Study author Outcomes reported	Inclusion criteria Numberof lesions; cases Algorithm used In‐person/image‐based	Clinicians recruited for training	Pre‐training	Training approach	Post‐training
Detection of Invasive melanoma or atypical intraepidermal melanocytic variants
Pagnanelli 2003 Pathway – unclear	Clinical and dermoscopic images of PSL from the training set of the Consensus Net Meeting on Dermoscopy (CNMD), selected by 2 experts N = 20; MEL 6 Dermoscopy (pattern analysis; Menzies; 7PCL; ABCD) Image‐based (clinical image)	Recruited 16 ‘colleagues’, including medical Students (n = 3), dermatology residents (n = 9) and dermatologists (n = 4) All reported limited personal experience of dermoscopy, no formal training and did not use dermoscopy in daily professional practice	After the 1‐hour lecture at the beginning of the study, lesion images were provided on CD‐Rom; participants asked to complete electronic data sheet listing criteria for diagnosing PSLs by pattern analysis and by the various algorithms and to offer a dermoscopic diagnosis for each case within 20 days	1‐hour lecture on basic principles of dermoscopy, the dermoscopic features of PSLs, pattern analysis and the diagnostic algorithms (ABCD rule, seven‐point checklist, Menzies’ method). + a web‐based tutorial (http://www.dermoscopy.org); participants requested to devote 1 hour per day, 5 days per week for 2 consecutive weeks.	Post‐training evaluation 5 weeks after initial evaluation Participants re‐evaluated the same 20 cases, again over a 20‐day period.
Piccolo 2014 Pathway – unclear	Dermoscopically atypical PSL N = 165; MM 23; MiS 10 Dermoscopy (ABCD) Image‐based (blinded) (Also evaluates CAD dermoscopy)	3 dermatologists and 1 GP scored according to number of years specializing in dermoscopy, number of PSLs assessed by dermoscopy on a daily basis, number of relevant workshops/seminars attended, and number of authored publications on dermoscopy: highly experienced (observer 1), moderately experienced, (observers 2 and 3); and minimally experienced (observer 4).	Digital dermoscopic images assessed by each observer using ABCD at T0	Between T0 and T1, Observer 4 underwent dermoscopic training by studying an interactive atlas of dermoscopy (Argenziano 2003; appears to be same as for Pagnanelli 2003)	The same digital dermoscopic images were assessed by each of the 4 observers using ABCD after 6 months (T1)
Stanganelli 1999 Pathway – unclear	PSL images selected from database for training study N = 30; MM 10; MiS 1 Dermoscopy (no algorithm) Image‐based (clinical image)	Of 223 dermatologists who participated in one of the six workshops, 83 (37%) were reported on; average of 10 years of general experience in dermatology (range 1‐22) with routine use of ELM by 52 individuals (conventional dermatoscope for 43 and digital equipment for 9)	Pre‐training test conducted after the opening lecture of each workshop (clinical classification of PSLs). Images projected onto a screen in pairs (clinical and ELM image); classified by as CMM, MN, NML, unclassifiable or equivocal; approximately 2.5 min per lesion	Nationwide educational programme in ELM; one‐day meetings and workshops (duration: 6 hours) held with free registration. Topics included: clinical classification and diagnosis of PSLs management of patients with PSLs; basic principles of ELM; ELM criteria ELM diagnosis; limitations of ELM	Same set of slides re‐evaluated at the end of the workshop. Slides and respective correct diagnosis were discussed only after the second test.
Tan 2009 Pathway – unclear	Test series of images of melanomas and benign lesions N = 30; MEL 15 Dermoscopy (pattern analysis modified) Image‐based (clinical image)	3 consultant dermatologists and 3 specialist registrars; none had routinely used a dermatoscope	Assessed 30 test cards consisting of 1 macroscopic and 1 dermatoscopic image of each lesion; printed on A4 laminated paper Participants classified images as ‘benign’, ‘malignant’ or ‘not known’, gave a diagnosis if known, and indicated whether they would excise the lesion.	Participants received an online tutorial (www.dermatoscopy.org) teaching the MPADA (Modified Pattern Analysis Diagnostic Algorithm), which could be referred to during the study period. Also each given a dermatoscope to use in clinical practice for 10 months.	10 months later, the test‐card questionnaire was repeated (test 2)
Detection of invasive melanoma alone
Troyanova 2003	Patients with atypical melanocytic lesions or suspected early malignant melanoma N = 50 1. VI (no algorithm) 2. Dermoscopy (no algorithm) Image‐based (clinical image)	Volunteer dermatologists (n = 32); experienced in clinical diagnosis of PSLs, but had no formal training in dermoscopy. ELM qualification based on good theoretical knowledge of the literature and on personal experience by trial and error.	50 clinical images displayed individually using slide projector; scored as melanoma or "not‐melanoma". 50 dermoscopy slides then presented and ELM diagnoses recorded. Each image shown for 30 seconds; no discussion of assumed diagnosis was permitted. None of the test slides used for training	6 hours of teaching daily for 2 consecutive days. Training was based on presentation of several hundred slides with oral explanation of the ELM criteria.	Tests were performed in the beginning and in the end of the teaching course Same test performed with slides of 50 different PSLs
Westerhoff 2000	Images of PSL selected for a dermoscopy training study N = 50; 50 MM 1. VI (no algorithm) 2. Dermoscopy (no algorithm; Menzies criteria) Image‐based (blinded)	GPs (n = 74) recruited by telephone from a list of current practitioners. Required to have no formal training in dermoscopy and did not use dermoscopy in their clinical practice. Participants randomised into an education intervention group or non‐education intervention group (each n = 37)	Lesions presented with the clinical photograph first, followed by dermoscopy image Participants given 4 options: melanoma; benign melanocytic lesion, benign non‐melanocytic lesion, 'other' (specify) Clinical diagnosis recorded prior to observation of dermoscopic image. Tests completed at participants' leisure	Supplied with pictorial atlas by Menzies 1996b and a 1‐hour presentation on dermoscopy that specifically reviewed the Menzies method and included a quiz with images of 25 different PSL (not used in test)	As for pre‐test
ABCD: asymmetry, border, colour, dimensions;AK: actinic keratosis; alg: algorithm; BD: Bowen’s disease; BCC: basal cell carcinoma; BN: benign naevi; BPC: between person comparison (of tests);CAD: computer‐assisted diagnosis; CCS: case‐control study; CMM: cutaneous malignant melanoma; CS: case series; cSCC: cutaneous squamous cell carcinoma; DF: dermatofibroma; ELM: epiluminescence microscopy; FU: follow‐up; LS: lentigo simplex; MEL: melanoma; MiS: melanoma in situ (or lentigo maligna); MM: malignant melanoma; NC: non‐comparative; NML: non melanocytic lesion; NR: not reported; Obs: observer; P: prospective; PLC: pigmented lesion clinic; PSL: pigmented skin lesion; R: retrospective; RCM: reflectance confocal microscopy; SK: seborrhoeic keratosis; SN: Spitz naevi; WPC: within person comparison (of tests)