Better measurement helps doctors know sooner whether a patient is actually benefiting.
At Brigham and Women's Hospital, a radiologist and his team have brought a measure of clarity to one of oncology's quieter dilemmas: how best to see whether a treatment is truly working. By comparing imaging methods used to track melanoma tumors in the brain, Raymond Y. Huang's research points toward two approaches — a modified measurement standard and three-dimensional volumetric analysis — that more faithfully predict how long patients will live and how long their disease will stay at bay. In a field where inconsistency between hospitals and clinical trials can obscure the truth of a patient's progress, the search for a common language of measurement is itself a form of care.
- Melanoma patients whose cancer has reached the brain face not only an aggressive disease but a fragmented system — doctors measuring tumors differently depending on where they practice, making it harder to know whether any treatment is truly succeeding.
- The absence of a universal imaging standard has quietly undermined clinical trials, making cross-study comparisons unreliable and leaving individual patients without clear answers about their own trajectories.
- Huang's team tested three competing measurement systems against real patient data from the multi-center CheckMate 204 immunotherapy trial, seeking the method that most accurately predicted survival — and two clear frontrunners emerged.
- Both mRECIST and 3D volumetric measurement outperformed traditional approaches, even for small tumors that older methods struggled to assess reliably.
- The remaining obstacle is practical: volumetric measurement is labor-intensive and error-prone when done by hand, so the team is now building automated segmentation software to make the method fast, consistent, and universally accessible.
A radiologist at Brigham and Women's Hospital has identified which imaging methods work best for tracking how melanoma brain tumors respond to immunotherapy — a finding with quiet but significant consequences for some of the sickest cancer patients. Raymond Y. Huang and his team published their results in the Journal of Clinical Oncology after analyzing MRI scans from the CheckMate 204 trial, a multi-center study testing the immunotherapy combination of nivolumab and ipilimumab in patients whose melanoma had spread to the brain.
The problem they set out to solve sounds technical but carries real human weight. When a patient with brain metastases begins treatment, doctors must determine whether the drugs are working — and they do so by measuring tumors on MRI scans. But there is no universal agreement on how that measuring should be done. Different hospitals and different trials use different methods, creating inconsistency that can obscure whether a patient is genuinely improving.
Huang's team compared three imaging measurement systems against each other, asking which best predicted survival and progression-free outcomes. Two emerged as superior: a modified version of the standard mRECIST criteria, which measures tumors in two dimensions, and volumetric measurement, a three-dimensional analysis that calculates actual tumor volume. Both outperformed traditional approaches and proved reliable even for small tumors — a known weakness of older methods.
The implications extend beyond any single patient. Inconsistent measurement standards make it harder to compare results across studies and harder to determine which treatments genuinely work best. Standardizing around mRECIST and volumetric analysis could resolve that problem — but volumetric measurement currently requires painstaking manual tracing of tumor boundaries on each scan, introducing both time costs and human error. To address this, Huang's team is developing automated segmentation software that would perform that work consistently, giving any hospital or oncology practice a reliable, shared tool.
For patients with melanoma that has reached the brain, better imaging assessment won't change the nature of their illness — but it could help their doctors make faster, more confident decisions about whether to stay the course or change direction. In oncology, that kind of clarity can matter in ways measured in months.
A radiologist at Brigham and Women's Hospital has identified which imaging methods work best for tracking how melanoma tumors in the brain respond to immunotherapy—a finding that could reshape how doctors monitor some of the sickest cancer patients. Raymond Y. Huang and his team published their work in the Journal of Clinical Oncology after analyzing scans from a clinical trial that tested a combination of two immunotherapy drugs, nivolumab and ipilimumab, in patients whose melanoma had spread to the brain.
The question sounds technical, but it matters enormously. When a patient with brain metastases—tumors that form when cancer cells migrate from the skin to the brain—starts treatment, doctors need to know whether the drugs are working. They look at MRI scans and measure the tumors. But there's no universal agreement on how to do that measuring. Some radiologists use one method, others use another. The result is inconsistency that can muddy the picture of whether a patient is actually improving.
Huang's team took a different approach. They gathered MRI scans from patients in the CheckMate 204 trial, a multi-center study that ran for two years or longer, and tested three different imaging measurement systems against each other. The question was simple but crucial: which method best predicted how long patients would survive, and how long they would go without their cancer getting worse?
Two approaches emerged as winners. The first was a modified version of a standard measurement system called mRECIST, which assesses tumor size in two dimensions. The second was volumetric measurement—essentially a three-dimensional analysis that calculates the actual volume of each tumor rather than just its width and length. Both methods proved significantly more accurate at predicting survival outcomes than the traditional approaches doctors had been using. Importantly, both worked well even for small tumors, which had been a weak point of older measurement systems.
The implications ripple outward. Right now, when oncologists evaluate whether a patient's brain metastases are responding to immunotherapy, they may be using different yardsticks depending on where they practice or which trial they're enrolled in. That inconsistency makes it harder to compare results across studies, harder to know which treatments truly work best, and potentially harder for individual patients to get clear answers about their own progress.
Huang's findings suggest that adopting mRECIST and volumetric measurement as standards could fix that problem. But there's a practical obstacle: volumetric measurement requires someone to carefully trace the boundaries of each tumor on each scan slice—tedious, time-consuming work that introduces room for human error. That's why Huang's team is now developing automated segmentation technology, software that would do that tracing automatically and consistently. The goal is to create a standardized tool that any hospital, any oncology practice, could use to measure brain metastases the same way.
For patients with melanoma that has reached the brain, the stakes are high. These are people with advanced cancer, facing aggressive treatment and uncertain outcomes. Better imaging assessment won't cure them, but it could help their doctors make faster, more confident decisions about whether their current treatment is working or whether they need to switch course. In oncology, that kind of clarity can mean the difference between months.
Citas Notables
Our study examines how different imaging criteria can be used to assess brain tumor responses in patients with melanoma treated with immunotherapy drugs. We identified one imaging method that was more accurate in predicting survival compared to others.— Raymond Y. Huang, MD, Ph.D., Brigham and Women's Hospital
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter which imaging method you use to measure these tumors? Aren't they all looking at the same thing?
They're looking at the same tumors, yes, but measuring them differently. One method might measure length and width only. Another measures volume in three dimensions. When you're trying to predict whether a patient will survive longer, those differences actually change the answer.
So the three-dimensional approach is always better?
It was in our study, yes—along with the modified two-dimensional method. But volumetric measurement has a problem: it requires a person to manually trace each tumor on every scan slice. That's labor-intensive and introduces inconsistency.
That's why you're building the automated software?
Exactly. If we can make a computer do that tracing consistently, we remove the human variability. Then every hospital, every trial, measures the same way.
What happens to patients in the meantime, while you're developing this tool?
They're still being evaluated with whatever method their hospital uses. Some get better assessment than others, depending on where they're treated. That's the problem we're trying to solve.
And the immunotherapy drugs themselves—did your study show they work?
The trial showed they work for some patients. Our study was about how to measure that response accurately. Better measurement helps doctors know sooner whether a patient is actually benefiting.