What happened
Pathwork Diagnostics developed the Tissue of Origin Test, which compares the RNA expression of a fresh-frozen tumor with a database of 15 tumor types. Its 1550 marker probesets were selected with machine learning, and its model was built on 2039 specimens split into training and test sets. Four processing laboratories ran the test on poorly differentiated, undifferentiated and metastatic tumor specimens. Of 659 specimens processed, 545 met the labeling limits and quality checks. FDA reviewers report that results agreed with the available diagnosis for 86.3% of metastatic and 92.0% of primary specimens. A similarity score of 30 or more marked the right tissue 92.9% of the time. FDA cleared the test on 30 July 2008. Tumors from tissues outside the panel often scored as a panel tissue, including 48 of 49 endometrial cancers. The test is not intended for tumors that current practice cannot diagnose.
Sources
-
Regulatory document, by authors with no tie to the system.
Based on the n = 545 results, the probability that a true positive tissue call was obtained when a Similarity Score of ≥ 30 was reported was 92.9%, 95% CI [90.3, 95.0], and the probability that a true negative tissue call was obtained when a Similarity Score of ≤ 5 was reported was 99.7%, 95% CI [99.6, 99.8].
Details
AI system
- Workflow
-
RNA from fresh-frozen tumor tissue was hybridized to the Pathchip microarray and scanned. Laboratories processed specimens as the Pathwork Specimen Processing Guide sets out.
Follow-up
- Conditions for replication
-
The tumor can be diagnosed according to current clinical and pathological practice. A qualified clinician reads the result in the context of the patient's clinical history and other tests.
Entry details
- Sector
-
Healthcare.
- Purpose
-
Pathwork Diagnostics built the Tissue of Origin Test to measure how closely a tumor's RNA expression pattern resembles each of 15 common tumor types, for use with other diagnostic tests.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). Pathwork Tissue of Origin Test matched the known origin of most tumors (AIBD-0808). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0808.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)