Facilitating clinically relevant skin tumor diagnostics with spectroscopy-driven machine learning.

In the dawning era of artificial intelligence (AI), health care stands to undergo a significant transformation with the increasing digitalization of patient data. Digital imaging, in particular, will serve as an important platform for AI to aid decision making and diagnostics. A growing number of studies demonstrate the potential of automatic pre-surgical skin tumor delineation, which could have tremendous impact on clinical practice. However, current methods rely on having ground truth images in which tumor borders are already identified, which is not clinically possible. We report a novel approach where hyperspectral images provide spectra from small regions representing healthy tissue and tumor, which are used to generate prediction maps using artificial neural networks (ANNs), after which a segmentation algorithm automatically identifies the tumor borders. This circumvents the need for ground truth images, since an ANN model is trained with data from each individual patient, representing a more clinically relevant approach.

Mild Cognitive Impairment, Reversion Rates, and Associated Factors: Comparison of Two Diagnostic Approaches.

BACKGROUND: As mild cognitive impairment (MCI) is typically used to identify prodromal stages of dementia, it is essential to identify MCI criteria with high diagnostic stability and prediction of dementia. Moreover, further investigation into pinpointing key factors for reversion is required to foresee future prognosis of MCI patients accurately. OBJECTIVE: To explore disparities in diagnostic stability by examining reversion rates produced by two operationalizations of the MCI definition: the widely applied Petersen criteria and a version of the Neuropsychological (NP) criteria and to identify cognitive, lifestyle, and health related factors for reversion. METHODS: MCI was retrospectively classified in a sample from the Swedish community-based study Good Aging in Skåne with the Petersen criteria (n = 744, median follow-up = 7.0 years) and the NP criteria (n = 375, median follow-up, 6.7 years), respectively. Poisson regression models estimated the effect of various factors on the likelihood of incident reversion. RESULTS: Reversion rates were 323/744 (43.4%, 95% confidence intervals (CI): 39.8; 47.0) and 181/375 (48.3% 95% CI: 43.2; 53.5) for the Petersen criteria and NP criteria, respectively. Participants with impairment in a single cognitive domain, regular alcohol consumption, living with someone, older age, and lower body mass index had a higher likelihood of reverting to normal. CONCLUSION: Reversion rates were similar for Petersen and NP criteria indicating that one definition is not superior to the other regarding diagnostic stability. Additionally, the results highlight important aspects such as multiple domain MCI, cohabitation, and the role of alcohol on predicting the trajectory of those diagnosed with MCI.