Digital Staining Instrumentation
Honorable Mention - Imaging InformaticsJared Marv Orrock ; Resident, Mayo Clinic ;
Content:
With light sources of increasingly broader ranges, spectral analysis of tissue sections has evolved from 2 wavelength image subtraction techniques to Raman near infrared micro-spectroscopic mapping permitting discrimination of cell types & tissue patterns.
Technology:
We are developing next-generation hyperspectral imaging (HSI) systems for use in the physical and life sciences, specifically histopathology using unstained sections of routinely formalin-fixed paraffin embedded tissue. Previous investigators have shown previous success with high diagnostic accuracy for algorithms derived from HSI in H&E stained sections. The use of unstained sections eliminates the cost and effort associated with staining techniques as well as artifacts that may be introduced by the staining process.
Design:
A dedicated light source is integrated with a standard commercial research grade microscope and cooled CCD camera for obtaining a hyperspectral cube with each stack on the order of 512 x 512 pixels. Wavelength range is on the order of 5 nanometers (nm) covering the full optical range of 400 to 700 nm.
Results:
Preliminary experience has shown that current available technologies can readily digitally reproduce a H&E stained section while creating a digital data set that can be subjected to algorithmic tools from these large data sets that may enhance the practice of pathology. We will show that gray areas of surgical pathology, subject to staining and intra-/interobserver variation, such as accurate grading of dysplasia in Barretts esophagus, determination of in-situ versus invasive in a wide variety of organ systems and rare event detection in lymph nodes or exfoliated specimens can be diagnosed with greater specificity beyond light microscope analysis alone.
Conclusion:
Hyperspectral image analysis may enable a pathologist to view an unstained sample and obtain the same information they would obtain if they looked at a stained sample under a traditional light microscope. A digital staining system would enable a pathologist to review a stained sample more effectively than one could visualize it using traditional light microscopy techniques. Image identification and classification could be used to prescreen samples or augment the screening capabilities of a pathologist. Digitally stained samples could yield information for pathologists to distinguish more details and enhance diagnostic accuracy.
