US-based physicians have engineered an artificial intelligence framework capable of preemptively gauging the likelihood of developing chronic graft-versus-host disease. This breakthrough was announced by the press office of the Medical University of South Carolina.
According to Professor Sofia Paczesny, this complication is typically identified months subsequent to its onset, by which point the inflammatory process has already inflicted considerable harm upon the body. The newly developed system offers the capability to detect the hazard before any symptoms manifest or irreversible damage occurs.
Chronic graft-versus-host disease emerges when the immune cells originating from the donor commence an assault on the tissues of the recipient. This process can result in afflictions to the liver, the gastrointestinal tract, the skin, and various other organ systems. The condition progresses slowly, meaning that during its initial phases, the patient may perceive no decline in their well-being.
To formulate this algorithm, experts scrutinized gene activity within 1,300 stem cell samples taken from patients following transplantation. Furthermore, they quantified the concentrations of proteins associated with inflammatory and immune responses at the 90- to 100-day mark post-surgery. The resulting AI model incorporated patient demographics such as age, the nature of the transplanted organ, and their prior medical histories.
During validation, the algorithm accurately pinpointed individuals within the high-risk category and successfully predicted the evolution of the complication over the subsequent eighteen months. The creators estimate that this technology will enable threat detection several months preceding the appearance of clinical indicators, allowing for timely adjustments to ongoing treatment regimens.
About the Author
