Scientists headed by a team at the Institute of Cancer Research, London (ICR) have used artificial intelligence (AI) and machine learning (ML) to discover five new subtypes of breast cancer that could help clinicians deliver the most effective therapies—including immunotherapy—for individual patients, as well as potentially direct the development of new anticancer drugs.
The computational tools find patterns in the genetic, molecular, and cellular make-up of primary luminal A-type breast tumors, which are analyzed alongside data on patient survival. The team had previously used the same approach to uncover subtypes of colorectal cancer (CRC). The latest study classifies the different subtypes of luminal A breast cancer identified according to similar features.
“Doctors have used the current classification of breast cancers as a guide for treatment for years, but it is quite crude and patients who seemingly have the same type of the disease often respond very differently to drugs,” commented Maggie Cheang, PhD, team leader of the Genomic Analysis Clinical Trials Team at the ICR, London, who is co-author of the researchers published paper in npj Breast Cancer. “Our study has used AI algorithms to spot patterns within breast cancers that human analysis had up to now missed—and found additional types of the disease that respond in very particular ways to treatment.”
The ICR researchers, headed by Anguraj Sadanandam, PhD, team leader in systems and precision cancer medicine and colleagues at the Royal Marsden Hospital, report their findings in a paper titled, “Heterocellular gene signatures reveal luminal-A breast cancer heterogeneity and differential therapeutic responses.”
Most breast cancers develop in the inner cells that line the mammary ducts and are estrogen-receptor and/or progesterone-receptor positive, and HER2-negative. These tumors, known as luminal-A-type breast cancers tend to have the best cure rates, but patients do respond differently to standard-of-care treatments such as tamoxifen, and to newer immunotherapies, which might be a recourse in the event of relapse. “Even this relatively well-characterized breast cancer subtype possesses heterogeneity at the levels of hormone receptor expression, treatment response, and genetic variability that requires further understanding,” the authors wrote.
Factors that impact on tumor heterogeneity even between different luminal-A-type tumors, are complex, and include genetic changes, the tumor microenvironment, and interactions between different cell types. And while immune-related genes are also often expressed in different subtypes of breast cancer, including luminal-A tumors, “… unlike in colorectal and pancreatic cancers, no exclusive immune-enriched breast cancer subtype has been reported (to our knowledge).”
To investigate differences between luminal-A breast tumors from different patients, the team turned to ML/AI tools that they had previously used to classify CRC into five heterocellular subtypes, which they termed inflammatory, enterocyte, globlet-like, stem-like, and transit-amplifying (TA). “… we sought to use our CRC heterocellular signatures as surrogates to re-characterize breast cancer subtypes, especially luminal-A breast cancers, and understand their phenotypes according to their differentiated, stem, fibroblast, and immune characteristics,” the investigators stated.
The results of their analyses showed that the inflammatory-type luminal-A tumors contained immune cells and high levels PD-L1, which suggested that they may likely respond to immunotherapies. “… we observed an enrichment of inflammatory heterocellular subtype samples in the luminal-A subtype harboring high expression of immune checkpoint genes,” the team noted. “A subset of the stem-like subtype also showed increased expression of immune genes.”
Cheang noted that one of the “exciting implications” of the research was its ability to identify women who might respond well to immunotherapy, “… even when the broad classification of their cancer would suggest that these treatments wouldn’t work for them.”
The researchers were additionally surprised to discover that the stem-like subtype of luminal-A breast tumors showed good recurrence free survival (RFS), “indicating that the presence of stem cells and fibroblasts (enriched in the stem-like subtype) does not indicate poor survival in differentiated luminal-A breast cancer patients.” The results also indicated that TA tumors were characterized by changes in chromosome 8, and patients with this type of tumor had worse survival than other groups when treated with tamoxifen. These patients also tended to relapse much earlier—after an average of 42 months compared to 83 months in patients who had the tumor type that contained high levels of stem cells. The findings indicated that patients with the chromosome 8 changes may benefit from an additional or new treatment to delay or prevent late relapse.
The researchers also looked at tumors from a group of patients with triple negative breast cancer (TNBC). This type of breast cancer doesn’t respond to standard hormone treatments. The analysis of samples from this group of patients suggested that their tumors might also respond to immunotherapy. “While no immunotherapy is yet approved, but with immune checkpoint inhibitors being tested clinically in patients with breast cancer, our association of a subset of basal breast cancers with the CRC inflammatory subtype suggests a means to identify patients who might respond to immunotherapy.”
The authors say their results shed new light on luminal-A breast cancer subtypes that could help personalize diagnosis and treatment of patients with different types of breast cancer. “Our new study has shown that AI is able to recognize patterns in breast cancer that are beyond the limit of the human eye, and to point us to new avenues of treatment among those who have stopped responding to standard hormone therapies,” Sadanandam stated. “We are at the cusp of a revolution in healthcare, as we really get to grips with the possibilities AI and machine learning can open up … AI has the capacity to be used much more widely, and we think we will be able to apply this technique across all cancers, even opening up new possibilities for treatment in cancers that are currently without successful options.”
In addition, the findings could help to direct the discovery of new drugs even for patients who might tend to relapse after many years. “The AI used in our study could also be used to discover new drugs for those most at risk of late relapse, beyond five years, which is common in oestrogen-linked breast cancers and can cause considerable anxiety for patients,” Cheang concluded.
The ICR is pioneering the use of AI to understand cancer’s complexity and evolution, and is raising the final £15 million of a £75 million investment in a new Centre for Cancer Drug Discovery, which will house what it describes as “a world-first program of ‘anti-evolution’ therapies.”