINGS: Current residential exposure to dampness and mould was significantly associated with flexural eczema in the previous year, with a stronger association seen in non-affluent than in affluent countries (adjusted odds ratio [OR] 1·96, 95% CI 1·62-2·37, vs 1·34, 1·18-1·51). Dampness and mould in the first year of life was also significantly associated with parent-reported eczema ever (1·94, 1·40-2·68, vs 1·43, 1·28-1·60). However, the association with flexural eczema on examination was not significant (0·93, 0·76-1·13). Risk estimates were similar in children positive and negative on skin prick testing, and were not appreciably altered by the effect modifiers, apart from parental allergic disease (parental allergies OR 1·35, 95% CI 1·18-1·54, vs no parental allergies 1·61, 1·37-1·90). INTERPRETATION: These data suggest an association between damp housing conditions and childhood eczema symptoms, which may be causal. Further work is needed to elucidate possible mechanisms. Modification to home environment to reduce dampness and mould could be harnessed to improve or even prevent this common and debilitating condition. FUNDING: None.

What's new in acne? An analysis of systematic reviews published in 2011-2012.

This review summarizes important clinical developments in acne vulgaris identified from 17 systematic reviews published between February 2011 and August 2012. Regarding causes, Demodex mites have been shown to be associated with both acne vulgaris and rosacea, although it is unclear if their eradication improves either disease. Some weak evidence has emerged that suggests a possible link between dairy produce and acne, which warrants further research. With reference to the effects of acne, there is good evidence that acne negatively affects quality of life, self-esteem and mood in adolescents. Acne is also associated with an increased risk of anxiety, depression and suicidal ideation, highlighting the importance of asking patients with acne directly about psychological issues in order to identify those who might benefit from early psychiatric support. Regarding treatment, there seems to be no additional benefit to using higher strengths of benzoyl peroxide, and lower strengths such as 2.5% have fewer side effects. Despite earlier concerns of increased mortality in those using topical tretinoin for skin cancer prevention, a systematic review on this topic has not found any convincing evidence of a link between such non-cutaneous events and once-daily application of 0.02-0.05% tretinoin. Combined oral contraceptives are of benefit in both inflammatory and non-inflammatory acne. Current surveys suggest that implementation of the pregnancy prevention programme for isotretinoin may not be stringent, and a high level of monitoring and audit is recommended. Ablative and non-ablative laser resurfacing for the treatment of acne scars may be beneficial, but further studies with a longer follow-up period are required.

Epidemiology of acne vulgaris.

Despite acne being an almost universal condition in younger people, relatively little is known about its epidemiology. We sought to review what is known about the distribution and causes of acne by conducting a systematic review of relevant epidemiological studies. We searched Medline and Embase to the end of November 2011. The role of Propionibacterium acnes in pathogenesis is unclear: antibiotics have a direct antimicrobial as well as an anti-inflammatory effect. Moderate-to-severe acne affects around 20% of young people and severity correlates with pubertal maturity. Acne may be presenting at a younger age because of earlier puberty. It is unclear if ethnicity is truly associated with acne. Black individuals are more prone to postinflammatory hyperpigmentation and specific subtypes such as 'pomade acne'. Acne persists into the 20s and 30s in around 64% and 43% of individuals, respectively. The heritability of acne is almost 80% in first-degree relatives. Acne occurs earlier and is more severe in those with a positive family history. Suicidal ideation is more common in those with severe compared with mild acne. In the U.S.A., the cost of acne is over 3 billion dollars per year in terms of treatment and loss of productivity. A systematic review in 2005 found no clear evidence of dietary components increasing acne risk. One small randomized controlled trial showed that low glycaemic index (GI) diets can lower acne severity. A possible association between dairy food intake and acne requires closer scrutiny. Natural sunlight or poor hygiene are not associated. The association between smoking and acne is probably due to confounding. Validated core outcomes in future studies will help in combining future evidence.

What's new in atopic eczema? An analysis of systematic reviews published in 2009-2010.

This review provides a summary of key findings from 18 systematic reviews on atopic eczema, published or indexed between January 2009 and 24 August 2010. There was no good evidence on the possible benefit of organic food consumption and eczema. Maternal intake of fish or fish oil may be associated with a reduced risk of eczema in offspring, although further studies are needed. There is some evidence that partially hydrolysed infant formulas rather than standard formulas may be associated with a reduced risk of eczema in infants, but there are shortcomings in the existing evidence. An inverse relationship has been found between gliomas/acute lymphoblastic leukaemia and allergic disease/eczema, but there appears to be no association between multiple sclerosis and eczema. Attention deficit hyperactivity disorder does appear to be associated with eczema, but there is no evidence of a causal link. The risk of eczema seems to be increased in urban compared with rural areas. Some new evidence has suggested superiority of 1% pimecrolimus over potent and mild corticosteroids at 6 months but not 12 months, and there is some evidence for superiority of 0.03% and 0.1% tacrolimus over 1% pimecrolimus. An updated Cochrane Review still found no evidence of a benefit from any form of antistaphylococcal treatment in managing clinically infected or uninfected eczema. The evidence base is poor for bath emollients, occlusive treatments (e.g., wet and dry wraps) and woven silk clothing in treating eczema. In general, the methods used in most systematic reviews of eczema need to be reported more clearly, especially with regard to a more vigorous quality assessment of included studies. Included studies are frequently heterogeneous, proxy reporting is common, and appropriate disease definitions are often lacking. Better adherence to existing guidance on trial reporting and prospective registration of clinical trials may help improve the quality of studies.

What's new in skin cancer? An analysis of guidelines and systematic reviews published in 2008-2009.

This review summarizes clinically important findings from 17 systematic reviews and 2 guidelines on skin cancer indexed between April 2008 and April 2009. Melanoma primary-prevention measures, such as education, are more likely to be successful in younger children than adolescents, and general population screening for melanoma by whole-body examination is not currently supported by the evidence. A large systematic review of melanoma and pregnancy concluded that pregnancy does not affect prognosis. Two systematic reviews imply that sunburn later in life also increases the risk of melanoma, and that it is just as important as sunburn early in life. Three systematic reviews discussed the role of positron emission tomography and sentinel lymph-node biopsy for melanoma staging, but produced conflicting results. Superior diagnostic accuracy of dermatoscopy over naked-eye examination for melanoma was found in one review, while a second implied nonsignificantly higher sensitivity of computer-based diagnostic methods over dermatoscopy for melanoma but with reduced specificity. There were no identified randomized controlled trials of treatments for unresectable recurrent melanoma, and a review of immunotherapy with vaccines for melanoma failed to prove improved overall and disease-free survival. Guidelines for the management of basal cell carcinoma call for risk stratification, based on numerous factors including tumour size, site and histological subtype. Squamous cell carcinoma of the ear has been shown to spread to regional lymph nodes more commonly than to other sites, and may be predicted by depth of invasion, tumour size, cellular differentiation and completeness of excision.

What's new in atopic eczema? An analysis of systematic reviews published in 2007 and 2008. Part 1. Definitions, causes and consequences of eczema.

This review summarizes clinically important findings from nine systematic reviews indexed in bibliographical databases between August 2007 and August 2008, dealing with the definitions, causes and consequences of atopic eczema (AE). One review of diagnostic criteria found that out of 10 sets of criteria, only the UK refinement of the Hanifin and Rajka criteria have been adequately tested (in 19 studies). Another review of 20 named outcome measures found that only three [SCORing Atopic Dermatitis (SCORAD), the Eczema Area and Severity Index (EASI) and the Patient Oriented Eczema Measure (POEM)] had been tested and found to perform adequately. In terms of risk factors for developing disease, a review found that birth by caesarean section increased the risk of asthma and hay fever but not eczema in offspring. A review of cohort studies also found evidence that adverse psychological factors in early life predispose to more atopic disease and a worse prognosis. Another review found that filaggrin gene mutations were a consistently strong risk factor for AE, with a person carrying one of these mutations being over three times more likely to exhibit eczema. It has been suggested that eczema might protect against some forms of cancer, and a detailed systematic review of brain cancers that included 53,233 participants from eight case-control and cohort studies found that having atopic disease was associated with a 39% reduction in glioma risk, a finding that was also present for just those with AE (odds ratio 0.69, 95% CI 0.58-0.82). A further review of case-control and cohort studies failed to find any association between keeping furry pets at birth and subsequent risk of eczema, although pet fur might still exacerbate established disease. In terms of disease consequences, a review found that eczema was the commonest cause of chronic sleep loss in young people, affected the whole family. A review of four economic studies from the US found that the annual cost of AE in the States was as high as $3.8 billion when indirect costs are included.

Management of acne vulgaris: an evidence-based update.

This review summarizes clinically important findings from 3 systematic reviews, 1 updated guideline and a selection from the 62 randomized controlled trials (RCTs) published between February 2007 and January 2009 on the topic of acne vulgaris. Low glycaemic-load diets might reduce acne severity but this remains unproven. Written patient information leaflets have not been surpassed by other communication methods. New combination topical treatments have not shown convincing advantages over current combination products such as clindamycin/benzoyl peroxide. Topical dapsone is superior to placebo but has yet to be compared with standard topical treatments. Long-term topical tretinoin to prevent nonmelanoma skin cancer in elderly men was associated with higher all-cause mortality, but there is currently no evidence of increased mortality for topical retinoid use when treating acne. All oral tetracyclines have similar efficacy, yet minocycline is the most costly. Oral isotretinoin monotherapy remains the gold-standard treatment for severe acne. Flutamide plus the oral contraceptive pill is beneficial for acne associated with polycystic ovary syndrome. Photodynamic therapy, phototherapy and laser therapy cannot be recommended universally for acne until minimal postinflammatory pigmentation and longer-term benefit can be shown, especially with current high costs. Development of non-antibiotic therapies is preferable to minimize the risk of community antibiotic resistance. Future trials should use active comparators at optimum doses and avoid noninferiority comparisons unless appropriately powered. Trials need to shift from using multiple, unvalidated outcome measures to including patient-reported and quality-of-life outcomes, and all trials should be registered on a public clinical-trials database.

What's new in atopic eczema? An analysis of the clinical significance of systematic reviews on atopic eczema published in 2006 and 2007.

This review summarizes clinically important findings from 19 systematic reviews published between January 2006 and August 2007 on the topic of atopic eczema (AE). The evidence suggests that avoidance of allergenic foods during pregnancy or the use of hydrolyzed or soy formula milks does not prevent eczema. Delayed introduction of solids may decrease eczema risk. Asthma typically develops in around a third of children with eczema, and wheezing in early infancy is a predictor of risk. Established topical corticosteroids such as betamethasone should be used just once daily. Topical tacrolimus and pimecrolimus can be used for people who become dependent on topical corticosteroids, especially on sensitive sites such as the face. Wet wraps are useful in secondary care for inducing remission in a child, but they are not a treatment for mild eczema and they should not be used long term. Oral ciclosporin can be used for inducing a remission in severe eczema, and azathioprine can be considered for maintenance treatment. Narrowband ultraviolet (UV)B phototherapy can be used for chronic AE, and UVA1 may be useful for acute eczema. There is little convincing evidence of a clinical benefit with evening primrose oil for eczema, but there is some good new evidence that educational support to eczema families is beneficial. Future trials need to be larger, and include active comparators, patient-reported outcomes and longer-term aspects of disease control. They should be better reported, and registered on a public clinical trials register.