image: An MSK researcher working in the lab.
Credit: Memorial Sloan Kettering Cancer Center
New research from Memorial Sloan Kettering Cancer Center (MSK) supported the approval of a new treatment for ER-positive, HER2-negative advanced breast cancer; suggests a new strategy for overcoming resistance to Akt inhibitors in patients with PTEN-mutated prostate cancers; provides new understanding of B cell lymphomas that could lead to new treatments; examines using ChatGPT to help capture patient-reported outcomes; and developed the first customizable, scalable model for palliative care training for emergency departments.
FDA approves new treatment for ER+, HER2- advanced breast cancer
On January 17, the U.S. Food and Drug Administration (FDA) approved datopotamab deruxtecan for people with metastatic ER-positive, HER2-negative breast cancer whose tumors cannot be removed with surgery. MSK medical oncologist Komal Jhaveri, MD, FACP, was one of the lead authors of the study, and MSK was among the institutions enrolling the most patients in the trial.
Datopotamab deruxtecan is an antibody-drug conjugate (ADC). This type of therapy consists of an antibody that binds to specific proteins or receptors on the surface of a cancer cell linked to a drug that is toxic to the cancer cell.
The approval was based on results from a phase 3 trial testing datopotamab deruxtecan compared with chemotherapy. The trial, called TROPION Breast 01, showed that people receiving the new therapy had better overall response and progression-free survival than those who received chemotherapy. Datopotamab deruxtecan also appeared safe, without serious side effects.
“The FDA approval gives doctors an important new treatment option for people with inoperable ER-positive, HER2-negative metastatic breast cancer — the most common subtype,” Dr. Jhaveri says. Read more in Journal of Clinical Oncology.
Discovering the role of mTOR in resistance to Akt-targeting drugs in PTEN-deficient prostate cancers
A team of MSK researchers led by surgeon-scientist Brett S. Carver, MD, have discovered how some metastatic prostate cancers develop resistance to targeted drugs called Akt inhibitors. Patients with prostate cancers that have a deletion or mutation in the PTEN gene are less responsive to androgen receptor (AR) inhibitors, the main treatment for metastatic prostate cancer. In recent years, two phase 3 clinical trials, based on previous findings from Dr. Carver and colleagues, showed improved clinical outcomes and survival for patients with the PTEN mutation who received treatment combining AR inhibitors with Akt inhibitors.
However, not all patients with PTEN alterations respond to Akt inhibition, and those who do sometimes develop resistance. To understand the mechanisms of response and resistance to Akt inhibitors, the researchers studied patient-derived PTEN-mutated prostate cancer organoids, developed in collaboration with Yu Chen, MD, PhD, and Charles Sawyers, MD, to model the resistance observed in the clinic.
They discovered that mTOR, a protein that helps control several key cellular functions, including cell division, survival, and protein synthesis, played a key role in driving resistance to Akt inhibition. Across multiple prostate cancer models, restoration of mTOR activity drove resistance to Akt inhibitors, while directly targeting mTOR restored sensitivity. The findings suggest a new strategy for overcoming resistance to Akt inhibitors in patients with PTEN-mutated prostate cancers through the addition of a drug targeting mTOR. Read more in Science Advances.
New understanding of B cell lymphomas provides potential targets for future treatments
An MSK research team investigating B cell lymphoma has identified new potential targets for future treatments. The study examined the role of a mechanism called ubiquitin-mediated degradation in two subtypes of B cell lymphoma — diffuse large B cell lymphoma and follicular lymphoma — identifying that the protein FBXO45 was involved in their development.
When FBXO45 was deleted in laboratory mice, it led to the formation of lymphoma, the researchers found. They also discovered that FBXO45 targets another protein called ARHGEF2/GEF-H1 for degradation, and that removing both FBXO45 and ARHGEF2 reduced lymphoma formation.
When the team — overseen by senior authors Kojo Elenitoba-Johnson, MD, and Megan Lim, MD, PhD, who have a joint lab in MSK’s Human Oncology and Pathogenesis Program — looked at human lymphoma samples, they found a common pattern of changes in the FBXO45 and ARHGEF2 genes, suggesting they play a role in the development of these cancers.
The study also found that when FBXO45 was silenced, lymphoma cells became more sensitive to a type of targeted therapy called MEK inhibition. “The new understanding of the role that FBXO45 and ARHGEF2 play in these cancers will help us explore new therapeutic strategies to improve outcomes for patients with B cell lymphomas,” Dr. Elenitoba-Johnson says. Read more in Cancer Discovery.
Harnessing ChatGPT to capture patient-reported outcomes
Patient-reported outcomes (PROs) are an important component of clinical research. They provide a patient’s perspective on their own health and help researchers understand how someone’s disease and treatment may be affecting their life. PROs are generally collected by asking patients to fill out standardized questionnaires that ask about things like their pain symptoms and how bothered they are by them. But in a new proof-of-concept study from MSK, researchers decided to look at whether large language models could help to capture PROs in a patient’s own words, allowing for a more nuanced interpretation of the patient experience.
The researchers, led by behavioral scientist Thomas Atkinson, PhD, used archived transcripts from interviews with nine patients and two caregivers about how they were coping with the uncertainty of living with cancer. The responses were put into a model called GPT-4o, which had not previously received any specific training in how to interpret this type of information. The researchers found that GPT-4o was able to classify the PROs into categories previously labeled by trained human coders based on text entries alone.
The research team, which also includes behavioral scientist Yuelin Li, PhD, and data scientist Aleksandr Petrov, MS, says these kinds of models eventually may be useful for developing interactive tools to help patients communicate more effectively. For example, if a patient reports, “I felt sick to my stomach,” this could be recorded as “mild nausea/vomiting.” The research suggests artificial intelligence–guided PRO reporting may expand the information obtained from questionnaires beyond the questionnaires’ fixed contents. Read more in the Journal of the National Cancer Institute.
Palliative care training for emergency departments can serve as a national model
Half of Americans 65 or older who are seriously ill will visit an emergency department in the last month of their life. But being admitted to the hospital is not always the safest option for these patients — and many can benefit from receiving care at home or in hospice. Palliative care physicians are not always available in emergency departments, putting the onus on emergency medicine physicians to make complex medical decisions for a population with unique needs.
A team led by palliative care specialists at MSK and the NYU Grossman School of Medicine aimed to equip emergency medicine practitioners with evidence-based palliative care training to help guide their decision-making.
Corita Grudzen MD, MSHS, Keith Goldfeld, DrPH, and their team conducted a clinical trial to evaluate whether such training could make a difference — ideally lowering the number of inpatient hospitalizations by increasing referrals to home care or hospice, and thus improving survival rates.
Investigators evaluated almost 99,000 initial patient visits to 29 U.S. emergency departments before and after palliative care training was implemented. Although results did not show a significant decrease in hospital admission rates following palliative care training, the effort established the first customizable, scalable model for palliative care training that could be used by hospitals nationwide. With continued development, these training resources have the potential to significantly improve tailored care for older patients with serious illness, the researchers say. Read more in JAMA.