Diagnostischer Nutzen der Konfokalmikroskopie als weiterführende Untersuchungsmethode von Gesichtsläsionen

Collective human intelligence outperforms artificial intelligence in a skin lesion classification task

BACKGROUND AND OBJECTIVES

Convolutional neural networks (CNN) enable accurate diagnosis of medical images and perform on or above the level of individual physicians. Recently, collective human intelligence (CoHI) was shown to exceed the diagnostic accuracy of individuals. Thus, diagnostic performance of CoHI (120 dermatologists) versus individual dermatologists versus two state-of-the-art CNN was investigated.

PATIENTS AND METHODS

Cross-sectional reader study with presentation of 30 clinical cases to 120 dermatologists. Six diagnoses were offered and votes collected via remote voting devices (quizzbox®, Quizzbox Solutions GmbH, Stuttgart, Germany). Dermatoscopic images were classified by a binary and multiclass CNN (FotoFinder Systems GmbH, Bad Birnbach, Germany). Three sets of diagnostic classifications were scored against ground truth: (1) CoHI, (2) individual dermatologists, and (3) CNN.

RESULTS

CoHI attained a significantly higher accuracy [95 % confidence interval] (80.0 % [62.7 %-90.5 %]) than individual dermatologists (75.7 % [73.8 %-77.5 %]) and CNN (70.0 % [52.1 %-83.3 %]; all P < 0.001) in binary classifications. Moreover, CoHI achieved a higher sensitivity (82.4 % [59.0 %-93.8 %]) and specificity (76.9 % [49.7 %-91.8 %]) than individual dermatologists (sensitivity 77.8 % [75.3 %-80.2 %], specificity 73.0 % [70.6 %-75.4 %]) and CNN (sensitivity 70.6 % [46.9 %-86.7 %], specificity 69.2 % [42.4 %-87.3 %]). The diagnostic accuracy of CoHI was superior to that of individual dermatologists (P < 0.001) in multiclass evaluation, with the accuracy of the latter comparable to multiclass CNN.

CONCLUSIONS

Our analysis revealed that the majority vote of an interconnected group of dermatologists (CoHI) outperformed individuals and CNN in a demanding skin lesion classification task.

Routine application of ex vivo confocal laser scanning microscopy with digital staining for examination of surgical margins in basal cell carcinomas

BACKGROUND AND OBJECTIVES

Ex vivo confocal laser scanning microscopy (CLSM) allows histologic examination of native tissue based on tissue reflection and nuclear fluorescence staining. The newly introduced digital staining process almost perfectly mimics conventional hematoxylin and eosin (HE) slides. The aim was to evaluate the new method in clinical routine, with regard to quality of findings and time requirements, in the examination of surgical margins of basal cell carcinomas.

PATIENTS AND METHODS

78 patients with 101 basal cell carcinomas were prospectively enrolled. Surgery was performed either with complete margin control (n = 60) or as elliptical excision (n = 41). Immediately after excision specimens were scanned with CLSM and then routinely processed by conventional histopathology. Blinded evaluation of images and slides was performed by a dermatopathologist.

RESULTS

Basal cell carcinomas were excellently recognizable by CLSM directly after excision, and the use of digital staining did not require any adjustment of the examiner's visualization preferences. CLSM images showed a sensitivity of 73.6 % and a specificity of 96.5 % compared to conventional HE stained slides. Erroneous findings were often due to limited assessment potential in cases where the epidermis could not be fully visualized.

CONCLUSIONS

CLSM with digital HE staining is very well suited to diagnose basal cell carcinomas and their incision margins even under routine conditions and thus represents a tissue-saving alternative to rapid cryostat sectioning.

Folliculotropism in head and neck lentigo maligna and lentigo maligna melanoma

BACKGROUND

Lentigo maligna (LM) and lentigo maligna-melanoma (LMM) are histotypes of melanoma arising in skin with cumulative solar radiation damage. The extension of atypical melanocytes to the hair follicle (folliculotropism) is a histopathological feature of LM/LMM. Its role has not been totally clarified, but it may be correlated to treatment response in LM or to progression in LMM.

OBJECTIVE

This retrospective, multicentric study aims to identify dermatoscopic features associated with folliculotropism in LMs/LMMs.

PATIENTS AND METHODS

We analyzed cases of head and neck LMs/LMMs diagnosed between 2005-2014 at Melanoma Units, University of Bologna/Modena/Florence/Siena (Italy), Nice (France): 25 LMs and 73 LMMs were included.

RESULTS

Grey circles (44 %) indicated an isthmic/bulb level of involvement, which were completely absent in the infundibular LM lesions (P = 0.041). In the group of LMMs, light/dark brown pseudonetwork and light brown structureless areas were an indicator of diffuse distribution of malignant melanocytes in the follicular units (P < 0.001 and P = 0.001, respectively), while grey circles indicated focal or diffuse distribution (P < 0.001).

CONCLUSIONS

A better understanding of the extension of malignant melanocytes is helpful, aiding clinicians in their decision to perform a radical excision or obtaining a biopsy in the most invasive area of the lesion, which includes potential folliculotropism.