One of the innovative features of the I-SPY trial is its extensive use of magnetic resonance imaging (MRI) to monitor tumor response to neoadjuvant therapy for breast cancer. Results of MRIs taken at various points before, during and after treatment provide key data for the randomization algorithm that fuels I-SPY 2’s adaptive design.

In a manuscript publishedSeptember 4, 2018 in the journal ‘RADIOLOGY,’ I-SPY investigators report that an advanced MRI technique known as “diffusion-weighted MRI” (DW MRI) may provide advantages over standard MRI methods. Using DW MRI measures fromI-SPY 2 participants who were half-way through treatment, investigators were able to predict with some accuracy whether they would achieve pathological complete response (pCR), or complete disappearance of their tumor, by the end of the 24 weeks of treatment. I-SPY has recently shown that pCR is very strongly indicative of excellent long-term outcome in early, high-risk breast cancer patients.

“Ultimately, we hope to use MRI to predict very early in the course of treatment if a therapy is going to get a woman to pCR”, said lead author Savannah Partridge, PhD, of University of Washington. “That way, if a treatment isn’t going to work, we will know quickly and can try something that may work better.”

The standard MRI method currently employed by I-SPY 2, known as "contrast enhanced" MRI, requires the intravenous injection of a contrast agent (gadolinium) to increase the visibility of tumors. The change in the functional volume of the tumor from one MRI to the next is used in I-SPY as a measure of treatment response.

DW MRI, however, essentially measures different properties of tissues. Effectively, it creates an image based on the amount ofBrownian (or random) motion of water molecules in the tissue, cells and interstitial space between cells. Changes in the number and/or condition of the cells (known as cellularity) due to the effects of treatment can therefore be detected and quantified.

“It is a very promising result, now the task at hand is to refine our methods further and to combine the valuable information from DW MRI along with other factors to predict response with greater accuracy and even earlier in treatment,” said Partridge.

The manuscript was accompanied by an editorial commissioned by the journal, which lauded I-SPY 2 for its efforts in quality control of imaging within a multicenter trial. Variability of MRI measures across different clinical centers has been a long-standing barrier to realizing the full potential of MRI in clinical trials.

The are many steps in the imaging procedures that could introduce slight variations that add up over time – differences in the equipment calibration, different operators, variations in analysis, etc. The editorial’s author refers to the quality control required as a ‘mammoth task’ and a ‘marathon’ undertaking.

Nola Hylton, PhD, principal investigator of the study and I-SPY’s imaging chair agrees.

“It has been over a decade of work, to define and incrementally refine the strict QA/QC program for imaging across I-SPY’s 16-20 different clinical centers. We’re always learning, always trying to do better.”

The first I-SPY trial, I-SPY 1 ran from 2003 through 2007, with the goal of determining whether MRI could accurately assess treatment response following neoadjuvant breast cancer treatment, and to establish standards for breast imaging that could be routinely applied across multiple centers.

“We’re very proud of what the I-SPY imaging group, researchers from centers around the country, have accomplished together,” said Hylton. “No one does this better than I-SPY.”

Original manuscript:  https://pubs.rsna.org/doi/10.1148/radiol.2018180273

Accompanying editorial: https://pubs.rsna.org/doi/10.1148/radiol.2018181717

‍