New data from Grail’s vast basket trial, CCGA, add to the evidence that its liquid biopsy has reasonable accuracy, but do not show marked improvement in sensitivity compared with data presented at Asco last year. An exception could be hepatic disease; in May the sensitivity for stage I-III liver cancer, at a preset specificity of more than 99%, was 68%. The new data, to be presented at the Asco-GI meeting tomorrow, show a sensitivity for liver and intrahepatic bile duct cancers combined of 79% for stages I-III. How much of this improvement might be down to the inclusion of patients with intrahepatic bile duct cancers is, however, unclear. The researchers also examined the test’s ability to predict each cancer’s tissue of origin, and the results here are comparable to, or possibly worse than, earlier data. At Asco the assay correctly identified tissue of origin in 90% of liver cancers and 100% of pancreatic cancers; the rates at Asco-GI were 78% for liver/intrahepatic bile duct cancers and 92% for pancreas/gallbladder/extrahepatic bile duct tumours. Again, these are not exact comparisons, and the new data cut could represent a tougher population. The CCGA trial is scheduled to conclude next month, according to clinicaltrials.gov.
|Grail's Asco-GI data|
|Sensitivity||Predicted TOO accuracy|
|Cancer type||All||Stage I-III||Stage IV||All||Stage I-III||Stage IV|
extrahepatic bile duct
|Liver/intrahepatic bile duct||86%||79%||100%||78%||85%||70%|
|All GI cancers||82%||72%||96%||92%||92%||92%|
|All at >99% specificity. TOO=tissue of origin. Source: Asco-GI abstract #283.|