Mini tumour replicas can transform lung cancer treatment

The treatment of lung cancer has improved over the years, but doctors can still’t predict what patients will do following surgery.

About half of the patients whose tumors were “advanced, but operable” will experience a recurrence of their disease within two years. After they’re discharged, we see them every six months in the clinic and they get a CT scan. That’s all,” said Professor Tony Ng, from King’s College London. He is also the Head of Oncology Translational Research at GSK.

We don’t have much information if they relapse. “This patient took drug X. Let’s now try drug Y.”

Ng and his colleagues are experimenting with a new, ambitious approach in their lab. In the first of its kind trial, they are growing millions miniature replicas of cancers from dozens of NHS patient who have had surgery for non-small cells lung cancer.

They hope that by studying these “organoid tumours” grown in the lab, they will be able predict accurately which patients will experience a relapse and forecast which combinations of treatment will be most effective.

Ng was shown a petri plate by a colleague during a visit to GSK’s Stevenage research centre. The petri dish contained 30 small blobs each the size of a lentil. Each blob was made up of tens thousands of organoids derived from stem cell taken from a surgically-removed tumour.

GSK’s Stevenage laboratory is studying the organoids.

Each organoid is unique in that it shares the genetic mutations of each patient’s cancer. This is the first time we have paired an organoid with a patient who had a primary surgery and then returned to the hospital with metastasis (where cancer has spread).

Organoids can be a kind of substitute for patients. Organoids can test many treatments, while a patient can only take part in one drug trial.

It is expensive and difficult to create organoids, so the ultimate aim is to study the ones grown in Stevenage, to reveal hidden characteristics that cancers from many patients share and to predict how they will respond to different treatments.

Researchers will sequence the DNA of organoid cell and collect “transcriptomic data”, which examines a genetic material known as RNA, and gives an insight into gene activity.

The data from the organoids is combined with “digital twins” of each patient’s disease.

They will analyze the proteome, which is the collection of proteins cancer cells depend on to survive. The researchers will use high-resolution imaging to track how immune and fibroblast cell behave in the presence of the organoids. These cells normally help the body defend itself and repair, but can become “pro-tumour cells” by cancer.

These data will be combined in order to create an “electronic twin” of the disease of each patient.

Humans will have too much data to process. Scientists will instead rely on a form of AI called machine learning to detect patterns.

It is hoped that AI will be able to pinpoint a few “biomarkers,” a combination or proteins, DNA, and RNA that can predict the effectiveness of a drug. The hope is that AI can identify a small number of “biomarkers” — a combination of proteins, DNA and RNA — which will predict how well a drug will perform.

Ng is convinced that the approach could be a “step-change” in cancer treatment. He said that by combining AI and organoid tumour models, they could improve their ability to predict if a cancer patient’s disease will relapse. They also hope to determine the best possible treatment for each patient. “There is the potential to change the entire field.”