To diagnose lung cancer and determine the optimal treatment strategy, physicians typically rely on tissue examination. However, taking biopsies is invasive and doesn't always provide sufficient information. In her research, PhD researcher Esther Visser shows that blood tests can offer a less taxing alternative.
Source: Cursor / Martina Silbrníková
When lung cancer is suspected, the patient first undergoes radiological examination, which results in images in the form of lung scans," explains Esther Visser. "If anything suspicious is seen on those, a tissue biopsy usually follows." This involves removing a piece of tissue from the body and examining it, allowing doctors to make the correct diagnosis and gather crucial information about the tumor. This is essential for determining the most effective treatment strategy.
Invasive procedure
A biopsy is an invasive procedure. "It can be done in different ways," Visser explains. "Through surgery, with a needle inserted through the chest wall into the lung, or with a tube passed through the airway via the nose or mouth." Not only are these procedures very taxing for the patient, but they also don't always provide sufficient information.
Sometimes the tumor is difficult to reach, making it difficult to extract suitable tissue, or the biopsy contains too few cancer cells, requiring the procedure to be repeated. In addition, not every patient is eligible for a biopsy; some are too fragile to undergo the procedure. "Complications can occur, such as pneumothorax in response to the insertion of the needle," says Visser.
Large-scale study
Consequently, there's a great need for alternative diagnostic methods that are accurate and at the same time less invasive for the patient.
Visser focused her research on a promising method in which crucial information is obtained through blood tests. Blood sampling is not only significantly less taxing than a tissue biopsy, but may also provide a solution for patients in whom a biopsy is too risky. "It would have great added value if we could obtain additional diagnostic information from blood samples from patients," the PhD candidate says.
To investigate this method, a large-scale study (lung marker study) was launched by a partnership of six different hospitals.
In this study, blood was drawn from patients suspected of having lung cancer and examined extensively. Visser thoroughly analyzed the data collected to better understand how blood testing can help make an accurate diagnosis, determine the most appropriate treatment, and monitor the disease.
Clearly defined limits
"The idea is that we can measure several biomarkers - measurable biological indicators - in the blood that tell us more about the diagnosis," Visser explains.
Protein tumor markers (TMs), which are associated with lung cancer, can for instance help diagnose the disease. "These proteins are also found in healthy people, but in patients with lung cancer we sometimes see an increased concentration in the blood," Visser explains.
Although this connection was already known from previous studies, the knowledge wasn't yet directly applicable in clinical practice. "For a biomarker, you actually have to determine a certain limit first, so that you can establish with certainty that everyone above that limit has lung cancer," Visser explains.
However, as these proteins also occur in other diseases, it's difficult to find such clearly defined limits.
AI models
By training self-learning AI models with some of the patient data, Visser developed a diagnostic algorithm that can detect lung cancer based on the concentrations of different TMs.
The models are designed to clearly demonstrate how the disease is related to different proteins and how certain conclusions are reached. "All this information is useful, so you don't want it to disappear into a black box," says the PhD candidate.
Visser then tested the model on the remaining patient data to validate its reliability. The results were impressive: the models diagnosed lung cancer in a subset of patients with 98 percent accuracy. Although the models need further validation in follow-up studies with more patient data, the initial results are promising.
Bits of DNA floating around
Detecting genetic mutations through blood can also provide valuable insights, Visser's research shows.
"In some patients with lung cancer, we find what is known as circulating tumor DNA (ctDNA) in the blood," she explains. "These are small pieces of DNA that come from tumor cells and float around in the blood." By examining the ctDNA, doctors can gain insight into the genetic changes in the tumor, which can help determine the best treatment.
"Nowadays, there are more treatment options than there used to be," Visser emphasizes. Targeted therapies that focus on DNA mutations are potentially much more effective than chemotherapy in patients with these genetic abnormalities. "For these treatments, however, we need more detailed information about the tumor, so examining the pieces of DNA in the blood is particularly valuable," she says.
Making a difference for patients
From the beginning, Visser's goal was to develop something that could be applied in clinical settings.
So she considers the fact that this method is now actually being used at the Catharina Hospital in Eindhoven as the biggest breakthrough. "When genetic mutations are found in the blood, the patient can be spared an invasive biopsy," she says.
The method has already helped real patients who weren't eligible for a biopsy. "Doctors couldn't do much for these patients anymore, but this way some of them were still able to receive targeted, life-extending treatment." Visser thought it was an extraordinary experience to closely collaborate with doctors and contribute to a method that's not only applicable in practice, but can really make a difference for patients. "The best part is that our research is now having a real impact on patients' lives."
