Research by Catharina Hospital and TU/e brings this possibility closer and has received international recognition.
Source: Catharina Hospital Eindhoven / Geert Piek
A patient has a suspicious lesion in their lung. To determine whether it is lung cancer, doctors want to examine a sample of tissue. But this takes time and is not always possible. Sometimes a patient is too frail to undergo the procedure. In other cases, tissue is collected, but the results do not provide enough clarity. The patient may then have to undergo the same invasive procedure again, sometimes more than once.
Researchers at Catharina Hospital and TU/e are investigating whether there is another way. By combining blood tests with information from CT scans, they hope to detect lung cancer more accurately and at an earlier stage, particularly in patients for whom a conventional biopsy is difficult or impossible. Their initial results are promising. PhD candidate Hao Cao (29) recently received an international ESMO Merit Award for this work at the ESMO Congress in Madrid , one of the world's leading scientific conferences on cancer.
Blood tests and scans complementing each other
During a biopsy, a doctor removes a small sample of tissue to check for the presence of cancer cells. This can be done, for example, by examining the airways using an endoscope or by inserting a needle through the chest wall to collect tissue. However, the procedure takes time, can be burdensome for patients and does not always provide a conclusive result.
Cao, who works at the General Clinical Laboratory of Catharina Hospital, has been investigating for some time whether blood tests can help. Certain substances in the blood can indicate the presence of cancer. These substances are known as tumor markers.
"Using blood tests alone, we were able to identify around 62 per cent of patients with lung cancer," Cao explains. "That is useful information, but it also means we still miss a proportion of patients. So we started looking at whether we could improve that figure by adding other information."
That information is already available at the hospital. Patients suspected of having lung cancer generally undergo a CT scan. Radiologists use these scans to assess the size, shape and other characteristics of suspicious lesions in the lungs.
The researchers combined the blood test results and CT scan data in a computer model. They used data from 346 patients, divided into two study groups. In the first group, the proportion of lung cancer cases correctly identified increased from 62 to approximately 78 per cent when the radiologist's assessment was added to the blood test results.
"That is a clear improvement," says Cao. "The blood tells us one thing, while the images provide different information. By combining the two, we get a more complete picture."
What can artificial intelligence add?
The research team took things a step further. Together with biomedical engineer Jasper Stiekema and TU/e experts PhD candidate Cris Claessens and Associate Professor Fons van der Sommen , they investigated whether artificial intelligence (AI) could also analyze the CT scans automatically. To do so, they used a so-called deep learning model: a computer system capable of recognizing patterns in medical images.
The results were striking. The combination of blood tests and AI correctly identified approximately 76 percent of lung cancer cases. This was comparable to the results achieved by combining blood tests with a radiologist's assessment of the scans. In a second, independent patient group, the AI model combined with blood tests achieved an identification rate of 73 percent.
Cao finds it just as interesting that AI did not outperform the radiologists. "Radiologists can already extract a tremendous amount of valuable information from a CT scan using their own knowledge and experience. AI can help, but our research also demonstrates just how valuable a radiologist's expertise is."
The researchers emphasize that the new method is not yet a substitute for a biopsy. A negative result does not rule out lung cancer completely. Also, the method cannot yet determine the exact type of lung cancer, even though this information is important when selecting the right treatment.
Nevertheless, Cao sees potential. "Imagine a patient who has already undergone two biopsies, but the results are still inconclusive. Combining blood tests and medical imaging could provide additional indications that something is wrong after all. This could help doctors decide whether further investigations are needed."
International recognition
The research was presented at the European Society for Medical Oncology (ESMO) Congress in Madrid at the end of October. Cao received a Merit Award for his scientific contribution. The award recognizes early-career researchers whose submitted research has been selected for its scientific quality.
Cao was informed that his research had received the award before the congress began. In Madrid, he presented the results during a poster session, where researchers from various countries had the opportunity to ask questions and exchange ideas.
"People came to see us from countries including the Netherlands, England and China," Cao says. "Some were surprised that we conduct so much scientific research at a non-university hospital. That was nice to hear. At Catharina Hospital, we naturally work closely with the university, but our research is also closely connected to everyday patient care."
According to Cao, collaboration is an important part of the project and involves a multidisciplinary team of healthcare professionals with diverse areas of expertise. Catharina Hospital contributes its knowledge of laboratory testing and clinical practice, while TU/e researchers focus on e.g. developing advanced computer models.
Those involved include supervisor and clinical chemist Volkher Scharnhorst, who is affiliated with Leiden University Medical Center (LUMC); TU/e researcher Federica Eduati ; pulmonologist-oncologists Christi Steendam and Ben van den Borne from Catharina Hospital; and thoracic radiologists Rianne Verhees and Astrid Donkers. Cao and Stiekema are also continuing to develop the research and prepare it for publication.
For now, the new method remains at the research stage. The next step is to validate the system further in everyday clinical practice, without using its results to make medical decisions at this stage.
"We are not going to be able to answer every question with this method," Cao emphasises. "But if we can provide doctors with additional reliable information, it could make a real difference for patients who have difficulty obtaining a tissue diagnosis."
Further research supported by the Catharina Health Foundation
This is not the first time Cao has received recognition for his research. Earlier, he won an award at Catharina Hospital's Science Evening for his research into the faster detection of cancer using fluid around the lungs .
He is also working on another lung cancer project, for which he received €45,000 from the Catharina Health Foundation . In this study, he investigates whether blood tests can detect additional abnormalities in tumor cell DNA. This could help doctors determine which medicines are most suitable for individual patients. This research is also nearing completion. Cao expects to publish the results later this year.
