Search

Counterfactual reasoning is often used in clinical settings to explain decisions or weigh alternatives. Therefore, for imaging based specialties such as ophthalmology, it would be beneficial to be able to create counterfactual images, illustrating answers to questions like "If the subject had had diabetic retinopathy, how would the fundus image have looked?". Here, we demonstrate that using a diffusion model in combination with an adversarially robust classifier trained on retinal disease classification tasks enables the generation of highly realistic counterfactuals of retinal fundus images and optical coherence tomography (OCT) B-scans. The key to the realism of counterfactuals is that these classifiers encode salient features indicative for each disease class and can steer the diffusion model to depict disease signs or remove disease-related lesions in a realistic way. In a user study, domain experts also found the counterfactuals generated using our method significantly more realistic than counterfactuals generated from a previous method, and even indistinguishable from real images.

Introduction Alectinib is an effective second-generation ALK tyrosine kinase inhibitor (TKI) used in the first-line treatment of patients with advanced ALK-positive NSCLC. Recent studies demonstrated that the percentage of ALK-positive tumor cells in patient groups receiving crizotinib might affect outcomes. This study aimed to investigate whether the percentage of ALK-positive cells had a predictive effect in patients with advanced NSCLC who received first-line Alectinib as ALK-TKI. Materials and methods This retrospective study included patients with advanced-stage NSCLC who received alectinib as a first-line ALK-TKI and whose percentage of ALK-positive cells was determined by FISH at 27 different centers. Patients who received any ALK-TKI before alectinib were not included in the study. Patients were separated into two groups according to the median (40%) value of the percentage of ALK-positive cells (high-positive group >= 40% and low-positive group < 40%). The primary endpoint was PFS, and the secondary endpoints were OS, ORR, and PFS of the subgroups based on different threshold values for the percentage of ALK-positive cells. Results 211 patients were enrolled (48.3% female, 51.7% male) to study. 37% (n = 78) of the patients had received chemotherapy previously. After a median of 19.4 months of follow-up, the median PFS was not reached in the high-positive group (n = 113), but it was 10.8 months in the low-positive group (n = 98) (HR 0.39; 95% CI 0.25-0.60, p < 0.001). The median OS in the high-positive group was not reached, whereas it was 22.8 months in the low-positive group (HR 0.37; 95% CI 0.22-0.63, p < 0.001). ORR was significantly higher in the high-positive group (87.2 vs. 68.5%; p = 0.002). According to the cut-off values of < 20%, 20-39%, 40-59%, and >= 60%, the median PFS was 4.5, 17.1, and 26 months, respectively, and could not be reached in the >= 60% group. Conclusion Our study demonstrated that the efficacy of alectinib varies significantly ac, NA

Alzheimer's disease is a major cause of dementia. Its diagnosis requires accurate biomarkers that are sensitive to disease stages. In this respect, we regard probabilistic classification as a method of designing a probabilistic biomarker for disease staging. Probabilistic biomarkers naturally support the interpretation of decisions and evaluation of uncertainty associated with them. In this paper, we obtain probabilistic biomarkers via Gaussian Processes. Gaussian Processes enable probabilistic kernel machines that offer flexible means to accomplish Multiple Kernel Learning. Exploiting this flexibility, we propose a new variation of Automatic Relevance Determination and tackle the challenges of high dimensionality through multiple kernels. Our research results demonstrate that the Gaussian Process models are competitive with or better than the well-known Support Vector Machine in terms of classification performance even in the cases of single kernel learning. Extending the basic scheme towards the Multiple Kernel Learning, we improve the efficacy of the Gaussian Process models and their interpretability in terms of the known anatomical correlates of the disease. For instance, the disease pathology starts in and around the hippocampus and entorhinal cortex. Through the use of Gaussian Processes and Multiple Kernel Learning, we have automatically and efficiently determined those portions of neuroimaging data. In addition to their interpretability, our Gaussian Process models are competitive with recent deep learning solutions under similar settings., Comment: The material presented here is to promote the dissemination of scholarly and technical work in a timely fashion. Data in this article are from ADNI (adni.loni.usc.edu). As such, ADNI provided data but did not participate in writing of this report