Evaluation of the clinical performance of Metagenomic Next-generation sequencing for diagnosis of invasive fungal diseases after CAR-T cells infusion

Abstract Introduction Chimeric antigen receptor T-cell (CAR-T) therapy is highly effective in treating hematological malignancies but increases the risk of invasive fungal diseases (IFDs) due to prolonged immunosuppression, driven by prior chemotherapy, lymphodepletion, and off-target effects of CAR-T cells. Diagnosing IFDs is challenging with conventional detection methods (CDM) due to low sensitivity, long turnaround times, and limited species identification. Metagenomic next-generation sequencing (mNGS), a culture-independent technique, offers rapid and comprehensive pathogen detection, potentially overcoming these limitations. This study evaluates mNGS's diagnostic performance for IFDs post-CAR-T therapy compared to CDM, examining pathogen distribution and the impact of mNGS-guided antifungal therapy. Methods Our retrospective study enrolled 25 patients who received CAR-T therapy (targeting CLL1, CD19, or CD38) for hematological malignancies (10 with AML, 11 with ALL, 3 with lymphoma, 1 with MDS) at Tianjin First Central Hospital between June 2020 and May 2024. IFD diagnoses were classified as proven, probable, or possible based on EORTC criteria. Samples (blood, n=15; bronchoalveolar lavage fluid [BALF], n=10) were collected during suspected infection. CDM included microscopy, culture, serology, and histopathology. For mNGS, DNA was extracted, host sequences filtered, and remaining reads aligned against microbial databases (bacteria, fungi, viruses, parasites) for pathogen identification. Statistical analyses used Mann-Whitney U, Pearson chi-square, or Fisher exact tests. Results mNGS detected fungal pathogens in 24/25 (96%) cases within 12-24 hours, significantly outperforming CDM (11/25, 44%) with a median 4-day turnaround (p<0.001). Aspergillus fumigatus (n=8, 32%) was the most common pathogen, followed by Rhizopus microsporus (n=3, 12%) and Aspergillus flavus (n=3, 12%). Most patients had co-infections: 6 with fungi alone, 8 with fungi+bacteria, 3 with fungi+viruses, and 8 with fungi+bacteria+viruses. All patients received antifungal prophylaxis; 23 adjusted therapy based on mNGS results, with 20 (80%) showing clinical improvement. Specifically, 77.8% of mucor-infected, 85.7% of Aspergillus-infected, and 50% of Candida-infected patients improved. Approximately 68% (n=17) received combination antifungal therapy. Conclusion mNGS demonstrates superior sensitivity, broad pathogen coverage, and faster turnaround compared to CDM for diagnosing IFDs post-CAR-T therapy, enabling early detection and targeted treatment. However, limitations include a small, single-center sample, potential selection bias, and challenges in distinguishing colonization from active infection. Further research with larger cohorts and standardized protocols is needed to validate mNGS's clinical utility, but it holds promise for enhancing IFD management and improving CAR-T therapy safety.