Premium
Efficacy of smartphone applications in high‐risk pigmented lesions
Author(s) -
Ngoo Alexander,
Finnane Anna,
McMeniman Erin,
Tan JeanMarie,
Janda Monika,
Soyer H Peter
Publication year - 2018
Publication title -
australasian journal of dermatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.67
H-Index - 53
eISSN - 1440-0960
pISSN - 0004-8380
DOI - 10.1111/ajd.12599
Subject(s) - medicine , melanoma , dermatology , smartphone app , smartphone application , dermatoscopy , lesion , melanoma diagnosis , surgery , multimedia , internet privacy , cancer research , computer science
Background/Objectives Melanoma apps are smartphone applications that assess risk of pigmented lesions using a smartphone camera and underlying algorithm. We aimed to assess the capability of melanoma smartphone applications (apps) in making clinical decisions about risk, compared with lesion assessment by specialist trained dermatologists. Methods A prospective study of 3 melanoma apps was conducted between 2015 and 2016, recruiting 30 patients with 57 pigmented lesions. Risk categories assigned by the apps were compared with the clinical decisions of two consultant dermatologists classifying lesions as ‘suspicious’ or ‘benign’. Results Of the 42 lesions deemed clinically suspicious to a dermatologist, from 9 to 26 were classified as suspicious by the apps; of the 15 clinically benign lesions 3 to 15 were correctly classified as benign by the apps. The apps’ sensitivity and specificity ranged from 21 to 72% and 27 to 100.0%, respectively, when compared with the specialists’ decisions. Two apps were unable to analyse 14 and 18% of lesions submitted, respectively. Interrater agreement between dermatologists and apps was poor (κ = −0.01 SE = 0.16; P = 0.97) to slight (κ = 0.16 SE = 0.09; P = 0.12). Conclusions None of the melanoma apps tested had high enough agreement with the dermatologist's clinical opinion to be considered to provide additional benefit to patients in assessing their skin for high‐risk pigmented lesions. The low sensitivity in detecting lesions that are suspicious to a trained specialist may mean false reassurance is being given to patients. Development of highly sensitive and specific melanoma apps remains a work in progress.