Probabilistic machine learning for breast cancer classification.

A probabilistic neural network has been implemented to predict the malignancy of breast cancer cells, based on a data set, the features of which are used for the formulation and training of a model for a binary classification problem. The focus is placed on considerations when building the model, in order to achieve not only accuracy but also a safe quantification of the expected uncertainty of the calculated network parameters and the medical prognosis. The source code is included to make the results reproducible, also in accordance with the latest trending in machine learning research, named Papers with Code. The various steps taken for the code development are introduced in detail but also the results are visually displayed and critically analyzed also in the sense of explainable artificial intelligence. In statistical-classification problems, the decision boundary is the region of the problem space in which the classification label of the classifier is ambiguous. Problem aspects and model parameters which influence the decision boundary are a special aspect of practical investigation considered in this work. Classification results issued by technically transparent machine learning software can inspire more confidence, as regards their trustworthiness which is very important, especially in the case of medical prognosis. Furthermore, transparency allows the user to adapt models and learning processes to the specific needs of a problem and has a boosting influence on the development of new methods in relevant machine learning fields (transfer learning).

Doxorubicin selects for fluconazole-resistant petite mutants in Candida glabrata isolates.

Candida infections are a permanent threat to immunocompromised individuals such as cancer patients, and Candida glabrata has emerged as a major problem in recent years. Resistance may develop during lengthy antifungal therapies and is often mediated by upregulation of fungal drug efflux pumps. During chemotherapy the yeast cell is also exposed to cytotoxic agents that may affect its drug susceptibility. Four C. glabrata isolates, three susceptible and one resistant to fluconazole (FLU), were incubated with 20 µg/ml of doxorubicin (DOX) for 90 min. In a second experiment, the isolates were cultured with DOX for ten days. Samples were taken on subsequent days to determine the minimal inhibitory concentration (MIC) of FLU and to analyze expression of CgCDR1, CgCDR2, CgSNQ2 and CgPDR1. Samples were also used to assess the petite phenotype. Short-term DOX exposure did not induce efflux pump gene expression, but genes were consistently overexpressed in FLU-susceptible isolates during long-term exposure. An increase in MIC values on day 6 in two of the isolates coincided with the first occurrence of petite mutants in all susceptible isolates. The respiratory deficiency of selected petite mutants was confirmed by culturing mutants on agar containing glycerol as the sole carbon source. FLU MIC values for respiratory-deficient clones were ≥64 µg/ml, and efflux pump gene expression was greatly increased. The resistant isolate did not develop mitochondrial dysfunction. In summary, the cytotoxic agent DOX selects for FLU-resistant respiratory-deficient C. glabrata mutants, which may affect antifungal therapy.

Influence of doxorubicin on fluconazole susceptibility and efflux pump gene expression of Candida dubliniensis.

The effect of doxorubicin (DOX) on the fluconazole (FLU) susceptibility of C. dubliniensis was investigated. Isolates were exposed to DOX and FLU in a chequerboard assay and resistance gene expressions were analysed after DOX exposure. The susceptibility of the yeast to FLU was decreased in the presence of DOX in the chequerboard assay with FIC indices suggesting an antagonistic effect. Gene expression analyses showed an overexpression of CdCDR2. Hence, DOX was found to have an impact on resistance mechanisms in C. dubliniensis isolates.

Difference in virulence between fluconazole-susceptible and fluconazole-resistant Candida albicans in a mouse model.

Two Candida albicans isolates were collected from a HIV-positive patient with recurrent oropharyngeal candidosis (OPC). One isolate was taken during the first episode of oral candidosis [fluconazole susceptible (FLU-S), minimal inhibitory concentration (MIC) = 0.25 mg l(-1) ] and the second after the patient developed refractory OPC and resistance to fluconazole (FLU-R, MIC = 64 mg l(-1)). Both isolates were clonally identical. Different in vitro studies were carried out to assess putative virulence factors of both isolates. Gene expressions of efflux pumps and CSH1 were determined as well as adherence to human epithelial cells, determination of proteinase secretion and biofilm formation activity. Virulence was studied using a disseminated mouse model. All mice challenged with the FLU-S isolate survived the experiment when FLU was given. However, when FLU was absent, the mortality of the FLU-S isolate was higher than that of the FLU-R isolate with no mice surviving the experiment. In vitro studies showed pronounced growth rates of the FLU-S isolate and a more intense biofilm-building activity compared with the FLU-R isolate. The FLU-R isolate highly up-regulated MDR1 and CSH1. This isolate also adhered stronger to the epithelial cell line. The results showed that FLU-S and FLU-R isolates exhibit different virulence factors, which enable the survival of both isolates in adapted environments.