Accuracy of a smartphone application for triage of skin lesions based on machine learning algorithms.

BACKGROUND: Machine learning algorithms achieve expert-level accuracy in skin lesion classification based on clinical images. However, it is not yet shown whether these algorithms could have high accuracy when embedded in a smartphone app, where image quality is lower and there is high variability in image taking scenarios by users. In the past, these applications were criticized due to lack of accuracy. OBJECTIVE: In this study, we evaluate the accuracy of the newest version of a smartphone application (SA) for risk assessment of skin lesions. METHODS: This SA uses a machine learning algorithm to compute a risk rating. The algorithm is trained on 131 873 images taken by 31 449 users in multiple countries between January 2016 and August 2018 and rated for risk by dermatologists. To evaluate the sensitivity of the algorithm, we use 285 histopathologically validated skin cancer cases (including 138 malignant melanomas), from two previously published clinical studies (195 cases) and from the SA user database (90 cases). We calculate the specificity on a separate set from the SA user database containing 6000 clinically validated benign cases. RESULTS: The algorithm scored a 95.1% (95% CI, 91.9-97.3%) sensitivity in detecting (pre)malignant conditions (93% for malignant melanoma and 97% for keratinocyte carcinomas and precursors). This level of sensitivity was achieved with a 78.3% (95% CI, 77.2-79.3%) specificity. CONCLUSIONS: This SA provides a high sensitivity to detect skin cancer; however, there is still room for improvement in terms of specificity. Future studies are needed to assess the impact of this SA on the health systems and its users.

Accuracy of a smartphone application using fractal image analysis of pigmented moles compared to clinical diagnosis and histological result.

BACKGROUND: Lately, various smartphone applications have been introduced as diagnostic self-monitoring tools in the evaluation of pigmented moles, but most of these techniques have not been evaluated systematically. OBJECTIVES: The purpose of this study was to evaluate prospectively the sensitivity and specificity of a recently developed smartphone application using fractal image analysis for the risk evaluation algorithm in the diagnosis of malignant melanoma compared to clinical diagnosis and histopathological result. METHODS: Consecutive patients with melanocytic lesions were recruited and clinical and dermoscopical diagnosis was documented by two dermatologists independently. Imaging and analysis with the smartphone application was performed prior to excision of lesions. The findings were compared to the histological results as gold standard. RESULTS: Of 195 included lesions histopathological analysis revealed 40 melanomas, 42 dysplastic nevi and 113 benign nevi. The sensitivity of the diagnosis melanoma by fractal image analysis using smartphone images was 73%, the specificity was 83% compared to a sensitivity of 88% and specificity of 97% regarding the clinical diagnosis by the dermatologists. CONCLUSION: The smartphone application using fractal analysis might be a promising tool in the pre-evaluation of pigmented moles by laypersons, while it is to date inferior to the diagnostic evaluation by a dermatologist.

Prognostic factors in retroperitoneal fibrosis.

UNLABELLED: The aim of this study is to evaluate effective prognostic factors in the evolution of patients with retroperitoneal fibrosis and to establish the validity of fractal analysis in determining the disease severity in these patients. MATERIAL AND METHODS: This study included 19 patients (M/F: 5/14) treated for idiopathic retroperitoneal fibrosis and bilateral obstructive renal failure between Jan 2004-Dec 2008. Patients were identified retrospectively, searching for patients diagnosed with IRF, after retroperitoneal biopsy or, in most cases the diagnosis rested on radiological findings, especially CT, with identification of a retroperitoneal mass, the absence of other demonstrable renal or ureteric disease or any other pathology that could explain the findings. CT was very useful in describing the retroperitoneal mass around the aorta and inferior vena cava, the extent of the lesion and for monitoring the response to surgical treatment during the follow-up. The data were evaluated about medical history, physical examination findings, laboratory tests (serum urea and creatinine, blood sugar, sodium, potassium, bicarbonate levels, serum pH, uric acid, haematocrit, white blood cell count), imaging methods (renal ultrasound, abdominal CT-scan, MRI). At admission all patients had active disease with obstructive renal failure and underwent bilateral ureteric stenting in order to normalize the BUN levels. After normalizing of BUN levels, ureterolysis and omental wrapping was performed. Postoperatively, ureteric stents were removed after 1 month and remission of renal disfunction was obtained in approximately 5 months (range 2-10 months). All patients were followed for at least 1 year. Patients were regularly checked every 3 months. RESULTS: Of the 19 patients, there were 5 men and 14 women. The median age at diagnosis of RF was 50 years (range 42-64 years). The most frequent presenting symptoms were back or abdominal pain, weakness, weight loss, oligoanuria, arterial hypertension and mild fever. The duration of symptoms before diagnosis ranged from 6 to 18 months. At presentation all patients had active disease, presenting renal dysfunction with a median serum creatinine of 5.18 mg/dl (range 1-15.4 mg/dl). Most of the patients had moderate bilateral hydronephrosis (2nd degree hydronephrosis). In our study, all patients had excellent prognosis, with full recovery of renal function in 78% of cases (15 patients). The fractal dimension of the fibrosis mass contour correlates with level of renal function impairment. Even more, the fractal dimension seems to slightly variate between CT evaluations (1.30 +/- 0.1), suggesting a non aggressive pattern of extension of the fibrotic mass characteristic for benign lesions. CONCLUSIONS: The imaging parameters did not predict the disease severity, except the increase in fractal dimension of fibrosis surface area. Efficacy of bilateral ureteric stenting in improving renal function is limited in most of the cases. Dispite the level of renal function impairment at admission, full recovery can be achieved after bilateral ureteric stenting/nephrostomy and ureterolisis.