Which medication works for which patient? This is one of the central questions in oncology \u2013 and one that remains difficult to answer. Machine learning opens up new possibilities here, but for clinical use, predictive performance alone is not enough: doctors need to be able to trust a model\u2019s output.<\/p>\n\n\n\n
This is precisely the challenge that researchers from Saarland University<\/a>, the University Medical Center G\u00f6ttingen<\/a>, and CAIMed<\/a> set out to address. Their newly developed method, MORGOTH, not only improves the accuracy of drug sensitivity predictions but also makes it transparent how confident the model is in any given result. In clinical decision-making, this kind of interpretability is at least as important as the prediction itself.<\/p>\n\n\n\n The study also surfaces a methodologically relevant finding: previous comparisons between different model types contained a systematic bias \u2013 when evaluated on equal footing, results differ from what had previously been assumed.<\/p>\n\n\n\n MORGOTH<\/em><\/strong> is freely available to the research community as an open-source package. The paper was published open access in the journal Digital Discovery<\/em>\u00a0by the Royal Society of Chemistry<\/a>.<\/p>\n\n\n\n