Massive Bio co-founder cautions that AI trial matches do not guarantee enrollment
In an oncology interview, Arturo Loaiza-Bonilla points to matching systems that found more candidates without improving enrollment—and describes the work left after an alert.
Massive Bio co-founder Arturo Loaiza-Bonilla used a November 13 CancerNetwork interview to draw a boundary around AI trial matching: identifying eligible patients does not ensure enrollment, and education and follow-through remain necessary. He cited alert-based systems, including one tested in a Harvard randomized study, that increased matches without improving enrollment; he was not describing a result from Massive Bio. Separately, OncLive reports that Massive Bio’s trial database was developed with OpenAI and that its platform reaches more than 200,000 patients in 17 countries, but its reported evaluation does not include an enrollment outcome.
01
OncLive says the platform can characterize patients and match trials in as little as 17 seconds; that figure is not enrollment time.
02
The published prospective evaluation involved 3,800 patients, but OncLive reports no enrollment result.
03
Loaiza-Bonilla says work with uninsured and underrepresented patients at the University of Miami helped motivate him to found Massive Bio.
Finding a cancer patient a potential clinical trial is not the same as getting them enrolled. In an interview featured by CancerNetwork on November 13, Massive Bio co-founder Arturo Loaiza-Bonilla described that gap as a limit on AI matching: patient engagement, education and follow-through remain essential after a system identifies an eligible candidate.
Speaking on The Moonlight Shift, Loaiza-Bonilla cited alert-based matching systems, including one tested in a randomized study at Harvard, that produced more matches but no improvement in enrollment. That was his account of those systems—not a finding he presented about Massive Bio’s own platform.
His distinction concerns what happens after the software finds a possible study. Identifying eligibility addresses one part of access; helping patients understand and pursue participation addresses another. Loaiza-Bonilla placed the remaining barriers in that latter work, rather than treating a larger list of matches as sufficient.
What the matching system does
Massive Bio’s platform draws structured patient information from U.S. electronic medical records and compares it with a clinical-trial database. OncLive’s November 18 profile describes that database as developed in collaboration with OpenAI, with trial requirements organized into parameters the system can use for matching.
The profile puts the platform’s reach at more than 200,000 patients across 17 countries. It also describes rapid patient characterization and matching, alongside a published prospective evaluation—one that followed patients forward—involving 3,800 people. It does not give an enrollment result from that evaluation.
Reported speed and evaluation size
As little as 17 secondsPatient characterization and trial matching
OncLive gives this as the platform’s end-to-end processing time, not the time needed to enroll a patient.
3,800 patientsProspective evaluation
The profile reports a published evaluation of this size; the figure describes participation in the evaluation.
The access problem behind the company
Loaiza-Bonilla traced his decision to build Massive Bio to clinical work at the University of Miami. He described serving a largely uninsured, underrepresented population whose patients could qualify for trials yet remain overlooked by the systems intended to identify them. The problem was not simply whether a suitable study existed.
At Penn Medicine, he encountered another information gap: colleagues working in the same department and disease areas did not always know which studies were open. He attributed that to limited time and misaligned incentives, not indifference. Those experiences led him to leave academia and co-found the company.
Asked whether AI would make cancer care more equitable, the company’s co-founder stopped short of a promise. “Jury still out,” he said.
Sources
onclive.comCan AI Fix Clinical Trial Access in Oncology? Arturo Loaiza-Bonilla, MD, Built the Answer With the Data to Back It | OncLive
cancernetwork.comCan AI Fix Clinical Trial Access in Oncology? Dr. Arturo Loaiza-Bonilla Built the Answer With the Data to Back It. | CancerNetwork
Reader comments
Newest comments first. Replies stay oldest first.