Clinical evaluation of cancer signal origin prediction and diagnostic resolution following multi-cancer early detection testing.

10520 Background: PATHFINDER (NCT04241796) was a prospective, interventional return of results study that enrolled 6,662 participants (pts) from 7 US institutions to assess clinical implementation of a blood-based multi-cancer early detection (MCED) test. Test results for those with a cancer signal detected also included a cancer signal origin prediction designed to help guide diagnostic evaluations. We analyzed whether the diagnostic evaluations in PATHFINDER were appropriately directed by cancer signal origin predictions, assessed the residual risk of cancer after negative initial evaluation, and examined the role of whole-body imaging (WBI, defined as PET/CT or CT chest/abdomen/pelvis) in diagnostic evaluation. Methods: Analysis included pts with a cancer signal detected result from the two different versions of an MCED test evaluated in PATHFINDER. Prespecified criteria blinded to study results were used to define expected diagnostic evaluations by cancer signal origin prediction. These criteria were then used to judge concordance between cancer signal origin predictions and the diagnostic evaluation actually performed. Definitive cancer or no cancer diagnosis (cancer status at end of study [EOS]) was assessed at 12 months after enrollment (EOS). Results: 39 pts had cancer signal detected results by both test versions. Diagnostic evaluations were consistent with cancer signal origin in 30/39 (77%) cases. Diagnostic resolution was achieved in 32/39 (82%) pts after initial evaluation, while 7/39 (18%) required additional workup. 25 of the 32 (78%) pts who achieved diagnostic resolution after initial evaluation had a cancer signal origin-guided workup. Additional evaluation of the 7 pts not achieving initial diagnostic resolution was based on persistent clinical suspicion of cancer due to prior cancer history (n = 1) or abnormal/equivocal findings on initial evaluation (n = 6). Additional workup led to diagnostic resolution in all 7 cases (3 with cancer and 4 with no cancer confirmed at EOS). WBI was ordered in 27 cases, of whom 9 had a prior cancer history, 10 had non-localizing cancer signal origin (hematologic or indeterminate), and 1 had both. WBI was equivocal or had no impact on diagnostic resolution in 8/27 (30%) cases. Conclusions: Most cases resulted in diagnostic resolution after cancer signal origin-directed diagnostic workup. In pts with a cancer history and negative initial evaluations, and those with equivocal initial evaluations, there was still a risk of cancer, suggesting the need for additional evaluation. WBI was used mostly in those with a cancer history or non-localizing cancer signal origins, and had no impact on diagnosis in almost one-third of pts. Clinical trial information: NCT04241796 .