Machine learning for precision dermatology: Advances, opportunities, and outlook

Seeing the random forest through the decision trees. Supporting learning health systems from histopathology with machine learning models: Challenges and opportunities

This paper discusses some overlooked challenges faced when working with machine learning models for histopathology and presents a novel opportunity to support "Learning Health Systems" with them. Initially, the authors elaborate on these challenges after separating them according to their mitigation strategies: those that need innovative approaches, time, or future technological capabilities and those that require a conceptual reappraisal from a critical perspective. Then, a novel opportunity to support "Learning Health Systems" by integrating hidden information extracted by ML models from digitalized histopathology slides with other healthcare big data is presented.

Histology-based classifier to distinguish early mycosis fungoides from atopic dermatitis

BACKGROUND

Histopathological differentiation of early mycosis fungoides (MF) from benign chronic inflammatory dermatoses remains difficult and often impossible, despite the inclusion of all available diagnostic parameters.

OBJECTIVE

To identify the most impactful histological criteria for a predictive diagnostic model to discriminate MF from atopic dermatitis (AD).

METHODS

In this multicentre study, two cohorts of patients with either unequivocal AD or MF were evaluated by two independent dermatopathologists. Based on 32 histological attributes, a hypothesis-free prediction model was developed and validated on an independent patient's cohort.

RESULTS

A reduced set of two histological features (presence of atypical lymphocytes in either epidermis or dermis) was trained. In an independent validation cohort, this model showed high predictive power (95% sensitivity and 100% specificity) to differentiate MF from AD and robustness against inter-individual investigator differences.

LIMITATIONS

The study investigated a limited number of cases and the classifier is based on subjectively evaluated histological criteria.

CONCLUSION

Aiming at distinguishing early MF from AD, the proposed binary classifier performed well in an independent cohort and across observers. Combining this histological classifier with immunohistochemical and/or molecular techniques (such as clonality analysis or molecular classifiers) could further promote differentiation of early MF and AD.

Conventional and three-dimensional photography as a tool to map distribution patterns of in-transit melanoma metastases on the lower extremity

Background

In melanoma, in-transit metastases characteristically occur at the lower extremity along lymphatic vessels.

Objectives

The objective of this study was to evaluate conventional or three-dimensional photography as a tool to analyze in-transit metastasis pattern of melanoma of the lower extremity. In addition, we assessed risk factors for the development of in-transit metastases in cutaneous melanoma.

Methods

In this retrospective, monocentric study first we compared the clinical data of all evaluable patients with in-transit metastases of melanoma on the lower extremity (n = 94) with melanoma patients without recurrence of disease (n = 288). In addition, based on conventional (n = 24) and three-dimensional photography (n = 22), we defined the specific distribution patterns of the in-transit metastases on the lower extremity.

Results

Using a multivariate analysis we identified nodular melanoma, tumor thickness, and ulceration as independent risk factors to develop in-transit metastases ITM (n = 94). In patients with melanoma on the lower leg (n = 31), in-transit metastases preferentially developed along anatomically predefined lymphatic pathways. In contrast when analyzing in-transit metastases of melanoma on the foot (n = 15) no clear pattern could be visualized. In addition, no difference in distance between in-transit metastases and primary melanoma on the foot compared to the lower leg was observed using three-dimensional photography (n = 22).

Conclusion

A risk-adapted follow-up of melanoma patients to detect in-transit metastases can be applied by knowledge of the specific lymphatic drainage of the lower extremity. Our current analysis suggests a more complex lymphatic drainage of the foot.

Predicting Mohs surgery complexity by applying machine learning to patient demographics and tumor characteristics

Mohs micrographic surgery (MMS) is considered the gold standard for difficult-to-treat malignant skin tumors, whose incidence is on the rise. Currently, there are no agreed upon classifiers to predict complex MMS procedures. Such classifiers could enable better patient scheduling, reduce staff burnout and improve patient education. Our goal was to create an accessible and interpretable classifier(s) that would predict complex MMS procedures. A retrospective study applying machine learning models to a dataset of 8644 MMS procedures to predict complex wound reconstruction and number of MMS procedure stages. Each procedure record contained preoperative data on patient demographics, estimated clinical tumor size prior to surgery (mean diameter), tumor characteristics and tumor location, and postoperative procedure outcomes included the wound reconstruction technique and the number of MMS stages performed in order to achieve tumor-free margins. For the number of stages complexity classification model, the area under the receiver operating characteristic curve (AUROC) was 0.79 (good performance), with model accuracy of 77%, sensitivity of 71%, specificity of 77%, positive prediction value (PPV) of 14% and negative prediction value (NPV) of 98%. The results for the wound reconstruction complexity classification model were 0.84 for the AUROC (excellent performance), with model accuracy of 75%, sensitivity of 72%, specificity of 76%, PPV of 39% and NPV of 93%. The ML models we created predict the complexity of the components that comprise the MMS procedure. Using the accessible and interpretable tool we provide online, clinicians can improve the management and well-being of their patients. Study limitation is that models are based on data generated from a single surgeon.

Expanding Personalized, Data-Driven Dermatology: Leveraging Digital Health Technology and Machine Learning to Improve Patient Outcomes

The current revolution of digital health technology and machine learning offers enormous potential to improve patient care. Nevertheless, it is essential to recognize that dermatology requires an approach different from those of other specialties. For many dermatological conditions, there is a lack of standardized methodology for quantitatively tracking disease progression and treatment response (clinimetrics). Furthermore, dermatological diseases impact patients in complex ways, some of which can be measured only through patient reports (psychometrics). New tools using digital health technology (e.g., smartphone applications, wearable devices) can aid in capturing both clinimetric and psychometric variables over time. With these data, machine learning can inform efforts to improve health care by, for example, the identification of high-risk patient groups, optimization of treatment strategies, and prediction of disease outcomes. We use the term personalized, data-driven dermatology to refer to the use of comprehensive data to inform individual patient care and improve patient outcomes. In this paper, we provide a framework that includes data from multiple sources, leverages digital health technology, and uses machine learning. Although this framework is applicable broadly to dermatological conditions, we use the example of a serious inflammatory skin condition, chronic cutaneous graft-versus-host disease, to illustrate personalized, data-driven dermatology.

Frequency of Publication of Dermoscopic Images in Inter-observer Studies: A Systematic Review

Research interest in dermoscopy is increasing, but the complete dermoscopic image sets used in inter-observer studies of skin tumours are not often shared in research publications. The aim of this systematic review was to analyse what proportion of images depicting skin tumours are published in studies investigating inter-observer variations in the assessment of dermoscopic features and/or patterns. Embase, MEDLINE and Scopus databases were screened for eligible studies published from inception to 2 July 2020. For included studies the proportion of lesion images presented in the papers and/or supplements was extracted. A total of 61 studies (53 original studies and 8 shorter reports (i.e. research letters or concise reports)). published in the period 1997 to 2020 were included. These studies combined included 14,124 skin tumours, of which 373 (3%) images were published. This systematic review highlights that the vast majority of images included in dermoscopy research are not published. Data sharing should be a requirement for future studies, and must be enabled and standardized by the dermatology research community and editorial offices.