Biomarker analysis in a phase III study of pemetrexed-carboplatin versus etoposide-carboplatin in chemonaive patients with extensive-stage small-cell lung cancer

BACKGROUND

Clinical results of a randomized phase III trial comparing pemetrexed-carboplatin (PC) with etoposide-carboplatin (EC) in chemonaive patients with extensive-stage disease small-cell lung cancer (ED-SCLC) resulted in trial closure for futility; biomarker analyses using immunohistochemistry (IHC) and single-nucleotide polymorphisms (SNPs) are described herein.

PATIENTS AND METHODS

Thymidylate synthase (TS), excision repair cross complementing-1 (ERCC1), glycinamide ribonucleotide formyltransferase (GARFT), and folylpolyglutamate synthetase (FPGS) were investigated using IHC (n=395). SNPs were genotyped for TS, FPGS, γ-glutamyl hydrolase (GGH), methylenetetrahydrofolate reductase (MTHFR), folate receptor-α FR-α, and solute carrier 19A1 (SLC19A1; n=611).

RESULTS

None of the IHC biomarkers (folate pathway or ERCC1) were found to be predictive or prognostic in this setting. rs2838952 (adjacent to SLC19A1) had significant treatment-independent association with overall survival (OS; hazard ratio 0.590, P=0.01). Nine GGH-associated SNPs interacted with rs3788205 (SLC19A1) for OS on the PC arm. rs12379987 (FPGS) interacted with treatment for OS (interaction P=0.036).

CONCLUSION

Potential ERCC1 and folate pathway IHC biomarkers failed to predict outcome in either study arm in ED-SCLC. SNPs in regions including FPGS and SLC19A1 and interacting SNPs in GGH and SLC19A1 were associated with differences in OS; however, none of these SNPs predicted for greater survival with PC over EC.

Lactose-binding lectin expression in human colorectal carcinomas. Relation to tumor progression

Lactose-binding lectins having Mr values of approximately 14,000 (L-14.5) and approximately 35,000 Da have been found in a variety of vertebrate tissues, including normal intestine and colon, and in several types of tumors such as colon carcinomas. To determine the clinical relevance of such lectins in human colon cancer, specimens from 46 patients with colorectal carcinoma of identified Dukes' stages were selected and analyzed for the presence and amount of lactose-binding lectins by immunoblotting using a polyclonal, rabbit anti-lectin antibody followed by binding of 125I-labeled anti-rabbit IgG. The amount of a lectin having an Mr value of approximately 31,000 Da (L-31) varied among the specimens. The levels of L-31 lectin in colorectal cancer specimens from primary tumors of patients with distant metastases (Dukes' stage D) were significantly higher than were those from patients without detectable metastases (Dukes' stages B1 and B2). In contrast, among the various specimens the variation in the level of the L-14.5 lectin was smaller, and there was no correlation between the amount of this lectin and cancer stage. Immunohistochemical staining of thin sections of colorectal tumor specimens using antibodies specific for either L-31 or L-14.5 lectin revealed that the two were located at different places, the L-31 lectin primarily within the cytoplasm of carcinoma cells, and the L-14.5 lectin associated with secreted material. These results indicated that the relative amount of the L-31 lectin increases as the colorectal cancer progresses to a more malignant stage.