Swiss pharmaceutical giant Roche announced a $750 million deal to acquire artificial intelligence startup PathAI on Tuesday, aiming to significantly accelerate the speed at which diagnostic software can suggest further testing is needed.
Swiss pharmaceutical giant Roche finalized a $750 million agreement to acquire artificial intelligence startup PathAI this week, seeking to accelerate the speed at which algorithms can cautiously suggest a statistical probability of cellular abnormalities. However, independent researchers note the merger’s efficacy at actually diagnosing diseases remains largely observational.
The technology, which utilizes machine learning to analyze digital pathology slides, was previously shown to reduce the time human pathologists spend looking at tissue samples. Yet experts warn that the sample size of one $750 million corporate acquisition is too small to draw definitive conclusions, emphasizing that the financial transaction has not yet been subjected to a double-blind, placebo-controlled trial.
While early press releases show a strong statistical association between Roche and the $750 million, we must be careful not to conflate a corporate buyout with a clinically significant improvement in patient outcomes.
A meta-analysis of the acquisition announcement reveals several potential methodological flaws, primarily that the $750 million figure was self-reported by executives in an uncontrolled financial environment. Furthermore, the algorithmic models utilized by PathAI were trained on retrospective medical records, meaning the software's ability to definitively state whether a patient has cancer remains subject to confounding variables, such as whether the patient actually has cancer.
In a statement, Roche executives recommended a baseline dosage of optimism regarding the integration of artificial intelligence into daily pathology workflows. Still, the company disclosed in its regulatory filings that the software may cause mild to moderate algorithmic hallucinations.
At press time, the medical community recommended that investors take a conservative dose of skepticism regarding the merger, at least until the Q3 earnings report can be replicated in a multi-center, randomized trial.