Abstract / Summary
The 90-gene expression assay is a molecular diagnostic tool developed to assist in identifying the tissue of origin in cancer of unknown primary. This study aimed to evaluate the performance of the 90-gene expression assay in predicting the tissue of origin for both primary and liver metastases in Egyptian patients. A total of 51 formalin-fixed, paraffin-embedded tumor specimens, comprising 34 primary tumors and 17 liver metastases with known origins, were analyzed using the 90-gene expression assay. Gene expression profiles were generated for each specimen and compared against a reference database of 21 tumor types to calculate similarity scores and predict the tissue of origin. The 90-gene expression assay achieved an overall agreement of 90.2% (46/51; 95% CI 78.6–96.7) with the reference diagnosis, correctly classifying 100% (34/34) of primary tumors and 70.6% (12/17) of liver metastases. Agreement was 97.8% (45/46) across primary tumors and liver metastases from non-pancreatic primaries combined, and 91.7% (11/12) for non-pancreatic liver metastases. Notably, four of the five discordant cases were pancreatic liver metastases, for which agreement was 20.0% (1/5); two of these were assigned to hepatocellular origin. The 90-gene expression assay showed high concordance with reference diagnoses in Egyptian patients, performing consistently across primary tumors and across liver metastases from breast, colorectal and gastroesophageal primaries. Lower agreement was confined to pancreatic liver metastases, a small subgroup in which the assay may assign tumors to hepatocellular origin. These findings support the assay as a practical adjunct for tissue-of-origin determination in this population, also indicate that a hepatocellular prediction obtained from a liver lesion should be corroborated before it guides management, and warrant larger prospective validation across diverse populations.