YEARS in Cancer Patients: What the Hydra Trial Really Means

Dr. James Whitfield

If, uh, if you are a clinician working in an emergency department or an oncology clinic, you know this scenario by heart. A patient with active lung cancer comes in, and they are just a little more short of breath than usual, or maybe their heart rate is slightly up. Your mind instantly jumps to, to pulmonary embolism. PE is a massive deal here, Elena. Active malignancy increases the risk of venous thromboembolism up to seven fold, and it is actually a leading non cancer cause of death in these patients.

Dr. Elena Rodriguez

Right, and because of that high risk, we have historically had a very low threshold to scan them. Like, a D dimer test is almost useless because cancer itself drives up inflammation and baseline D dimer levels, so we just bypass the clinical decision rules entirely and send them straight to CTPA. The, the direct to scanner standard.

Dr. James Whitfield

Exactly. But that standard, as common as it is, is really not benign. I mean, we are talking about repetitive radiation, intravenous contrast with risks of acute kidney injury or even anaphylaxis, not to mention the ED bottlenecks. And, and then there is the problem of finding those tiny, incidental subsegmental clots. We find them, and then we are forced to commit these patients to lifelong, bleeding prone anticoagulation therapy, even when we are not sure it helps.

Dr. Elena Rodriguez

It is a huge clinical dilemma. Which is why everyone was so excited about the YEARS diagnostic shortcut. For people who do not use it daily, YEARS simplifies things down to just three questions. Clinical signs of deep vein thrombosis, hemoptysis, and whether PE is the most likely diagnosis. If a patient meets zero of those criteria, the D dimer threshold to rule out a PE safely doubles from the standard five hundred nanograms per milliliter to one thousand. If they meet one or more, it drops back to five hundred.

Dr. James Whitfield

Right, but we never knew if we could trust that in cancer patients. Until now, or at least, until the Hydra trial was published in JAMA on July twelve, two thousand twenty six. This was an open label, randomized, noninferiority trial across twenty one hospitals in Europe. They took six hundred ninety eight active cancer patients and randomized them to either the YEARS algorithm or the standard direct to CTPA care.

Dr. Elena Rodriguez

And the results were pretty striking, right? In the YEARS group, twenty two percent of patients, that is seventy seven out of three hundred fifty two people, safely avoided CT scans entirely. And they met their noninferiority safety margin. The ninety day rate of subsequent symptomatic VTE or PE related death was one point eight percent in the YEARS group compared to five point five percent in the CTPA only group. The absolute risk difference was minus three point seven percent.

Dr. James Whitfield

Yes, although that, that five point five percent versus one point eight percent comparison actually brings up a really fascinating, almost bizarre paradox when you look closely at the math.

Dr. Elena Rodriguez

Wait, how so? If the YEARS group had fewer subsequent clots, does that not mean it was actually safer than scanning everyone?

Dr. James Whitfield

Well, that is the trap. On the surface, it looks like YEARS was superior, not just noninferior. But biologically, how does *not* scanning someone prevent a clot ninety days later? It does not. Remember, seventy eight point one percent of the patients in the YEARS arm still ended up getting a CTPA because they either had a positive clinical sign or their D dimer was above the threshold. So these two groups heavily overlapped. That difference between five point five percent and one point eight percent is almost certainly statistical noise, just random sampling error, rather than a true biological advantage.

Dr. Elena Rodriguez

Ah, okay, I see. So it is not that skipping the scan protected them. It is just that the trial was small enough that the random variation swung that way. But, James, there is a deeper mathematical issue here when we talk about prior probabilities. It is a classic Bayesian trap.

Dr. James Whitfield

Go on. Lay out the Bayes' theorem angle, because this is where the clinical nuance gets really critical.

Dr. Elena Rodriguez

Okay, so clinical decision rules like YEARS work by identifying a group with a very low pre test probability. In the original general population YEARS trial back in two thousand seventeen, if a patient met zero criteria, their baseline risk of PE was tiny, like, three point two percent. At that low starting point, a negative D dimer is incredibly reassuring. But in this Hydra cancer cohort, even if a patient had zero YEARS criteria, their baseline PE prevalence was still ten point five percent. That is a massive difference.

Dr. James Whitfield

Ten point five percent? That is, that is actually a high risk group in any other context.

Dr. Elena Rodriguez

Exactly! Ten point five percent versus fifteen point one percent in the group with one or more criteria. The YEARS score barely separates the two subgroups. So if you start with a ten point five percent baseline risk, even with a D dimer negative likelihood ratio of, say, zero point three, your post test probability is still around three to four percent. That actually exceeds the two percent failure rate that clinicians traditionally accept as safe for ruling out a PE.

Dr. James Whitfield

Huh. So, dynamically, we are accepting a higher residual risk of missed PE in these cancer patients just to save twenty two percent of them from a scan. It is a trade off.

Dr. Elena Rodriguez

Precisely. It is a trade off. It means the Hydra trial gives us permission to use YEARS to reduce unnecessary scans and contrast exposure in about one in five cancer patients, which is a real win for quality of life and ED flow. But it does not change the biology of cancer. A YEARS negative cancer patient is still fundamentally different, and higher risk, than a YEARS negative general patient.

Dr. James Whitfield

So the takeaway for the bedside is, yes, use the algorithm, save the scan if they meet the criteria, but do not let your guard down. If that patient still looks wrong, trust your clinical instinct over the algorithm. Alright, that seems like a good place to stop. Good chatting, Elena.

Dr. Elena Rodriguez

Yeah, good chatting. See you next time.