Deciding which cancer patients need immediate treatment and who can safely wait is an uncomfortable assessment for cancer clinicians during the COVID-19 pandemic.
In early April, as the COVID-19 surge was bearing down on New York City, those treatment decisions were “a juggling act every single day,” Jonathan Yang, MD, PhD, a radiation oncologist from New York’s Memorial Sloan Kettering Cancer Center, told Medscape Medical News.
Eventually, a glut of guidelines, recommendations, and expert opinions aimed at helping oncologists emerged. The tools help navigate the complicated risk-benefit analysis of their patient’s risk of infection by SARS-CoV-2 and delaying therapy.
Now, a new tool, which appears to be the first of its kind, quantifies that risk-benefit analysis. But its presence immediately raises the question: can it help?
Three-Tier Systems Are Not Very Sophisticated
OncCOVID, a free tool that was launched May 26 by the University of Michigan, allows physicians to individualize risk estimates for delaying treatment of up to 25 early- to late-stage cancers. It includes more than 45 patient characteristics, such as age, location, cancer type, cancer stage, treatment plan, underlying medical conditions, and proposed length of delay in care.
Combining these personal details with data from the National Cancer Institute’s SEER (Surveillance, Epidemiology, and End Results) registry and the National Cancer Database, the Michigan app then estimates a patient’s 5- or 10-year survival with immediate vs delayed treatment and weighs that against their risk for COVID-19 using data from the Johns Hopkins Coronavirus Resource Center.
“We thought, isn’t it better to at least provide some evidence-based quantification, rather than a back-of-the-envelope three-tier system that is just sort of ‘made up’?“ explained one of the developers, Daniel Spratt, MD, associate professor of radiation oncology at Michigan Medicine.
Spratt explained that almost every organization, professional society, and government has created something like a three-tier system. Tier 1 represents urgent cases and patients who need immediate treatment. For tier 2, treatment can be delayed weeks or a month, and with tier 3, it can be delayed until the pandemic is over or it’s deemed safe.
“[This system] sounds good at first glance, but in cancer, we’re always talking about personalized medicine, and it’s mind-blowing that these tier systems are only based on urgency and prognosis,” he told Medscape Medical News.
Spratt offered an example. Consider a patient with a very aggressive brain tumor ― that patient is in tier 1 and should undergo treatment immediately. But will the treatment actually help? And how helpful would the procedure be if, say, the patient is 80 years old and, if infected, would have a 30% to 50% chance of dying from the coronavirus?
“If the model says this guy has a 5% harm and this one has 30% harm, you can use that to help prioritize,” summarized Spratt.
The app can generate risk estimates for patients living anywhere in the world and has already been accessed by people from 37 countries. However, Spratt cautions that it is primarily “designed and calibrated for the US.
“The estimates are based on very large US registries, and though it’s probably somewhat similar across much of the world, there’s probably certain cancer types that are more region specific ― especially something like stomach cancer or certain types of head and neck cancer in parts of Asia, for example,” he said.
Although the app’s COVID-19 data are specific to the county level in the United States, elsewhere in the world, it is only country specific.
“We’re using the best data we have for coronavirus, but everyone knows we still have large data gaps,” he acknowledged.
