Bladder cancer is one of the top 10 frequently occurring cancers and leads to most cancer deaths worldwide. Recently, blue light (BL) cystoscopy-based photodynamic diagnosis was introduced as a unique technology to enhance the detection of bladder cancer, particularly for the detection of flat and small lesions. Here, we aim to demonstrate a BL image-based artificial intelligence (AI) diagnostic platform using 216 BL images, that were acquired in four different urological departments and pathologically identified with respect to cancer malignancy, invasiveness, and grading. Thereafter, four pre-trained convolution neural networks were utilized to predict image malignancy, invasiveness, and grading. The results indicated that the classification sensitivity and specificity of malignant lesions are 95.77% and 87.84%, while the mean sensitivity and mean specificity of tumor invasiveness are 88% and 96.56%, respectively. This small multicenter clinical study clearly shows the potential of AI based classification of BL images allowing for better treatment decisions and potentially higher detection rates.
The goal of sample-size planning (SSP) is to determine the number of measurements needed for statistical analysis. This SSP is necessary to achieve robust and significant results with a minimal number of measurements that need to be collected. SSP is a common procedure for univariate measurements, whereas for multivariate measurements, like spectra or time traces, no general sample-size-planning method exists. Sample-size planning becomes more important for biospectroscopic data because the data generation is time-consuming and costly. Additionally, ethical reasons do not allow the use of unnecessary samples and the measurement of unnecessary spectra. In this paper, a general sample-size-planning algorithm is presented that is based on learning curves. The learning curve quantifies the improvement of a classifier for an increasing training-set size. These curves are fitted by the inverse-power law, and the parameters of this fit can be utilized to predict the necessary training-set size. Sample-size planning is demonstrated for a biospectroscopic task of differentiating six different bacterial species, including Escherichia coli, Klebsiella terrigena, Pseudomonas stutzeri, Listeria innocua, Staphylococcus warneri, and Staphylococcus cohnii, on the basis of their Raman spectra. Thereby, we estimate the required number of Raman spectra and biological replicates to train a classification model, which consists of principal-component analysis (PCA) combined with linear-discriminant analysis (LDA). The presented algorithm revealed that 142 Raman spectra per species and seven biological replicates are needed for the above-mentioned biospectroscopicclassification task. Even though it was not demonstrated, the learning-curve-based sample-size-planning algorithm can be applied to any multivariate data and in particular to biospectroscopic-classification tasks.
Breast cancer is the main cause of all female cancer deaths worldwide. Because of the lack of early symptoms, the early detection of breast cancer becomes challenging. This detection is performed by screening techniques in organized preventive examinations. A promising imaging technology that can detect biomolecular alterations and can support the screening technologies by enhancing their low sensitivity, is nonlinear multimodal imaging. To detect these biomolecular alterations, machine‐learning algorithms are utilized. Our analysis started by preprocessing the images and comparing them to the pathological diagnosis. We trained two classification models utilizing the deep convolutional neural network ResNet50. This network was either used as feature extractor or to be fine‐tuned as a classification model. Beside these two classification approaches, two data validation techniques were investigated: the leave‐one‐patient‐out cross‐validation (LOPO‐CV) and the training‐test validation. The best reported result of breast cancer detection was introduced by the fine‐tuned ResNet50 network and LOPO‐CV accounting to 86.23% mean‐sensitivity.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.