Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 11.
Published in final edited form as: Bone. 2021 Nov 4;154:116254. doi: 10.1016/j.bone.2021.116254

Rapid histological imaging of bone without microtome sectioning using nonlinear microscopy

Tadayuki Yoshitake 1, Seymour Rosen 2, Lucas C Cahill 1,3, Simon Lamothe 2, Ashley Ward 2, James G Fujimoto 1
PMCID: PMC9832301  NIHMSID: NIHMS1858866  PMID: 34743041

Abstract

Tissue preparation for histologic evaluation of bone is particularly lengthy, limiting its use in intraoperative or intraprocedural histological evaluation. Nonlinear microscopy (NLM) is an optical sectioning microscopy method that can visualize pathology in freshly excised tissue without requiring physical microtome sectioning. This study describes a rapid protocol for NLM imaging of bone and associated cartilage. NLM imaging was performed on 71 specimens of normal bone as well as arthritic, malignant and inflammatory bone tissue from 40 patients who underwent joint replacement, amputation, bone marrow biopsy or autopsy. Specimens ranged in size from core needle biopsies to transections of entire femoral heads. Specimens were stained with acridine orange and sulforhodamine 101, nuclear and cytoplasmic/stromal fluorescent dyes, for 5 minutes, then rinsed for 30 seconds. NLM fluorescent images were displayed using colors analogous to hematoxylin and eosin (H&E) to facilitate interpretation. Pathologists examined NLM images of the specimens in real time by rapidly translating the specimen to areas of interest, similar to a standard transmission light microscope. By adjusting the NLM focus depth, images from a few-μm-thick layer could be obtained down to ~100 μm below the tissue surface, analogous to serial sectioning. Following real-time NLM imaging, the tissue was processed for conventional paraffin histology, and H&E slides were compared to recorded NLM images. Similarities and differences between NLM and paraffin H&E were assessed. NLM enabled visualization of normal bone architecture, including the lamellar matrix and osteocytes of trabecular bone, articular cartilage, as well as osteoarthritis, osteomyelitis, and malignancy mimicking the paraffin H&E. Differences such as changes in cell border sharpness, cellular and nucleolar size, and color patterns were noted, suggesting that training is required for accurate evaluation of bone pathology with NLM. Irregular surface contours and debris generated by gross tissue preparation of bone can make some regions difficult to evaluate with NLM, but the ability to perform rapid three-dimensional translation and sub-surface imaging reduced these problems. NLM is a promising technique for rapid evaluation of bone pathology. Further studies assessing diagnostic performance are warranted.

Keywords: Nonlinear microscopy, Two photon excitation microscopy, Histology of bone, Rapid tissue examination

2. INTRODUCTION

Histologic evaluation of bone typically requires fixation, decalcification, and paraffin processing to enable microtome sectioning [1]. If decalcification is required, the procedure is especially time intensive and rapid analysis using frozen section is not feasible. The ability to rapidly process and visualize bone tissue pathology in such cases would enable intraoperative consultation for surgical margin evaluation [2-11], intraprocedural analysis of biopsies to assess adequacy [12-16], tissue triaging for specific assays, as well as near real-time access to diagnostic information. Optical sectioning fluorescence microscopy technologies, including two-photon excitation nonlinear microscopy (NLM) [17-21], confocal fluorescence [22-25], structured illumination [26,27], and light sheet microscopy [28,29] are being investigated for rapid histologic evaluation of freshly excised tissue. NLM is of particular interest because blind reading studies of NLM compared to paraffin embedded hematoxylin and eosin (H&E) histology demonstrated high sensitivity and specificity for detecting breast and prostate cancer [17,20]. The NLM protocol enables rapid staining of tissue with nuclear and stromal/cytoplasmic fluorophores and images can be displayed in an H&E color scale to facilitate interpretation. Specimens can be rapidly imaged without microtome sectioning using a scanned, focused laser beam which excites fluorescence from a few micron axial range of tissue. Furthermore, by adjusting the focus depth, NLM can image up to ~100 μm below the tissue surface, analogous to serial sections. In this study, we describe a method for rapid tissue preparation and visualization of bone using NLM. We describe imaging of normal bone and cartilaginous tissue, as well as orthopedic surgical and biopsy specimens with osteoarthritis, osteomyelitis and malignancy.

3. METHODS

NLM imaging was performed on 71 specimens of discarded bone (not required for clinical diagnosis) from 40 patients who underwent joint replacement (hip or knee), amputation (hand, arm or foot), bone marrow biopsy, or autopsy. The protocols were approved by the Beth Israel Deaconess Medical Center Committee on Clinical Investigations and Institutional Review Board and the Massachusetts Institute of Technology Committee on the Use of Humans as Experimental Subjects. Informed consent was waived by both committees.

The bone specimens, with the exception of biopsies, were cut to expose the surface of interest using a scalpel, saw or hammer (Fig. 1a) in the same manner as conventional histology processing. The surface was gently brushed and/or rinsed with water to remove debris from cutting. Biopsy specimens were taken using a 16-gauge needle during autopsy or excess tissue from abundant bone marrow biopsy specimens. Additional cutting, brushing and rinsing were not performed on the biopsy specimens because of their small size. The specimens were stained in acridine orange (40 μg/ml) and sulforhodamine 101 (40 μg/ml) in a 1:1 ethanol:water solution for 5 minutes and rinsed for 30 seconds in saline to remove excess staining solution. Acridine orange provides nuclear contrast, while sulforhodamine 101 provides stromal and cytoplasmic contrast, analogous to hematoxylin and eosin. The protocol is similar to a previously published NLM staining protocol for breast and prostate tissue [18,19], but requires longer staining time because both cortical and cancellous bone were found to be stained weakly compared with soft tissues. Immediately after staining, the specimens were placed on a glass specimen holder, covered with a lid to block room light, and imaged using an inverted NLM microscope (Fig. 1b) [30].

Figure 1. Tissue processing and NLM imaging.

Figure 1.

Open in a new tab

(a). Specimens were cut to expose surfaces of interest, washed to remove surface debris, stained in a solution of acridine orange (AO) and sulforhodamine 101 (SR101) for 5 minutes, and rinsed for 30 seconds in saline to remove excess stain. (b). The specimens were placed on a glass window of a specimen holder in an inverted NLM instrument. The specimens were translated in the horizontal (x-y) plane to regions of interest and the microscope focus depth, vertical (z-axis) translation, could be adjusted to image irregular surfaces and subsurface features. A position indicator was displayed on a white-light image of the gross specimen to guide NLM image positioning. The specimen holder allows imaging of up to 7 cm x 10 cm specimens. (c). The NLM images were displayed in an H&E color scale at 16 frames per second.

The NLM instrument had a 10x, 0.45 numerical aperture (NA) air objective with a 2 mm field of view (FOV) (CFI Plan Apo Lambda, Nikon) and a 5x, 0.25 NA air objective with a 3.4 mm FOV (Fluar, Carl Zeiss Microscopy) that could be rapidly interchanged under servo control [30]. Specimen holders for the wide FOV (5x and 10x) objectives had rigid glass windows (400 μm thickness) that accommodate up to 7 cm x 10 cm size specimens. Alternatively, the NLM instrument had an option to manually change to a higher magnification using a 20x, 0.75 NA air objective with a 1 mm FOV (CFI Plan Apo Lambda, Nikon). Specimen holders for the 20x high magnification objective required thin glass windows (170 μm, standard cover slip thickness) to avoid aberration and achieve the desired resolution. The thin glass is fragile and easily bendable, so it is difficult to prepare large windows. The specimen holders for the 20x objective were therefore designed to accommodate specimens with the size of a histology cassette or biopsy.

A femtosecond laser at 1 μm wavelength was used to excite two-photon fluorescence at the focus of a raster scanned beam. Digital magnification up to 40x (or 80x for the 20x objective) was available by changing the scan area and speed. The fluorescent signals from acridine orange and sulforhodamine 101 were collected and separated using a low-pass dichroic beam splitter (T588lpxr, Chroma Technology) and two emission filters (ET540/40m, Chroma Technology and FF01–650/60, Semrock), and then detected using two photomultiplier tubes (7422-40p, Hamamatsu). NLM images were displayed at 16 frames per second in a color scale that resembles H&E histology (Fig. 1c) [31].

The specimens were imaged in real time on NLM and evaluated in a manner similar to standard transmission light microscopy of histology slides. A pathologist examined the specimen by translating it horizontally (x-y plane) to regions of interest and changing magnification. A white-light image of the gross specimen surface was used to guide positioning (Fig. 1b). The microscope focus can be vertically adjusted (z-axis) to image tissue below the surface, analogous to serial sectioning. The rapid focus depth adjustment was also important for imaging the three-dimensional structure of bone as well as the irregular specimen surfaces caused by preparation, saw marks and cracks.

During real-time evaluation, the NLM images were recorded along with position data for subsequent analysis. The NLM images were then stitched together using commercial software (Microsoft Image Composite Editor) to provide a single 2D image of the entire real-time NLM imaging sequence for review/publication. After NLM imaging, the specimens were processed for standard histology and paraffin H&E slides were digitally scanned (20x magnification, Leica Aperio slide scanner). A pathologist performed an unblinded comparison of the digital NLM images and corresponding paraffin H&E slides to identify similarities and differences in normal tissue and pathology.

4. RESULTS

4.1. Imaging large specimens vs. biopsies

NLM enabled the visualization of histological features of bone such as bone architecture, degenerative changes, infection, and malignant alterations in specimens ranging from biopsies to multi-centimeter femoral heads.

Figure 2 shows a transected, entire femoral head cross section (4.5 cm width) imaged with NLM prior to decalcification and paraffin processing for standard histology. The white-light image of the gross specimen (Fig. 2a) was used to guide NLM imaging. The articular cartilage at the periphery was imaged to assess osteoarthritis, and the internal trabecular bone network and bone marrow could also be imaged. The specimen surface was irregular and could not be placed flat against the glass NLM specimen holder. However, the NLM focus depth could be adjusted (vertical or z-axis specimen position) to compensate for surface contour variations up to ~ 500 μm and to image up to ~100 μm deep in tissue. The stitched sequence of real-time images in Figure 2b shows the areas imaged by NLM with focus depth adjustment. The regions of the femoral head shown (periphery cartilage and trabecular network) were evaluated in <5 minutes. Corresponding paraffin H&E histology required 4 separate slides (Fig. 2c). The dissection and decalcification/paraffin processing for histology slide preparation resulted in partial detachment of fragile tissue regions.

Figure 2. NLM imaging of large specimens.

Figure 2.

Open in a new tab

An example showing NLM imaging of a transected femoral head (4.5 cm width). (a) A white-light image of the gross specimen was used to guide NLM positioning. (b) Stitched sequence of NLM images showing the pathologist examination which included the articular cartilage and the trabecular network. Digital resolution of the figure is limited because the original data size is too large for publication. (c) Corresponding paraffin H&E histology. The specimen was dissected into quadrants for 4 processing (30 x 26 mm) cassettes. Detachment of tissue caused by tissue processing for slide preparation is seen (top left corner).

NLM can rapidly image small specimens such as biopsies. Figure 3 shows 16-gauge needle core biopsies of bone marrow obtained from discarded tissue. Although the biopsy specimens were small, their cylindrical shape required adjusting the focus depth to image regions of interest. The ~2-mm-wide x 2-cm-long specimen was imaged in 2 minutes (Fig. 3a and 3b). Red bone marrow with granulopoiesis and erythropoiesis having an approximate cellularity of 50% are shown in Figure 3c, as well as in the corresponding paraffin H&E histology slide (Fig. 3d). Endosteal lining along with trabecular bone are also seen in Figure 3e. The white area of trabecular bone shows imaging deep inside of trabecular bone, where fluorescent signals are limited because of limited fluorophores and/or excitation laser light penetration into tissue. Bone marrow in some of the marrow spaces is missing because of aspiration artifacts due to the needle biopsy procedure. Nuclear details are well represented in NLM, but cellular borders are less sharp. Cell cytoplasm is darker and granulocyte granules are more distinctive than in paraffin H&E.

Figure 3. Needle core biopsy imaging with NLM.

Figure 3.

Open in a new tab

A white-light gross image (a) and corresponding composite NLM image generated from the real-time image sequence (b) of a 16-gauge core needle biopsy from pelvic bone marrow. The biopsy surface was irregular and necessitated adjustment of the focus depth during imaging. Out of focus frames were excluded from the image composition to minimize image processing artifact. Red bone marrow with granulopoiesis and erythropoiesis with an approximate cellularity of 50% are shown in images with NLM (c) and paraffin H&E histology (d). NLM has less sharp cell boundaries, darker cytoplasm, and more distinctive granulocyte granules than paraffin H&E images. (e). NLM shows endsteal lining. NLM with a 20x objective enables visualization of sub-cellular details more clearly such as (f) osteocytes embedded in trabecular bone. (Scale bars = (a, b). 3 mm, (c-f). 50 μm) T: Trabecular bone; Ed: Endothelial cells; MS: Marrow space; Oc: Osteocyte; Nb: Normoblast.

Higher resolution imaging using a high NA objective and a thin imaging window are shown in Figures 3f and 4. The NLM images of core needle biopsy specimens visualize sub-cellular details with the high resolution 20x objective. The thin optical sectioning with the high NA 20x objective enables visualization of osteocytes embedded in trabecular bone with its lacunar and canalicular spaces (Fig. 3f). Figure 4 demonstrates NLM imaging of bone marrow biopsy specimens, visualizing different types/stages of hematopoietic cells. Figure 4a shows hematopoietic marrow. All three cell lineages, myeloid, erythroid and megakaryocytic, are represented. Neutrophils and eosinophils are readily identified, including different maturational stages of neutrophils such as myelocytes, metamyelocytes, band forms, and mature neutrophils. Numerous erythroblasts at all stages of maturation are present as well. The cells with abundant cytoplasm and clock face chromatin are plasma cells. Stromal cells, endothelial cells and a venous sinus are clearly visualized. Figure 4b shows a group of eosinophils and neutrophils at different stages of maturation, and a lymphocyte. Eosinophils are distinguishable from neutrophils by their larger and redder “eosinophilic” cytoplasmic granules. Orthochromatic erythroblasts are also easily distinguishable. Figure 4c shows a multinucleated (normal) megakaryocyte.

Figure 4. High resolution bone marrow imaging with NLM.

Figure 4.

Open in a new tab

20x objective enables visualization of finer details of hematopoietic cells near the marrow such as (a) cells associated with all three cell lineages, myeloid, erythroid and megakaryocytic, (b). eosinophils and neutrophils at different stages of maturation, and (c). megakaryocytes. (Scale bars = (a-c). 50 μm) Nt: Neutrophils; Eo: eosinophils; My: myelocytes; Mm: metamyelocytes; Bn: band forms neutrophils; Eb: erythroblasts; Pc: plasma cells; St: stromal cells; Ed: endothelial cells; Vs: venous sinus; Ly: lymphocyte; Mk: Megakaryocytes.

4.2. Bone architecture

NLM can visualize the characteristic architecture of trabecular bone from the femoral head and knee joint, as shown in Figure 5. Figure 5a is a stitched sequence of NLM images showing the trabecular network and displaying trabecular lamellar bone with associated osteocytes. The layered lamellae structure and osteocytes are clearly recognized within the trabecular bone (Fig. 5b). In articular cartilage, NLM can visualize the basophilic matrix, clusters of chondrocytes, and the tidemark (Fig. 5c). NLM imaging of bone marrow enables identification of maturing trilineage hematopoiesis (Fig. 5d). Synovium is also readily visualized, including synovial epithelium and small vessels (Fig. 5e). A corresponding paraffin H&E histology slide image is shown in Figure 5f. Lacunae around osteocytes are white and much more distinct in paraffin H&E than in NLM. The adipose tissue in the marrow spaces is also better visualized in paraffin H&E.

Figure 5. NLM of bone architecture.

Figure 5.

Open in a new tab

(a) NLM images of bone architecture that includes lamellar structure of trabecular bone and osteocytes. The bone surface is irregular and requires images from different focus depth positions to be combined in order to generate a composite image. (b) High magnification NLM image showing trabeculae containing osteocyte nuclei. (c) NLM image of basophilic matrix and chondrocyte clusters of articular cartilage. (d) NLM image of bone marrow. Megakaryocytes, granulopoiesis, and erythropoiesis are visualized. (e) NLM image of synovium. (f) Corresponding paraffin H&E histology of trabecular bone specimen (a). (Scale bars = (a). 300 μm, (b). 100 μm, (c). 100 μm, (d). 50 μm, (e). 200 μm) MS: Marrow space; RC: Resorption cavity; T: Trabecular bone; CC: Chondrocytes; TM: Tidemark; AP: Adipose; MK: Megakaryocytes.

4.2. Osteoarthritis

Figures 6a and 6b show NLM images and corresponding paraffin H&E histology of articular cartilage from a knee. NLM images have more pronounced eosinophilic and basophilic staining when compared with paraffin H&E. Irregularity of the cartilage surface, which is a feature of osteoarthritis, is apparent in both the NLM and paraffin H&E. Chondrocytes proliferation and tidemark duplication are also more prominent in the NLM image (Fig. 6a, toward left) than in paraffin H&E histology.

Figure 6. NLM of articular cartilage.

Figure 6.

Open in a new tab

Both NLM (a) and paraffin H&E histology (b) show surface irregularity/clefts with the surface cartilage replaced by fibroconnective tissue (center). The staining is more pronounced in NLM than H&E histology, with basal cartilage calcification evident. (Scale bar = 500 μm)

4.3. Osteomyelitis

Figure 7 shows NLM images and corresponding paraffin H&E histology of osteomyelitis in a knee. NLM shows neutrophils with multilobed nuclei on the bone surface (Fig. 7a). Corresponding H&E images (Fig. 7b) show the margin of bone and infection. Figure 7c and 7d show NLM images with focus depth settings 50 μm apart, which visualize reactive changes with partial trabecular dissolution (black arrow). Osteoblasts lining the edge of trabecular bone, multinucleated osteoclasts, fibroblasts and histiocytes are seen inside the cavity. (Fig. 7e, corresponding H&E histology)

Figure 7. NLM of osteomyelitis.

Figure 7.

Open in a new tab

The NLM image (a) shows large numbers of neutrophils within the bony space (red square). Corresponding paraffin H&E histology (b) includes the bony margin. The NLM images (c, d) show reactive changes with partial trabecular dissolution (black arrow). The two images are 50 μm apart in depth. The bony edge is seen in both NLM and corresponding H&E histology (e) but the osteoblastic proliferation and occasional osteoclasts are more distinct in the NLM image. (Scale bars = 100 μm) T: Trabecular bone; Oc: Osteoclast; Ob: Osteoblast.

4.4. Malignancy in bone

Many cancers can spread to bone and bone metastases are more common than primary bone cancer in adults. Figure 8 shows example NLM images and paraffin H&E histology of bone metastases and a primary bone tumor. Figure 8a shows an NLM image of squamous cell carcinoma metastatic to the bone. Cellular aggregates of atypical epithelial cells with keratin pearl formations, which are characteristic of squamous cell carcinoma, are visualized. Figure 8c shows NLM of a renal carcinoma metastasis. Nested cells with cleared out cytoplasm are characteristic of clear cell renal cell carcinoma. Figure 8e shows NLM of a pancreatic cancer metastasis. Glandular formation can be identified. Figure 8g shows NLM of a primary bone sarcoma depicting a cellular spindle cell neoplasm, which is consistent with a high-grade sarcoma. Figure 8b, d, f, and h show corresponding paraffin H&E histology images. Typical appearance of malignancies, such as more prominent nuclei and darker cytoplasm, are similar in NLM to paraffin H&E. Differences between NLM and paraffin H&E include visualization of the stroma, which is more prominent in paraffin H&E than NLM images and visualization of erythrocytes, which are not clearly defined in NLM images.

Figure 8. NLM of representative malignancies.

Figure 8.

Open in a new tab

(a) NLM of metastatic squamous cell carcinoma showing atypical cells with abundant cytoplasm and keratin pearl formation, and (b) corresponding paraffin H&E histology. (c) NLM of clear cell renal cell carcinoma metastatic to the femoral head, and (d) corresponding paraffin H&E histology. Clusters of cells with cleared out cytoplasm are seen in both H&E and NLM, but erythrocytes are defined only in H&E. (e) NLM of pancreatic cancer metastatic to the femoral head, and (f) corresponding paraffin H&E histology. A cellular proliferation with enlarged pleomorphic nuclei and glandular structures can be seen. (g) NLM of a primary bone sarcoma of the femoral head showing an atypical spindle cell proliferation, and (h) corresponding paraffin H&E histology. (Scale bars = 100 μm)

5. Discussion

NLM enabled visualization of bone within minutes, without requiring decalcification or microtome sectioning. NLM could visualize normal bone architecture, articular cartilage, osteoarthritis, osteomyelitis and malignancy involving bone.

The structural context of trabecular bone containing osteocytes with intervening bone marrow was readily distinguishable by NLM. The NLM images of cartilage, including the basophilic matrix and chondrocytes, were remarkably equivalent to paraffin H&E. Neutrophils with multilobed nuclei on the bone surface were clearly visualized on NLM. In carcinomas and sarcomas, the morphologic features of malignancy are maintained. The clear cell appearance of renal cell carcinoma was similar in NLM to paraffin H&E and the membranes of the poorly differentiated sarcoma were well delineated by NLM. High magnification NLM images of small needle biopsy specimens clearly visualized sub-cellular details of pathology, including characteristic nuclear patterns and cytoplasmic staining of hematopoietic cells and osteocytes embedded in trabecular bone with lacunar and canalicular spaces.

Despite the similarities between NLM images and paraffin H&E histology, several differences were noted. Cellular borders were less sharp in NLM, cells appeared rounder and larger and nucleoli were more prominent than in paraffin H&E histology. Erythrocytes are not clearly visualized in NLM, but prominent in paraffin H&E histology. Some histologic features that rely on well-establish color patterns in paraffin H&E histology (such as the color of cytoplasm or the “blandness” of nuclei) were less evident on NLM. These differences are due to a combination of factors: dehydration, decalcification, and paraffin processing used in H&E histology result in cellular shrinkage and cytoplasmic contraction. Differences in staining specificity of acridine orange and sulforhodamine 101 in fresh tissue vs. hematoxylin and eosin staining of microtomed tissue specimens in histology slides were also observed. For example, cytoplasmic and nuclear elements can be more darkly stained and marrow granulocytes staining is distinctive and bright using acridine orange and sulforhodamine 101 in NLM.

NLM enabled real-time imaging of large specimens by rapidly translating the specimen horizontally (x-y plane) and adjusting the focus depth (z-axis) using 5x and 10x objectives and specimen holders with a rigid large glass window. Large specimens could be imaged without dissection into the smaller sizes typically required for microtoming, enabling imaging of intact bone with attached soft tissue as well as whole tumor margins. Real-time adjustment of the focus depth enabled imaging of subsurface tissue and complex architectural structures such as trabecular bone, glandular structures, branching components, and cavities.

All NLM images except Figs. 3f and 4 were acquired using a 10x, 0.45 NA microscope objective and a 400 μm thick glass for the imaging window. The 10x wide-field imaging objective has an axial resolution 8.7 μm [30], which is thicker than histology slide sections (5 μm). Therefore the 10x NLM images appear to have lower resolution compared with the high magnification (20x) view of the corresponding H&E histology slides. However, NLM can achieve higher resolution imaging by using a high NA 20x objective and thin glass window as shown in Figs. 3f and 4. Standard cover slips are too fragile and bendable to hold multi-centimeter square large specimens, limiting the size of imaging window for high resolution NLM imaging. Biopsy or histology cassette size specimens are easily accommodated by specimen holders with thin glass. Large soft tissue specimens can be easily cut into small pieces without appreciably increasing imaging time, if high resolution is required. However, large tissue specimens involving bone are not more time consuming to cut into smaller sizes.

NLM generates images at a given focal depth position. A flat and clean tissue surface, which can be placed flat against the specimen holder glass, maximizes NLM image quality. However, bone must be coarsely cut with tools like saws and hammers, which create surface fractures and debris. Brushing or washing the surface prior to NLM reduced surface debris and irregularity, while real-time focus depth adjustment can keep regions of interest in focus. However, imaging bone with NLM is still more challenging than soft tissue.

The ability of NLM to acquire images from different focal depths can pose a challenge for achieving exact correspondence between NLM and histology slides. Mismatches in imaging/sectioning plane between NLM and histology can cause discordance, if pathologies are focal. However, NLM can acquire multiple images at continuously adjustable focal depths and therefore achieve more comprehensive sampling than possible with isolated histological sections. For example, NLM can visualize the curved surface of bone, three dimensional structure of trabecular networks, and bone marrow embedded deep inside of marrow spaces. Serial sectioning can provide denser sampling for standard histology, but is not commonly performed because it is time-consuming and costly. The diagnostic performance of NLM will need to be validated in blinded reading studies compared with standard histology. However, if NLM diagnostic accuracy is high, it has the potential to achieve higher accuracy than histology because sampling errors can be reduced.

The NLM images are generated using fluorescent signals from acridine orange / sulforhodamine 101 and are displayed in the color scale analogous to H&E histology to facilitate pathologist interpretation and minimize the training required. As an example, previous blinded NLM reading study of prostate cancer by three pathologists achieved pooled sensitivity and specificity of 97.3% (93.7–99.1%; 95% confidence interval) and 100.0% (97.0–100.0%) with only three hours of training [20]. Bone tissue imaging involves more frequent focus depth adjustment because of surface irregularities, therefore more training will likely be necessary. However, we expect that the total training time required will be on the order of hours rather than days.

Further technological improvements will be necessary to develop ergonomic and robust NLM evaluation protocols for bone. Cutting tools like diamond bone saws or ultrasonic bone scalpels will be helpful to minimize debris, saw marks and cracks. Automated z-axis focus adjustment can compensate the surface irregularity from grossing as well as the innate three-dimensional structure of bone without delays associated with manual focus adjustments. Using a low magnification (low NA) objective, NLM imaging is less sensitive to surface irregularity because the optical section is thick (e.g., 27 μm, with 5x objective [30]). However, the low magnification reduces cellular/subcellular resolution and rapidly interchangeable higher magnification objectives are needed to achieve finer resolutions to visualize details of pathology.

The NLM tissue processing and imaging protocol is non-destructive, and does not consume or alter tissue by decalcification and physical sectioning. After grossing to expose areas of interest, the fluorescent staining is rapid, requiring only ~5 minutes. After NLM imaging, the tissue can be used for standard assays such as H&E histology, immunohistochemistry and fluorescence in situ hybridization [18].

The NLM system and tissue processing protocol were originally developed for rapid H&E-like histological evaluation of fresh tissue for use in intraoperative evaluation applications. Acridine orange and sulforhodamine 101 were chosen because they can label fresh surgical tissue specimens rapidly, with specificities very similar to H&E histology. However, NLM is a versatile technology and multiple contrast agents with different specificities have been demonstrated in other biological/biomedical studies [32-35]. Future studies will investigate other contrast agents, which may further extend NLM imaging capability by improving visualization and diagnostic accuracy for specific pathologies.

NLM imaging might benefit applications where rapid histological evaluation is needed, such as intraoperative surgical margin evaluation of bone resection, rapid and non-destructive determination of biopsy adequacy, and triaging fresh tissue for specific assays. For example, intraoperative surgical margin evaluation for osteomyelitis surgery may decrease positive margin rates by guided additional resection. The positive margin rate of amputation for osteomyelitis is currently high (40.7%) and associated with higher likelihood of poor outcome (81.8% for positive margin vs 25% for negative margin) [6]. Postoperative antibiotic treatment can be longer and more intense for patients with positive surgical margin. If NLM can provide accurate osteomyelitis margin assessment intraoperatively, NLM-guided resection may reduce positive margin rates, decrease the burden of postoperative care (antibiotic treatment and hospitalization), and improve patient outcome.

Achieving clear margins is also critical in oncological surgery. For example, surgical margins of head and neck cancers are often intraoperatively evaluated with frozen sections. However, head and neck cancer specimens often involve bone, making sampling difficult and limiting accuracy of tissue orientation for frozen section [36]. NLM can evaluate intact surgical margins from tissue specimens including bone and may improve sampling and tissue orientation accuracy compared to current frozen section protocols.

Non-destructive adequacy testing for bone needle biopsy is another potential application of NLM. A retrospective study which investigated CT-guided bony needle biopsy found a high frequency of cases with non-diagnostic cores (21.6%) [16]. Amongst these cases, ~40% were followed up with surgical biopsy, but the majority of biopsies were benign (>85%). Other, low risk patients (~50%) had clinical watchful follow-up, and most patients did not developed cancer (~98%). Repeat core needle biopsy is shown to be effective to improve diagnostic yield (38%, converted from non-diagnostic to diagnostic). NLM can non-destructively evaluate needle biopsy specimens, and has the potential to perform intraprocedural adequacy testing to reduce sampling error and increase diagnostic yield. This will benefit patients by decreasing surgical biopsies or reducing anxiety caused by a long clinical follow-up with unknown diagnosis.

NLM also may be useful for triaging large specimens. For example, histological examination of arthroplasty specimens is costly and sometimes only gross examination is performed because of the large specimen size and low incidence of unexpected findings [37]. However, the discrepancy rate between preoperative diagnosis and postoperative histology is shown to be small, but not negligible [38]. The example in Fig. 2 shows NLM imaging of an entire femoral head cross section can be performed in ~10 minutes (5 minutes staining + 5 minutes imaging, excluding standard grossing/sawing time). Therefore, NLM might be able to decrease the cost of histological examination while maintaining the diagnostic accuracy of standard histology.

NLM might also be a useful clinical research tool which is complementary to radiographic analysis of bony microarchitecture or other biological studies. NLM could be used to guide sampling/microdissection of viable cells from targeted regions of interest. Bone growth and regeneration have been investigated with the ingestion of fluorophores such as calcein or tetracycline [39,40], which may have synergetic benefit when combined with NLM imaging technology.

NLM is extensively used in biological/biomedical research on cell cultures or small animals. However, the NLM instrument in this study was custom designed for rapid imaging of pathology specimens [30]. Research NLM instruments use high-cost Ti: Sapphire femtosecond laser light sources, however our pathology imaging protocol is designed to be compatible with Yb-fiber femtosecond lasers at ~1 μm wavelength which have much lower cost [18].

This study has several limitations. The range of pathologies and sample size are limited. For example, two of the most common cancers metastasize to bone, breast and prostate cancers, are not included because of limited tissue availability, although NLM imaging of breast and prostate cancer was previously investigated and had high diagnostic accuracy [17,20]. NLM images were read unblinded to paraffin H&E in order to assess similarities and differences. Therefore, the ability to diagnose bone pathology on NLM was not demonstrated. Sensitivity and specificity were not measured, and statistical analysis was not performed. More comprehensive studies are needed to evaluate NLM performance for specific applications, and specimen preparation and imaging times must be within clinically acceptable limits.

In summary, NLM can provide rapid histological visualization of pathology in bone tissue, and may enable rapid diagnosis and intraoperative/procedural consultation. Rapid histological evaluation of bone may have applications in orthopedic surgery, bone biopsy, and cancer management. Future larger scale studies on specific pathologies which compare blinded evaluation of NLM with paraffin histology to assess sensitivity, specificity, and inter-reader agreement are warranted.

Highlights.

  • Nonlinear microscopy enables imaging of bone without requiring microtome sectioning

  • Tissue surfaces can be examined after rapid fluorescent staining

  • Images are displayed in a color scale similar to H&E to facilitate interpretation

ACKNOWLEDGEMENTS

We gratefully acknowledge Dr. Jeff Broker and Alex Cable at Thorlabs for helpful scientific advice and providing microscope hardware. We also greatly thank Dr. German Pihan for insightful comments related to hematopathology. This work was supported in part by the National Institutes of Health programs R01-CA178636 and R01-CA252216.

Glossary

NLM

nonlinear microscopy

H&E

hematoxylin and eosin

NA

numerical aperture

FOV

field of view

Footnotes

COMPETING FINANCIAL INTERESTS

JGF, TY, and LCC are inventors on US patent No. 10416434: Method and apparatus for imaging unsectioned tissue specimens. The intellectual property is owned by M.I.T. and is not currently licensed. The remaining authors declare no conflict of interest.

REFERENCES

  • [1].Donath K, Rohrer M, Bone Sectioning Using the Exakt System, in: Handb. Histol. Methods Bone Cartil, Humana Press, 2003: pp. 243–252. 10.1007/978-1-59259-417-7_16. [DOI] [Google Scholar]
  • [2].Kawaguchi N, Matumoto S, Manabe J, New method of evaluating the surgical margin and safety margin for musculoskeletal sarcoma, analysed on the basis of 457 surgical cases, J. Cancer Res. Clin. Oncol 121 (1995) 555–563. 10.1007/BF01197769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Kawaguchi N, Ahmed AR, Matsumoto S, Manabe J, Matsushita Y, The Concept of Curative Margin in Surgery for Bone and Soft Tissue Sarcoma, Clin Orthop. 419 (2004) 165–172. 10.1097/01.blo.0000116308.86761.f5. [DOI] [PubMed] [Google Scholar]
  • [4].Bilodeau EA, Chiosea S, Oral Squamous Cell Carcinoma with Mandibular Bone Invasion: Intraoperative Evaluation of Bone Margins by Routine Frozen Section, Head Neck Pathol. 5 (2011) 216–220. 10.1007/s12105-011-0264-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Lue N, Kang JW, Yu CC, Barman I, Dingari NC, Feld MS, Dasari RR, Fitzmaurice M, Portable optical fiber probe-based spectroscopic scanner for rapid cancer diagnosis: A new tool for intraoperative margin assessment, PLoS One. 7 (2012). 10.1371/journal.pone.0030887. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Atway S, Nerone VS, Springer KD, Woodruff DM, Rate of Residual Osteomyelitis after Partial Foot Amputation in Diabetic Patients: A Standardized Method for Evaluating Bone Margins with Intraoperative Culture, J. Foot Ankle Surg 51 (2012) 749–752. 10.1053/j.jfas.2012.06.017. [DOI] [PubMed] [Google Scholar]
  • [7].Nieberler M, Häusler P, Drecoll E, Stoeckelhuber M, Deppe H, Hölzle F, Kolk A, Wolff K-D, Resting MR, Weirich G, Evaluation of intraoperative cytological assessment of bone resection margins in patients with oral squamous cell carcinoma, Cancer Cytopathol. 122 (2014) 646–656. 10.1002/cncy.21428. [DOI] [PubMed] [Google Scholar]
  • [8].Qaisi M, Montague L, Bone Margin Analysis for Osteonecrosis and Osteomyelitis of the Jaws, Oral Maxillofac. Surg. Clin. North Am 29 (2017) 301–313. 10.1016/j.coms.2017.03.007. [DOI] [PubMed] [Google Scholar]
  • [9].Krishnan CK, F Han HS Kim, Outcome after surgery for metastases to the pelvic bone: A single institutional experience, CiOS Clin. Orthop. Surg 9 (2017) 116–125. 10.4055/cios.2017.9.1.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Lubek JE, Magliocca KR, Evaluation of the Bone Margin in Oral Squamous Cell Carcinoma, Oral Maxillofac. Surg. Clin. North Am 29 (2017) 281–292. 10.1016/j.coms.2017.03.005. [DOI] [PubMed] [Google Scholar]
  • [11].Zeng B, Yang L, Liang YJ, Lao XM, Mei XY, Liao GQ, Diagnostic value of intraoperative bone marrow assessment for bone margins in patients with head and neck squamous cell carcinoma: a systematic review and meta-analysis, Int. J. Oral Maxillofac. Surg (2020) 1–7. 10.1016/j.ijom.2019.11.014. [DOI] [PubMed] [Google Scholar]
  • [12].Wu JS, Goldsmith JD, Horwich PJ, Shetty SK, Hochman MG, Bone and Soft-Tissue Lesions: What Factors Affect Diagnostic Yield of Image-guided Core-Needle Biopsy?, Radiology. 248 (2008) 962–970. 10.1148/radiol.2483071742. [DOI] [PubMed] [Google Scholar]
  • [13].Blondiaux E, Laurent M, Audureau E, Boudjemaa S, Sileo C, Lenoir M, Dainese L, Garel C, Coulomb A, Ducou le Pointe H, Factors influencing the diagnostic yield and accuracy of image-guided percutaneous needle biopsy of pediatric tumors: single-center audit of a 26-year experience, Pediatr. Radiol 46 (2016) 372–382. 10.1007/s00247-015-3484-8. [DOI] [PubMed] [Google Scholar]
  • [14].Taupin T, Decouvelaere AV, Vaz G, Thiesse P, Accuracy of core needle biopsy for the diagnosis of osteosarcoma: A retrospective analysis of 73 patients, Diagn. Interv. Imaging 97 (2016) 327–331. 10.1016/j.diii.2015.09.013. [DOI] [PubMed] [Google Scholar]
  • [15].Chadayammuri V, Herbert B, Hao J, Mavrogenis A, Quispe JC, Kim JW, Young H, Hake M, Mauffrey C, Diagnostic accuracy of various modalities relative to open bone biopsy for detection of long bone posttraumatic osteomyelitis, Eur. J. Orthop. Surg. Traumatol 27 (2017) 871–875. 10.1007/s00590-017-1976-y. [DOI] [PubMed] [Google Scholar]
  • [16].Wu JS, McMahon CJ, Lozano-Calderon S, Kung JW, Utility of repeat core needle biopsy of musculoskeletal lesions with initially nondiagnostic findings, Am. J. Roentgenol 208 (2017) 609–616. 10.2214/AJR.16.16220. [DOI] [PubMed] [Google Scholar]
  • [17].Tao YK, Shen D, Sheikine Y, Ahsen OO, Wang HH, Schmolze DB, Johnson NB, Brooker JS, Cable AE, Connolly JL, Fujimoto JG, Assessment of breast pathologies using nonlinear microscopy, Proc. Natl. Acad. Sci. U. S. A 111 (2014) 15304–15309. 10.1073/pnas.1416955111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Cahill LC, Giacomelli MG, Yoshitake T, Vardeh H, Faulkner-Jones BE, Connolly JL, Sun C-K, Fujimoto JG, Rapid virtual hematoxylin and eosin histology of breast tissue specimens using a compact fluorescence nonlinear microscope, Lab. Investig 98 (2018) 150–160. 10.1038/labinvest.2017.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Cahill LC, Fujimoto JG, Giacomelli MG, Yoshitake T, Wu Y, Lin DI, Ye H, Carrasco-Zevallos OM, Wagner AA, Rosen S, Comparing histologic evaluation of prostate tissue using nonlinear microscopy and paraffin H&E: a pilot study, Mod. Pathol 32 (2019) 1158–1167. 10.1038/s41379-019-0250-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Cahill LC, Wu Y, Yoshitake T, Ponchiardi C, Giacomelli MG, Wagner AA, Rosen S, Fujimoto JG, Nonlinear microscopy for detection of prostate cancer: analysis of sensitivity and specificity in radical prostatectomies, Mod. Pathol (2019). 10.1038/s41379-019-0408-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Giacomelli MG, Faulkner-Jones BE, Cahill LC, Yoshitake T, Do D, Fujimoto JG, Comparison of nonlinear microscopy and frozen section histology for imaging of Mohs surgical margins, Biomed. Opt. Express 10 (2019) 4249. 10.1364/boe.10.004249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Karen JK, Gareau DS, Dusza SW, Tudisco M, Rajadhyaksha M, Nehal KS, Detection of basal cell carcinomas in Mohs excisions with fluorescence confocal mosaicing microscopy, Br. J. Dermatol 160 (2009) 1242–1250. 10.1111/j.1365-2133.2009.09141.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Longo C, Ragazzi M, Rajadhyaksha M, Nehal K, Bennassar A, Pellacani G, Malvehy Guilera J, In Vivo and Ex Vivo Confocal Microscopy for Dermatologic and Mohs Surgeons., Dermatol. Clin 34 (2016) 497–504. 10.1016/j.det.2016.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Yoshitake T, Giacomelli MG, Cahill LC, Schmolze DB, Vardeh H, Faulkner-Jones BE, Connolly JL, Fujimoto JG, Direct comparison between confocal and multiphoton microscopy for rapid histopathological evaluation of unfixed human breast tissue, J. Biomed. Opt 21 (2016) 126021. 10.1117/1.JBO.21.12.126021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Krishnamurthy S, Cortes A, Lopez M, Wallace M, Sabir S, Shaw K, Mills G, Ex vivo confocal fluorescence microscopy for rapid evaluation of tissues in surgical pathology practice, Arch. Pathol. Lab. Med 142 (2018) 396–401. 10.5858/arpa.2017-0164-OA. [DOI] [PubMed] [Google Scholar]
  • [26].Wang M, Kimbrell HZ, Sholl AB, Tulman DB, Elfer KN, Schlichenmeyer TC, Lee BR, Lacey M, Brown JQ, High-Resolution Rapid Diagnostic Imaging of Whole Prostate Biopsies Using Video-Rate Fluorescence Structured Illumination Microscopy, Cancer Res. 75 (2015) 4032–4041. 10.1158/0008-5472.CAN-14-3806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Wang M, Tulman DB, Sholl AB, Kimbrell HZ, Mandava SH, Elfer KN, Luethy S, Maddox MM, Lai W, Lee BR, Brown JQ, Gigapixel surface imaging of radical prostatectomy specimens for comprehensive detection of cancer-positive surgical margins using structured illumination microscopy, Sci. Rep 6 (2016) 27419. 10.1038/srep27419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Glaser AK, Reder NP, Chen Y, McCarty EF, Yin C, Wei L, Wang Y, True LD, Liu JTC, Light-sheet microscopy for slide-free non-destructive pathology of large clinical specimens, Nat. Biomed. Eng 1 (2017) 0084. 10.1038/s41551-017-0084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Glaser AK, Reder NP, Chen Y, Yin C, Wei L, Kang S, Barner LA, Xie W, McCarty EF, Mao C, Halpern AR, Stoltzfus CR, Daniels JS, Gerner MY, Nicovich PR, Vaughan JC, True LD, Liu JTC, Multi-immersion open-top light-sheet microscope for high-throughput imaging of cleared tissues, Nat. Commun 2019 101. 10 (2019) 1–8. 10.1038/s41467-019-10534-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Giacomelli MG, Yoshitake T, Cahill LC, Vardeh H, Quintana LM, Faulkner-Jones BE, Brooker J, Connolly JL, Fujimoto JG, Multiscale nonlinear microscopy and widefield white light imaging enables rapid histological imaging of surgical specimen margins, Biomed. Opt. Express 9 (2018) 2457. 10.1364/BOE.9.002457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Giacomelli MG, Husvogt L, Vardeh H, Faulkner-Jones BE, Hornegger J, Connolly JL, Fujimoto JG, Virtual Hematoxylin and Eosin Transillumination Microscopy Using Epi-Fluorescence Imaging, PLoS One. 11 (2016) e0159337. 10.1371/journal.pone.0159337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [32].Kowada T, Kikuta J, Kubo A, Ishii M, Maeda H, Mizukami S, Kikuchi K, In Vivo Fluorescence Imaging of Bone-Resorbing Osteoclasts, J. Am. Chem. Soc 133 (2011) 17772–17776. 10.1021/JA2064582. [DOI] [PubMed] [Google Scholar]
  • [33].Sano H, Kikuta J, Furuya M, Kondo N, Endo N, Ishii M, Intravital bone imaging by two-photon excitation microscopy to identify osteocytic osteolysis in vivo, Bone. 74 (2015) 134–139. 10.1016/J.BONE.2015.01.013. [DOI] [PubMed] [Google Scholar]
  • [34].Li Y, Chen X, Watkins B, Saini N, Gannon S, Nadeau E, Reeves R, Gao B, Pelligrini V, Yao H, Mercuri J, Ye T, Nonlabeling and quantitative assessment of chondrocyte viability in articular cartilage with intrinsic nonlinear optical signatures:, Exp. Biol. Med 245 (2020) 348–359. 10.1177/1535370219896545. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Chen X, Li Y, Wyman N, Zhang Z, Fan H, Le M, Gannon S, Rose C, Zhang Z, Mercuri J, Yao H, Gao B, Woolf S, Pécot T, Tong Ye A, Deep learning provides high accuracy in automated chondrocyte viability assessment in articular cartilage using nonlinear optical microscopy, Biomed. Opt. Express 12 (2021) 2759. 10.1364/BOE.417478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Black C, Marotti J, Zarovnaya E, Paydarfar J, Critical evaluation of frozen section margins in head and neck cancer resections, Cancer. 107 (2006) 2792–2800. 10.1002/cncr.22347. [DOI] [PubMed] [Google Scholar]
  • [37].Lin MM, Goldsmith JD, Resch SC, DeAngelis JP, Ramappa AJ, Histologic examinations of arthroplasty specimens are not cost-effective: a retrospective cohort study., Clin. Orthop. Relat. Res 470 (2012) 1452–60. 10.1007/s11999-011-2149-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].DiCarlo EF, Klein MJ, Comparison of Clinical and Histologic Diagnoses in 16,587 Total Joint Arthroplasties, Am. J. Clin. Pathol 141 (2014) 111–118. 10.1309/AJCPDMFQK6QZK9NN. [DOI] [PubMed] [Google Scholar]
  • [39].Weibrich G, Hansen T, Kleis W, Buch R, Hitzler WE, Effect of platelet concentration in platelet-rich plasma on peri-implant bone regeneration, Bone. 34 (2004) 665–671. 10.1016/J.BONE.2003.12.010. [DOI] [PubMed] [Google Scholar]
  • [40].Pautke C, Vogt S, Tischer T, Wexel G, Deppe H, Milz S, Schieker M, Kolk A, Polychrome labeling of bone with seven different fluorochromes: Enhancing fluorochrome discrimination by spectral image analysis, Bone. 37 (2005) 441–445. 10.1016/J.BONE.2005.05.008. [DOI] [PubMed] [Google Scholar]

RESOURCES