Effects of formalin fixation on ex vivo confocal microscopy imaging: towards rapid pathology

Abstract

Ex vivo confocal microscopy (EVCM) rapidly images excised tissues at near-histopathological resolution, bypassing the time-consuming and labour-intensive tissue processing required for conventional histopathological evaluation. Thus far, the technique has been used to image only fresh or frozen tissues, which are challenging to transport, limiting its use to specialized centres. However, this technique could be extremely beneficial for primary health centres without histopathology laboratories, providing access to rapid pathology. For this, tissues would need to be transported in a preservative to maintain their integrity. Formalin is an excellent preservative that is readily available in clinics. We tested its effect on tissues evaluated by EVCM at 30 min, 1 h, 2 h, 4 h, 6 h, 24 h, 3 days, 5 days and 7 days after fixation and compared the images acquired with images from fresh tissue (baseline). Sixty images were acquired from normal skin tissues and tumours and analysed by a dermatopathologist blinded to the histopathology diagnosis. Normal tissue structures and tumours were correctly identified at all timepoints. Tissue became easier to handle after fixation, and photobleaching was not seen in fixed tissues. Using formalin would also preserve tissue morphology for ancillary studies. Our pilot work sets the stage for future large-scale studies and aids in widespread adaptation of this technique beyond specialized centres.

Ex vivo confocal microscopy (EVCM) rapidly images tissues at near-histopathological resolution, bypassing the time-consuming tissue processing required for conventional histopathological evaluation. Access to this technique can benefit centres without histopathology laboratories, but tissues must be transported in a preservative to maintain tissue integrity. Formalin is a readily available tissue preservative, and here we show that it has no adverse effect on tissue imaging and diagnosis using EVCM.

Ex vivo confocal microscopy (EVCM) is a novel technique in which tissues are rapidly imaged at histological resolution without the need for time-consuming tissue processing. The technique captures images in reflectance and fluorescent modes, with the reflectance signal (backscattered light) mimicking in vivo reflectance confocal microscopy images used in clinics in the USA.¹ The fluorescent signal originates from nuclei labelled with acridine orange (AO) dye. The reflectance and fluorescent signals are digitally converted into pink and purple colouring, respectively, resembling haematoxylin and eosin (H&E)-stained images. This rapid imaging process takes only a few minutes, allowing for bedside diagnosis.

In select European academic centres, EVCM complements frozen section analysis during Mohs surgery for basal cell carcinoma (BCC) margin assessment, demonstrating high sensitivity and specificity and reducing surgery time significantly.^2,3 Beyond Mohs surgery, EVCM facilitates rapid evaluation of biopsies from various organs.⁴ In the USA, the technique remains primarily constrained to research settings.

Despite its success, the limited use of this technique in select health centres is primarily a consequence of its high cost. However, its potential benefits for primary health centres, particularly those lacking laboratory infrastructure, are considerable, offering rapid pathology (within a few hours). For its wider adoption, a practical solution could involve transferring tissues from primary or remote health centres to select centres for imaging. To ensure the integrity of tissues during transfer and imaging, the use of a tissue fixative such as formalin, which is readily available in clinics, would be required. Tissue collected in formalin can be later subjected to ancillary analysis such as immunohistochemistry (IHC).

So far, only fresh or frozen tissues have been imaged by EVCM.^5,6 Fresh tissue degrades quickly, and freezing tissue requires costly, specialized laboratory setups and trained staff. Formalin fixation is therefore a potentially valuable approach for expanding the reach of advanced imaging technologies to a broader range of healthcare settings, but the effects of formalin fixation on freshly excised tissues imaged with EVCM have not been studied. Understanding these effects is crucial for optimizing tissue preservation and imaging quality, especially in scenarios where immediate imaging is not feasible.

Report

This study, conducted in the Department of Dermatology at Memorial Sloan Kettering Cancer Center, New York, involved samples of fresh normal skin (n = 2) and samples from skin tumour debulking procedures (n = 3), imaged by EVCM over 7 days. Tissues were dipped in AO for 30 s before imaging. The tissues were initially imaged in a fresh state [timepoint (TP) 0 (TP0)], to establish a baseline, before being restained and imaged after formalin fixation at the following TPs: 30 min (TP1), 1 h (TP2), 2 h (TP3), 4 h (TP4), 6 h (TP5), 24 h (TP6), 3 days (TP7), 5 days (TP8) and 7 days (TP9). To assess the impact of fixation, mosaics of the entire tissue and smaller submosaic images of normal structures and tumours were acquired in the same areas for a given tissue. The reflectance and fluorescent laser intensities were established at TP0 and were adjusted for optimal image quality at each TP. A pathologist and an EVCM expert (M.J.) guided image acquisition. After 7 days the tissues were processed for histopathology evaluation.

At each TP, tissue measurements and laser intensities were recorded to evaluate the impact of the formalin fixation. There was a slight increase (average 1 mm) in tissue size in normal fatty tissues but no change in tumour tissues. No tissue shrinkage was noted. Laser intensities had to be raised as time increased for better digital H&E (DHE) image quality, particularly for the fluorescent signal (laser intensity increase +2.1 to +6.1 nm). Minimal increases were required for the reflectance signal (ranging from +0.125 to +1 nm). Notably, tissue samples containing tumours showed greater increases in fluorescent intensities [+6.1 nm for invasive squamous cell carcinoma (iSCC), + 2.5 nm for superficial BCC (sBCC) and +2.1 nm for nodular BCC (nBCC)] (Table S1; see Supporting Information).

In total, 60 DHE images were generated from all TPs, including 30 from normal skin structures (Figure 1) and 30 from tumours [10 each of sBCC (Figure 2), nBCC and iSCC (Figure 3)]. These images were reviewed by a dermatopathologist (M.P.) with little experience of reading DHE images who was blinded to the TPs and histopathological diagnosis. A Google form was created to record the presence of normal skin structures and tumours with their diagnostic features and subtypes. Normal structures and tumours were correctly identified in all images across all TPs. No normal structure was misidentified as a tumour. Tumour typing was correct throughout, with the exception of one image of iSCC at TP0, where typing was inconclusive. Tumour features were generally identified, although peritumoral clefting and palisading were not identified in one sBCC image at TP0 (Figure 2a).

Figure 1.

Ex vivo confocal microscopy images from normal skin at (a) baseline [timepoint (TP) 0] before formalin fixation and following formalin fixation at (b) TP1 30 min; (c) TP2 1 h; (d) TP5 6 h; (e) TP6 24 h; and (f) TP9 7 days. The epidermis (arrows) and dermis (*) are clearly identified at all TPs.

Figure 2.

Ex vivo confocal microscopy images of a superficial basal cell carcinoma at (a) baseline [timepoint (TP) 0] before formalin fixation and following formalin fixation at (b) TP1, 30 min; (c) TP2, 1 h; (d) TP5, 6 h; (e) TP6, 24 h; and (f) TP9, 7 days. A tumour nodule (black arrow) with nuclear crowding (yellow asterisk) is seen at all TPs. At TP0 there is fading of the fluorescent signal (photobleaching) within the tumour nodule. Palisading (green asterisk) and clefting (red asterisk) are more prominent in (b–f) the fixed tissue than in (a) the fresh tissue.

Figure 3.

Ex vivo confocal microscopy images of invasive squamous cell carcinoma at (a) baseline [timepoint (TP) 0] before formalin fixation and following formalin fixation at (b) TP1, 30 min; (c) TP2, 1 h; (d) TP5, 6 h; (e) TP6, 24 h; and (f) TP9, 7 days. Invasion (arrows) becomes more evident in (b–f) the formalin-fixed tissue but is barely visible in (a) the fresh tissue.

Formalin fixation had positive effects, such as making the fatty tissues easier to handle, reducing tissue folding and preventing photobleaching, thus enhancing EVCM imaging quality and speed.

To our knowldege, we have successfully demonstrated for the first time the feasibility of imaging formalin-fixed skin samples (normal and tumours) using EVCM. Fixation improved tissue handling, particularly for fatty tissues, reducing movement susceptibility. It also reduced photobleaching, a common issue with confocal imaging of fluorescent-labelled fresh tissues.⁷ Photobleaching degrades image quality and necessitates restaining, delaying the imaging procedure. We hypothesize that formalin fixation stabilized the fluorescent dye,⁸ which may lessen the fluorescent laser intensity fading associated with photobleaching.

Although tissue shrinkage was expected as a result of formalin fixation,⁸ we noticed a slight increase in tissue size, likely because of tissue pressing during imaging to capture the entire epidermal margin.⁹ We needed to increase the laser intensities with increasing TPs, but the diagnostic capability remained unaffected. Normal skin structures and tumours were correctly identified at all TPs and tumour subtyping (sBCC, nBCC and iSCC), and diagnostic features (basaloid tumour nodules and cords, peritumoral clefting, palisading, keratinocyte atypia) were accurately identified in formalin-fixed tissues. However, subtype uncertainty (invasive or not) was noted in one iSCC image at TP0, and peritumoral clefting and palisading were missed in one sBCC image at TP0. Typically, biopsies are sent in formalin for pathology to maintain the integrity of the tissue, and the improved tumour typing and feature identification after fixation may be attributed to the preservation of tissue morphology. However, the small sample size in this study limits definitive conclusions.

The novice reader, despite minimal training in DHE image interpretation, readily analysed the images, showcasing the user-friendliness of EVCM-generated images for trained histopathologists.

While frozen section analysis is currently used for rapid analysis, it presents challenges such as freezing artefacts and loss of tissue during cutting.¹⁰ Formalin-fixed tissue, however, allows for ancillary studies such as IHC and molecular analysis. Access to EVCM could therefore be valuable in remote areas or in resource-poor countries lacking histopathology laboratories and tissue-freezing capabilities. At these sites, DHE images can be easily acquired by any technician and can be digitally stored and transferred to pathologists anywhere in the world for interpretation.

Our study was limited by the small sample size and the inclusion of skin samples only. Future studies are warranted to confirm the results of our pilot study using a larger sample size and studying the impact of formalin on various tissue types.

In summary, we have shown that formalin has no adverse effect on tissue imaging and diagnosis using EVCM. Our approach makes the use of EVCM analysis more accessible to health centres without a confocal microscope and to remote areas without histopathology laboratories. Collecting tissues in formalin will not only preserve tissue morphology but could also aid in subsequent histopathological analysis, something that is not feasible with fresh and/or frozen tissues.

Learning points

• It is feasible to use ex vivo confocal microscopy (EVCM) to image formalin-fixed skin (normal and tumour) samples.

• Formalin has no adverse effect on image quality and diagnosis of tissues imaged by EVCM.

• Formalin-fixed digital haematoxylin and eosin images generated by EVCM can be readily interpreted by a trained histopathologist.

• Formalin improves tissue handling and imaging by EVCM.

• This approach can widen the uptake of EVCM in remote areas without histopathology laboratories, enabling rapid histopathology diagnosis.

Funding sources

This research was funded by a grant from the National Cancer Institute/National Institutes of Health (P30-CA008748) made to the Memorial Sloan Kettering Cancer Center.

Data availability

The data that support the findings of this study can be found within the article.

Ethics statement

This study was reviewed and approved by the Institutional Review Board of Memorial Sloan Kettering Cancer Center (IRB protocols #08-006 and #06-107).

Patient consent

Consent for the publication of recognizable patient photographs or other identifiable material was obtained and all patients gave consent with the understanding that this information may be publicly available.

Supporting Information

Additional Supporting Information may be found in the online version of this article at the publisher’s website.