A non-immunocompromised host with nontuberculous mycobacteria-associated tubulointerstitial nephritis.

A 50-year-old man was admitted to our hospital with the complaints of fever and general malaise. He had no history of human immunodeficiency virus (HIV) infection or treatment with immunosuppressive agents. We performed renal biopsy to investigate possible acute kidney injury. Pathological findings showed inflammatory cell infiltration, including granulomatous lesions in the interstitium. We diagnosed the patient with acute granulomatous tubulointerstitial nephritis. We initiated prednisolone (PSL) 40 mg/day (0.6 mg/kg), in combination with isoniazid for a latent tuberculosis infection, because of positive results in interferon-γ release assays. The patient's fever and malaise promptly disappeared, and his renal function improved. After the patient had been discharged, Mycobacterium intracellulare grew in cultures of his renal tissue and urine. We gradually reduced the dose of PSL; we initiated combination therapy with ethambutol, clarithromycin, and rifampin. After 2 years of follow-up, the patient continued treatment for chronic kidney disease; it has since enabled him to avoid renal replacement therapy. This report describes a rare instance of nontuberculous mycobacteria-associated tubulointerstitial nephritis in a patient without a history of HIV infection or organ transplantation. In differential diagnosis of granulomatous tubulointerstitial nephritis, clinicians should consider drugs, sarcoidosis, tubulointerstitial nephritis and uveitis syndrome, vasculitis, and infections (e.g., involving mycobacteria). Prompt microbiological examinations, especially of urine or biopsy cultures, are vital for diagnosis.

Vulnerability of deep neural networks for detecting COVID-19 cases from chest X-ray images to universal adversarial attacks.

Owing the epidemic of the novel coronavirus disease 2019 (COVID-19), chest X-ray computed tomography imaging is being used for effectively screening COVID-19 patients. The development of computer-aided systems based on deep neural networks (DNNs) has become an advanced open source to rapidly and accurately detect COVID-19 cases because the need for expert radiologists, who are limited in number, forms a bottleneck for screening. However, thus far, the vulnerability of DNN-based systems has been poorly evaluated, although realistic and high-risk attacks using universal adversarial perturbation (UAP), a single (input image agnostic) perturbation that can induce DNN failure in most classification tasks, are available. Thus, we focus on representative DNN models for detecting COVID-19 cases from chest X-ray images and evaluate their vulnerability to UAPs. We consider non-targeted UAPs, which cause a task failure, resulting in an input being assigned an incorrect label, and targeted UAPs, which cause the DNN to classify an input into a specific class. The results demonstrate that the models are vulnerable to non-targeted and targeted UAPs, even in the case of small UAPs. In particular, the 2% norm of the UAPs to the average norm of an image in the image dataset achieves >85% and >90% success rates for the non-targeted and targeted attacks, respectively. Owing to the non-targeted UAPs, the DNN models judge most chest X-ray images as COVID-19 cases. The targeted UAPs allow the DNN models to classify most chest X-ray images into a specified target class. The results indicate that careful consideration is required in practical applications of DNNs to COVID-19 diagnosis; in particular, they emphasize the need for strategies to address security concerns. As an example, we show that iterative fine-tuning of DNN models using UAPs improves the robustness of DNN models against UAPs.