Genetic Determinants of Outcome in Intrahepatic Cholangiocarcinoma
Thomas Boerner,Esther Drill,Linda M. Pak,Bastien Nguyen,Carlie S. Sigel,Alexandre Doussot,Paul Shin,Debra A. Goldman,Mithat Gonen,Peter J. Allen,Vinod P. Balachandran,Andrea Cercek,James Harding,David B. Solit,Nikolaus Schultz,Ritika Kundra,Henry Walch,Michael I. D’Angelica,Ronald P. DeMatteo,Jeffrey Drebin,Nancy E. Kemeny,T. Peter Kingham,Amber L. Simpson,Jaclyn F. Hechtman,Efsevia Vakiani,Maeve A. Lowery,J.N.M. Ijzermans,S. Buettner,B. Groot Koerkamp,M. Doukas,Rohit Chandwani,William R. Jarnagin
DOI: https://doi.org/10.1002/hep.31829
IF: 17.298
2021-09-01
Hepatology
Abstract:<section class="article-section__content"><h3 class="article-section__sub-title section1"> Background/Aims</h3><p>Genetic alterations in intrahepatic cholangiocarcinoma (iCCA) are increasingly well‐characterized, but their impact on outcome and prognosis remain unknown.</p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Approach/Results</h3><p>This bi‐institutional study of patients with confirmed iCCA (n=412) used targeted next‐generation sequencing of primary tumors to define associations among genetic alterations, clinicopathological variables, and outcome. The most common oncogenic alterations were <i>IDH1</i> (20%), <i>ARID1A</i> (20%), <i>TP53</i> (17%), <i>CDKN2A</i> (15%), <i>BAP1</i> (15%), <i>FGFR2</i> (15%), <i>PBRM1</i> (12%), and <i>KRAS</i> (10%). <i>IDH1/2</i> mutations (<i>mut</i>) were mutually exclusive with <i>FGFR2</i> fusions (<i>fus</i>), but neither was associated with outcome. For all patients, <i>TP53</i> (p<0.0001), <i>KRAS</i> (p=0.0001), and <i>CDKN2A</i> (p<0.0001) alterations predicted worse overall survival (OS). These high‐risk alterations were enriched in advanced disease but adversely impacted survival across all stages, even when controlling for known correlates of outcome (multifocal disease, lymph node involvement, bile duct type, periductal infiltration). In resected patients (n=209), <i>TP53mut</i> (HR=1.82, 95%CI=1.08‐3.06, p=0.03) and <i>CDKN2A</i> deletions (<i>del</i>) (HR=3.40, 95%CI=1.95‐5.94, p<0.001) independently predicted shorter OS, as did high‐risk clinical variables (multifocal liver disease [p<0.001]; regional lymph node metastases [p<0.001]), whereas <i>KRASmut</i> (HR=1.69, 95%CI=0.97‐2.93, p=0.06) trended toward statistical significance. The presence of both or neither high‐risk clinical or genetic factors represented outcome extremes (median OS=18.3 vs. 74.2 months, p<0.001), with high‐risk genetic alterations alone (median OS=38.6 months, 95%CI=28.8‐73.5) or high‐risk clinical variables alone (median OS=37.0 months, 95%CI=27.6‐NA) associated with intermediate outcome. <i>TP53mut, KRASmut,</i> and <i>CDKN2Adel</i> similarly predicted worse outcome in patients with unresectable iCCA. <i>CDKN2Adel</i> tumors with high‐risk clinical features were notable for limited survival and no benefit of resection over chemotherapy. </p></section><section class="article-section__content"><h3 class="article-section__sub-title section1"> Conclusions</h3><p><i>TP53</i>, <i>KRAS,</i> and <i>CDKN2A</i> alterations were independent prognostic factors in iCCA when controlling for clinical and pathologic variables, disease stage, and treatment. Since genetic profiling can be integrated into pre‐treatment therapeutic decision‐making, combining clinical variables with targeted tumor sequencing may identify patient subgroups with poor outcome irrespective of treatment strategy. </p></section>
gastroenterology & hepatology