LB955 Predicting metastatic melanoma from melanoma pathological images using a Convolutional Neural Network: A Multicenter Study

Abstract

Cutaneous Malignant Melanoma (CMM) incidence has been rising around the world and over the last three decades at rates greater than for any other malignancy. Our objective was to describe geographic trends in incidence and mortality of CMM in Russia between 2001 and 2017 using geo-informatics technique (mapping) and descriptive statistical analysis. Additionally, we aimed to study the associations between ethnicity, geographic latitude/longitude and CMM incidence/mortality rates. We retrospectively analyzed the data from the Moscow Oncology Research Institute, Ministry of Health of Russian Federation for the period of the study. International Classification of Diseases (ICD) C43 code (comprising C43.0-C.43.9) was used to identify cutaneous melanoma cases. Routine methods of descriptive epidemiology were used to study incidence and mortality rates by age groups, years, and jurisdictions (i.e., Federal Districts and Federal Subjects of Russia). In total 141,597 patients were diagnosed with melanoma in Russia over the period 2001-2017, of which 62% were women (p<0.001). The overall age-standardized incidence and mortality rates were 4.27/100,000 and 1.62/ 100,000, respectively. Geographic mapping revealed North-to-South and East-to-West gradients across the country. Intrinsic patient characteristics such as the skin phenotype and the climate zones of the country could be an important risk factors for melanoma development. This study, for the first time, reports the burden and geographic distribution of CMM in Russia and the trends correlate with observations in countries with similar geography.

Image analysis of cutaneous melanoma histology: a systematic review and meta-analysis

“…Overall, 4,888 specimens were included, of which at least 2,715 were melanoma specimens. The diagnostic entities within the datasets varied between studies, with some only containing melanoma deposits 12,21,[24][25][26]33 and others containing more than one pathology 10,11,22,[27][28][29][30][31][32] .…”

“…There was between-study variation in terms of intended use of the IA. Most studies focused on a binary classification task, with some focussing on detection and localisation of melanoma deposits in WSIs containing melanoma (melanoma versus not melanoma) 12,25,26,32 and others performing diagnostic classifications including melanomas versus naevi 11,22,31 and primary melanoma versus metastatic melanoma 33 . Five studies addressed more complex classifications into three or more diagnostics entities 10,23,[28][29][30] .…”

“…There was some degree of variation in the IT used. Most employed the use of a convolutional neural networks (CNN), with the architecture differing considerably [10][11][12]22,23,26,[29][30][31]33 . Studies using CNNs were more recently conducted.…”

“…Some studies used a combination of these approaches 23,32 . Two studies did not detail the reference standard used 29,33 .…”

“…Some studies reported pixel-based outcomes 25,26 or cell-level outcomes 24,27,28 , whereas others focussed on patch-level 12,22,31,32 or slide-level classifications 10,11,30,32 . Three studies did not report on their unit for analysis 23,29,33 , whilst one appeared to be at the WSI-level 23 .…”