Signature to predict risk in patients with and without adjuvant therapy. WinterChristof KristiansenGlen KerstingStephan RoyJanine AustDaniela KnöselThomas RümmelePetra JahnkeBeatrix HentrichVera RückertFelix NiedergethmannMarco WeichertWilko BahraMarcus J. SchlittHans SettmacherUtz FriessHelmut BüchlerMarkus SaegerHans-Detlev SchroederMichael PilarskyChristian GrützmannRobert 2012 <p>(<b>A</b>) Signature to predict risk in patients with adjuvant therapy. The signature was developed with patients receiving adjuvant therapy separated by their median survival into two groups, a high risk group with shorter survival and a low risk group with longer survival. A classifier trained with the signature using leave-one-out cross-validation shows a significant difference between the predicted low and high risk group (, logrank test). (<b>B</b>) Signature to predict risk in patients without adjuvant therapy. The signature was developed with patients not receiving adjuvant therapy separated by their median survival into two groups, a high risk group with shorter survival and a low risk group with longer survival. A classifier trained with the signature using leave-one-out cross-validation shows a significant difference between the predicted low and high risk group (, logrank test).</p>