Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1585
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-16T10:43:11.082860Z
status: published
lens: g_space
sector: health
headline: A multimodal AI biomarker PATH-ORACLE improves prediction of recurrence in stage I lung adenocarcinoma
dek: Background: Surgical resection is the standard treatment for stage I lung adenocarcinoma, in most cases without additional systemic adjuvant treatment. A substantial proportion of stage I cases recur, with a 5-year survival rate <50%. Clinical data suggest that adjuvant treatment, including immune checkpoint inhibitor therapy, may improve survival in such recurrent cases. Previously evaluated predictors, including the International Association for the Study of Lung Cancer (IASLC) grading system applied to histol…
gain_title: Combining routine histology slides with a transcriptomic score predicted 1-year recurrence more accurately than either source alone, across two independent validation cohorts.
problem_title: (none)
trace_subject: (none)
gain_reading: Combining routine histology slides with a transcriptomic score predicted 1-year recurrence more accurately than either source alone, across two independent validation cohorts.
gain_evidence: PATH-ORACLE exceeded both constituent modalities, predicting 1-year recurrence with an area under the curve of 0.86 and 0.84 in the two independent validation cohorts
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers built PATH-ORACLE, a multimodal deep learning biomarker that adds routine histological sections, read by pretrained foundation models, to an existing validated transcriptomic risk score for stage I lung adenocarcinoma. It was tested in two independent cohorts and outperformed each of its component predictors on recurrence risk.

Better risk stratification matters because many stage I patients recur after surgery, yet adjuvant treatment is not routinely given. The tool is so far only prognostic, and the authors say extensive further validation is needed before it could guide who receives adjuvant therapy.
limitation: PATH-ORACLE is currently a prognostic biomarker; extensive validation is still required before it can serve as a predictive biomarker for choosing adjuvant therapy.
tag: Evidence-backed gain
key_points: PATH-ORACLE builds on the prospectively validated transcriptomic ORACLE score and adds histology analyzed by pretrained foundation models. | The histology-only predictor outperformed automated IASLC grading and stayed prognostic after adjustment for tumor size, grade and T-stage. | One-year recurrence AUC was 0.86 and 0.84 in the two independent validation cohorts. | A single prespecified cut-off separated high-risk from low-risk groups.
rundown: Surgery is the standard treatment for stage I lung adenocarcinoma and usually involves no additional systemic adjuvant therapy. The source notes that a substantial share of cases recur, with 5-year survival below 50% in those cases, and that clinical data suggest adjuvant treatment such as immune checkpoint inhibitors may improve survival.

Prior predictors, including IASLC grading of histological sections and transcriptomic profiles, were described as not sufficiently accurate or consistent for risk stratification or treatment guidance. The authors hypothesized that these measurements carry complementary information that gains power when combined.

The authors envision eventual use to prioritize stage IB patients for adjuvant targeted therapy, chemotherapy or immune checkpoint inhibitor therapy, with or without liquid biopsy-based monitoring, but that would require converting it from a prognostic to a predictive biomarker.
sources:
- peer_reviewed | Immuno-Oncology and Technology | https://doi.org/10.1016/j.iotech.2026.101614 | 2026-09-17
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

dd16cabad881acc430bbc708e4c466ccb3be52eacddaec4760697ff42fbbb46a

previous

0000000000000000000000000000000000000000000000000000000000000000