It’s difficult to cure cancer once it spreads. But AI-directed precise, individualized treatments may successfully raise the odds for curing metastatic cancer.
The primary reason cancer can be so serious is the ability of the disease to spread (metastasize) inside the body, according to the National Cancer Institute (NCI). Malignant cells may travel to nearby normal lymph nodes, tissues, or organs, and migrate to parts of the body far removed from the initial cancer.
When this happens, metastatic cancer (often called stage IV cancer), is usually very difficult to treat successfully. However, there’s reason to hope this situation may change, thanks to advances in technology driven by artificial intelligence (AI).
By harnessing AI, researchers are working on ways to improve the treatment of metastatic cancer, even, potentially, finding cures for cancer that has spread throughout the body.
How AI could target metastasized cancer
Current treatments can stop or slow the growth of metastatic cancer and prolong life in some cancer patients whose malignancies have spread. However, the majority of people with metastatic cancer cannot be cured, the NCI points out.
Treatment for metastatic cancer may include surgery, radiation, hormone therapy, and chemotherapy. But finding a therapy or combination of therapies to effectively treat metastatic cancer can be difficult. That’s because cancers resulting from metastasis often differ from the original tumor at the molecular and genetic level. So, the same treatment used for the original cancerous tumor won’t necessarily be effective for new tumors resulting from metastasis, according to the American Society of Clinical Oncology (ASCO).
In addition, one type of treatment may work for a while to halt metastasized cancer growth, but then stop. When that happens, oncologists often try another type of treatment, then another. Adding to the difficulty in treating metastatic cancer is the fact certain therapies must be curtailed if a patient can’t physically tolerate them, even though the chemotherapy or other treatments might be effective at slowing or stopping a tumor’s growth.
However, there’s good news: Researchers are hopeful AI may find a way around these problems.
Using AI to help treat metastatic cancer looks promising
By using computerized systems to analyze enormous stores of information and scan complex data, AI can recognize and document patterns which scientists and doctors can then use to determine the best treatment for specific metastatic cancers for individual patients.
Research is still early, but so far the “deep learning” provided by AI is showing promise in finding ways oncologists can successfully treat metastatic cancer using “precision (or personalized) medicine” — a term that describes personalized treatments based on an individual’s genetic makeup, sensitivity to medications and exact state of health, and a tumor’s response to specific therapies.
Multiple studies have already shown ways AI can help extract characteristics and details from medical images of cancerous tumors and uncover disease characteristics the human eye could never catch, a method called radiomics.
For example, Columbia University researchers used radiomics to zero in on CT scans of gastric cancer tumors. The AI-powered results predicted, with high accuracy, which malignant tumors had characteristics that would respond to certain chemotherapies and which would not.
AI-directed personalized treatment for metastatic cancer
No one is claiming AI can cure metastatic cancer yet, but better treatments, which might totally stop cancer progression, could be in the not too distant future.
A case in point: A team of National University of Singapore (NUS) and University of California, Los Angeles (UCLA) researchers harnessed a powerful AI platform called CURATE.AI at the UCLA Institute of Urologic Oncology, to find an effective treatment for a man with advanced, stage IV prostate cancer.
The patient was diagnosed with metastatic castration-resistant prostate cancer (MCRPC), meaning his cancer had progressed despite androgen depletion therapy (ADT) to stop the male hormone, androgen, from feeding tumor growth.
To treat the man, the researchers turned to combination therapy — multiple drugs designed to work together, in synergy, to improve efficacy. This approach, by itself, wasn’t new. Combination therapy is a cornerstone in 21st century cancer treatment.
What was groundbreaking: using AI to find a precision treatment the man could tolerate and halt the malignancies in his body.
This is important because the standard combination approach, although it has improved overall survival for many cancers, comes with a significant problem — the combination of drugs used to treat metastatic cancer are usually given in doses that are fixed and high. Some patients may not be able to physically tolerate the treatment.
"Patients respond to chemotherapy differently from one another. Even a single patient's response to therapy can vary substantially over the course of treatment. In fact, many patients do not respond at all to the drug combination because the dosages, which can profoundly impact efficacy, are not suitable for them,” explained Professor Dean Ho, director of the Singapore Institute for Neurotechnology at NUS, who helped lead the study.
While fixed-dose combination therapy represents the standard of care in today’s treatment of cancer, it can also serve as a barrier to optimal treatment for individual patients suffering from metastatic cancer, Ho and colleagues explained in a research paper published in Advanced Therapeutics. What’s needed, they concluded, is highly personalized but effective treatment.
To that end, NUS engineers developed CURATE.AI. This artificial intelligence platform compiles detailed medical data about a metastatic cancer patient, including chemotherapy types and amounts used to treat the person’s cancer and the corresponding changes to tumor sizes or levels of cancer biomarkers in the blood. Then the AI program uses this information to create a CURATE.AI profile, or “map,” for the detailed treatment of a patient’s metastatic cancer.
The treatment map provides an AI-calibrated plan, showing a cancer patient’s individualized, expected response to a treatment or combination of treatments, and identifies specific drug doses that may be changed frequently to enable the best possible treatment outcome at any given point in time.
"No two patients' profiles are alike, and as a patient's body and the cancer itself evolve during treatment, the CURATE.AI profile evolves as well, enabling the clinical and engineering teams to optimize care for the entire duration of treatment, an unprecedented advance for combination therapy," Ho explained.
Evidence AI may help treat metastatic cancer successfully
The UCLA/NUS team tested the CURATE.AI map to see how well it worked, using it to treat the patient with advanced, stage IV, treatment-resistant prostate cancer. Based on the AI calibrations, the man was given a novel drug combination consisting of an investigational drug, ZEN-3694, and enzalutamide, a widely used prostate cancer drug.
The research team used the CURATE.AI patient treatment map to continuously identify and raise and lower the optimal doses of each drug, a strategy known as “dynamic dosing.”
"Dynamic dosing in cancer therapy is not commonly used. In fact, drug dosing changes in oncology are typically performed only to reduce toxicity. CURATE.AI uniquely modifies drug dosing to increase efficacy,” Ho explained. “Our clinical study has shown that dosing can profoundly affect the efficacy and safety of treatment. A patient's clinical profile changes over time. The unique ability for CURATE.AI to rapidly identify the drug doses that result in the best possible treatment outcomes allows for actionable and perpetually optimized personalized medicine,"
The amounts of chemotherapy the prostate cancer research subject received were shifted as needed, so he could tolerate the treatment while the therapy was monitored and adjusted as the metastatic cancer responded to varying combinations of the drugs.
And it worked. The man’s metastatic cancer was halted, and he returned to his previously active life.
What’s more, the study provides evidence cancer treatments that may be discontinued because they no longer work according to standard, fixed doses, might actually still be effective ways to treat and potentially cure malignancies, if specific dose changes are geared to individual patients.
"With fixed dose therapy, patients are often switched to other drugs when they no longer respond to treatment. However, CURATE.AI has shown that patients can still respond to the therapies that have seemingly stopped working by continuously identifying the patient's optimal dosing parameters,” UCLA urologist and prostate cancer expert Allan Pantuck, MD, the lead clinician of the study, pointed out.
While more work and research are needed, discovering and documenting how AI may help treat metastatic cancer could be a game changer for stage IV cancer treatment in the future, individualizing treatment as the patient, and his or her metastatic cancer, change.
"No two patients' profiles are alike, and as a patient's body and the cancer itself evolve during treatment, the CURATE.AI profile evolves as well, enabling the clinical and engineering teams to optimize care for the entire duration of treatment, an unprecedented advance for combination therapy," researcher Ho said.
May 18, 2020
Janet O'Dell, RN