Scar tissue patterns may help predict pancreatic cancer outcomes, guide treatment
Click Here to Manage Email Alerts
Key takeaways:
- Researchers developed an algorithm that identified correlations between scar architecture and survival.
- Future research aims to explore the role of scar architecture in other cancer types.
The organization of cells in scar tissue around pancreatic tumors may help predict survival, according to researchers at Stanford Medicine.
Pancreatic ductal adenocarcinoma (PDAC) triggers a strong response by fibroblasts found in the connective tissue around the pancreas.
These fibroblasts discharge collagen and other components of the extracellular matrix to fix the tumor in a web of cancer-related scar tissue called desmoplasia. Prior research suggested that this scar tissue may predict prognosis; however, those analyses had been restricted to gross measurements of scar quantity.
A new study published in Cell Reports Medicine suggests the geometric patterns of the desmoplasia are the second-most valuable prognostic factors for PDAC, behind only to tumor stage at diagnosis.
“One of the big ideas from this work is that it’s not just about the pure quantity of fibrosis when we look at both fibrotic diseases, it’s about the organization of the tissue that is impacted by fibrotic development,” researcher Jason L. Guo, PhD, postdoctoral scholar at Stanford Medicine, told Healio. “If we look at the architecture and the organization of scarring as opposed to just the quantity, we might find more well-defined granular drivers of the underlying disease.”
Significant survival differential
Guo and colleagues — including Michael T. Longaker, MD, Deane P. and Louise Mitchell professor in Stanford School of Medicine, and Shamik Mascharak, MD, PhD, a former graduate student — evaluated hundreds of PDAC tissue samples collected during surgeries.
Investigators used a computer algorithm to identify and characterize nearly 300 distinct features of the desmoplasia, including length, width, alignment and density of the fibers.
They used a technique called CODEX to evaluate the prevalence and location of immune cells and other types of cells in the desmoplasia.
“We dove deeper into that question of what cells might be associated with different architectures because, in addition to the scar itself, we want to understand the underlying biological drivers,” Guo said. “We looked at cell spatial organization within these tumors, and we correlated that with scar architecture to determine what drives this matrix phenotype or this tissue-level architecture based on the cells present within different scar architectures.”
Researchers next used a machine-learning approach to combine the data about cellular interactions and fiber patterns with clinical patient information to identify factors associated with survival.
Results showed the presence of cytotoxic T cells appeared associated with longer OS, whereas enrichment of tumor cells that modulate the immune system, as well as activated B cells, correlated with poorer survival.
The data allowed computers to categorize patients with PDAC into two key groups. Those in one group survived a median 655 days longer than those in the other group.
“Jason took a very rigorous approach, saying ‘OK, we have this tool, this algorithm that doesn’t look at the cancer cells at all, only the fibrosis and the desmoplasia around it,’” Longaker said. “It’s shocking that it’s almost 2 years’ difference just based on what the scar architecture and features are.”
Researchers also found thinner fibers predicted better survival compared with the more compressed fibers observed in samples from patients who died shortly after diagnosis.
‘A new field’
Longaker said he hopes these findings will draw more attention to fibrosis and scar architecture as a potential predictor of cancer survival.
“Most people have no idea that 45% of Americans die of something that Jason and I would call fibrosis,” Longaker said. “If you have a stroke, that part of your brain dies and is replaced by a scar. If that’s your breathing center, you die. Three million people have pulmonary fibrosis, millions of people have liver cirrhosis, but we don’t think of it in those terms.”
Although the investigators developed their algorithm based on trichrome staining, the team hopes to make the tool more translatable to common histologic stains, Guo said.
“We have ongoing efforts to adapt this algorithm in this predictive capacity to very common stains like hemotoxin and eosin staining,” he said. “The vision is to make a pipeline where we could take commonly acquired routine histologic stains that are already part of clinical workflows and feed that into a prognostic algorithm.”
Longaker said he hopes the information gained from this tool might be used to help clinicians decide which patients with PDAC would benefit from more aggressive treatment, and who should be spared the potential toxicities associated with these treatments.
The research team hopes to explore the role played by scar architecture in other cancers, particularly those that present with a palpable lump.
“You can find a lung tumor with your eyes closed — the lung is soft, the tumor is hard, just as the pancreas is soft and the tumor is hard,” Longaker said. “Based on the fact that we now know this [about desmoplasia architecture] in a lethal cancer like PDAC, I believe we should be exploring this in other cancers to see how widespread it is.”
Longaker said the group is interested in investigating the role of scar architecture in breast and colon cancer in the future.
“These are other things we’re committed to exploring,” he said. “Jason’s coding expertise and his molecular expertise have really opened up a new field of tumor desmoplasia architecture.”
References:
- Mascharak S, et al. Cell Rep Med. 2023;doi:10.1016/j.xcrm.2023.101248.
- Stanford Medicine. Scar tissue holds hints about pancreatic cancer outcome, Stanford Medicine-led research finds (press release). Available at: https://med.stanford.edu/news/all-news/2023/11/pancreatic-cancer.html. Posted Nov. 22, 2023. Accessed Feb. 21, 2024.
For more information:
Jason L. Guo, PhD, can be reached at jasonguo@stanford.edu.
Michael T. Longaker, MD, can be reached at longaker@stanford.edu.
Collapse
Healio Interviews
Click Here to Manage Email Alerts