Deep learning‐based diagnosis models for onychomycosis in dermoscopy

Zhu, Xianzhong; Zheng, Bowen; Cai, Wenying; Zhang, Jing; Lu, Sha; Li, Xiqing; Xi, Liyan; Kong, Yinying

doi:10.1111/myc.13427

Cited by 15 publications

(14 citation statements)

References 17 publications

(13 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Youden index score (sensitivity+specificity-100%) of the AI (0.429) was also comparable to the mean Youden index score of the five board-certified dermatologists (0.230±0.176) (Wilcoxon rank-sum test; p = 0.667). In a study evaluating 1155 dermoscopic images (603 onychomycosis, 227 normal nails, 221 nail psoriasis, 104 traumatic onychodystrophy), 85 the diagnostic performances between the AI and 54 dermatologists were compared. There was high specificity (>82%) of diagnosing onychomycosis for five dermoscopic nail plate patterns, including jagged edge, longitudinal striae, marble-like turbid areas, distal irregular termination, and cone-shaped keratosis.…”

Section: Diagnostic Testingmentioning

confidence: 99%

“…[79][80][81][82][83][84] AI has also been applied for diagnosis of onychomycosis. 78,85,86 In a prospective study of 90 patients with onychodystrophy of the toenails, 78 clinical photographs of the nails were taken by nonphysicians and evaluated clinically by five board-certified dermatologists (mean 5.6 years of experience), as well as by two board-certified dermatologists using dermoscopy and compared to AI to diagnose onychomycosis. KOH microscopy or fungal culture was used to confirm the diagnosis in all cases.…”

Section: Pcrmentioning

confidence: 99%

See 1 more Smart Citation

Updated Perspectives on the Diagnosis and Management of Onychomycosis

Falotico¹,

Lipner²

2022

CCID

View full text Add to dashboard Cite

Onychomycosis is the most common nail disease encountered in clinical practice and can cause pain, difficulty with ambulation, and psycho-social problems. A thorough history and physical examination, including dermoscopy, should be performed for each patient presenting with nail findings suggestive of onychomycosis. Several approaches are available for definitive diagnostic testing, including potassium hydroxide and microscopy, fungal culture, histopathology, polymerase chain reaction, or a combination of techniques. Confirmatory testing should be performed for each patient prior to initiating any antifungal therapies. There are several different therapeutic options available, including oral and topical medications as well as device-based treatments. Oral antifungals are generally recommended for moderate to severe onychomycosis and have higher cure rates, while topical antifungals are recommended for mild to moderate disease and have more favorable safety profiles. Oral terbinafine, itraconazole, and griseofulvin and topical ciclopirox 8% nail lacquer, efinaconazole 10% solution, and tavaborole 5% solution are approved by the Food and Drug Administration for treatment of onychomycosis in the United States and amorolfine 5% nail lacquer is approved in Europe. Laser treatment is approved in the United States for temporary increases in clear nail, but clinical results are suboptimal. Oral fluconazole is not approved in the United States for onychomycosis treatment, but is frequently used off-label with good efficacy. Several novel oral, topical, and over-the-counter therapies are currently under investigation. Physicians should consider the disease severity, infecting pathogen, medication safety, efficacy and cost, and patient age, comorbidities, medication history, and likelihood of compliance when determining management plans. Onychomycosis is a chronic disease with high recurrence rates and patients should be counseled on an appropriate plan to minimize recurrence risk following effective antifungal therapy.

show abstract

Section: Diagnostic Testingmentioning

confidence: 99%

Section: Pcrmentioning

confidence: 99%

Updated Perspectives on the Diagnosis and Management of Onychomycosis

Falotico¹,

Lipner²

2022

CCID

View full text Add to dashboard Cite

show abstract

“…12 To the best of our knowledge, three studies have investigated the use of AI for the diagnosis of onychomycosis so far. [13][14][15] These studies were carried out on the clinical and dermoscopic photographs of the patients. Achieving standardisation in clinical and dermoscopic photography is very difficult.…”

Section: Introductionmentioning

confidence: 99%

“…There are many studies investigating its use in diagnosis and treatment monitoring of dermatological disorders 12 . To the best of our knowledge, three studies have investigated the use of AI for the diagnosis of onychomycosis so far 13–15 . These studies were carried out on the clinical and dermoscopic photographs of the patients.…”

Section: Introductionmentioning

confidence: 99%

“…respectively. The mean accuracy, sensitivity, specificity, precision and NPV values of 5 dermatologists who have(12)(13)(14)(15)(16)(17)(18)(19)(20) year experience are 73.30 ± 6.50%, 87.60 ± 7.66%, 59.00 ± 16.80%, 69.46 ± 8.09% and 83.56 ± 7.09%, respectively. The mean accuracy, sensitivity, specificity, precision and NPV values of 5 dermatologists who have(20)(21)(22)(23)(24)(25)(26) year experience are 74.20 ± 8.86%, 63.40 ± 21.27%, 85.00 ± 6.93%, 81.42 ± 6.38% and 72.15 ± 11.27%, respectively.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Deep convolutional neural networks for onychomycosis detection using microscopic images with KOH examination

Yılmaz

Goktay

Varol

et al. 2022

Mycoses

View full text Add to dashboard Cite

Background The diagnosis of superficial fungal infections is still mostly based on direct microscopic examination with potassium hydroxide solution. However, this method can be time consuming, and its diagnostic accuracy rates vary widely depending on the clinician's experience. Objectives This study presents a deep neural network structure that enables the rapid solutions for these problems and can perform automatic fungi detection in grayscale images without dyes. Methods One hundred sixty microscopic full field photographs containing the fungal element, obtained from patients with onychomycosis, and 297 microscopic full field photographs containing dissolved keratin obtained from normal nails were collected. Smaller patches containing fungi (n = 1835) and keratin (n = 5238) were extracted from these full field images. In order to detect fungus and keratin, VGG16 and InceptionV3 models were developed by the use of these patches. The diagnostic performance of models was compared with 16 dermatologists by using 200 test patches. Results For the VGG16 model, the InceptionV3 model and 16 dermatologists, mean accuracy rates were 88.10 ± 0.8%, 88.78 ± 0.35% and 74.53 ± 8.57%, respectively; mean sensitivity rates were 75.04 ± 2.73%, 74.93 ± 4.52% and 74.81 ± 19.51%, respectively; and mean specificity rates were 92.67 ± 1.17%, 93.78 ± 1.74% and 74.25 ± 18.03%, respectively. The models were statistically superior to dermatologists according to rates of accuracy and specificity but not to sensitivity (p < .0001, p < .005 and p > .05, respectively). Area under curve values of the VGG16 and InceptionV3 models were 0.9339 and 0.9292, respectively. Conclusion Our research demonstrates that it is possible to build an automated system capable of detecting fungi present in microscopic images employing the proposed deep learning models. It has great potential for fungal detection applications based on AI.

show abstract