Research tests a new way to use data that better predicts the spread and outcomes of colorectal cancer
October 7, 2019
Research tests a new way to use data that better predicts the spread and outcomes of colorectal cancer
Summary
This research built on earlier work to test a new statistical method, called ΔPC1.EMT, to predict patient survival and the spread (metastasis) of colorectal cancer.
- ΔPC1.EMT uses complex patterns in genetic cancer data and gives a score to predict metastasis and survival rates
- Researchers tested ΔPC1.EMT with 6 different colorectal cancer datasets
What were the main results?
Researchers found the new ΔPC1.EMT score better predicted survival and changes of cancer spreading (metastasis) for patients with stage 1, 2, and 3 colorectal cancer. The researchers also found that ΔPC1.EMT was more accurate than 10 other cancer-predicting methods.
Researchers believe ΔPC1.EMT may be able to identify patients with colorectal cancer who may benefit from receiving chemotherapy shortly after surgery (adjuvant chemotherapy).
What did researchers study?
Researchers calculated ΔPC1.EMT score using genomic data. They tested this score to see if it better predicted patient survival and the spread of colorectal cancer and compared it to other cancer-predicting scores.
How many patients' data did researchers study?
The researchers began with genomic data from 468 patients with colorectal cancer who joined Total Cancer Care® (TCC) between 2006 and 2010. Then, the researchers further tested ΔPC1.EMT using data from 3,697 patients with colorectal cancer from 5 more datasets.
What kind of research was this?
This was an observational study. Researchers looked at (observed) genomic data donated by patients with colorectal cancer tumors.
The results of this research alone should not be used to make health decisions. It takes many studies for researchers to confirm findings and use them in patient care.
Why was this research needed?
Colorectal cancer is the second leading cause of death from cancer in the United States. The spread of colorectal cancer plays a major role in these deaths, but it’s difficult to predict if colorectal cancer will spread. If researchers could predict if colorectal cancer will spread, they could identify patients who may benefit from adjuvant chemotherapy (chemotherapy shortly after surgery).
Researchers think that patterns of changes in a cancer cell’s genes can predict if cancer is likely to grow or spread. Previous research found 2 related patterns in colorectal cancer cells that grow or spread:
- PC1 – an overall pattern found in cancer cells that grow or spread
- EMT – a pattern for a cell during epithelial-to-mesenchymal transition (EMT), a normal process that cancer cells may use to spread to other parts of the body
Previous research used statistical methods to calculate EMT and PC1 scores that measure how close a cancer cell aligns with these 2 patterns.
Using certain statistics, researchers found that when a cancer cell has either a higher PC1 score or a higher EMT score, a patient was less likely to survive. While either score alone can predict survival, researchers were surprised to find that they could predict survival dramatically better when they combined the scores by subtracting the EMT score from the PC1 score. This suggested that the non-EMT part of the pattern is the most important. They called the new score ΔPC1.EMT.
In this study, researchers tested ΔPC1.EMT to see if it better predicted patient survival and the spread of colorectal cancer compared to PC1 alone, EMT alone, and several other known methods.
More about the statistics in this research
Statistics is a part of math that focuses on data collection, organization, analysis, interpretation, and presentation.
In this study, researchers used a statistical method called principal component analysis (PCA) to identify important factors. Researchers used PCA to connect certain genes or groups of genes to patterns found in certain tumors. Researchers named the main pattern they found PC1.
-
How many patient's data did researchers study?
The researchers began with data from 468 patients with colorectal cancer who donated their data between 2006 and 2010 as part of TCC. The data included genetic information from:
- 367 primary tumors (the original site of cancer)
- 101 metastatic tumors (where the cancer had spread in the body)
Then, the researchers further tested ΔPC1.EMT using cancer data from 3,697 patients with colorectal cancer from 5 more publicly available datasets: PETACC3, ALMAC, LNCC, GEO41258, and GSE1433.
Researchers used the following information from the datasets:
- Genetic sequencing from the tumor and specific genes that researchers believe may have a role in colorectal cancer - researchers use sequencing to know the order of the DNA building blocks in genes. It allows them to look for differences in genes and gene mutations (changes).
- Possible or present genetic mutations within the cells’ DNA
- Cancer stage – researchers included tumors at all 4 cancer stages
- Clinical information, such as a patient’s survival
-
What kind of research was this?
This was an observational study. Researchers looked at (observed) genomic data donated by patients with colorectal cancer tumors.
This kind of study can further cancer research aimed at developing new treatments, tailoring treatments to patients, and identifying a patient’s cancer sooner.
-
What happened during this research?
Researchers first found the EMT and PC1 scores using each patient’s genetic data. Then, researchers subtracted the EMT score from the PC1 score to calculate a new score for each patient called ΔPC1.EMT:
ΔPC1.EMT score = PC1.score – EMT.score
Researchers used statistical methods to learn if the ΔPC1.EMT score for each colorectal cancer patient had any relation to their survival rates or to specific cancer gene mutations linked to metastasis. Researchers also compared the ΔPC1.EMT score to 10 other cancer-predicting scoring systems.
-
How researchers designed this study
Researchers started with data from 468 patients with colorectal cancer. Researchers calculated ΔPC1.EMT scores to predict cancer outcomes from patterns in the data.
Researchers tested ΔPC1.EMT further with data from 3,697 patients with colorectal cancer. Researchers compared ΔPC1.EMT scores with 10 other methods to see if it better predicted cancer outcomes.
-
What were the main results?
Researchers learned that the combined ΔPC1.EMT score better predicts metastasis and survival outcomes in colorectal cancer.
By testing the ΔPC1.EMT score on the TCC dataset, researchers found that it was more accurate than other methods in identifying:
- If colorectal cancer tumors were metastatic
- How long patients survived
- How colorectal cancer tumors grew and spread
For each of these, they found it more accurately predicted for colorectal cancer in stages 1, 2, and 3 than in stage 4.
In testing the ΔPC1.EMT score on the 5 other datasets, researchers found similar results as with the TCC dataset. The ΔPC1.EMT was more accurate than both PC1 and EMT scores in most tests.
Other results
The ΔPC1.EMT scored also showed a potential to identify different types of gene mutations within colorectal cancer tumors.
When compared to 5 different patterns, the ΔPC1.EMT was more accurate at predicting patient survival and the spread of colorectal cancer.
-
How has this research helped?
This research found that the new ΔPC1.EMT score more accurately predicts overall survival and relapse-free survival from colorectal cancer than other methods. It may also help predict metastasis in patients with stage 2 and 3 colorectal cancer. Researchers suggest it may help identify which patients with colorectal cancer may benefit from adjuvant chemotherapy.
Researchers suggest further study on this method in a clinical trial with patients diagnosed with colorectal cancer.
