Role of corneal biopsy in the management of infectious keratitis

Abstract

Purpose of Review:

To review the existing literature and investigate the role of microbiologic culture and histopathologic examination of corneal biopsies in the management of infectious keratitis.

Recent Findings:

Corneal biopsy continues to be a significantly useful tool in the diagnosis and tailored management of infectious keratitis. Several techniques can be employed for tissue collection, handling, and processing to optimize diagnostic yield and maximize safety, including emerging femtosecond laser-assisted biopsy.

Summary:

Corneal opacities represent a significant cause of global blindness, and infectious keratitis is the most common etiology. Organism identification in progressive infectious keratitis is essential for proper management. However, microbiological culture alone has a high rate of false-negative results. Records from the Bascom Palmer Eye Institute were retrospectively searched for patients between January 1, 2015, and December 31, 2019, who underwent corneal biopsy, therapeutic keratoplasty, or endothelial graft removal for infectious keratitis and had specimens bisected and submitted for evaluation with both microbiologic culture and histopathologic examination. Detection of bacteria, fungus, and mycobacteria was not statistically different between culture and histopathology. Microbiology and histopathology are complementary methods for the identification of causative microorganisms in corneal specimens with presumed infectious keratitis.

Introduction

Infectious keratitis (IK) is a common yet vision-threatening ophthalmic emergency worldwide. Even with timely management, IK can cause significant complications such as corneal scarring perforation, endophthalmitis, and even permanent vision loss. Corneal stromal scarring is a common result of the healing process resulting in corneal opacities that can cause significant visual morbidity. Corneal opacities are the 5^th leading cause of global blindness [1]; and out of all the etiologies of corneal opacities, IK is the most common, highlighting its global visual burden [2]. In the United States, the incidence of IK is estimated to be at around 27 per 100,000 person-years in non-contact lens wearers, and 130 per 100,000 person-years in contact lens wearers [3]. These numbers are estimated to be increasing in the developed world, largely due to the increased usage of contact lenses.

Successful management of IK depends on prompt diagnosis and elimination of the causative organism. However, identification of this organism can be difficult. Presently, microbiologic culture is the gold standard used to detect corneal microorganisms. It provides speciation and antibiotic sensitivity information imperative to individualized therapy. However, specimens have a relatively high rate of negative culture results, ranging from 32 to 53% in the literature [4–9]. This high culture-negative rate has been attributed to sampling error, depth of infiltrates, prior use of antibiotics, sterile or hypersensitivity infiltrates, and slow-growing organisms [6]. Moreover, cultures can take several days to grow, with bacteria requiring 3–5 days and fungi requiring at least 5–7 days to culture [7].

When initial cultures are inconclusive in the setting of progressive disease, corneal biopsy has been recommended to access deeper stromal infiltrates and additional tissue for microbial and histopathologic analysis [10, 11]. Therapeutic keratoplasty (TPK) is sometimes needed to restore the integrity of the globe, decrease the microbial load, prevent scleral extension of the infection, and provide additional tissue for analysis [10, 12]. It is important to highlight that despite microbiological analysis of corneal scrapings is the gold standard for diagnosis, many centers do not have access to these services. Meanwhile, access to histopathological analyses is more common, and can also help guide management.

Another diagnostic modality that may be complementary to microbiological and histopathological studies is molecular analysis. Polymerase chain reaction (PCR) has been used to accurately identify the offending organism and tailor management [13]. The sample preparation requires that once scrapings or a biopsy has been obtained, some is placed in a PCR tube with balanced salt solution (BSS). This is then homogenized and processed according to the laboratory protocol and analyzed with the appropriate primers for the suspected organism, in cases where molecular diagnosis will be used.

There are several methods and settings to obtain corneal biopsies [14]. The most common settings include the slit lamp, minor procedure room with patient supine using a surgical microscope, or the operating room. Deciding which setting is most appropriate depends on several factors, chief among them are physician experience and expertise, as well as patient comfort and cooperation. Regardless of the clinical setting, the surgical principles of a corneal biopsy are the same. The surgeon should try to avoid disrupting the visual axis as much as possible and minimize the risk of perforation. The tissue obtained is then sent in the proper media to the microbiology and pathology laboratories. A favored site for sampling is at the leading edge of the lesion. This allows for excision of infected and non-infected tissue and avoids the necrotic material that is usually present in the central area of the ulcer. Attempting to biopsy this necrotic area can yield a lower microorganism load and increase the risk of perforation [15]. Some of the techniques used to obtain a corneal biopsy include:

Freehand Lamellar Dissection

This technique employs a metal or diamond blade to manually dissect the selected area to be biopsied. Due to the freehand nature of this technique, there is less precise control of the depth of dissection, which theoretically increases the risk of perforation. Additionally, depending on the size of the biopsied area, this technique can induce significant corneal scarring and irregularity.

Trephination

This technique involves the use of either a dermatologic trephine or a microtrephine. The device is advanced at the selected biopsy site until the desired depth is reached, typically at the anterior stroma. At this point, a blade is used to perform a lamellar dissection and free the excised tissue. Trephination can allow for precise control of the depth of excision, therefore minimizing the risks of perforation [16]. However, larger trephines can create scars that have potential for causing visually significant refractive changes.

Lamellar Flap

This method allows for sampling of deeper infiltrates characteristically seen in mycotic infections. The process involves the formation of a V-shaped flap to expose the plane where the infiltrate is found and sampling of the deep inner-facing portion of the flap [17].

Suture Pass

This technique employs a filamentous suture to gather material from the core of the infiltrate. Braided silk is commonly used for this purpose. This is helpful in cases of deep stromal infiltrates. The suture is passed directly through the core of the infiltrate, allowing the thread to collect material. This thread can then be sent for microbiological and PCR analysis. However, since this technique does not yield a discrete tissue sample, histopathological analysis cannot be performed.

Reverse-Cutting Blade

This technique uses a reverse-cutting angled blade, which is dragged backward multiple times over the surface of the area of interest. With each pass, the tool collects epithelial and superficial stromal tissue, which remains attached at the angle of the blade and is later dissolved in phosphate-buffered saline (PBS) in a micro tissue grinder [18].

Femtosecond Laser-Assisted

This technique involves the use of the femtosecond laser to fully dissect the tissue, creating a blade-free method for corneal biopsy. Given the high precision of the femtosecond laser, this method allows for increased safety, particularly in cases with deeper infiltrates. Additionally, there is a high degree of customizability regarding the size and depth of dissection. There is also higher tissue integrity with this method due to the decreased manual manipulation. The smooth surface left behind by the femtosecond laser may also optimize healing and minimize the risk for scarring-induced refractive changes. Certain disadvantages associated with this technique include the low cost-effectiveness, and the theoretical risk of laser-induced tissue sterilization at the biopsy margins [19, 20].

Optical Coherence Tomography (OCT)-Assisted

Intraoperative anterior segment optical coherent tomography (AS-OCT) has been used in several procedures such as endothelial keratoplasties, cataract surgery, and others, due to its utility in visualizing corneal anatomy in real time. There are reports of its utility in guiding manual corneal biopsies. AS-OCT can help localize the infiltrate as well as control the depth and extent of dissection, especially in deep infiltrates [21*, 22]. This technique can be applied in the clinic setting by using AS-OCT images and three-dimensional reconstructions to map out the depth and width of the infiltrate. It can also be employed in real-time intraoperatively when using OCT-equipped microscopes.

All in all, histopathology of corneal specimens in IK can serve as an adjunct to microbiology in identifying causative organisms. These results are also sometimes available more quickly than with microbiology, allowing for rapid adjustment of patient management. Although Alexandrakis et al. and Rosa et al. compared the diagnostic ability of histopathologic and microbiologic examination of corneal specimens in a subset of 33 and 11 patients, respectively, this has not been evaluated in a large patient population [10, 23].

In this review, we compared the diagnostic yield of a large series of patients with IK who had corneal specimens submitted for both microbiologic and histopathologic evaluation at the Bascom Palmer Eye Institute.

Methods

The study was approved by the Institutional Review Board at the University of Miami. Records from the Bascom Palmer Eye Institute ocular microbiology laboratory and ocular histopathology were retrospectively searched for patients who had pathological specimens submitted from corneal biopsies or therapeutic corneal transplants between January 1, 2015, and December 31, 2019.

Cases were excluded if active IK was not suspected such as those who underwent penetrating keratoplasty for corneal graft failure, corneal scar after IK, or neurotrophic ulcers. Additionally, specimens must have been submitted for both microbiologic and histopathologic analysis.

The portion of tissue submitted for microbiologic examination was prepared as previously described [23]. The specimen was brought to the microbiology laboratory in a sterile petri dish with a few drops of BSS. Under a sterile hood, the specimen was homogenized in a mortar with sterile trypticase soy broth and plated on blood agar, chocolate agar, selective aerobic and anaerobic media, Sabouraud media, and thioglycolate broth. Selected specimens were also plated on Lowenstein-Jensen and non-nutrient agar with Escherichia coli overlay. Smears of homogenized corneal tissue were performed in all cases.

The portion of tissue submitted for histologic examination was fixed in 4% buffered formaldehyde and embedded in paraffin. Sections were cut at 6 μm, mounted on glass slides, and stained with hematoxylin and eosin, as well as special stains for bacteria (Brown and Hopps), fungi (Gomori methenamine silver), acid-fast bacilli (Ziehl-Neelsen), and Acanthamoeba (periodic acid-Schiff). Slides were evaluated using light microscopy (Olympus).

McNemar’s test was used to compare sample positivity between microbiologic and histopathologic analysis and between corneal specimen types. Fisher’s exact test was used to compare sample yield based on the type of microorganism identified. A p-value <0.05 was considered to be significant. Statistical analysis was carried out using StataIC 15.1 (StataCorp, LLC, College Station, TX).

Results

In total 213 corneal specimens were identified, including 93 corneal biopsies, 113 therapeutic penetrating keratoplasties, and 4 corneal endothelial graft specimens, from 200 patients. Patients ranged in age from 23 days to 95 years with a mean age of 59 years. 116 (58.6%) were female and 82 (41%) were male. 15 specimens were repeat samples from patients who had previously undergone biopsy.

Medication history was provided for 212 of the 213 specimens; 8 (3.8%) samples were from corneas that were not treated with anti-microbial therapy before biopsy or transplantation. Vancomycin was the most used anti-microbial (122 cases, 57.5%), followed by tobramycin (119 cases, 56.1%), fluoroquinolones (74 cases, 34.9%), acyclovir/valacyclovir (51 cases, 24.1%), natamycin (44 cases, 20.8%), and anti-amoebic therapy (26 cases, 12.3%).

The frequency of detection of microorganisms by culture and histopathology are shown in Table 1. No organism was identified in 51% of cases (n=108). There were 105 cases (49%) with an organism identified on either microbiology, histology, or both. Among these positive cultures, microbiology identified an organism, but histopathologic analysis did not identify organisms in 29% of cases (30 of 105). Meanwhile, organisms were identified on histopathology, but cultures were negative in 32% of cases (34 of 105). The remaining 39% (41 cases) were positive by both methods. There was no statistically significant difference between the overall detection rate of microbiologic culture (33.8%; 72 of 213) versus histopathologic evaluation (35%; 75 of 213) (p=0.62).

Table 1.

Detection of Microorganisms in Infectious Keratitis by Microbiological Culture and Histopathologic Examination

	Histopathology Positive	Histopathology Negative
Culture Positive	41 (19.2%)	30 (14.1%)	71 (33%)
Culture Negative	34 (15.9%)	108 (50.7%)	142 (67%)
	75 (35%)	138 (65%)	Total: 213
			P Value 0.7080a

^aMcNemar test to evaluate if the techniques perform differently on the same sample. This does not account for accuracy as one cannot assess which of the reported distributions is more accurate.

The yield of organism identification based on the type of corneal specimen is shown in Table 2. Positive results were found in 37.6% (35 of 93) of corneal biopsies, 59.3% (67 of 113) of TPKs, and 50% (2 of 4) of endothelial specimens (3 combined samples were excluded). There was no significant difference in the yield of microorganisms on histopathology versus microbiology based on the type of corneal specimen (p=0.83 for corneal biopsies, p=0.63 for TPKs, and p=1 for endothelial specimens).

Table 2.

Detection of Microorganisms in Infectious Keratitis in Microbiologic Culture and Histopathology by Specimen Type

Type of Corneal Specimen	Microbiological Culture Positive Only	Histopathologic Examination Positive Only	Both Positive	Either Positive	Both Negative	P Valueb
Corneal Biopsy (n= 93)	12 (12.9%)	11 (18.8%)	12 (12.9%)	35 (37.6%)	58 (62.4%)	1.0000
Penetrating Keratoplasty (n= 113)	18 (15.9%)	21 (18.6%)	28 (24.8%)	67 (59.3%)	46 (40.7%)	0.7493
Corneal Endothelium (n= 4)	0 (0%)	1 (25%)	1 (25.0%)	2 (50%)	2 (50.0%)	1.0000
Total (n= 210a)	30 (14.3%)	33 (15.7%)	41 (19.5%)	105 (49.3%)	106 (50.5%)

^aThree combined samples were excluded from this analysis: a corneal and conjunctival biopsy, a corneal biopsy and PK, and a PK with other ocular tissue.

^bExact McNemar significance probability

The rate of detection on microbiology versus histopathology based on the type of microorganism was also examined (Table 3). Of positive cases, 35% were bacteria, 43% fungal, 18% acanthamoeba, and 6% mycobacteria. Two cases grew multiple organisms: one case grew both fungus and bacteria while the other grew two bacteria. Detection of bacteria, fungus, and mycobacteria were not statistically different between culture and histopathology (p=0.15, p=0.80, p=1, respectively). However, histopathology detected a significantly greater number of acanthamoeba cases than culture (p= 0.03). The isolated microorganisms in the present study are listed in Table 4, with 25 different species identified.

Table 3.

Detection of Microorganisms in Infectious Keratitis in Microbiologic Culture and Histopathology by Organism Type

Organism Identified		Microbiological Culture	Histopathologic Examination	Both	P Value
Acanthamoeba sp.	n =19	10 (52.6%)	17 (89.4%)	8 (42.1%)	0.0293
Bacteria	n = 37	26 (70.2%)	20 (54.1%)	9 (24.3%)	0.1504
Fungus	n = 45	34 (75.6%)	35 (77.8%)	24 (53.3%)	0.0621
Mycobacteria	n = 6	3 (50.0%)	4 (66.7%)	1 (16.7%)	1.0000
Total	n= 107a	73 (67.0%)	76 (69.7%)	42 (38.5%)

^aTwo samples grew multiple organisms. One grew bacteria and fungus and the other grew two bacteria.

Table 4.

Identified Microorganisms

Microorganism Group		Number
Gram-positive		24
	NOS1	8
	Granulicatella adiacens	1
	Nocardia 2	1
	Staph. aureus	5
	Staph. epidermidis	3
	Strep. viridans group	3
	Strep. agalactiae	1
	Strep. pneumoniae	2
Gram-negative		13
	NOS	2
	Alcaligenes xylosoxidans	1
	Burkholderia cepacia	1
	Citrobacter freundii	1
	Capnocytophagea	1
	Haemophilus Influenzae	1
	Klebsiella pneumoniae	1
	Pseudomonas aeruginosa	4
	Serratia marcesans	1
Gram-positive rods, AFB		6
	NOS	3
	Mycobacterium abscessus	3
Fungi		45
	NOS	11
	Aspergillus	3
	Candida albicans	6
	Candida parapsilosis	1
	Culvularia species	1
	Fusarium	18
	Graphium basitruncatum	1
	Microsporidia	1
	Paecilomyces	3
Acanthamoeba		19

¹NOS = not otherwise speciated

²Partially acid-fast

Discussion

In 2000, Alexandrakis et al. reviewed the role of corneal biopsies in the management of progressive IK in 33 patients. Rosa et al. and Robaei et al. performed similar analyses in 11 and 38 patients, respectively [10, 23, 24]. To our knowledge, the present study represents the largest study of the yield of microbiologic culture and histopathologic analysis of corneal specimens from patients with clinically presumed IK to date.

In this study, neither modality was found to be superior. Rather, the two diagnostic modalities were found to be complementary. Performing microbiology or histopathology alone would potentially miss 29% or 32% of causative organisms, respectively. Our results are in agreement with similar smaller series. Younger et al. demonstrated positive discordant results between histopathology and culture in 16 of 48 cases (33%), with 10 cases positive on histopathology only. Similarly, Robaei et al showed positive discordant results with 47% (7/15 cases) positive on culture-only and 27% (4/15 cases) positive on histopathology only [24].

Detection of bacteria, fungus, and mycobacteria was not statistically different between culture and histopathology. Prior studies have shown either greater detection of bacteria by culture as compared to histology [10, 24], or equivalent results between culture and histology [11]. One reason that may explain why histopathology was as effective as microbiology in our study is the presence of a dedicated ocular pathologist with experience reading a high volume of corneal specimens. Additionally, given that the corneal specimens were taken after failed initial scrapings and the initiation of antibiotics in all but 8 specimens, the bacterial load in the specimens may have been decreased leading to a lower detection rate. Meanwhile, histopathology detected more acanthamoeba compared to microbiologic culture in the present study, which has also been reported in prior studies [11, 24]. This finding may be due to the difficulty in culturing acanthamoeba after prior anti-amoeba topical therapy as compared to the readily visible stromal cysts on histology [25].

A sub-analysis examining the yield of histopathology and microbiology by the type of corneal specimen submitted did not demonstrate that either modality was superior for corneal biopsy or TPKs. These results suggest that whether a biopsy is performed or a transplant, tissue should be submitted for evaluation utilizing both microbiology and histopathology.

Our study shares the limitations of retrospective studies. The treating physician may have a proclivity for sending the more involved half of the corneal specimen to either histology or microbiology based on personal biases as to which would yield a positive result. Additionally, patients were not randomized to each type of corneal procedure (TPK versus corneal biopsy). Finally, this series of cases also represents a selected population at a tertiary referral center in South Florida and limits the generalizability of the results to the larger population of patients with IK.

Conclusion

The present study demonstrates that corneal biopsy continues to be a significantly useful tool in the diagnosis and management of progressive IK. Microbiological and histopathological analyses performed together to evaluate corneal specimens from patients with presumed IK led to higher rates of microorganism detection compared to either modality alone. Based on these results, if the decision is made to perform a corneal biopsy or therapeutic transplant, we recommend bisecting the corneal specimen and submitting portions to both microbiology and histology to maximize their diagnostic yield and subsequent effect on patient management.

Key Points.

• IK represents a significant global cause of visual morbidity.

• Successful management and resolution of IK depends on prompt diagnosis and elimination of the causative organism.

• Microbiological analysis of corneal scrapings is the gold standard for initial diagnosing IK; however, due to its high rate of false negativity, corneal biopsy continues to be a significantly useful tool in the identification of the causative organisms and management of progressive IK.

• There are several techniques to maximize yield and safety while obtaining corneal tissue specimens.

• Microbiological and histopathological analyses performed together lead to higher rates of microorganism detection compared to either modality alone.