Preoperative prediction of microsatellite instability in rectal cancer using multiparametric MRIderived intratumoral heterogeneity and exploration of biological mechanisms: a multicenter study

Cancer is rarely a single, uniform mass. Inside a tumor, cells can vary wildly from one spot to another, a phenomenon scientists call intratumoral heterogeneity. This internal diversity is a major reason why cancers are so difficult to treat; a drug that kills one group of cells might leave another group untouched, allowing the disease to return. One specific type of colorectal cancer, known as microsatellite instability, behaves differently from the rest. Patients with this type often respond very well to immunotherapy but poorly to standard chemotherapy. Currently, doctors must wait until after surgery or perform a biopsy to determine if a patient has this specific type of cancer. These methods are invasive and only sample a tiny piece of the tumor, potentially missing the complex variations happening elsewhere in the mass.

A new study from a team of researchers in China offers a way to see these hidden differences before surgery even begins. By using magnetic resonance imaging, the same technology used to create detailed pictures of the body's soft tissues, the researchers developed a method to map the internal landscape of rectal tumors. They found that the way these tumors scatter their signals on an MRI scan can reveal whether they are the microsatellite instability type. This approach not only helps predict the cancer's behavior without cutting into the patient but also points toward specific biological mechanisms that drive the disease, bridging the gap between what a camera sees and what is happening inside the cells.

The researchers began by gathering data from 464 patients with rectal cancer across three different hospitals. They focused on patients who had undergone a specific type of MRI scan before any treatment was given. These scans included two main sequences: one that showed the structure of the tissue in high detail and another that measured how water molecules moved within the cells. The team then used a computer to break down the tumor in these images into thousands of tiny three-dimensional blocks. Instead of treating the tumor as one solid object, they analyzed how the signals varied from block to block. They calculated a score that represented how scattered or mixed these signals were across the entire tumor. A high score meant the tumor was very diverse internally, while a low score suggested it was more uniform.

To test if this method worked, the team built a computer model that combined these imaging scores with standard patient information, such as age, tumor size, and location. They trained the model on data from one group of patients and then tested it on two other groups to see if it could make accurate predictions on new, unseen data. The results were promising. The combined model, which used both the imaging scores and the patient details, correctly identified the microsatellite instability status in about 86 percent of the training cases and maintained strong accuracy in the external groups. The imaging scores alone were not enough to make a perfect prediction, but when added to the clinical data, they significantly improved the model's ability to distinguish between the two types of cancer. This suggests that the internal chaos of the tumor, visible on the scan, carries unique information that standard clinical factors cannot capture on their own.

The study did not stop at prediction; the researchers wanted to understand why these tumors looked different on the scans. They turned to a massive public database of genetic information from thousands of cancer patients to look for molecular clues. By comparing the genes active in microsatellite instability tumors against those in standard tumors, they identified a specific protein called AKT1 that behaved differently. In standard tumors, this protein was highly active, but in the microsatellite instability type, it was much less active. To confirm this finding, the team examined actual tissue samples from ten patients in their own hospital. Using a staining technique that makes proteins visible under a microscope, they measured the amount of AKT1 present. The results matched the genetic data: the standard tumors had significantly higher levels of this protein than the microsatellite instability tumors.

This discovery provides a potential biological explanation for what the MRI scans were seeing. The protein AKT1 is part of a pathway that drives cell growth and division. In standard tumors, where this protein is abundant, the cells may grow in a more uniform, aggressive manner, leading to the consistent signals seen on the scans. In contrast, the microsatellite instability tumors, with their lower levels of this protein, appear to have a different internal structure, perhaps driven by immune system activity or other factors that create the patchwork of signals the researchers measured. While the study does not prove a direct cause-and-effect link between the protein levels and the specific image patterns, it offers a strong hypothesis that the visual differences on the MRI reflect real, measurable differences in the tumor's biology.

The researchers acknowledge that their work has limits. The study was retrospective, meaning it looked back at data that had already been collected, and the number of tissue samples used for the protein verification was small. They also noted that their method relied on specific MRI sequences and did not include contrast dye, which might provide even more detail in the future. Despite these constraints, the study demonstrates that non-invasive imaging can do more than just locate a tumor; it can reveal the complex, heterogeneous nature of the disease inside. By quantifying this internal diversity, doctors may soon be able to tailor treatment plans earlier and more accurately, sparing patients from unnecessary procedures and ensuring they receive the therapy most likely to work for their specific type of cancer.