ESMO Congress 2026

As an asset class, ctDNA-based liquid biopsy rests on a series of key studies: in colorectal cancer, a personalized genetic approach showed ctDNA could reliably monitor tumor dynamics during surgery and chemotherapy ^[3]; CAPP-Seq pushed sensitivity down to mutant allele fractions of ~0.02% and enabled earlier response assessment than imaging ^[4]; TRACERx used a phylogenetic approach on postoperative plasma to identify early-stage lung cancer patients at high risk of recurrence ^[5]; and reviews have framed ctDNA applications as prognostication, monitoring of response and resistance, and MRD quantification ^[2][6]. Notably, prostate is a low-ctDNA-shedding tumor type: among 640 patients across cancer types, ctDNA detection exceeded 75% in advanced pancreatic, ovarian, colorectal and other cancers but was below 50% in prostate cancer ^[1], so ctDNA-based stratification in prostate cancer requires dedicated validation.

The #ESMO26 posts cover two distinct items. First, the ctDNA tumor fraction analysis from TALAPRO-3: in HRR-deficient mCSPC, baseline ctDNA positivity (TF ≥1%) was prognostic for rPFS in both arms; at week 5, 82% of ctDNA-positive patients cleared on TALA + ENZA versus 68% on PBO + ENZA, and clearance tracked with longer rPFS (HR 0.36 versus persistent ctDNA+). Second, Myriad announced new MONSTAR-SCREEN-3 data, stating that its WGS-based Precise MRD™ platform demonstrated the potential for ultrasensitive MRD detection after surgery in resectable colorectal and pancreatic cancers.

Each moves one step beyond prior knowledge. TALAPRO-3 delivers, in a prostate population with two randomized treatment arms, both the between-arm difference in week-5 clearance (82% vs 68%) and the rPFS difference between cleared and persistently positive patients (HR 0.36), suggesting early on-treatment ctDNA could become a real tool for risk stratification and monitoring; the post explicitly notes the next question is whether it adds value beyond PSA. On the Myriad side, evidence that postoperative MRD predicts recurrence was previously established mainly in early-stage lung cancer ^[5]; MONSTAR-SCREEN-3 extends WGS-based ultrasensitive detection into the postoperative colorectal and pancreatic cancer settings.

The limitations are equally clear: these figures come from social media disclosure during the conference, with full statistical details not disclosed; in TALAPRO-3, the HR 0.36 derives from a comparison grouped by ctDNA clearance status, and whether it can directly guide treatment decisions, along with any incremental value beyond PSA, remains to be confirmed; MONSTAR-SCREEN-3 currently demonstrates detection potential, and whether it translates into clinical utility for guiding postoperative treatment decisions is not addressed in the material.