BIOMEDICAL OPTICAL IMAGING AND BIOPHOTONICS GROUP
Professor James G. Fujimoto

Nonlinear Microscopy – rapid histopathological imaging of fresh tissue

The only person that can tell you if you have cancer is a pathologist” (a quote from Dr. Donald Coffey, [1])

Guiding cancer surgery using nonlinear microscopy for real time pathology

Pathologists diagnose cancer by microscopic examination (histology) of surgical and biopsy specimens. However, preparing histology is time consuming. Tissue is processed, physically sliced into sections thin enough to transmit light, then stained and examined using light microscopy. Frozen section histology can be used for intraoperative evaluation, but the process is slow and labor intensive. Only small areas of tissue can be examined.

The aim of this research is to develop and demonstrate a new imaging technology (nonlinear microscopy) which enables pathologists to examine large tissue specimens in real time without the need for physical sectioning. This would enable real time guidance of cancer surgeries, reducing the need for repeat surgery and improving patient outcomes.  The challenge in this research is to develop new imaging technology, methods for intraoperatively imaging cancer specimens and integrating this into the pathology and surgical workflow.  This research involves advanced imaging technology, software display and image analysis, working with cancer specimens and pathology interpretation, and collaborating with pathologists, radiologists, and cancer surgeons.

Nonlinear (two-photon fluorescence) microscopy (NLM) has revolutionized fundamental scientific research and biological imaging. NLM performs optical sectioning using a femtosecond laser that is focused and scanned on tissue stained with fluorophores. Nonlinear excitation of the fluorescence occurs only in the laser focal region, generating an image from a thin depth section of tissue non-destructively and without the need for physical sectioning.

We, with collaborators at Beth Isreal Deaconess Hospital, have developed tissue processing protocols, rapid NLM imaging systems, and image processing algorithms that enable pathologists to evaluate freshly excised, multicentimeter specimens in real-time with minimal training. Our tissue processing protocol enables NLM evaluation for fresh tissue after only <2 minutes of staining. Furthermore, the simple staining protocol enables large size/ numbers of specimens to be processed in parallel without delay or the infrastructural requirements of frozen sections. Our real-time NLM imaging system enables pathologists to rapidly evaluate the region of interest, and our histology-like visualization minimizes the additional training required for pathologists to evaluate NLM images. These novel features enable throughputs magnitudes higher than conventional histopathological evaluation [2].

Currently we are focusing on testing the clinical impact our NLM protocol/system potentially can make. We have demonstrated that our NLM imaging protocol/system can distinguish malignant cancer vs benign with high accuracy for breast [3, 4] and prostate [5] compared to standard histology. There are many clinical scenarios where immediate histological evaluation during surgery would improve clinical outcomes but not currently available/utilized. We are trying to provide the evidence that our NLM technology can improve current clinical care/workflow and is readily transferable to clinical practice, through multiple observational/interventional clinical studies.

We are currently running a randomized controlled interventional clinical trial for intrasurgical NLM evaluation of excised breast tissue during lumpectomy [6]. Lumpectomy (breast conserving surgery) is a standard surgical treatment for local/early stage breast cancer. However, up to 30% of patients with lumpectomy receive a second surgery because of incomplete resection of cancer. Our clinical trial is to demonstrate the feasibility that intrasurgical NLM can provide comprehensive margin assessment of excised tissue without substantial surgical delay, resulting in decreased rate of second surgery after lumpectomy.

(1). Rapid fluorescent staining can be performed on fresh tissue immediately after excision (lumpectomy as example in this figure). (2). Schematic of NLM system and H&E-resembling color mapping. The stained fresh, intact tissue can be imaged with NLM. (3).  Conventional histopathological tissue processing. Tissue imaged with NLM protocol is compatible with standard histopathological evaluation.

Example of NLM images. Invasive lobular breast cancer (Left a. NLM, Left b. H&E), and prostate cancer and extracapsular extention invading to nerve (Right, NLM).

References:

[1] In Memoriam: Donald S. Coffey, PhD https://www.youtube.com/watch?v=iZqfy776OaI

[2] Multiscale nonlinear microscopy and widefield white light imaging enables rapid histological imaging of surgical specimen margins, Michael G. Giacomelli, et al, Biomedical Optics Express, 2018 https://www.osapublishing.org/boe/abstract.cfm?uri=boe-9-5-2457

[3] Assessment of breast pathologies using nonlinear microscopy, Yuankai K. Tao, et al, Proceeding of the National Academy of Sciences, 2014  https://www.pnas.org/content/111/43/15304

[4] Rapid virtual hematoxylin and eosin histology of breast tissue specimens using a compact fluorescence nonlinear microscope Lucas C Cahill, et al. Laboratory Investigation, 2018  https://www.nature.com/articles/labinvest2017116

[5] Nonlinear microscopy for detection of prostate cancer: analysis of sensitivity and specificity in radical prostatectomies, Lucas C. Cahill, et al, Modern Pathology, 2019 https://www.nature.com/articles/s41379-019-0408-4

[6] Real-Time Assessment Of Breast Cancer Lumpectomy Specimen Margins With Nonlinear Microscopy, Clinical Trials (.gov), NCT02926729  https://clinicaltrials.gov/ct2/show/NCT02926729?term=NCT02926729&draw=2&rank=1