# Scan-based index helps predict which breast cancer patients will respond to drug

> Hong Kong researchers have devised an MRI-based score that can show, before treatment starts, which breast cancer patients are most likely to benefit from an AKT inhibitor.

- Published: 2026-11-01 by Tickety-Boo News
- Section: Health
- Location: Hong Kong
- Web page: https://ticketyboonews.co.uk/stories/scan-based-index-helps-predict-which-breast-cancer-patients-will/
- Original report: [Tumor index can identify breast cancer patients' drug response](https://medicalxpress.com/news/2026-10-tumor-index-breast-cancer-patients.html) (Medical Xpress)

## In brief

- PolyU researchers created the Tumour Heterogeneity Index, an MRI-based score predicting breast cancer response to AKT inhibitors.
- Patients with a high index had an estimated complete response rate of up to 56.1%.
- The method is noninvasive, uses existing scans, and was validated across different imaging equipment and institutions.

Researchers in Hong Kong have developed a scan-based score that can indicate, before treatment begins, which breast cancer patients are likely to respond to a type of drug known as an AKT inhibitor. The tool, called the Tumour Heterogeneity Index (THI), comes from a team at The Hong Kong Polytechnic University (PolyU).

THI is built from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), a scan already widely used to diagnose cancer and plan treatment. Using radiomics techniques, the team turned that imaging into a measure of how complex a tumour's internal structure is. Because it relies on scans rather than extra procedures, the assessment is noninvasive.

The study looked at patients with high-risk, early-stage breast cancer who were given the investigational AKT inhibitor MK2206 alongside chemotherapy before surgery. The drug had previously improved pathological complete response in a large trial, but its overall effect was limited and results varied widely between patients. Patients with a high THI had an estimated complete response rate of up to 56.1%, significantly higher than those with a low THI.

Promising results were also seen in triple-negative breast cancer, a subtype that is typically harder to treat. THI also did better than conventional receptor testing at picking out suitable patients, and it shed light on the immune cells and blood vessels around the tumour.

Unlike a biopsy, which samples only a small part of a tumour, THI assesses the whole thing. The team says it has tested the method for repeatability and reproducibility, and that it performed reliably across different imaging equipment and health care institutions, which is a cracking sign for wider use.

The researchers have also created a second model, DCE-VASC-TRACK, which follows changes in tumour blood vessels over repeated scans. In patients with HER2-negative breast cancer, it predicted response to chemotherapy more effectively than tumour size alone. The two studies appear in Breast Cancer Research and Radiology: Artificial Intelligence.

Chair professor Cai Jing said the approach could spare patients treatments with limited effectiveness and cut unnecessary side effects. The team hopes to extend it to other cancers in future.

## Why it matters

Cancer drugs often work well for some patients and barely at all for others, so knowing in advance who will benefit can mean fewer needless side effects and a quicker route to the right treatment. Using scans that hospitals already perform, rather than costly genetic sequencing, could make that kind of tailored care easier to offer more widely. The work is still at the research stage, but it points towards more personalised cancer treatment.