A pilot study of micro-CT-based whole tissue imaging (WTI) on endoscopic submucosal dissection (ESD) specimens

To the best of our knowledge, this is the first case series in which ESD specimens were scanned and evaluated using micro-CT. The WTI and WBI obtained from the micro-CT scan, while incomplete, were able to confirm the detailed tissue structures and point out the extent of each lesion. Since the image resolution of micro-CT is defined by the distance between the source and the specimen10, ESD specimens, which are smaller than surgical specimens of GI cancers, were considered to be suitable for evaluation by micro-CT imaging.

Detailed pathological examination of superficial GI lesions resected en bloc by ESD is important for assessing the risk of subsequent recurrence, especially whether R0 resection has been achieved, which is essential for assessing the risk of residual recurrence11,12,13. However, since the pathological diagnosis is based on slides made every few millimeters of ESD resected specimen, it is difficult to completely exclude the possibility that a positive margin was present in a section where no slides were made. In this study, both WTI and WBI were able to assess margins in all specimens but did not find additional information that might change the diagnosis made with WSI.

The distance of submucosal invasion is also an important finding in assessing the risk of recurrence. When the submucosal invasion is more than 500 μm in gastric cancer, and more than 1000 μm in colorectal cancer, the risk of lymph node metastasis increases and additional surgery is recommended11,12,13. Diagnosis of submucosal invasion based on surface structure is mainly done by magnified observation with an endoscope14,15, but it is not easy to accurately identify the deepest invasion site in a specimen and to make an H&E slide of its cross-section. In fact, cases of lymph node metastasis recurrence despite the diagnosis of R0 resection of intramucosal carcinoma after ESD has been reported. This suggests that submucosal invasion existed in the section where H&E slides were not made16,17. WTI and WBI with micro-CT scans can observe findings that cannot be confirmed by conventional pathology slides and may lead to a more accurate diagnosis. However, in one of the two lesions in which submucosal invasion was observed by WSI, the invasion site was pointed out on WBI but could not be clearly identified on WTI. The major reason for this was that the deep part of the mucosal layer was poorly stained.

Evaluation of the presence of lymphatic and vascular invasion is also essential for pathological diagnosis of ESD specimens11,12,13. In the specimen examined in this study, the lymphatic invasion was observed in one lesion at the time of pathological diagnosis, but the area where the lymphatic invasion was observed did not remain in the WBI performed afterward, and the lymphatic invasion could not be confirmed in the WSI prepared after WBI imaging. In addition, in submucosal invasive colorectal cancer, the risk of lymph node metastasis recurrence is also evaluated by the presence or absence of budding11,12,13, but there was no specimen with budding on WSI in this study. Therefore, it is necessary to collect specimens with lymphatic invasion, vascular invasion, and budding for further analysis. Moreover, in order to objectively evaluate the usefulness of micro-CT imaging, a comparative study in which H&E and Micro-CT images are evaluated blindly is necessary.

Next to Lugol’s solution, several other contrast agents have been reported to be candidates for micro-CT18. Müller et al. reported that a hematein-based X-ray staining method enables micro-CT imaging of cell nuclei in mouse liver19. However, this method is difficult to use for clinical specimens at this stage because it requires about 10 days for staining and cannot depict structures other than the nucleus. Therefore, Micro-CT cannot completely replace the conventional pathological diagnosis by H&E slides. However, since it is thought to be possible to accurately grasp the extent of the tumor, it is expected that the deepest part of the tumor can be evaluated by WTI before sectioning the specimen and that a higher quality pathological diagnosis can be made by sectioning that specific area for H&E slide. To achieve this, it is necessary to produce a clear WTI with no staining defects. Xia et al. reported that favorable Micro-CT images were obtained by staining tongue squamous cell carcinoma specimens with 3% Lugol’s iodine solution for 12 h20. The ESD specimens used in this study were collected in routine clinical practice, and we cannot perform any protocol that would affect the conventional pathological diagnosis. In a study using test samples before the start of this study, 10% Lugol’s iodine solution caused unacceptable damage to the specimens when staining was performed for more than 180 s, so it was difficult to extend the time further. Prolonged staining with diluted Lugol’s iodine solution was also difficult because it would have prolonged the time to pathological diagnosis. However, some specimens could not be evaluated sufficiently with the staining method used in this study, and the staining method needs to be improved. It is also necessary to evaluate the effect of staining with Lugol’s iodine solution on immunostaining. Although it has been reported that staining with Lugol’s iodine solution does not affect immunostaining for neuronal-related proteins21, we have experienced poor immunostaining for mismatch repair proteins due to staining with Lugol’s iodine solution.

If the method of WTI by micro-CT scan in ESD specimen is established, it can be applied to other specimens. It may be possible to derive a more accurate pathological diagnosis by clarifying the most important cross-section by WTI before sectioning. Furthermore, since a much larger number of micro-CT images can be produced per specimen compared to conventional pathology images, it can save labor in the evaluation of obtaining images for clinical application. Ongoing development of artificial intelligence technologies has the potential to develop micro-CT image interpretation support systems that directly lead to pathological diagnosis by combining deep machine learning with biomedical image processing algorithms such as texture matching and non-rigid registration algorithms22,23,24.

In conclusion, WTI and WBI using micro-CT could delineate the extent of the lesion in ESD specimens. Thus, WTI and WBI have the potential to compensate for the limitations of conventional pathological diagnosis and help reduce the labor required for pathological diagnosis. However, improvement of image quality is essential for the diagnosis of ESD specimens with micro-CT in clinical practice.

Source link

Back to top button