Limits...
Three-dimensional reconstruction of oral tongue squamous cell carcinoma at invasion front.

Kudo T, Shimazu Y, Yagishita H, Izumo T, Soeno Y, Sato K, Taya Y, Aoba T - Int J Dent (2013)

Bottom Line: Serial sections (4  μ m thick) were double immunostained with pan-cytokeratin and Ki67 antibodies and digitized images were acquired using virtual microscopy.Direct visualization and quantitative assessment of the parenchymal-stromal border provide a new dimension in our understanding of OTSCC architecture.These 3D morphometric analyses also ascertained that cell invasion (individually and collectively) occurs at the deep invasive front of the OTSCC.

View Article: PubMed Central - PubMed

Affiliation: Department of Pathology, Hyogo College of Medicine, 1-1 Mukogawa-cho, Nishinomiya, Hyogo 663-8501, Japan ; Department of Pathology, School of Life Dentistry at Tokyo, The Nippon Dental University, 1-9-20 Fujimi, Chiyoda-ku, Tokyo 102-8159, Japan.

ABSTRACT
We conducted three-dimensional (3D) reconstruction of oral tongue squamous cell carcinoma (OTSCC) using serial histological sections to visualize the architecture of invasive tumors. Fourteen OTSCC cases were collected from archival paraffin-embedded specimens. Based on a pathodiagnostic survey of whole cancer lesions, a core tissue specimen (3 mm in diameter) was dissected out from the deep invasion front using a paraffin tissue microarray. Serial sections (4  μ m thick) were double immunostained with pan-cytokeratin and Ki67 antibodies and digitized images were acquired using virtual microscopy. For 3D reconstruction, image registration and RGB color segmentation were automated using ImageJ software to avoid operator-dependent subjective errors. Based on the 3D tumor architecture, we classified the mode of invasion into four types: pushing and bulky architecture; trabecular architecture; diffuse spreading; and special forms. Direct visualization and quantitative assessment of the parenchymal-stromal border provide a new dimension in our understanding of OTSCC architecture. These 3D morphometric analyses also ascertained that cell invasion (individually and collectively) occurs at the deep invasive front of the OTSCC. These results demonstrate the advantages of histology-based 3D reconstruction for evaluating tumor architecture and its potential for a wide range of applications.

No MeSH data available.


Related in: MedlinePlus

(a) Image registration of consecutive histological images in the z-axis. (b) Sagittal cross-cut view through the center of an image stack comprising 108 serial sections, showing successful superimposition without marked irregularity or distortion in interior tissue geometry. Bar = 0.5 mm.((c)–(e)) Segmentation of cancer cells and Ki67-positive nuclei: (c) original double-immunostained image with DAB and Vector SG as chromogens; (d) segmentation of CK-positive cancer cell cytoplasm, shown as purple pseudocolor (note that Ki67-negative nuclei remained open within the segmented cytoplasm); and (e) stromal region (yellow) after subtraction of the tumor parenchyma (see text for details); (f) Ki67-positive nuclei (red) attained by ImageJ SG color segmentation. Bar = 50 μm.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3818895&req=5

fig2: (a) Image registration of consecutive histological images in the z-axis. (b) Sagittal cross-cut view through the center of an image stack comprising 108 serial sections, showing successful superimposition without marked irregularity or distortion in interior tissue geometry. Bar = 0.5 mm.((c)–(e)) Segmentation of cancer cells and Ki67-positive nuclei: (c) original double-immunostained image with DAB and Vector SG as chromogens; (d) segmentation of CK-positive cancer cell cytoplasm, shown as purple pseudocolor (note that Ki67-negative nuclei remained open within the segmented cytoplasm); and (e) stromal region (yellow) after subtraction of the tumor parenchyma (see text for details); (f) Ki67-positive nuclei (red) attained by ImageJ SG color segmentation. Bar = 50 μm.

Mentions: In this study, superimposition of 80–108 consecutively captured images (Figure 2(a)) was accomplished successfully in an automated, algorithm-driven manner. By browsing through an image stack, it was verified that the tumor architecture was smoothly continuous between adjacent slices with no marked irregularity or distortion in interior tumor geometry and surface contour delineation (Figure 2(b)). One crucial condition for accomplishing the automated image registration is the preparation of regularly-oriented good-quality serial sections, since computation for alignment of serial sections depends on the performance of an approximate initial registration to correct for shifts and rotations of sections on the glass slides. It is also notable that we used no algorithmic image transformation, such as stretching or shrinking images in the x-y dimension to adjust for small misalignments between sections; in our experience of image registration, the introduction of the ImageJ affine image transformer actually results in deterioration of the quality of the image stack rather than improvement.


Three-dimensional reconstruction of oral tongue squamous cell carcinoma at invasion front.

Kudo T, Shimazu Y, Yagishita H, Izumo T, Soeno Y, Sato K, Taya Y, Aoba T - Int J Dent (2013)

(a) Image registration of consecutive histological images in the z-axis. (b) Sagittal cross-cut view through the center of an image stack comprising 108 serial sections, showing successful superimposition without marked irregularity or distortion in interior tissue geometry. Bar = 0.5 mm.((c)–(e)) Segmentation of cancer cells and Ki67-positive nuclei: (c) original double-immunostained image with DAB and Vector SG as chromogens; (d) segmentation of CK-positive cancer cell cytoplasm, shown as purple pseudocolor (note that Ki67-negative nuclei remained open within the segmented cytoplasm); and (e) stromal region (yellow) after subtraction of the tumor parenchyma (see text for details); (f) Ki67-positive nuclei (red) attained by ImageJ SG color segmentation. Bar = 50 μm.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3818895&req=5

fig2: (a) Image registration of consecutive histological images in the z-axis. (b) Sagittal cross-cut view through the center of an image stack comprising 108 serial sections, showing successful superimposition without marked irregularity or distortion in interior tissue geometry. Bar = 0.5 mm.((c)–(e)) Segmentation of cancer cells and Ki67-positive nuclei: (c) original double-immunostained image with DAB and Vector SG as chromogens; (d) segmentation of CK-positive cancer cell cytoplasm, shown as purple pseudocolor (note that Ki67-negative nuclei remained open within the segmented cytoplasm); and (e) stromal region (yellow) after subtraction of the tumor parenchyma (see text for details); (f) Ki67-positive nuclei (red) attained by ImageJ SG color segmentation. Bar = 50 μm.
Mentions: In this study, superimposition of 80–108 consecutively captured images (Figure 2(a)) was accomplished successfully in an automated, algorithm-driven manner. By browsing through an image stack, it was verified that the tumor architecture was smoothly continuous between adjacent slices with no marked irregularity or distortion in interior tumor geometry and surface contour delineation (Figure 2(b)). One crucial condition for accomplishing the automated image registration is the preparation of regularly-oriented good-quality serial sections, since computation for alignment of serial sections depends on the performance of an approximate initial registration to correct for shifts and rotations of sections on the glass slides. It is also notable that we used no algorithmic image transformation, such as stretching or shrinking images in the x-y dimension to adjust for small misalignments between sections; in our experience of image registration, the introduction of the ImageJ affine image transformer actually results in deterioration of the quality of the image stack rather than improvement.

Bottom Line: Serial sections (4  μ m thick) were double immunostained with pan-cytokeratin and Ki67 antibodies and digitized images were acquired using virtual microscopy.Direct visualization and quantitative assessment of the parenchymal-stromal border provide a new dimension in our understanding of OTSCC architecture.These 3D morphometric analyses also ascertained that cell invasion (individually and collectively) occurs at the deep invasive front of the OTSCC.

View Article: PubMed Central - PubMed

Affiliation: Department of Pathology, Hyogo College of Medicine, 1-1 Mukogawa-cho, Nishinomiya, Hyogo 663-8501, Japan ; Department of Pathology, School of Life Dentistry at Tokyo, The Nippon Dental University, 1-9-20 Fujimi, Chiyoda-ku, Tokyo 102-8159, Japan.

ABSTRACT
We conducted three-dimensional (3D) reconstruction of oral tongue squamous cell carcinoma (OTSCC) using serial histological sections to visualize the architecture of invasive tumors. Fourteen OTSCC cases were collected from archival paraffin-embedded specimens. Based on a pathodiagnostic survey of whole cancer lesions, a core tissue specimen (3 mm in diameter) was dissected out from the deep invasion front using a paraffin tissue microarray. Serial sections (4  μ m thick) were double immunostained with pan-cytokeratin and Ki67 antibodies and digitized images were acquired using virtual microscopy. For 3D reconstruction, image registration and RGB color segmentation were automated using ImageJ software to avoid operator-dependent subjective errors. Based on the 3D tumor architecture, we classified the mode of invasion into four types: pushing and bulky architecture; trabecular architecture; diffuse spreading; and special forms. Direct visualization and quantitative assessment of the parenchymal-stromal border provide a new dimension in our understanding of OTSCC architecture. These 3D morphometric analyses also ascertained that cell invasion (individually and collectively) occurs at the deep invasive front of the OTSCC. These results demonstrate the advantages of histology-based 3D reconstruction for evaluating tumor architecture and its potential for a wide range of applications.

No MeSH data available.


Related in: MedlinePlus