Monitoring Treatment Response in Radiation Therapy: A Scoping Review of Dosimetric, Imaging, and Artificial Intelligence-Based Approaches
This scoping review synthesizes current methodologies for monitoring radiation therapy response, highlighting the complementary roles of in vivo dosimetry, functional imaging, and emerging artificial intelligence-driven predictive models while emphasizing the need for further standardization and clinical validation.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Cancer is a disease of uncontrolled growth, where cells multiply without stopping and spread to other parts of the body. To stop this, doctors often use radiation therapy, a treatment that beams high-energy particles at a tumor to damage the DNA inside the cancer cells, causing them to die. This process is a delicate balancing act: the beam must be strong enough to destroy the tumor but precise enough to spare the healthy tissue surrounding it. For decades, doctors have relied on standard scans to see if the tumor is shrinking, much like checking the size of a balloon after letting some air out. However, a tumor can change its internal chemistry or blood flow long before it changes its physical size, meaning a patient might be getting better or worse without the doctor knowing until it is too late. The question of how to watch this process in real time, to see exactly how much radiation the patient is actually receiving and how the tumor is reacting to it, has become a critical challenge in modern medicine.
A team of researchers from Ghana, led by Emmanuel Amponsah and colleagues at the Kwame Nkrumah University of Science and Technology, set out to map the current landscape of tools used to answer this question. They conducted a broad review of scientific literature published between 2010 and 2024, searching for studies that looked at how doctors monitor radiation therapy. They focused on two main ways radiation is delivered: beams from outside the body and radioactive sources placed directly inside or near the tumor. After sifting through hundreds of reports, they selected twenty-one studies that met their strict criteria, including clinical trials, computer simulations, and doctoral research. Their goal was not to invent a new machine, but to understand what tools currently exist, how well they work, and where the gaps remain in the effort to track treatment success.
The researchers found that the current approach relies on three distinct pillars working together. The first pillar is a method called in vivo dosimetry. This involves placing small detectors directly on or inside a patient to measure the actual amount of radiation hitting the body during the treatment. Instead of just trusting a computer plan, these detectors tell the doctor exactly what dose was delivered. The review highlighted that while this method is powerful for ensuring safety, it is not yet used in every hospital. Some detectors work instantly, allowing doctors to stop a treatment session if the dose is wrong, while others require a delay before they can be read. The studies showed that these tools can catch small errors, with some measurements differing from the plan by less than one percent, but the technology is still being refined to work smoothly in routine care.
The second pillar is imaging, which looks at the tumor itself. The review distinguished between standard scans, which show the shape and size of a tumor, and functional scans, which reveal how the tumor is working. Standard scans, like computed tomography, are common and good for seeing if a mass is getting smaller. However, the researchers found that functional magnetic resonance imaging is far more sensitive to early changes. These advanced scans can detect shifts in how water moves through the tumor or how blood flows to it, providing a glimpse into the biological response before the tumor physically shrinks. For example, one study noted that specific measurements from these scans could distinguish between patients who achieved a complete metabolic response and those who did not. Another study found that combining different types of scans, such as those that show both structure and metabolism, offered a clearer picture of how the treatment was affecting the tumor's environment.
The third pillar is the use of artificial intelligence and radiomics. Radiomics is a technique that uses computers to pull out thousands of tiny, invisible details from medical images—details too subtle for the human eye to see. Artificial intelligence then analyzes these details to predict how a patient will respond to treatment. The review found that these computer models are showing great promise. In studies involving cancers of the cervix, prostate, and rectum, these models could predict outcomes with high accuracy, sometimes matching or exceeding the performance of traditional methods. For instance, one study tested twelve different computer algorithms and found that one could predict treatment success with ninety-three percent accuracy. These tools allow doctors to look at the data and foresee whether a tumor is likely to shrink or if the treatment needs to be adjusted, potentially saving time and improving results.
Despite these advances, the researchers identified significant gaps in how these tools are used. They created a map of the evidence and found that while some cancers, like those of the cervix and prostate, are well-studied, others like breast and lung cancer have very little research on how to monitor them with these specific tools. More importantly, they found that no study had successfully combined all three pillars—measuring the dose, scanning the tumor, and using artificial intelligence—into a single, unified system. Currently, these methods often operate in isolation. A doctor might check the dose, then look at a scan, and then use a computer model separately, rather than having a system that integrates all this information at once to guide the treatment in real time.
The authors conclude that while the technology to monitor treatment response is advancing rapidly, it is not yet ready for widespread, standardized use in every clinic. The tools exist, from the tiny detectors that measure radiation to the smart computers that analyze images, but they need to be tested more thoroughly and made to work together seamlessly. The review suggests that the future of radiation therapy lies in bringing these three areas together into a closed loop, where the dose, the image, and the prediction constantly inform one another. Until then, the medical community must continue to refine these methods, ensuring that the promise of seeing exactly how a tumor is reacting to treatment becomes a reality for every patient, not just those in specialized research centers.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.