*7.4.2 Next-generation sequencing (NGS)*

NGS encompasses an evolving group of high-throughput sequencing technologies which allow massive sequencing of nucleic acid. The Sanger (1970s), a precursor to NGS, is a first-generation sequencing platform with high accuracy when dealing with one bacterium. In fact, the Human Genome Project (2003) was completed with the automatization of this technique. Isolated bacterial sequencing required multiple reactions with the Sanger platform, and thus, it was complex and time consuming [102]. Second-generation platforms (Illumina HiSeq 2500), although able to generate high sequence throughput data in a single reaction, they only sequenced part of the 16S gene [94, 103, 104]. Current third-generation platforms use nanopore sequencing technology directly from clinical samples to diagnose bacterial keratitis in real time and with higher accuracy [98].

Metagenomic NGS (mNGS) is an emerging approach that analyzes microbial, and host's genetic material (DNA and RNA) in samples from patients [105]. mNGS may detect all potential pathogens (bacteria, fungi, parasites, and viruses) in a clinical or environmental sample and simultaneously interrogate host responses by performing billions of reads in a single run [105, 106]. Unfortunately, the untargeted nature of this approach most likely results in host-derived reads [102].

Obtaining a rapid, real-time diagnosis of the causative microbe in bacterial keratitis will allow the ophthalmologist to initiate prompt and adequate antibiotic therapy; thus, improving the visual outcome and reducing antibiotic resistance [107]. However, test validation, reproducibility, high costs, among others, are significant drawbacks for the routine use of NGS and mNGS in clinical settings. Nevertheless, they must be considered in refractory difficult-to-identify cases of infection.

#### **7.5 In vivo confocal microscopy (IVCM)**

IVCM is a non-invasive imaging technique that allows dissection of the corneal architecture at a cellular level, providing real-time images equivalent to those obtained from ex-vivo histopathological techniques (tissue biopsy) [108]. It is currently used to evaluate corneal nerves in healthy eyes and those affected by ectatic corneal diseases, neurotrophic keratopathy, corneal dystrophies, ocular surface inflammation, contact lens wear, and infectious keratitis [108–110].

The role of IVCM in CLAIK relies on the identification of fungal hyphae and *Acanthamoeba* cysts; bacteria are too small to be visualized by IVCM [111]. Chidambaram et al. evaluated the IVCM cellular features in patients with bacterial, fungal, and *Acanthamoeba* keratitis [112]. A honeycomb-like distribution of anterior inflammatory cells in the corneal stroma was distinctive of fungal keratitis. *Aspergillus* and *Fusarium* ulcers were also associated with stromal dendritiform cells and interconnected cell processes with a stellate appearance, respectively. Bacterial keratitis was significantly associated with anterior stromal bullae and basal dendritiform cells. Normal keratocyte-like morphology was found in most eyes with both bacterial and fungal keratitis. Distinguishing features of *Acanthamoeba* included double-walled cysts, bright spots, and clusters after topical steroid use. While the keratocyte morphology was altered in 82% of bacterial (82%) and 77% of fungal keratitis, it was only abnormal in 39% of *Acanthamoeba* cases [112].

Although IVCM may be used in culture-negative cases or when the clinical diagnosis is unclear, this technique requires an experienced examiner. The rearmost since cellular features exhibited in microbial keratitis may be easily confused with nerve fibers, activated stromal keratocytes, and Langerhans cells [111]. Moreover, its small field of view precludes fair dismissal of *Acanthamoeba* cysts [113].
